This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Silence two warnings from gcc when being -pedantic
[perl5.git] / sv.c
CommitLineData
a0d0e21e 1/* sv.c
79072805 2 *
4bb101f2 3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
b94e2f88 4 * 2000, 2001, 2002, 2003, 2004, 2005, 2006, by Larry Wall and others
79072805
LW
5 *
6 * You may distribute under the terms of either the GNU General Public
7 * License or the Artistic License, as specified in the README file.
8 *
a0d0e21e 9 * "I wonder what the Entish is for 'yes' and 'no'," he thought.
645c22ef
DM
10 *
11 *
5e045b90
AMS
12 * This file contains the code that creates, manipulates and destroys
13 * scalar values (SVs). The other types (AV, HV, GV, etc.) reuse the
14 * structure of an SV, so their creation and destruction is handled
15 * here; higher-level functions are in av.c, hv.c, and so on. Opcode
16 * level functions (eg. substr, split, join) for each of the types are
17 * in the pp*.c files.
79072805
LW
18 */
19
20#include "EXTERN.h"
864dbfa3 21#define PERL_IN_SV_C
79072805 22#include "perl.h"
d2f185dc 23#include "regcomp.h"
79072805 24
51371543 25#define FCALL *f
2c5424a7 26
2f8ed50e
OS
27#ifdef __Lynx__
28/* Missing proto on LynxOS */
29 char *gconvert(double, int, int, char *);
30#endif
31
e23c8137 32#ifdef PERL_UTF8_CACHE_ASSERT
ab455f60 33/* if adding more checks watch out for the following tests:
e23c8137
JH
34 * t/op/index.t t/op/length.t t/op/pat.t t/op/substr.t
35 * lib/utf8.t lib/Unicode/Collate/t/index.t
36 * --jhi
37 */
6f207bd3 38# define ASSERT_UTF8_CACHE(cache) \
ab455f60
NC
39 STMT_START { if (cache) { assert((cache)[0] <= (cache)[1]); \
40 assert((cache)[2] <= (cache)[3]); \
41 assert((cache)[3] <= (cache)[1]);} \
42 } STMT_END
e23c8137 43#else
6f207bd3 44# define ASSERT_UTF8_CACHE(cache) NOOP
e23c8137
JH
45#endif
46
f8c7b90f 47#ifdef PERL_OLD_COPY_ON_WRITE
765f542d 48#define SV_COW_NEXT_SV(sv) INT2PTR(SV *,SvUVX(sv))
607fa7f2 49#define SV_COW_NEXT_SV_SET(current,next) SvUV_set(current, PTR2UV(next))
b5ccf5f2 50/* This is a pessimistic view. Scalar must be purely a read-write PV to copy-
765f542d 51 on-write. */
765f542d 52#endif
645c22ef
DM
53
54/* ============================================================================
55
56=head1 Allocation and deallocation of SVs.
57
d2a0f284
JC
58An SV (or AV, HV, etc.) is allocated in two parts: the head (struct
59sv, av, hv...) contains type and reference count information, and for
60many types, a pointer to the body (struct xrv, xpv, xpviv...), which
61contains fields specific to each type. Some types store all they need
62in the head, so don't have a body.
63
64In all but the most memory-paranoid configuations (ex: PURIFY), heads
65and bodies are allocated out of arenas, which by default are
66approximately 4K chunks of memory parcelled up into N heads or bodies.
93e68bfb
JC
67Sv-bodies are allocated by their sv-type, guaranteeing size
68consistency needed to allocate safely from arrays.
69
d2a0f284
JC
70For SV-heads, the first slot in each arena is reserved, and holds a
71link to the next arena, some flags, and a note of the number of slots.
72Snaked through each arena chain is a linked list of free items; when
73this becomes empty, an extra arena is allocated and divided up into N
74items which are threaded into the free list.
75
76SV-bodies are similar, but they use arena-sets by default, which
77separate the link and info from the arena itself, and reclaim the 1st
78slot in the arena. SV-bodies are further described later.
645c22ef
DM
79
80The following global variables are associated with arenas:
81
82 PL_sv_arenaroot pointer to list of SV arenas
83 PL_sv_root pointer to list of free SV structures
84
d2a0f284
JC
85 PL_body_arenas head of linked-list of body arenas
86 PL_body_roots[] array of pointers to list of free bodies of svtype
87 arrays are indexed by the svtype needed
93e68bfb 88
d2a0f284
JC
89A few special SV heads are not allocated from an arena, but are
90instead directly created in the interpreter structure, eg PL_sv_undef.
93e68bfb
JC
91The size of arenas can be changed from the default by setting
92PERL_ARENA_SIZE appropriately at compile time.
645c22ef
DM
93
94The SV arena serves the secondary purpose of allowing still-live SVs
95to be located and destroyed during final cleanup.
96
97At the lowest level, the macros new_SV() and del_SV() grab and free
98an SV head. (If debugging with -DD, del_SV() calls the function S_del_sv()
99to return the SV to the free list with error checking.) new_SV() calls
100more_sv() / sv_add_arena() to add an extra arena if the free list is empty.
101SVs in the free list have their SvTYPE field set to all ones.
102
ff276b08 103At the time of very final cleanup, sv_free_arenas() is called from
645c22ef 104perl_destruct() to physically free all the arenas allocated since the
6a93a7e5 105start of the interpreter.
645c22ef
DM
106
107Manipulation of any of the PL_*root pointers is protected by enclosing
108LOCK_SV_MUTEX; ... UNLOCK_SV_MUTEX calls which should Do the Right Thing
109if threads are enabled.
110
111The function visit() scans the SV arenas list, and calls a specified
112function for each SV it finds which is still live - ie which has an SvTYPE
113other than all 1's, and a non-zero SvREFCNT. visit() is used by the
114following functions (specified as [function that calls visit()] / [function
115called by visit() for each SV]):
116
117 sv_report_used() / do_report_used()
f2524eef 118 dump all remaining SVs (debugging aid)
645c22ef
DM
119
120 sv_clean_objs() / do_clean_objs(),do_clean_named_objs()
121 Attempt to free all objects pointed to by RVs,
122 and, unless DISABLE_DESTRUCTOR_KLUDGE is defined,
123 try to do the same for all objects indirectly
124 referenced by typeglobs too. Called once from
125 perl_destruct(), prior to calling sv_clean_all()
126 below.
127
128 sv_clean_all() / do_clean_all()
129 SvREFCNT_dec(sv) each remaining SV, possibly
130 triggering an sv_free(). It also sets the
131 SVf_BREAK flag on the SV to indicate that the
132 refcnt has been artificially lowered, and thus
133 stopping sv_free() from giving spurious warnings
134 about SVs which unexpectedly have a refcnt
135 of zero. called repeatedly from perl_destruct()
136 until there are no SVs left.
137
93e68bfb 138=head2 Arena allocator API Summary
645c22ef
DM
139
140Private API to rest of sv.c
141
142 new_SV(), del_SV(),
143
144 new_XIV(), del_XIV(),
145 new_XNV(), del_XNV(),
146 etc
147
148Public API:
149
8cf8f3d1 150 sv_report_used(), sv_clean_objs(), sv_clean_all(), sv_free_arenas()
645c22ef 151
645c22ef
DM
152=cut
153
154============================================================================ */
155
4561caa4
CS
156/*
157 * "A time to plant, and a time to uproot what was planted..."
158 */
159
77354fb4
NC
160/*
161 * nice_chunk and nice_chunk size need to be set
162 * and queried under the protection of sv_mutex
163 */
164void
165Perl_offer_nice_chunk(pTHX_ void *chunk, U32 chunk_size)
166{
97aff369 167 dVAR;
77354fb4
NC
168 void *new_chunk;
169 U32 new_chunk_size;
170 LOCK_SV_MUTEX;
171 new_chunk = (void *)(chunk);
172 new_chunk_size = (chunk_size);
173 if (new_chunk_size > PL_nice_chunk_size) {
174 Safefree(PL_nice_chunk);
175 PL_nice_chunk = (char *) new_chunk;
176 PL_nice_chunk_size = new_chunk_size;
177 } else {
178 Safefree(chunk);
179 }
180 UNLOCK_SV_MUTEX;
181}
cac9b346 182
fd0854ff 183#ifdef DEBUG_LEAKING_SCALARS
22162ca8 184# define FREE_SV_DEBUG_FILE(sv) Safefree((sv)->sv_debug_file)
fd0854ff
DM
185#else
186# define FREE_SV_DEBUG_FILE(sv)
187#endif
188
48614a46
NC
189#ifdef PERL_POISON
190# define SvARENA_CHAIN(sv) ((sv)->sv_u.svu_rv)
191/* Whilst I'd love to do this, it seems that things like to check on
192 unreferenced scalars
7e337ee0 193# define POSION_SV_HEAD(sv) PoisonNew(sv, 1, struct STRUCT_SV)
48614a46 194*/
7e337ee0
JH
195# define POSION_SV_HEAD(sv) PoisonNew(&SvANY(sv), 1, void *), \
196 PoisonNew(&SvREFCNT(sv), 1, U32)
48614a46
NC
197#else
198# define SvARENA_CHAIN(sv) SvANY(sv)
199# define POSION_SV_HEAD(sv)
200#endif
201
053fc874
GS
202#define plant_SV(p) \
203 STMT_START { \
fd0854ff 204 FREE_SV_DEBUG_FILE(p); \
48614a46
NC
205 POSION_SV_HEAD(p); \
206 SvARENA_CHAIN(p) = (void *)PL_sv_root; \
053fc874
GS
207 SvFLAGS(p) = SVTYPEMASK; \
208 PL_sv_root = (p); \
209 --PL_sv_count; \
210 } STMT_END
a0d0e21e 211
fba3b22e 212/* sv_mutex must be held while calling uproot_SV() */
053fc874
GS
213#define uproot_SV(p) \
214 STMT_START { \
215 (p) = PL_sv_root; \
bb7bbd9c 216 PL_sv_root = (SV*)SvARENA_CHAIN(p); \
053fc874
GS
217 ++PL_sv_count; \
218 } STMT_END
219
645c22ef 220
cac9b346
NC
221/* make some more SVs by adding another arena */
222
223/* sv_mutex must be held while calling more_sv() */
224STATIC SV*
225S_more_sv(pTHX)
226{
97aff369 227 dVAR;
cac9b346
NC
228 SV* sv;
229
230 if (PL_nice_chunk) {
231 sv_add_arena(PL_nice_chunk, PL_nice_chunk_size, 0);
bd61b366 232 PL_nice_chunk = NULL;
cac9b346
NC
233 PL_nice_chunk_size = 0;
234 }
235 else {
236 char *chunk; /* must use New here to match call to */
d2a0f284 237 Newx(chunk,PERL_ARENA_SIZE,char); /* Safefree() in sv_free_arenas() */
2e7ed132 238 sv_add_arena(chunk, PERL_ARENA_SIZE, 0);
cac9b346
NC
239 }
240 uproot_SV(sv);
241 return sv;
242}
243
645c22ef
DM
244/* new_SV(): return a new, empty SV head */
245
eba0f806
DM
246#ifdef DEBUG_LEAKING_SCALARS
247/* provide a real function for a debugger to play with */
248STATIC SV*
249S_new_SV(pTHX)
250{
251 SV* sv;
252
253 LOCK_SV_MUTEX;
254 if (PL_sv_root)
255 uproot_SV(sv);
256 else
cac9b346 257 sv = S_more_sv(aTHX);
eba0f806
DM
258 UNLOCK_SV_MUTEX;
259 SvANY(sv) = 0;
260 SvREFCNT(sv) = 1;
261 SvFLAGS(sv) = 0;
fd0854ff
DM
262 sv->sv_debug_optype = PL_op ? PL_op->op_type : 0;
263 sv->sv_debug_line = (U16) ((PL_copline == NOLINE) ?
264 (PL_curcop ? CopLINE(PL_curcop) : 0) : PL_copline);
265 sv->sv_debug_inpad = 0;
266 sv->sv_debug_cloned = 0;
fd0854ff 267 sv->sv_debug_file = PL_curcop ? savepv(CopFILE(PL_curcop)): NULL;
fd0854ff 268
eba0f806
DM
269 return sv;
270}
271# define new_SV(p) (p)=S_new_SV(aTHX)
272
273#else
274# define new_SV(p) \
053fc874
GS
275 STMT_START { \
276 LOCK_SV_MUTEX; \
277 if (PL_sv_root) \
278 uproot_SV(p); \
279 else \
cac9b346 280 (p) = S_more_sv(aTHX); \
053fc874
GS
281 UNLOCK_SV_MUTEX; \
282 SvANY(p) = 0; \
283 SvREFCNT(p) = 1; \
284 SvFLAGS(p) = 0; \
285 } STMT_END
eba0f806 286#endif
463ee0b2 287
645c22ef
DM
288
289/* del_SV(): return an empty SV head to the free list */
290
a0d0e21e 291#ifdef DEBUGGING
4561caa4 292
053fc874
GS
293#define del_SV(p) \
294 STMT_START { \
295 LOCK_SV_MUTEX; \
aea4f609 296 if (DEBUG_D_TEST) \
053fc874
GS
297 del_sv(p); \
298 else \
299 plant_SV(p); \
300 UNLOCK_SV_MUTEX; \
301 } STMT_END
a0d0e21e 302
76e3520e 303STATIC void
cea2e8a9 304S_del_sv(pTHX_ SV *p)
463ee0b2 305{
97aff369 306 dVAR;
aea4f609 307 if (DEBUG_D_TEST) {
4633a7c4 308 SV* sva;
a3b680e6 309 bool ok = 0;
3280af22 310 for (sva = PL_sv_arenaroot; sva; sva = (SV *) SvANY(sva)) {
53c1dcc0
AL
311 const SV * const sv = sva + 1;
312 const SV * const svend = &sva[SvREFCNT(sva)];
c0ff570e 313 if (p >= sv && p < svend) {
a0d0e21e 314 ok = 1;
c0ff570e
NC
315 break;
316 }
a0d0e21e
LW
317 }
318 if (!ok) {
0453d815 319 if (ckWARN_d(WARN_INTERNAL))
9014280d 320 Perl_warner(aTHX_ packWARN(WARN_INTERNAL),
472d47bc
SB
321 "Attempt to free non-arena SV: 0x%"UVxf
322 pTHX__FORMAT, PTR2UV(p) pTHX__VALUE);
a0d0e21e
LW
323 return;
324 }
325 }
4561caa4 326 plant_SV(p);
463ee0b2 327}
a0d0e21e 328
4561caa4
CS
329#else /* ! DEBUGGING */
330
331#define del_SV(p) plant_SV(p)
332
333#endif /* DEBUGGING */
463ee0b2 334
645c22ef
DM
335
336/*
ccfc67b7
JH
337=head1 SV Manipulation Functions
338
645c22ef
DM
339=for apidoc sv_add_arena
340
341Given a chunk of memory, link it to the head of the list of arenas,
342and split it into a list of free SVs.
343
344=cut
345*/
346
4633a7c4 347void
864dbfa3 348Perl_sv_add_arena(pTHX_ char *ptr, U32 size, U32 flags)
463ee0b2 349{
97aff369 350 dVAR;
0bd48802 351 SV* const sva = (SV*)ptr;
463ee0b2
LW
352 register SV* sv;
353 register SV* svend;
4633a7c4
LW
354
355 /* The first SV in an arena isn't an SV. */
3280af22 356 SvANY(sva) = (void *) PL_sv_arenaroot; /* ptr to next arena */
4633a7c4
LW
357 SvREFCNT(sva) = size / sizeof(SV); /* number of SV slots */
358 SvFLAGS(sva) = flags; /* FAKE if not to be freed */
359
3280af22
NIS
360 PL_sv_arenaroot = sva;
361 PL_sv_root = sva + 1;
4633a7c4
LW
362
363 svend = &sva[SvREFCNT(sva) - 1];
364 sv = sva + 1;
463ee0b2 365 while (sv < svend) {
48614a46 366 SvARENA_CHAIN(sv) = (void *)(SV*)(sv + 1);
03e36789 367#ifdef DEBUGGING
978b032e 368 SvREFCNT(sv) = 0;
03e36789
NC
369#endif
370 /* Must always set typemask because it's awlays checked in on cleanup
371 when the arenas are walked looking for objects. */
8990e307 372 SvFLAGS(sv) = SVTYPEMASK;
463ee0b2
LW
373 sv++;
374 }
48614a46 375 SvARENA_CHAIN(sv) = 0;
03e36789
NC
376#ifdef DEBUGGING
377 SvREFCNT(sv) = 0;
378#endif
4633a7c4
LW
379 SvFLAGS(sv) = SVTYPEMASK;
380}
381
055972dc
DM
382/* visit(): call the named function for each non-free SV in the arenas
383 * whose flags field matches the flags/mask args. */
645c22ef 384
5226ed68 385STATIC I32
055972dc 386S_visit(pTHX_ SVFUNC_t f, U32 flags, U32 mask)
8990e307 387{
97aff369 388 dVAR;
4633a7c4 389 SV* sva;
5226ed68 390 I32 visited = 0;
8990e307 391
3280af22 392 for (sva = PL_sv_arenaroot; sva; sva = (SV*)SvANY(sva)) {
53c1dcc0 393 register const SV * const svend = &sva[SvREFCNT(sva)];
a3b680e6 394 register SV* sv;
4561caa4 395 for (sv = sva + 1; sv < svend; ++sv) {
055972dc
DM
396 if (SvTYPE(sv) != SVTYPEMASK
397 && (sv->sv_flags & mask) == flags
398 && SvREFCNT(sv))
399 {
acfe0abc 400 (FCALL)(aTHX_ sv);
5226ed68
JH
401 ++visited;
402 }
8990e307
LW
403 }
404 }
5226ed68 405 return visited;
8990e307
LW
406}
407
758a08c3
JH
408#ifdef DEBUGGING
409
645c22ef
DM
410/* called by sv_report_used() for each live SV */
411
412static void
acfe0abc 413do_report_used(pTHX_ SV *sv)
645c22ef
DM
414{
415 if (SvTYPE(sv) != SVTYPEMASK) {
416 PerlIO_printf(Perl_debug_log, "****\n");
417 sv_dump(sv);
418 }
419}
758a08c3 420#endif
645c22ef
DM
421
422/*
423=for apidoc sv_report_used
424
425Dump the contents of all SVs not yet freed. (Debugging aid).
426
427=cut
428*/
429
8990e307 430void
864dbfa3 431Perl_sv_report_used(pTHX)
4561caa4 432{
ff270d3a 433#ifdef DEBUGGING
055972dc 434 visit(do_report_used, 0, 0);
96a5add6
AL
435#else
436 PERL_UNUSED_CONTEXT;
ff270d3a 437#endif
4561caa4
CS
438}
439
645c22ef
DM
440/* called by sv_clean_objs() for each live SV */
441
442static void
e15faf7d 443do_clean_objs(pTHX_ SV *ref)
645c22ef 444{
97aff369 445 dVAR;
823a54a3
AL
446 if (SvROK(ref)) {
447 SV * const target = SvRV(ref);
448 if (SvOBJECT(target)) {
449 DEBUG_D((PerlIO_printf(Perl_debug_log, "Cleaning object ref:\n "), sv_dump(ref)));
450 if (SvWEAKREF(ref)) {
451 sv_del_backref(target, ref);
452 SvWEAKREF_off(ref);
453 SvRV_set(ref, NULL);
454 } else {
455 SvROK_off(ref);
456 SvRV_set(ref, NULL);
457 SvREFCNT_dec(target);
458 }
645c22ef
DM
459 }
460 }
461
462 /* XXX Might want to check arrays, etc. */
463}
464
465/* called by sv_clean_objs() for each live SV */
466
467#ifndef DISABLE_DESTRUCTOR_KLUDGE
468static void
acfe0abc 469do_clean_named_objs(pTHX_ SV *sv)
645c22ef 470{
97aff369 471 dVAR;
f7877b28 472 if (SvTYPE(sv) == SVt_PVGV && isGV_with_GP(sv) && GvGP(sv)) {
c69033f2
NC
473 if ((
474#ifdef PERL_DONT_CREATE_GVSV
475 GvSV(sv) &&
476#endif
477 SvOBJECT(GvSV(sv))) ||
645c22ef
DM
478 (GvAV(sv) && SvOBJECT(GvAV(sv))) ||
479 (GvHV(sv) && SvOBJECT(GvHV(sv))) ||
480 (GvIO(sv) && SvOBJECT(GvIO(sv))) ||
481 (GvCV(sv) && SvOBJECT(GvCV(sv))) )
482 {
483 DEBUG_D((PerlIO_printf(Perl_debug_log, "Cleaning named glob object:\n "), sv_dump(sv)));
ec5f3c78 484 SvFLAGS(sv) |= SVf_BREAK;
645c22ef
DM
485 SvREFCNT_dec(sv);
486 }
487 }
488}
489#endif
490
491/*
492=for apidoc sv_clean_objs
493
494Attempt to destroy all objects not yet freed
495
496=cut
497*/
498
4561caa4 499void
864dbfa3 500Perl_sv_clean_objs(pTHX)
4561caa4 501{
97aff369 502 dVAR;
3280af22 503 PL_in_clean_objs = TRUE;
055972dc 504 visit(do_clean_objs, SVf_ROK, SVf_ROK);
4561caa4 505#ifndef DISABLE_DESTRUCTOR_KLUDGE
2d0f3c12 506 /* some barnacles may yet remain, clinging to typeglobs */
055972dc 507 visit(do_clean_named_objs, SVt_PVGV, SVTYPEMASK);
4561caa4 508#endif
3280af22 509 PL_in_clean_objs = FALSE;
4561caa4
CS
510}
511
645c22ef
DM
512/* called by sv_clean_all() for each live SV */
513
514static void
acfe0abc 515do_clean_all(pTHX_ SV *sv)
645c22ef 516{
97aff369 517 dVAR;
645c22ef
DM
518 DEBUG_D((PerlIO_printf(Perl_debug_log, "Cleaning loops: SV at 0x%"UVxf"\n", PTR2UV(sv)) ));
519 SvFLAGS(sv) |= SVf_BREAK;
0e705b3b 520 if (PL_comppad == (AV*)sv) {
7d49f689 521 PL_comppad = NULL;
4608196e 522 PL_curpad = NULL;
0e705b3b 523 }
645c22ef
DM
524 SvREFCNT_dec(sv);
525}
526
527/*
528=for apidoc sv_clean_all
529
530Decrement the refcnt of each remaining SV, possibly triggering a
531cleanup. This function may have to be called multiple times to free
ff276b08 532SVs which are in complex self-referential hierarchies.
645c22ef
DM
533
534=cut
535*/
536
5226ed68 537I32
864dbfa3 538Perl_sv_clean_all(pTHX)
8990e307 539{
97aff369 540 dVAR;
5226ed68 541 I32 cleaned;
3280af22 542 PL_in_clean_all = TRUE;
055972dc 543 cleaned = visit(do_clean_all, 0,0);
3280af22 544 PL_in_clean_all = FALSE;
5226ed68 545 return cleaned;
8990e307 546}
463ee0b2 547
5e258f8c
JC
548/*
549 ARENASETS: a meta-arena implementation which separates arena-info
550 into struct arena_set, which contains an array of struct
551 arena_descs, each holding info for a single arena. By separating
552 the meta-info from the arena, we recover the 1st slot, formerly
553 borrowed for list management. The arena_set is about the size of an
554 arena, avoiding the needless malloc overhead of a naive linked-list
555
556 The cost is 1 arena-set malloc per ~320 arena-mallocs, + the unused
557 memory in the last arena-set (1/2 on average). In trade, we get
558 back the 1st slot in each arena (ie 1.7% of a CV-arena, less for
d2a0f284
JC
559 smaller types). The recovery of the wasted space allows use of
560 small arenas for large, rare body types,
5e258f8c 561*/
5e258f8c 562struct arena_desc {
398c677b
NC
563 char *arena; /* the raw storage, allocated aligned */
564 size_t size; /* its size ~4k typ */
565 int unit_type; /* useful for arena audits */
5e258f8c
JC
566 /* info for sv-heads (eventually)
567 int count, flags;
568 */
569};
570
e6148039
NC
571struct arena_set;
572
573/* Get the maximum number of elements in set[] such that struct arena_set
574 will fit within PERL_ARENA_SIZE, which is probabably just under 4K, and
575 therefore likely to be 1 aligned memory page. */
576
577#define ARENAS_PER_SET ((PERL_ARENA_SIZE - sizeof(struct arena_set*) \
578 - 2 * sizeof(int)) / sizeof (struct arena_desc))
5e258f8c
JC
579
580struct arena_set {
581 struct arena_set* next;
582 int set_size; /* ie ARENAS_PER_SET */
583 int curr; /* index of next available arena-desc */
584 struct arena_desc set[ARENAS_PER_SET];
585};
586
645c22ef
DM
587/*
588=for apidoc sv_free_arenas
589
590Deallocate the memory used by all arenas. Note that all the individual SV
591heads and bodies within the arenas must already have been freed.
592
593=cut
594*/
4633a7c4 595void
864dbfa3 596Perl_sv_free_arenas(pTHX)
4633a7c4 597{
97aff369 598 dVAR;
4633a7c4
LW
599 SV* sva;
600 SV* svanext;
93e68bfb 601 int i;
4633a7c4
LW
602
603 /* Free arenas here, but be careful about fake ones. (We assume
604 contiguity of the fake ones with the corresponding real ones.) */
605
3280af22 606 for (sva = PL_sv_arenaroot; sva; sva = svanext) {
4633a7c4
LW
607 svanext = (SV*) SvANY(sva);
608 while (svanext && SvFAKE(svanext))
609 svanext = (SV*) SvANY(svanext);
610
611 if (!SvFAKE(sva))
1df70142 612 Safefree(sva);
4633a7c4 613 }
93e68bfb 614
5e258f8c
JC
615 {
616 struct arena_set *next, *aroot = (struct arena_set*) PL_body_arenas;
617
618 for (; aroot; aroot = next) {
96a5add6 619 const int max = aroot->curr;
5e258f8c
JC
620 for (i=0; i<max; i++) {
621 assert(aroot->set[i].arena);
622 Safefree(aroot->set[i].arena);
623 }
624 next = aroot->next;
625 Safefree(aroot);
626 }
627 }
dc8220bf 628 PL_body_arenas = 0;
fdda85ca 629
232d1c15 630 for (i=0; i<PERL_ARENA_ROOTS_SIZE; i++)
93e68bfb 631 PL_body_roots[i] = 0;
93e68bfb 632
43c5f42d 633 Safefree(PL_nice_chunk);
bd61b366 634 PL_nice_chunk = NULL;
3280af22
NIS
635 PL_nice_chunk_size = 0;
636 PL_sv_arenaroot = 0;
637 PL_sv_root = 0;
4633a7c4
LW
638}
639
bd81e77b
NC
640/*
641 Here are mid-level routines that manage the allocation of bodies out
642 of the various arenas. There are 5 kinds of arenas:
29489e7c 643
bd81e77b
NC
644 1. SV-head arenas, which are discussed and handled above
645 2. regular body arenas
646 3. arenas for reduced-size bodies
647 4. Hash-Entry arenas
648 5. pte arenas (thread related)
29489e7c 649
bd81e77b
NC
650 Arena types 2 & 3 are chained by body-type off an array of
651 arena-root pointers, which is indexed by svtype. Some of the
652 larger/less used body types are malloced singly, since a large
653 unused block of them is wasteful. Also, several svtypes dont have
654 bodies; the data fits into the sv-head itself. The arena-root
655 pointer thus has a few unused root-pointers (which may be hijacked
656 later for arena types 4,5)
29489e7c 657
bd81e77b
NC
658 3 differs from 2 as an optimization; some body types have several
659 unused fields in the front of the structure (which are kept in-place
660 for consistency). These bodies can be allocated in smaller chunks,
661 because the leading fields arent accessed. Pointers to such bodies
662 are decremented to point at the unused 'ghost' memory, knowing that
663 the pointers are used with offsets to the real memory.
29489e7c 664
bd81e77b
NC
665 HE, HEK arenas are managed separately, with separate code, but may
666 be merge-able later..
667
668 PTE arenas are not sv-bodies, but they share these mid-level
669 mechanics, so are considered here. The new mid-level mechanics rely
670 on the sv_type of the body being allocated, so we just reserve one
671 of the unused body-slots for PTEs, then use it in those (2) PTE
672 contexts below (line ~10k)
673*/
674
bd26d9a3 675/* get_arena(size): this creates custom-sized arenas
5e258f8c
JC
676 TBD: export properly for hv.c: S_more_he().
677*/
678void*
679Perl_get_arena(pTHX_ int arena_size)
680{
7a89be66 681 dVAR;
5e258f8c 682 struct arena_desc* adesc;
476a1e16 683 struct arena_set *newroot, **aroot = (struct arena_set**) &PL_body_arenas;
5e258f8c
JC
684 int curr;
685
476a1e16
JC
686 /* shouldnt need this
687 if (!arena_size) arena_size = PERL_ARENA_SIZE;
688 */
5e258f8c
JC
689
690 /* may need new arena-set to hold new arena */
476a1e16 691 if (!*aroot || (*aroot)->curr >= (*aroot)->set_size) {
5e258f8c
JC
692 Newxz(newroot, 1, struct arena_set);
693 newroot->set_size = ARENAS_PER_SET;
476a1e16
JC
694 newroot->next = *aroot;
695 *aroot = newroot;
ca0270c4 696 DEBUG_m(PerlIO_printf(Perl_debug_log, "new arenaset %p\n", (void*)*aroot));
5e258f8c
JC
697 }
698
699 /* ok, now have arena-set with at least 1 empty/available arena-desc */
476a1e16
JC
700 curr = (*aroot)->curr++;
701 adesc = &((*aroot)->set[curr]);
5e258f8c
JC
702 assert(!adesc->arena);
703
5e258f8c
JC
704 Newxz(adesc->arena, arena_size, char);
705 adesc->size = arena_size;
d2a0f284
JC
706 DEBUG_m(PerlIO_printf(Perl_debug_log, "arena %d added: %p size %d\n",
707 curr, adesc->arena, arena_size));
5e258f8c
JC
708
709 return adesc->arena;
5e258f8c
JC
710}
711
53c1dcc0 712
bd81e77b 713/* return a thing to the free list */
29489e7c 714
bd81e77b
NC
715#define del_body(thing, root) \
716 STMT_START { \
00b6aa41 717 void ** const thing_copy = (void **)thing;\
bd81e77b
NC
718 LOCK_SV_MUTEX; \
719 *thing_copy = *root; \
720 *root = (void*)thing_copy; \
721 UNLOCK_SV_MUTEX; \
722 } STMT_END
29489e7c 723
bd81e77b 724/*
d2a0f284
JC
725
726=head1 SV-Body Allocation
727
728Allocation of SV-bodies is similar to SV-heads, differing as follows;
729the allocation mechanism is used for many body types, so is somewhat
730more complicated, it uses arena-sets, and has no need for still-live
731SV detection.
732
733At the outermost level, (new|del)_X*V macros return bodies of the
734appropriate type. These macros call either (new|del)_body_type or
735(new|del)_body_allocated macro pairs, depending on specifics of the
736type. Most body types use the former pair, the latter pair is used to
737allocate body types with "ghost fields".
738
739"ghost fields" are fields that are unused in certain types, and
740consequently dont need to actually exist. They are declared because
741they're part of a "base type", which allows use of functions as
742methods. The simplest examples are AVs and HVs, 2 aggregate types
743which don't use the fields which support SCALAR semantics.
744
745For these types, the arenas are carved up into *_allocated size
746chunks, we thus avoid wasted memory for those unaccessed members.
747When bodies are allocated, we adjust the pointer back in memory by the
748size of the bit not allocated, so it's as if we allocated the full
749structure. (But things will all go boom if you write to the part that
750is "not there", because you'll be overwriting the last members of the
751preceding structure in memory.)
752
753We calculate the correction using the STRUCT_OFFSET macro. For
754example, if xpv_allocated is the same structure as XPV then the two
755OFFSETs sum to zero, and the pointer is unchanged. If the allocated
756structure is smaller (no initial NV actually allocated) then the net
757effect is to subtract the size of the NV from the pointer, to return a
758new pointer as if an initial NV were actually allocated.
759
760This is the same trick as was used for NV and IV bodies. Ironically it
761doesn't need to be used for NV bodies any more, because NV is now at
762the start of the structure. IV bodies don't need it either, because
763they are no longer allocated.
764
765In turn, the new_body_* allocators call S_new_body(), which invokes
766new_body_inline macro, which takes a lock, and takes a body off the
767linked list at PL_body_roots[sv_type], calling S_more_bodies() if
768necessary to refresh an empty list. Then the lock is released, and
769the body is returned.
770
771S_more_bodies calls get_arena(), and carves it up into an array of N
772bodies, which it strings into a linked list. It looks up arena-size
773and body-size from the body_details table described below, thus
774supporting the multiple body-types.
775
776If PURIFY is defined, or PERL_ARENA_SIZE=0, arenas are not used, and
777the (new|del)_X*V macros are mapped directly to malloc/free.
778
779*/
780
781/*
782
783For each sv-type, struct body_details bodies_by_type[] carries
784parameters which control these aspects of SV handling:
785
786Arena_size determines whether arenas are used for this body type, and if
787so, how big they are. PURIFY or PERL_ARENA_SIZE=0 set this field to
788zero, forcing individual mallocs and frees.
789
790Body_size determines how big a body is, and therefore how many fit into
791each arena. Offset carries the body-pointer adjustment needed for
792*_allocated body types, and is used in *_allocated macros.
793
794But its main purpose is to parameterize info needed in
795Perl_sv_upgrade(). The info here dramatically simplifies the function
796vs the implementation in 5.8.7, making it table-driven. All fields
797are used for this, except for arena_size.
798
799For the sv-types that have no bodies, arenas are not used, so those
800PL_body_roots[sv_type] are unused, and can be overloaded. In
801something of a special case, SVt_NULL is borrowed for HE arenas;
802PL_body_roots[SVt_NULL] is filled by S_more_he, but the
803bodies_by_type[SVt_NULL] slot is not used, as the table is not
804available in hv.c,
805
806PTEs also use arenas, but are never seen in Perl_sv_upgrade.
807Nonetheless, they get their own slot in bodies_by_type[SVt_NULL], so
808they can just use the same allocation semantics. At first, PTEs were
809also overloaded to a non-body sv-type, but this yielded hard-to-find
810malloc bugs, so was simplified by claiming a new slot. This choice
811has no consequence at this time.
812
29489e7c
DM
813*/
814
bd81e77b 815struct body_details {
0fb58b32 816 U8 body_size; /* Size to allocate */
10666ae3 817 U8 copy; /* Size of structure to copy (may be shorter) */
0fb58b32 818 U8 offset;
10666ae3
NC
819 unsigned int type : 4; /* We have space for a sanity check. */
820 unsigned int cant_upgrade : 1; /* Cannot upgrade this type */
821 unsigned int zero_nv : 1; /* zero the NV when upgrading from this */
822 unsigned int arena : 1; /* Allocated from an arena */
823 size_t arena_size; /* Size of arena to allocate */
bd81e77b 824};
29489e7c 825
bd81e77b
NC
826#define HADNV FALSE
827#define NONV TRUE
29489e7c 828
d2a0f284 829
bd81e77b
NC
830#ifdef PURIFY
831/* With -DPURFIY we allocate everything directly, and don't use arenas.
832 This seems a rather elegant way to simplify some of the code below. */
833#define HASARENA FALSE
834#else
835#define HASARENA TRUE
836#endif
837#define NOARENA FALSE
29489e7c 838
d2a0f284
JC
839/* Size the arenas to exactly fit a given number of bodies. A count
840 of 0 fits the max number bodies into a PERL_ARENA_SIZE.block,
841 simplifying the default. If count > 0, the arena is sized to fit
842 only that many bodies, allowing arenas to be used for large, rare
843 bodies (XPVFM, XPVIO) without undue waste. The arena size is
844 limited by PERL_ARENA_SIZE, so we can safely oversize the
845 declarations.
846 */
95db5f15
MB
847#define FIT_ARENA0(body_size) \
848 ((size_t)(PERL_ARENA_SIZE / body_size) * body_size)
849#define FIT_ARENAn(count,body_size) \
850 ( count * body_size <= PERL_ARENA_SIZE) \
851 ? count * body_size \
852 : FIT_ARENA0 (body_size)
853#define FIT_ARENA(count,body_size) \
854 count \
855 ? FIT_ARENAn (count, body_size) \
856 : FIT_ARENA0 (body_size)
d2a0f284 857
bd81e77b 858/* A macro to work out the offset needed to subtract from a pointer to (say)
29489e7c 859
bd81e77b
NC
860typedef struct {
861 STRLEN xpv_cur;
862 STRLEN xpv_len;
863} xpv_allocated;
29489e7c 864
bd81e77b 865to make its members accessible via a pointer to (say)
29489e7c 866
bd81e77b
NC
867struct xpv {
868 NV xnv_nv;
869 STRLEN xpv_cur;
870 STRLEN xpv_len;
871};
29489e7c 872
bd81e77b 873*/
29489e7c 874
bd81e77b
NC
875#define relative_STRUCT_OFFSET(longer, shorter, member) \
876 (STRUCT_OFFSET(shorter, member) - STRUCT_OFFSET(longer, member))
29489e7c 877
bd81e77b
NC
878/* Calculate the length to copy. Specifically work out the length less any
879 final padding the compiler needed to add. See the comment in sv_upgrade
880 for why copying the padding proved to be a bug. */
29489e7c 881
bd81e77b
NC
882#define copy_length(type, last_member) \
883 STRUCT_OFFSET(type, last_member) \
884 + sizeof (((type*)SvANY((SV*)0))->last_member)
29489e7c 885
bd81e77b 886static const struct body_details bodies_by_type[] = {
10666ae3
NC
887 { sizeof(HE), 0, 0, SVt_NULL,
888 FALSE, NONV, NOARENA, FIT_ARENA(0, sizeof(HE)) },
d2a0f284
JC
889
890 /* IVs are in the head, so the allocation size is 0.
891 However, the slot is overloaded for PTEs. */
892 { sizeof(struct ptr_tbl_ent), /* This is used for PTEs. */
893 sizeof(IV), /* This is used to copy out the IV body. */
10666ae3 894 STRUCT_OFFSET(XPVIV, xiv_iv), SVt_IV, FALSE, NONV,
d2a0f284
JC
895 NOARENA /* IVS don't need an arena */,
896 /* But PTEs need to know the size of their arena */
897 FIT_ARENA(0, sizeof(struct ptr_tbl_ent))
898 },
899
bd81e77b 900 /* 8 bytes on most ILP32 with IEEE doubles */
10666ae3 901 { sizeof(NV), sizeof(NV), 0, SVt_NV, FALSE, HADNV, HASARENA,
d2a0f284
JC
902 FIT_ARENA(0, sizeof(NV)) },
903
904 /* RVs are in the head now. */
10666ae3 905 { 0, 0, 0, SVt_RV, FALSE, NONV, NOARENA, 0 },
d2a0f284 906
bd81e77b 907 /* 8 bytes on most ILP32 with IEEE doubles */
d2a0f284
JC
908 { sizeof(xpv_allocated),
909 copy_length(XPV, xpv_len)
910 - relative_STRUCT_OFFSET(xpv_allocated, XPV, xpv_cur),
911 + relative_STRUCT_OFFSET(xpv_allocated, XPV, xpv_cur),
10666ae3 912 SVt_PV, FALSE, NONV, HASARENA, FIT_ARENA(0, sizeof(xpv_allocated)) },
d2a0f284 913
bd81e77b 914 /* 12 */
d2a0f284
JC
915 { sizeof(xpviv_allocated),
916 copy_length(XPVIV, xiv_u)
917 - relative_STRUCT_OFFSET(xpviv_allocated, XPVIV, xpv_cur),
918 + relative_STRUCT_OFFSET(xpviv_allocated, XPVIV, xpv_cur),
10666ae3 919 SVt_PVIV, FALSE, NONV, HASARENA, FIT_ARENA(0, sizeof(xpviv_allocated)) },
d2a0f284 920
bd81e77b 921 /* 20 */
10666ae3 922 { sizeof(XPVNV), copy_length(XPVNV, xiv_u), 0, SVt_PVNV, FALSE, HADNV,
d2a0f284
JC
923 HASARENA, FIT_ARENA(0, sizeof(XPVNV)) },
924
bd81e77b 925 /* 28 */
10666ae3 926 { sizeof(XPVMG), copy_length(XPVMG, xmg_stash), 0, SVt_PVMG, FALSE, HADNV,
d2a0f284
JC
927 HASARENA, FIT_ARENA(0, sizeof(XPVMG)) },
928
bd81e77b 929 /* 36 */
10666ae3 930 { sizeof(XPVBM), sizeof(XPVBM), 0, SVt_PVBM, TRUE, HADNV,
d2a0f284
JC
931 HASARENA, FIT_ARENA(0, sizeof(XPVBM)) },
932
bd81e77b 933 /* 48 */
10666ae3 934 { sizeof(XPVGV), sizeof(XPVGV), 0, SVt_PVGV, TRUE, HADNV,
d2a0f284
JC
935 HASARENA, FIT_ARENA(0, sizeof(XPVGV)) },
936
bd81e77b 937 /* 64 */
10666ae3 938 { sizeof(XPVLV), sizeof(XPVLV), 0, SVt_PVLV, TRUE, HADNV,
d2a0f284
JC
939 HASARENA, FIT_ARENA(0, sizeof(XPVLV)) },
940
941 { sizeof(xpvav_allocated),
942 copy_length(XPVAV, xmg_stash)
943 - relative_STRUCT_OFFSET(xpvav_allocated, XPVAV, xav_fill),
944 + relative_STRUCT_OFFSET(xpvav_allocated, XPVAV, xav_fill),
10666ae3 945 SVt_PVAV, TRUE, HADNV, HASARENA, FIT_ARENA(0, sizeof(xpvav_allocated)) },
d2a0f284
JC
946
947 { sizeof(xpvhv_allocated),
948 copy_length(XPVHV, xmg_stash)
949 - relative_STRUCT_OFFSET(xpvhv_allocated, XPVHV, xhv_fill),
950 + relative_STRUCT_OFFSET(xpvhv_allocated, XPVHV, xhv_fill),
10666ae3 951 SVt_PVHV, TRUE, HADNV, HASARENA, FIT_ARENA(0, sizeof(xpvhv_allocated)) },
d2a0f284 952
c84c4652 953 /* 56 */
4115f141 954 { sizeof(xpvcv_allocated), sizeof(xpvcv_allocated),
c84c4652 955 + relative_STRUCT_OFFSET(xpvcv_allocated, XPVCV, xpv_cur),
10666ae3 956 SVt_PVCV, TRUE, NONV, HASARENA, FIT_ARENA(0, sizeof(xpvcv_allocated)) },
d2a0f284 957
4115f141 958 { sizeof(xpvfm_allocated), sizeof(xpvfm_allocated),
3038937b 959 + relative_STRUCT_OFFSET(xpvfm_allocated, XPVFM, xpv_cur),
10666ae3 960 SVt_PVFM, TRUE, NONV, NOARENA, FIT_ARENA(20, sizeof(xpvfm_allocated)) },
d2a0f284
JC
961
962 /* XPVIO is 84 bytes, fits 48x */
10666ae3 963 { sizeof(XPVIO), sizeof(XPVIO), 0, SVt_PVIO, TRUE, HADNV,
d2a0f284 964 HASARENA, FIT_ARENA(24, sizeof(XPVIO)) },
bd81e77b 965};
29489e7c 966
d2a0f284
JC
967#define new_body_type(sv_type) \
968 (void *)((char *)S_new_body(aTHX_ sv_type))
29489e7c 969
bd81e77b
NC
970#define del_body_type(p, sv_type) \
971 del_body(p, &PL_body_roots[sv_type])
29489e7c 972
29489e7c 973
bd81e77b 974#define new_body_allocated(sv_type) \
d2a0f284 975 (void *)((char *)S_new_body(aTHX_ sv_type) \
bd81e77b 976 - bodies_by_type[sv_type].offset)
29489e7c 977
bd81e77b
NC
978#define del_body_allocated(p, sv_type) \
979 del_body(p + bodies_by_type[sv_type].offset, &PL_body_roots[sv_type])
29489e7c 980
29489e7c 981
bd81e77b
NC
982#define my_safemalloc(s) (void*)safemalloc(s)
983#define my_safecalloc(s) (void*)safecalloc(s, 1)
984#define my_safefree(p) safefree((char*)p)
29489e7c 985
bd81e77b 986#ifdef PURIFY
29489e7c 987
bd81e77b
NC
988#define new_XNV() my_safemalloc(sizeof(XPVNV))
989#define del_XNV(p) my_safefree(p)
29489e7c 990
bd81e77b
NC
991#define new_XPVNV() my_safemalloc(sizeof(XPVNV))
992#define del_XPVNV(p) my_safefree(p)
29489e7c 993
bd81e77b
NC
994#define new_XPVAV() my_safemalloc(sizeof(XPVAV))
995#define del_XPVAV(p) my_safefree(p)
29489e7c 996
bd81e77b
NC
997#define new_XPVHV() my_safemalloc(sizeof(XPVHV))
998#define del_XPVHV(p) my_safefree(p)
29489e7c 999
bd81e77b
NC
1000#define new_XPVMG() my_safemalloc(sizeof(XPVMG))
1001#define del_XPVMG(p) my_safefree(p)
29489e7c 1002
bd81e77b
NC
1003#define new_XPVGV() my_safemalloc(sizeof(XPVGV))
1004#define del_XPVGV(p) my_safefree(p)
29489e7c 1005
bd81e77b 1006#else /* !PURIFY */
29489e7c 1007
bd81e77b
NC
1008#define new_XNV() new_body_type(SVt_NV)
1009#define del_XNV(p) del_body_type(p, SVt_NV)
29489e7c 1010
bd81e77b
NC
1011#define new_XPVNV() new_body_type(SVt_PVNV)
1012#define del_XPVNV(p) del_body_type(p, SVt_PVNV)
29489e7c 1013
bd81e77b
NC
1014#define new_XPVAV() new_body_allocated(SVt_PVAV)
1015#define del_XPVAV(p) del_body_allocated(p, SVt_PVAV)
645c22ef 1016
bd81e77b
NC
1017#define new_XPVHV() new_body_allocated(SVt_PVHV)
1018#define del_XPVHV(p) del_body_allocated(p, SVt_PVHV)
645c22ef 1019
bd81e77b
NC
1020#define new_XPVMG() new_body_type(SVt_PVMG)
1021#define del_XPVMG(p) del_body_type(p, SVt_PVMG)
645c22ef 1022
bd81e77b
NC
1023#define new_XPVGV() new_body_type(SVt_PVGV)
1024#define del_XPVGV(p) del_body_type(p, SVt_PVGV)
1d7c1841 1025
bd81e77b 1026#endif /* PURIFY */
93e68bfb 1027
bd81e77b 1028/* no arena for you! */
93e68bfb 1029
bd81e77b 1030#define new_NOARENA(details) \
d2a0f284 1031 my_safemalloc((details)->body_size + (details)->offset)
bd81e77b 1032#define new_NOARENAZ(details) \
d2a0f284
JC
1033 my_safecalloc((details)->body_size + (details)->offset)
1034
0b2d3faa 1035#if defined(DEBUGGING) && !defined(PERL_GLOBAL_STRUCT_PRIVATE)
10666ae3
NC
1036static bool done_sanity_check;
1037#endif
1038
d2a0f284
JC
1039STATIC void *
1040S_more_bodies (pTHX_ svtype sv_type)
1041{
1042 dVAR;
1043 void ** const root = &PL_body_roots[sv_type];
96a5add6 1044 const struct body_details * const bdp = &bodies_by_type[sv_type];
d2a0f284
JC
1045 const size_t body_size = bdp->body_size;
1046 char *start;
1047 const char *end;
1048
1049 assert(bdp->arena_size);
10666ae3 1050
0b2d3faa
JH
1051#if defined(DEBUGGING) && !defined(PERL_GLOBAL_STRUCT_PRIVATE)
1052 /* PERL_GLOBAL_STRUCT_PRIVATE cannot coexist with global
1053 * variables like done_sanity_check. */
10666ae3 1054 if (!done_sanity_check) {
ea471437 1055 unsigned int i = SVt_LAST;
10666ae3
NC
1056
1057 done_sanity_check = TRUE;
1058
1059 while (i--)
1060 assert (bodies_by_type[i].type == i);
1061 }
1062#endif
1063
d2a0f284
JC
1064 start = (char*) Perl_get_arena(aTHX_ bdp->arena_size);
1065
1066 end = start + bdp->arena_size - body_size;
1067
d2a0f284
JC
1068 /* computed count doesnt reflect the 1st slot reservation */
1069 DEBUG_m(PerlIO_printf(Perl_debug_log,
1070 "arena %p end %p arena-size %d type %d size %d ct %d\n",
0e84aef4
JH
1071 start, end,
1072 (int)bdp->arena_size, sv_type, (int)body_size,
1073 (int)bdp->arena_size / (int)body_size));
d2a0f284
JC
1074
1075 *root = (void *)start;
1076
1077 while (start < end) {
1078 char * const next = start + body_size;
1079 *(void**) start = (void *)next;
1080 start = next;
1081 }
1082 *(void **)start = 0;
1083
1084 return *root;
1085}
1086
1087/* grab a new thing from the free list, allocating more if necessary.
1088 The inline version is used for speed in hot routines, and the
1089 function using it serves the rest (unless PURIFY).
1090*/
1091#define new_body_inline(xpv, sv_type) \
1092 STMT_START { \
1093 void ** const r3wt = &PL_body_roots[sv_type]; \
1094 LOCK_SV_MUTEX; \
11b79775
DD
1095 xpv = (PTR_TBL_ENT_t*) (*((void **)(r3wt)) \
1096 ? *((void **)(r3wt)) : more_bodies(sv_type)); \
d2a0f284
JC
1097 *(r3wt) = *(void**)(xpv); \
1098 UNLOCK_SV_MUTEX; \
1099 } STMT_END
1100
1101#ifndef PURIFY
1102
1103STATIC void *
1104S_new_body(pTHX_ svtype sv_type)
1105{
1106 dVAR;
1107 void *xpv;
1108 new_body_inline(xpv, sv_type);
1109 return xpv;
1110}
1111
1112#endif
93e68bfb 1113
bd81e77b
NC
1114/*
1115=for apidoc sv_upgrade
93e68bfb 1116
bd81e77b
NC
1117Upgrade an SV to a more complex form. Generally adds a new body type to the
1118SV, then copies across as much information as possible from the old body.
1119You generally want to use the C<SvUPGRADE> macro wrapper. See also C<svtype>.
93e68bfb 1120
bd81e77b 1121=cut
93e68bfb 1122*/
93e68bfb 1123
bd81e77b 1124void
42d0e0b7 1125Perl_sv_upgrade(pTHX_ register SV *sv, svtype new_type)
cac9b346 1126{
97aff369 1127 dVAR;
bd81e77b
NC
1128 void* old_body;
1129 void* new_body;
42d0e0b7 1130 const svtype old_type = SvTYPE(sv);
d2a0f284 1131 const struct body_details *new_type_details;
bd81e77b
NC
1132 const struct body_details *const old_type_details
1133 = bodies_by_type + old_type;
cac9b346 1134
bd81e77b
NC
1135 if (new_type != SVt_PV && SvIsCOW(sv)) {
1136 sv_force_normal_flags(sv, 0);
1137 }
cac9b346 1138
bd81e77b
NC
1139 if (old_type == new_type)
1140 return;
cac9b346 1141
bd81e77b
NC
1142 if (old_type > new_type)
1143 Perl_croak(aTHX_ "sv_upgrade from type %d down to type %d",
1144 (int)old_type, (int)new_type);
cac9b346 1145
cac9b346 1146
bd81e77b 1147 old_body = SvANY(sv);
de042e1d 1148
bd81e77b
NC
1149 /* Copying structures onto other structures that have been neatly zeroed
1150 has a subtle gotcha. Consider XPVMG
cac9b346 1151
bd81e77b
NC
1152 +------+------+------+------+------+-------+-------+
1153 | NV | CUR | LEN | IV | MAGIC | STASH |
1154 +------+------+------+------+------+-------+-------+
1155 0 4 8 12 16 20 24 28
645c22ef 1156
bd81e77b
NC
1157 where NVs are aligned to 8 bytes, so that sizeof that structure is
1158 actually 32 bytes long, with 4 bytes of padding at the end:
08742458 1159
bd81e77b
NC
1160 +------+------+------+------+------+-------+-------+------+
1161 | NV | CUR | LEN | IV | MAGIC | STASH | ??? |
1162 +------+------+------+------+------+-------+-------+------+
1163 0 4 8 12 16 20 24 28 32
08742458 1164
bd81e77b 1165 so what happens if you allocate memory for this structure:
30f9da9e 1166
bd81e77b
NC
1167 +------+------+------+------+------+-------+-------+------+------+...
1168 | NV | CUR | LEN | IV | MAGIC | STASH | GP | NAME |
1169 +------+------+------+------+------+-------+-------+------+------+...
1170 0 4 8 12 16 20 24 28 32 36
bfc44f79 1171
bd81e77b
NC
1172 zero it, then copy sizeof(XPVMG) bytes on top of it? Not quite what you
1173 expect, because you copy the area marked ??? onto GP. Now, ??? may have
1174 started out as zero once, but it's quite possible that it isn't. So now,
1175 rather than a nicely zeroed GP, you have it pointing somewhere random.
1176 Bugs ensue.
bfc44f79 1177
bd81e77b
NC
1178 (In fact, GP ends up pointing at a previous GP structure, because the
1179 principle cause of the padding in XPVMG getting garbage is a copy of
1180 sizeof(XPVMG) bytes from a XPVGV structure in sv_unglob)
30f9da9e 1181
bd81e77b
NC
1182 So we are careful and work out the size of used parts of all the
1183 structures. */
bfc44f79 1184
bd81e77b
NC
1185 switch (old_type) {
1186 case SVt_NULL:
1187 break;
1188 case SVt_IV:
1189 if (new_type < SVt_PVIV) {
1190 new_type = (new_type == SVt_NV)
1191 ? SVt_PVNV : SVt_PVIV;
bd81e77b
NC
1192 }
1193 break;
1194 case SVt_NV:
1195 if (new_type < SVt_PVNV) {
1196 new_type = SVt_PVNV;
bd81e77b
NC
1197 }
1198 break;
1199 case SVt_RV:
1200 break;
1201 case SVt_PV:
1202 assert(new_type > SVt_PV);
1203 assert(SVt_IV < SVt_PV);
1204 assert(SVt_NV < SVt_PV);
1205 break;
1206 case SVt_PVIV:
1207 break;
1208 case SVt_PVNV:
1209 break;
1210 case SVt_PVMG:
1211 /* Because the XPVMG of PL_mess_sv isn't allocated from the arena,
1212 there's no way that it can be safely upgraded, because perl.c
1213 expects to Safefree(SvANY(PL_mess_sv)) */
1214 assert(sv != PL_mess_sv);
1215 /* This flag bit is used to mean other things in other scalar types.
1216 Given that it only has meaning inside the pad, it shouldn't be set
1217 on anything that can get upgraded. */
00b1698f 1218 assert(!SvPAD_TYPED(sv));
bd81e77b
NC
1219 break;
1220 default:
1221 if (old_type_details->cant_upgrade)
c81225bc
NC
1222 Perl_croak(aTHX_ "Can't upgrade %s (%" UVuf ") to %" UVuf,
1223 sv_reftype(sv, 0), (UV) old_type, (UV) new_type);
bd81e77b 1224 }
2fa1109b 1225 new_type_details = bodies_by_type + new_type;
645c22ef 1226
bd81e77b
NC
1227 SvFLAGS(sv) &= ~SVTYPEMASK;
1228 SvFLAGS(sv) |= new_type;
932e9ff9 1229
ab4416c0
NC
1230 /* This can't happen, as SVt_NULL is <= all values of new_type, so one of
1231 the return statements above will have triggered. */
1232 assert (new_type != SVt_NULL);
bd81e77b 1233 switch (new_type) {
bd81e77b
NC
1234 case SVt_IV:
1235 assert(old_type == SVt_NULL);
1236 SvANY(sv) = (XPVIV*)((char*)&(sv->sv_u.svu_iv) - STRUCT_OFFSET(XPVIV, xiv_iv));
1237 SvIV_set(sv, 0);
1238 return;
1239 case SVt_NV:
1240 assert(old_type == SVt_NULL);
1241 SvANY(sv) = new_XNV();
1242 SvNV_set(sv, 0);
1243 return;
1244 case SVt_RV:
1245 assert(old_type == SVt_NULL);
1246 SvANY(sv) = &sv->sv_u.svu_rv;
1247 SvRV_set(sv, 0);
1248 return;
1249 case SVt_PVHV:
bd81e77b 1250 case SVt_PVAV:
d2a0f284 1251 assert(new_type_details->body_size);
c1ae03ae
NC
1252
1253#ifndef PURIFY
1254 assert(new_type_details->arena);
d2a0f284 1255 assert(new_type_details->arena_size);
c1ae03ae 1256 /* This points to the start of the allocated area. */
d2a0f284
JC
1257 new_body_inline(new_body, new_type);
1258 Zero(new_body, new_type_details->body_size, char);
c1ae03ae
NC
1259 new_body = ((char *)new_body) - new_type_details->offset;
1260#else
1261 /* We always allocated the full length item with PURIFY. To do this
1262 we fake things so that arena is false for all 16 types.. */
1263 new_body = new_NOARENAZ(new_type_details);
1264#endif
1265 SvANY(sv) = new_body;
1266 if (new_type == SVt_PVAV) {
1267 AvMAX(sv) = -1;
1268 AvFILLp(sv) = -1;
1269 AvREAL_only(sv);
1270 }
aeb18a1e 1271
bd81e77b
NC
1272 /* SVt_NULL isn't the only thing upgraded to AV or HV.
1273 The target created by newSVrv also is, and it can have magic.
1274 However, it never has SvPVX set.
1275 */
1276 if (old_type >= SVt_RV) {
1277 assert(SvPVX_const(sv) == 0);
1278 }
aeb18a1e 1279
bd81e77b
NC
1280 /* Could put this in the else clause below, as PVMG must have SvPVX
1281 0 already (the assertion above) */
6136c704 1282 SvPV_set(sv, NULL);
93e68bfb 1283
bd81e77b 1284 if (old_type >= SVt_PVMG) {
e736a858 1285 SvMAGIC_set(sv, ((XPVMG*)old_body)->xmg_u.xmg_magic);
bd81e77b 1286 SvSTASH_set(sv, ((XPVMG*)old_body)->xmg_stash);
bd81e77b
NC
1287 }
1288 break;
93e68bfb 1289
93e68bfb 1290
bd81e77b
NC
1291 case SVt_PVIV:
1292 /* XXX Is this still needed? Was it ever needed? Surely as there is
1293 no route from NV to PVIV, NOK can never be true */
1294 assert(!SvNOKp(sv));
1295 assert(!SvNOK(sv));
1296 case SVt_PVIO:
1297 case SVt_PVFM:
1298 case SVt_PVBM:
1299 case SVt_PVGV:
1300 case SVt_PVCV:
1301 case SVt_PVLV:
1302 case SVt_PVMG:
1303 case SVt_PVNV:
1304 case SVt_PV:
93e68bfb 1305
d2a0f284 1306 assert(new_type_details->body_size);
bd81e77b
NC
1307 /* We always allocated the full length item with PURIFY. To do this
1308 we fake things so that arena is false for all 16 types.. */
1309 if(new_type_details->arena) {
1310 /* This points to the start of the allocated area. */
d2a0f284
JC
1311 new_body_inline(new_body, new_type);
1312 Zero(new_body, new_type_details->body_size, char);
bd81e77b
NC
1313 new_body = ((char *)new_body) - new_type_details->offset;
1314 } else {
1315 new_body = new_NOARENAZ(new_type_details);
1316 }
1317 SvANY(sv) = new_body;
5e2fc214 1318
bd81e77b 1319 if (old_type_details->copy) {
f9ba3d20
NC
1320 /* There is now the potential for an upgrade from something without
1321 an offset (PVNV or PVMG) to something with one (PVCV, PVFM) */
1322 int offset = old_type_details->offset;
1323 int length = old_type_details->copy;
1324
1325 if (new_type_details->offset > old_type_details->offset) {
d4c19fe8 1326 const int difference
f9ba3d20
NC
1327 = new_type_details->offset - old_type_details->offset;
1328 offset += difference;
1329 length -= difference;
1330 }
1331 assert (length >= 0);
1332
1333 Copy((char *)old_body + offset, (char *)new_body + offset, length,
1334 char);
bd81e77b
NC
1335 }
1336
1337#ifndef NV_ZERO_IS_ALLBITS_ZERO
f2524eef 1338 /* If NV 0.0 is stores as all bits 0 then Zero() already creates a
e5ce394c
NC
1339 * correct 0.0 for us. Otherwise, if the old body didn't have an
1340 * NV slot, but the new one does, then we need to initialise the
1341 * freshly created NV slot with whatever the correct bit pattern is
1342 * for 0.0 */
1343 if (old_type_details->zero_nv && !new_type_details->zero_nv)
bd81e77b 1344 SvNV_set(sv, 0);
82048762 1345#endif
5e2fc214 1346
bd81e77b 1347 if (new_type == SVt_PVIO)
f2524eef 1348 IoPAGE_LEN(sv) = 60;
bd81e77b 1349 if (old_type < SVt_RV)
6136c704 1350 SvPV_set(sv, NULL);
bd81e77b
NC
1351 break;
1352 default:
afd78fd5
JH
1353 Perl_croak(aTHX_ "panic: sv_upgrade to unknown type %lu",
1354 (unsigned long)new_type);
bd81e77b 1355 }
73171d91 1356
d2a0f284
JC
1357 if (old_type_details->arena) {
1358 /* If there was an old body, then we need to free it.
1359 Note that there is an assumption that all bodies of types that
1360 can be upgraded came from arenas. Only the more complex non-
1361 upgradable types are allowed to be directly malloc()ed. */
bd81e77b
NC
1362#ifdef PURIFY
1363 my_safefree(old_body);
1364#else
1365 del_body((void*)((char*)old_body + old_type_details->offset),
1366 &PL_body_roots[old_type]);
1367#endif
1368 }
1369}
73171d91 1370
bd81e77b
NC
1371/*
1372=for apidoc sv_backoff
73171d91 1373
bd81e77b
NC
1374Remove any string offset. You should normally use the C<SvOOK_off> macro
1375wrapper instead.
73171d91 1376
bd81e77b 1377=cut
73171d91
NC
1378*/
1379
bd81e77b
NC
1380int
1381Perl_sv_backoff(pTHX_ register SV *sv)
1382{
96a5add6 1383 PERL_UNUSED_CONTEXT;
bd81e77b
NC
1384 assert(SvOOK(sv));
1385 assert(SvTYPE(sv) != SVt_PVHV);
1386 assert(SvTYPE(sv) != SVt_PVAV);
1387 if (SvIVX(sv)) {
1388 const char * const s = SvPVX_const(sv);
1389 SvLEN_set(sv, SvLEN(sv) + SvIVX(sv));
1390 SvPV_set(sv, SvPVX(sv) - SvIVX(sv));
1391 SvIV_set(sv, 0);
1392 Move(s, SvPVX(sv), SvCUR(sv)+1, char);
1393 }
1394 SvFLAGS(sv) &= ~SVf_OOK;
1395 return 0;
1396}
73171d91 1397
bd81e77b
NC
1398/*
1399=for apidoc sv_grow
73171d91 1400
bd81e77b
NC
1401Expands the character buffer in the SV. If necessary, uses C<sv_unref> and
1402upgrades the SV to C<SVt_PV>. Returns a pointer to the character buffer.
1403Use the C<SvGROW> wrapper instead.
93e68bfb 1404
bd81e77b
NC
1405=cut
1406*/
93e68bfb 1407
bd81e77b
NC
1408char *
1409Perl_sv_grow(pTHX_ register SV *sv, register STRLEN newlen)
1410{
1411 register char *s;
93e68bfb 1412
5db06880
NC
1413 if (PL_madskills && newlen >= 0x100000) {
1414 PerlIO_printf(Perl_debug_log,
1415 "Allocation too large: %"UVxf"\n", (UV)newlen);
1416 }
bd81e77b
NC
1417#ifdef HAS_64K_LIMIT
1418 if (newlen >= 0x10000) {
1419 PerlIO_printf(Perl_debug_log,
1420 "Allocation too large: %"UVxf"\n", (UV)newlen);
1421 my_exit(1);
1422 }
1423#endif /* HAS_64K_LIMIT */
1424 if (SvROK(sv))
1425 sv_unref(sv);
1426 if (SvTYPE(sv) < SVt_PV) {
1427 sv_upgrade(sv, SVt_PV);
1428 s = SvPVX_mutable(sv);
1429 }
1430 else if (SvOOK(sv)) { /* pv is offset? */
1431 sv_backoff(sv);
1432 s = SvPVX_mutable(sv);
1433 if (newlen > SvLEN(sv))
1434 newlen += 10 * (newlen - SvCUR(sv)); /* avoid copy each time */
1435#ifdef HAS_64K_LIMIT
1436 if (newlen >= 0x10000)
1437 newlen = 0xFFFF;
1438#endif
1439 }
1440 else
1441 s = SvPVX_mutable(sv);
aeb18a1e 1442
bd81e77b
NC
1443 if (newlen > SvLEN(sv)) { /* need more room? */
1444 newlen = PERL_STRLEN_ROUNDUP(newlen);
1445 if (SvLEN(sv) && s) {
1446#ifdef MYMALLOC
1447 const STRLEN l = malloced_size((void*)SvPVX_const(sv));
1448 if (newlen <= l) {
1449 SvLEN_set(sv, l);
1450 return s;
1451 } else
1452#endif
10edeb5d 1453 s = (char*)saferealloc(s, newlen);
bd81e77b
NC
1454 }
1455 else {
10edeb5d 1456 s = (char*)safemalloc(newlen);
bd81e77b
NC
1457 if (SvPVX_const(sv) && SvCUR(sv)) {
1458 Move(SvPVX_const(sv), s, (newlen < SvCUR(sv)) ? newlen : SvCUR(sv), char);
1459 }
1460 }
1461 SvPV_set(sv, s);
1462 SvLEN_set(sv, newlen);
1463 }
1464 return s;
1465}
aeb18a1e 1466
bd81e77b
NC
1467/*
1468=for apidoc sv_setiv
932e9ff9 1469
bd81e77b
NC
1470Copies an integer into the given SV, upgrading first if necessary.
1471Does not handle 'set' magic. See also C<sv_setiv_mg>.
463ee0b2 1472
bd81e77b
NC
1473=cut
1474*/
463ee0b2 1475
bd81e77b
NC
1476void
1477Perl_sv_setiv(pTHX_ register SV *sv, IV i)
1478{
97aff369 1479 dVAR;
bd81e77b
NC
1480 SV_CHECK_THINKFIRST_COW_DROP(sv);
1481 switch (SvTYPE(sv)) {
1482 case SVt_NULL:
1483 sv_upgrade(sv, SVt_IV);
1484 break;
1485 case SVt_NV:
1486 sv_upgrade(sv, SVt_PVNV);
1487 break;
1488 case SVt_RV:
1489 case SVt_PV:
1490 sv_upgrade(sv, SVt_PVIV);
1491 break;
463ee0b2 1492
bd81e77b
NC
1493 case SVt_PVGV:
1494 case SVt_PVAV:
1495 case SVt_PVHV:
1496 case SVt_PVCV:
1497 case SVt_PVFM:
1498 case SVt_PVIO:
1499 Perl_croak(aTHX_ "Can't coerce %s to integer in %s", sv_reftype(sv,0),
1500 OP_DESC(PL_op));
42d0e0b7 1501 default: NOOP;
bd81e77b
NC
1502 }
1503 (void)SvIOK_only(sv); /* validate number */
1504 SvIV_set(sv, i);
1505 SvTAINT(sv);
1506}
932e9ff9 1507
bd81e77b
NC
1508/*
1509=for apidoc sv_setiv_mg
d33b2eba 1510
bd81e77b 1511Like C<sv_setiv>, but also handles 'set' magic.
1c846c1f 1512
bd81e77b
NC
1513=cut
1514*/
d33b2eba 1515
bd81e77b
NC
1516void
1517Perl_sv_setiv_mg(pTHX_ register SV *sv, IV i)
1518{
1519 sv_setiv(sv,i);
1520 SvSETMAGIC(sv);
1521}
727879eb 1522
bd81e77b
NC
1523/*
1524=for apidoc sv_setuv
d33b2eba 1525
bd81e77b
NC
1526Copies an unsigned integer into the given SV, upgrading first if necessary.
1527Does not handle 'set' magic. See also C<sv_setuv_mg>.
9b94d1dd 1528
bd81e77b
NC
1529=cut
1530*/
d33b2eba 1531
bd81e77b
NC
1532void
1533Perl_sv_setuv(pTHX_ register SV *sv, UV u)
1534{
1535 /* With these two if statements:
1536 u=1.49 s=0.52 cu=72.49 cs=10.64 scripts=270 tests=20865
d33b2eba 1537
bd81e77b
NC
1538 without
1539 u=1.35 s=0.47 cu=73.45 cs=11.43 scripts=270 tests=20865
1c846c1f 1540
bd81e77b
NC
1541 If you wish to remove them, please benchmark to see what the effect is
1542 */
1543 if (u <= (UV)IV_MAX) {
1544 sv_setiv(sv, (IV)u);
1545 return;
1546 }
1547 sv_setiv(sv, 0);
1548 SvIsUV_on(sv);
1549 SvUV_set(sv, u);
1550}
d33b2eba 1551
bd81e77b
NC
1552/*
1553=for apidoc sv_setuv_mg
727879eb 1554
bd81e77b 1555Like C<sv_setuv>, but also handles 'set' magic.
9b94d1dd 1556
bd81e77b
NC
1557=cut
1558*/
5e2fc214 1559
bd81e77b
NC
1560void
1561Perl_sv_setuv_mg(pTHX_ register SV *sv, UV u)
1562{
1563 sv_setiv(sv, 0);
1564 SvIsUV_on(sv);
1565 sv_setuv(sv,u);
1566 SvSETMAGIC(sv);
1567}
5e2fc214 1568
954c1994 1569/*
bd81e77b 1570=for apidoc sv_setnv
954c1994 1571
bd81e77b
NC
1572Copies a double into the given SV, upgrading first if necessary.
1573Does not handle 'set' magic. See also C<sv_setnv_mg>.
954c1994
GS
1574
1575=cut
1576*/
1577
63f97190 1578void
bd81e77b 1579Perl_sv_setnv(pTHX_ register SV *sv, NV num)
79072805 1580{
97aff369 1581 dVAR;
bd81e77b
NC
1582 SV_CHECK_THINKFIRST_COW_DROP(sv);
1583 switch (SvTYPE(sv)) {
79072805 1584 case SVt_NULL:
79072805 1585 case SVt_IV:
bd81e77b 1586 sv_upgrade(sv, SVt_NV);
79072805 1587 break;
ed6116ce 1588 case SVt_RV:
79072805 1589 case SVt_PV:
79072805 1590 case SVt_PVIV:
bd81e77b 1591 sv_upgrade(sv, SVt_PVNV);
79072805 1592 break;
bd4b1eb5 1593
bd4b1eb5 1594 case SVt_PVGV:
bd81e77b
NC
1595 case SVt_PVAV:
1596 case SVt_PVHV:
79072805 1597 case SVt_PVCV:
bd81e77b
NC
1598 case SVt_PVFM:
1599 case SVt_PVIO:
1600 Perl_croak(aTHX_ "Can't coerce %s to number in %s", sv_reftype(sv,0),
1601 OP_NAME(PL_op));
42d0e0b7 1602 default: NOOP;
2068cd4d 1603 }
bd81e77b
NC
1604 SvNV_set(sv, num);
1605 (void)SvNOK_only(sv); /* validate number */
1606 SvTAINT(sv);
79072805
LW
1607}
1608
645c22ef 1609/*
bd81e77b 1610=for apidoc sv_setnv_mg
645c22ef 1611
bd81e77b 1612Like C<sv_setnv>, but also handles 'set' magic.
645c22ef
DM
1613
1614=cut
1615*/
1616
bd81e77b
NC
1617void
1618Perl_sv_setnv_mg(pTHX_ register SV *sv, NV num)
79072805 1619{
bd81e77b
NC
1620 sv_setnv(sv,num);
1621 SvSETMAGIC(sv);
79072805
LW
1622}
1623
bd81e77b
NC
1624/* Print an "isn't numeric" warning, using a cleaned-up,
1625 * printable version of the offending string
1626 */
954c1994 1627
bd81e77b
NC
1628STATIC void
1629S_not_a_number(pTHX_ SV *sv)
79072805 1630{
97aff369 1631 dVAR;
bd81e77b
NC
1632 SV *dsv;
1633 char tmpbuf[64];
1634 const char *pv;
94463019
JH
1635
1636 if (DO_UTF8(sv)) {
396482e1 1637 dsv = sv_2mortal(newSVpvs(""));
94463019
JH
1638 pv = sv_uni_display(dsv, sv, 10, 0);
1639 } else {
1640 char *d = tmpbuf;
551405c4 1641 const char * const limit = tmpbuf + sizeof(tmpbuf) - 8;
94463019
JH
1642 /* each *s can expand to 4 chars + "...\0",
1643 i.e. need room for 8 chars */
ecdeb87c 1644
00b6aa41
AL
1645 const char *s = SvPVX_const(sv);
1646 const char * const end = s + SvCUR(sv);
1647 for ( ; s < end && d < limit; s++ ) {
94463019
JH
1648 int ch = *s & 0xFF;
1649 if (ch & 128 && !isPRINT_LC(ch)) {
1650 *d++ = 'M';
1651 *d++ = '-';
1652 ch &= 127;
1653 }
1654 if (ch == '\n') {
1655 *d++ = '\\';
1656 *d++ = 'n';
1657 }
1658 else if (ch == '\r') {
1659 *d++ = '\\';
1660 *d++ = 'r';
1661 }
1662 else if (ch == '\f') {
1663 *d++ = '\\';
1664 *d++ = 'f';
1665 }
1666 else if (ch == '\\') {
1667 *d++ = '\\';
1668 *d++ = '\\';
1669 }
1670 else if (ch == '\0') {
1671 *d++ = '\\';
1672 *d++ = '0';
1673 }
1674 else if (isPRINT_LC(ch))
1675 *d++ = ch;
1676 else {
1677 *d++ = '^';
1678 *d++ = toCTRL(ch);
1679 }
1680 }
1681 if (s < end) {
1682 *d++ = '.';
1683 *d++ = '.';
1684 *d++ = '.';
1685 }
1686 *d = '\0';
1687 pv = tmpbuf;
a0d0e21e 1688 }
a0d0e21e 1689
533c011a 1690 if (PL_op)
9014280d 1691 Perl_warner(aTHX_ packWARN(WARN_NUMERIC),
94463019
JH
1692 "Argument \"%s\" isn't numeric in %s", pv,
1693 OP_DESC(PL_op));
a0d0e21e 1694 else
9014280d 1695 Perl_warner(aTHX_ packWARN(WARN_NUMERIC),
94463019 1696 "Argument \"%s\" isn't numeric", pv);
a0d0e21e
LW
1697}
1698
c2988b20
NC
1699/*
1700=for apidoc looks_like_number
1701
645c22ef
DM
1702Test if the content of an SV looks like a number (or is a number).
1703C<Inf> and C<Infinity> are treated as numbers (so will not issue a
1704non-numeric warning), even if your atof() doesn't grok them.
c2988b20
NC
1705
1706=cut
1707*/
1708
1709I32
1710Perl_looks_like_number(pTHX_ SV *sv)
1711{
a3b680e6 1712 register const char *sbegin;
c2988b20
NC
1713 STRLEN len;
1714
1715 if (SvPOK(sv)) {
3f7c398e 1716 sbegin = SvPVX_const(sv);
c2988b20
NC
1717 len = SvCUR(sv);
1718 }
1719 else if (SvPOKp(sv))
83003860 1720 sbegin = SvPV_const(sv, len);
c2988b20 1721 else
e0ab1c0e 1722 return SvFLAGS(sv) & (SVf_NOK|SVp_NOK|SVf_IOK|SVp_IOK);
c2988b20
NC
1723 return grok_number(sbegin, len, NULL);
1724}
25da4f38 1725
19f6321d
NC
1726STATIC bool
1727S_glob_2number(pTHX_ GV * const gv)
180488f8
NC
1728{
1729 const U32 wasfake = SvFLAGS(gv) & SVf_FAKE;
1730 SV *const buffer = sv_newmortal();
1731
1732 /* FAKE globs can get coerced, so need to turn this off temporarily if it
1733 is on. */
1734 SvFAKE_off(gv);
1735 gv_efullname3(buffer, gv, "*");
1736 SvFLAGS(gv) |= wasfake;
1737
675c862f
AL
1738 /* We know that all GVs stringify to something that is not-a-number,
1739 so no need to test that. */
1740 if (ckWARN(WARN_NUMERIC))
1741 not_a_number(buffer);
1742 /* We just want something true to return, so that S_sv_2iuv_common
1743 can tail call us and return true. */
19f6321d 1744 return TRUE;
675c862f
AL
1745}
1746
1747STATIC char *
19f6321d 1748S_glob_2pv(pTHX_ GV * const gv, STRLEN * const len)
675c862f
AL
1749{
1750 const U32 wasfake = SvFLAGS(gv) & SVf_FAKE;
1751 SV *const buffer = sv_newmortal();
1752
1753 /* FAKE globs can get coerced, so need to turn this off temporarily if it
1754 is on. */
1755 SvFAKE_off(gv);
1756 gv_efullname3(buffer, gv, "*");
1757 SvFLAGS(gv) |= wasfake;
1758
1759 assert(SvPOK(buffer));
a6d61a6c
NC
1760 if (len) {
1761 *len = SvCUR(buffer);
1762 }
675c862f 1763 return SvPVX(buffer);
180488f8
NC
1764}
1765
25da4f38
IZ
1766/* Actually, ISO C leaves conversion of UV to IV undefined, but
1767 until proven guilty, assume that things are not that bad... */
1768
645c22ef
DM
1769/*
1770 NV_PRESERVES_UV:
1771
1772 As 64 bit platforms often have an NV that doesn't preserve all bits of
28e5dec8
JH
1773 an IV (an assumption perl has been based on to date) it becomes necessary
1774 to remove the assumption that the NV always carries enough precision to
1775 recreate the IV whenever needed, and that the NV is the canonical form.
1776 Instead, IV/UV and NV need to be given equal rights. So as to not lose
645c22ef 1777 precision as a side effect of conversion (which would lead to insanity
28e5dec8
JH
1778 and the dragon(s) in t/op/numconvert.t getting very angry) the intent is
1779 1) to distinguish between IV/UV/NV slots that have cached a valid
1780 conversion where precision was lost and IV/UV/NV slots that have a
1781 valid conversion which has lost no precision
645c22ef 1782 2) to ensure that if a numeric conversion to one form is requested that
28e5dec8
JH
1783 would lose precision, the precise conversion (or differently
1784 imprecise conversion) is also performed and cached, to prevent
1785 requests for different numeric formats on the same SV causing
1786 lossy conversion chains. (lossless conversion chains are perfectly
1787 acceptable (still))
1788
1789
1790 flags are used:
1791 SvIOKp is true if the IV slot contains a valid value
1792 SvIOK is true only if the IV value is accurate (UV if SvIOK_UV true)
1793 SvNOKp is true if the NV slot contains a valid value
1794 SvNOK is true only if the NV value is accurate
1795
1796 so
645c22ef 1797 while converting from PV to NV, check to see if converting that NV to an
28e5dec8
JH
1798 IV(or UV) would lose accuracy over a direct conversion from PV to
1799 IV(or UV). If it would, cache both conversions, return NV, but mark
1800 SV as IOK NOKp (ie not NOK).
1801
645c22ef 1802 While converting from PV to IV, check to see if converting that IV to an
28e5dec8
JH
1803 NV would lose accuracy over a direct conversion from PV to NV. If it
1804 would, cache both conversions, flag similarly.
1805
1806 Before, the SV value "3.2" could become NV=3.2 IV=3 NOK, IOK quite
1807 correctly because if IV & NV were set NV *always* overruled.
645c22ef
DM
1808 Now, "3.2" will become NV=3.2 IV=3 NOK, IOKp, because the flag's meaning
1809 changes - now IV and NV together means that the two are interchangeable:
28e5dec8 1810 SvIVX == (IV) SvNVX && SvNVX == (NV) SvIVX;
d460ef45 1811
645c22ef
DM
1812 The benefit of this is that operations such as pp_add know that if
1813 SvIOK is true for both left and right operands, then integer addition
1814 can be used instead of floating point (for cases where the result won't
1815 overflow). Before, floating point was always used, which could lead to
28e5dec8
JH
1816 loss of precision compared with integer addition.
1817
1818 * making IV and NV equal status should make maths accurate on 64 bit
1819 platforms
1820 * may speed up maths somewhat if pp_add and friends start to use
645c22ef 1821 integers when possible instead of fp. (Hopefully the overhead in
28e5dec8
JH
1822 looking for SvIOK and checking for overflow will not outweigh the
1823 fp to integer speedup)
1824 * will slow down integer operations (callers of SvIV) on "inaccurate"
1825 values, as the change from SvIOK to SvIOKp will cause a call into
1826 sv_2iv each time rather than a macro access direct to the IV slot
1827 * should speed up number->string conversion on integers as IV is
645c22ef 1828 favoured when IV and NV are equally accurate
28e5dec8
JH
1829
1830 ####################################################################
645c22ef
DM
1831 You had better be using SvIOK_notUV if you want an IV for arithmetic:
1832 SvIOK is true if (IV or UV), so you might be getting (IV)SvUV.
1833 On the other hand, SvUOK is true iff UV.
28e5dec8
JH
1834 ####################################################################
1835
645c22ef 1836 Your mileage will vary depending your CPU's relative fp to integer
28e5dec8
JH
1837 performance ratio.
1838*/
1839
1840#ifndef NV_PRESERVES_UV
645c22ef
DM
1841# define IS_NUMBER_UNDERFLOW_IV 1
1842# define IS_NUMBER_UNDERFLOW_UV 2
1843# define IS_NUMBER_IV_AND_UV 2
1844# define IS_NUMBER_OVERFLOW_IV 4
1845# define IS_NUMBER_OVERFLOW_UV 5
1846
1847/* sv_2iuv_non_preserve(): private routine for use by sv_2iv() and sv_2uv() */
28e5dec8
JH
1848
1849/* For sv_2nv these three cases are "SvNOK and don't bother casting" */
1850STATIC int
645c22ef 1851S_sv_2iuv_non_preserve(pTHX_ register SV *sv, I32 numtype)
28e5dec8 1852{
97aff369 1853 dVAR;
b57a0404 1854 PERL_UNUSED_ARG(numtype); /* Used only under DEBUGGING? */
3f7c398e 1855 DEBUG_c(PerlIO_printf(Perl_debug_log,"sv_2iuv_non '%s', IV=0x%"UVxf" NV=%"NVgf" inttype=%"UVXf"\n", SvPVX_const(sv), SvIVX(sv), SvNVX(sv), (UV)numtype));
28e5dec8
JH
1856 if (SvNVX(sv) < (NV)IV_MIN) {
1857 (void)SvIOKp_on(sv);
1858 (void)SvNOK_on(sv);
45977657 1859 SvIV_set(sv, IV_MIN);
28e5dec8
JH
1860 return IS_NUMBER_UNDERFLOW_IV;
1861 }
1862 if (SvNVX(sv) > (NV)UV_MAX) {
1863 (void)SvIOKp_on(sv);
1864 (void)SvNOK_on(sv);
1865 SvIsUV_on(sv);
607fa7f2 1866 SvUV_set(sv, UV_MAX);
28e5dec8
JH
1867 return IS_NUMBER_OVERFLOW_UV;
1868 }
c2988b20
NC
1869 (void)SvIOKp_on(sv);
1870 (void)SvNOK_on(sv);
1871 /* Can't use strtol etc to convert this string. (See truth table in
1872 sv_2iv */
1873 if (SvNVX(sv) <= (UV)IV_MAX) {
45977657 1874 SvIV_set(sv, I_V(SvNVX(sv)));
c2988b20
NC
1875 if ((NV)(SvIVX(sv)) == SvNVX(sv)) {
1876 SvIOK_on(sv); /* Integer is precise. NOK, IOK */
1877 } else {
1878 /* Integer is imprecise. NOK, IOKp */
1879 }
1880 return SvNVX(sv) < 0 ? IS_NUMBER_UNDERFLOW_UV : IS_NUMBER_IV_AND_UV;
1881 }
1882 SvIsUV_on(sv);
607fa7f2 1883 SvUV_set(sv, U_V(SvNVX(sv)));
c2988b20
NC
1884 if ((NV)(SvUVX(sv)) == SvNVX(sv)) {
1885 if (SvUVX(sv) == UV_MAX) {
1886 /* As we know that NVs don't preserve UVs, UV_MAX cannot
1887 possibly be preserved by NV. Hence, it must be overflow.
1888 NOK, IOKp */
1889 return IS_NUMBER_OVERFLOW_UV;
1890 }
1891 SvIOK_on(sv); /* Integer is precise. NOK, UOK */
1892 } else {
1893 /* Integer is imprecise. NOK, IOKp */
28e5dec8 1894 }
c2988b20 1895 return IS_NUMBER_OVERFLOW_IV;
28e5dec8 1896}
645c22ef
DM
1897#endif /* !NV_PRESERVES_UV*/
1898
af359546
NC
1899STATIC bool
1900S_sv_2iuv_common(pTHX_ SV *sv) {
97aff369 1901 dVAR;
af359546 1902 if (SvNOKp(sv)) {
28e5dec8
JH
1903 /* erm. not sure. *should* never get NOKp (without NOK) from sv_2nv
1904 * without also getting a cached IV/UV from it at the same time
1905 * (ie PV->NV conversion should detect loss of accuracy and cache
af359546
NC
1906 * IV or UV at same time to avoid this. */
1907 /* IV-over-UV optimisation - choose to cache IV if possible */
25da4f38
IZ
1908
1909 if (SvTYPE(sv) == SVt_NV)
1910 sv_upgrade(sv, SVt_PVNV);
1911
28e5dec8
JH
1912 (void)SvIOKp_on(sv); /* Must do this first, to clear any SvOOK */
1913 /* < not <= as for NV doesn't preserve UV, ((NV)IV_MAX+1) will almost
1914 certainly cast into the IV range at IV_MAX, whereas the correct
1915 answer is the UV IV_MAX +1. Hence < ensures that dodgy boundary
1916 cases go to UV */
cab190d4
JD
1917#if defined(NAN_COMPARE_BROKEN) && defined(Perl_isnan)
1918 if (Perl_isnan(SvNVX(sv))) {
1919 SvUV_set(sv, 0);
1920 SvIsUV_on(sv);
fdbe6d7c 1921 return FALSE;
cab190d4 1922 }
cab190d4 1923#endif
28e5dec8 1924 if (SvNVX(sv) < (NV)IV_MAX + 0.5) {
45977657 1925 SvIV_set(sv, I_V(SvNVX(sv)));
28e5dec8
JH
1926 if (SvNVX(sv) == (NV) SvIVX(sv)
1927#ifndef NV_PRESERVES_UV
1928 && (((UV)1 << NV_PRESERVES_UV_BITS) >
1929 (UV)(SvIVX(sv) > 0 ? SvIVX(sv) : -SvIVX(sv)))
1930 /* Don't flag it as "accurately an integer" if the number
1931 came from a (by definition imprecise) NV operation, and
1932 we're outside the range of NV integer precision */
1933#endif
1934 ) {
1935 SvIOK_on(sv); /* Can this go wrong with rounding? NWC */
1936 DEBUG_c(PerlIO_printf(Perl_debug_log,
7234c960 1937 "0x%"UVxf" iv(%"NVgf" => %"IVdf") (precise)\n",
28e5dec8
JH
1938 PTR2UV(sv),
1939 SvNVX(sv),
1940 SvIVX(sv)));
1941
1942 } else {
1943 /* IV not precise. No need to convert from PV, as NV
1944 conversion would already have cached IV if it detected
1945 that PV->IV would be better than PV->NV->IV
1946 flags already correct - don't set public IOK. */
1947 DEBUG_c(PerlIO_printf(Perl_debug_log,
7234c960 1948 "0x%"UVxf" iv(%"NVgf" => %"IVdf") (imprecise)\n",
28e5dec8
JH
1949 PTR2UV(sv),
1950 SvNVX(sv),
1951 SvIVX(sv)));
1952 }
1953 /* Can the above go wrong if SvIVX == IV_MIN and SvNVX < IV_MIN,
1954 but the cast (NV)IV_MIN rounds to a the value less (more
1955 negative) than IV_MIN which happens to be equal to SvNVX ??
1956 Analogous to 0xFFFFFFFFFFFFFFFF rounding up to NV (2**64) and
1957 NV rounding back to 0xFFFFFFFFFFFFFFFF, so UVX == UV(NVX) and
1958 (NV)UVX == NVX are both true, but the values differ. :-(
1959 Hopefully for 2s complement IV_MIN is something like
1960 0x8000000000000000 which will be exact. NWC */
d460ef45 1961 }
25da4f38 1962 else {
607fa7f2 1963 SvUV_set(sv, U_V(SvNVX(sv)));
28e5dec8
JH
1964 if (
1965 (SvNVX(sv) == (NV) SvUVX(sv))
1966#ifndef NV_PRESERVES_UV
1967 /* Make sure it's not 0xFFFFFFFFFFFFFFFF */
1968 /*&& (SvUVX(sv) != UV_MAX) irrelevant with code below */
1969 && (((UV)1 << NV_PRESERVES_UV_BITS) > SvUVX(sv))
1970 /* Don't flag it as "accurately an integer" if the number
1971 came from a (by definition imprecise) NV operation, and
1972 we're outside the range of NV integer precision */
1973#endif
1974 )
1975 SvIOK_on(sv);
25da4f38 1976 SvIsUV_on(sv);
1c846c1f 1977 DEBUG_c(PerlIO_printf(Perl_debug_log,
57def98f 1978 "0x%"UVxf" 2iv(%"UVuf" => %"IVdf") (as unsigned)\n",
56431972 1979 PTR2UV(sv),
57def98f
JH
1980 SvUVX(sv),
1981 SvUVX(sv)));
25da4f38 1982 }
748a9306
LW
1983 }
1984 else if (SvPOKp(sv) && SvLEN(sv)) {
c2988b20 1985 UV value;
504618e9 1986 const int numtype = grok_number(SvPVX_const(sv), SvCUR(sv), &value);
af359546 1987 /* We want to avoid a possible problem when we cache an IV/ a UV which
25da4f38 1988 may be later translated to an NV, and the resulting NV is not
c2988b20
NC
1989 the same as the direct translation of the initial string
1990 (eg 123.456 can shortcut to the IV 123 with atol(), but we must
1991 be careful to ensure that the value with the .456 is around if the
1992 NV value is requested in the future).
1c846c1f 1993
af359546 1994 This means that if we cache such an IV/a UV, we need to cache the
25da4f38 1995 NV as well. Moreover, we trade speed for space, and do not
28e5dec8 1996 cache the NV if we are sure it's not needed.
25da4f38 1997 */
16b7a9a4 1998
c2988b20
NC
1999 /* SVt_PVNV is one higher than SVt_PVIV, hence this order */
2000 if ((numtype & (IS_NUMBER_IN_UV | IS_NUMBER_NOT_INT))
2001 == IS_NUMBER_IN_UV) {
5e045b90 2002 /* It's definitely an integer, only upgrade to PVIV */
28e5dec8
JH
2003 if (SvTYPE(sv) < SVt_PVIV)
2004 sv_upgrade(sv, SVt_PVIV);
f7bbb42a 2005 (void)SvIOK_on(sv);
c2988b20
NC
2006 } else if (SvTYPE(sv) < SVt_PVNV)
2007 sv_upgrade(sv, SVt_PVNV);
28e5dec8 2008
f2524eef 2009 /* If NVs preserve UVs then we only use the UV value if we know that
c2988b20
NC
2010 we aren't going to call atof() below. If NVs don't preserve UVs
2011 then the value returned may have more precision than atof() will
2012 return, even though value isn't perfectly accurate. */
2013 if ((numtype & (IS_NUMBER_IN_UV
2014#ifdef NV_PRESERVES_UV
2015 | IS_NUMBER_NOT_INT
2016#endif
2017 )) == IS_NUMBER_IN_UV) {
2018 /* This won't turn off the public IOK flag if it was set above */
2019 (void)SvIOKp_on(sv);
2020
2021 if (!(numtype & IS_NUMBER_NEG)) {
2022 /* positive */;
2023 if (value <= (UV)IV_MAX) {
45977657 2024 SvIV_set(sv, (IV)value);
c2988b20 2025 } else {
af359546 2026 /* it didn't overflow, and it was positive. */
607fa7f2 2027 SvUV_set(sv, value);
c2988b20
NC
2028 SvIsUV_on(sv);
2029 }
2030 } else {
2031 /* 2s complement assumption */
2032 if (value <= (UV)IV_MIN) {
45977657 2033 SvIV_set(sv, -(IV)value);
c2988b20
NC
2034 } else {
2035 /* Too negative for an IV. This is a double upgrade, but
d1be9408 2036 I'm assuming it will be rare. */
c2988b20
NC
2037 if (SvTYPE(sv) < SVt_PVNV)
2038 sv_upgrade(sv, SVt_PVNV);
2039 SvNOK_on(sv);
2040 SvIOK_off(sv);
2041 SvIOKp_on(sv);
9d6ce603 2042 SvNV_set(sv, -(NV)value);
45977657 2043 SvIV_set(sv, IV_MIN);
c2988b20
NC
2044 }
2045 }
2046 }
2047 /* For !NV_PRESERVES_UV and IS_NUMBER_IN_UV and IS_NUMBER_NOT_INT we
2048 will be in the previous block to set the IV slot, and the next
2049 block to set the NV slot. So no else here. */
2050
2051 if ((numtype & (IS_NUMBER_IN_UV | IS_NUMBER_NOT_INT))
2052 != IS_NUMBER_IN_UV) {
2053 /* It wasn't an (integer that doesn't overflow the UV). */
3f7c398e 2054 SvNV_set(sv, Atof(SvPVX_const(sv)));
28e5dec8 2055
c2988b20
NC
2056 if (! numtype && ckWARN(WARN_NUMERIC))
2057 not_a_number(sv);
28e5dec8 2058
65202027 2059#if defined(USE_LONG_DOUBLE)
c2988b20
NC
2060 DEBUG_c(PerlIO_printf(Perl_debug_log, "0x%"UVxf" 2iv(%" PERL_PRIgldbl ")\n",
2061 PTR2UV(sv), SvNVX(sv)));
65202027 2062#else
1779d84d 2063 DEBUG_c(PerlIO_printf(Perl_debug_log, "0x%"UVxf" 2iv(%"NVgf")\n",
c2988b20 2064 PTR2UV(sv), SvNVX(sv)));
65202027 2065#endif
28e5dec8 2066
28e5dec8 2067#ifdef NV_PRESERVES_UV
af359546
NC
2068 (void)SvIOKp_on(sv);
2069 (void)SvNOK_on(sv);
2070 if (SvNVX(sv) < (NV)IV_MAX + 0.5) {
2071 SvIV_set(sv, I_V(SvNVX(sv)));
2072 if ((NV)(SvIVX(sv)) == SvNVX(sv)) {
2073 SvIOK_on(sv);
2074 } else {
6f207bd3 2075 NOOP; /* Integer is imprecise. NOK, IOKp */
af359546
NC
2076 }
2077 /* UV will not work better than IV */
2078 } else {
2079 if (SvNVX(sv) > (NV)UV_MAX) {
2080 SvIsUV_on(sv);
2081 /* Integer is inaccurate. NOK, IOKp, is UV */
2082 SvUV_set(sv, UV_MAX);
af359546
NC
2083 } else {
2084 SvUV_set(sv, U_V(SvNVX(sv)));
2085 /* 0xFFFFFFFFFFFFFFFF not an issue in here, NVs
2086 NV preservse UV so can do correct comparison. */
2087 if ((NV)(SvUVX(sv)) == SvNVX(sv)) {
2088 SvIOK_on(sv);
af359546 2089 } else {
6f207bd3 2090 NOOP; /* Integer is imprecise. NOK, IOKp, is UV */
af359546
NC
2091 }
2092 }
4b0c9573 2093 SvIsUV_on(sv);
af359546 2094 }
28e5dec8 2095#else /* NV_PRESERVES_UV */
c2988b20
NC
2096 if ((numtype & (IS_NUMBER_IN_UV | IS_NUMBER_NOT_INT))
2097 == (IS_NUMBER_IN_UV | IS_NUMBER_NOT_INT)) {
af359546 2098 /* The IV/UV slot will have been set from value returned by
c2988b20
NC
2099 grok_number above. The NV slot has just been set using
2100 Atof. */
560b0c46 2101 SvNOK_on(sv);
c2988b20
NC
2102 assert (SvIOKp(sv));
2103 } else {
2104 if (((UV)1 << NV_PRESERVES_UV_BITS) >
2105 U_V(SvNVX(sv) > 0 ? SvNVX(sv) : -SvNVX(sv))) {
2106 /* Small enough to preserve all bits. */
2107 (void)SvIOKp_on(sv);
2108 SvNOK_on(sv);
45977657 2109 SvIV_set(sv, I_V(SvNVX(sv)));
c2988b20
NC
2110 if ((NV)(SvIVX(sv)) == SvNVX(sv))
2111 SvIOK_on(sv);
2112 /* Assumption: first non-preserved integer is < IV_MAX,
2113 this NV is in the preserved range, therefore: */
2114 if (!(U_V(SvNVX(sv) > 0 ? SvNVX(sv) : -SvNVX(sv))
2115 < (UV)IV_MAX)) {
32fdb065 2116 Perl_croak(aTHX_ "sv_2iv assumed (U_V(fabs((double)SvNVX(sv))) < (UV)IV_MAX) but SvNVX(sv)=%"NVgf" U_V is 0x%"UVxf", IV_MAX is 0x%"UVxf"\n", SvNVX(sv), U_V(SvNVX(sv)), (UV)IV_MAX);
c2988b20
NC
2117 }
2118 } else {
2119 /* IN_UV NOT_INT
2120 0 0 already failed to read UV.
2121 0 1 already failed to read UV.
2122 1 0 you won't get here in this case. IV/UV
2123 slot set, public IOK, Atof() unneeded.
2124 1 1 already read UV.
2125 so there's no point in sv_2iuv_non_preserve() attempting
2126 to use atol, strtol, strtoul etc. */
40a17c4c 2127 sv_2iuv_non_preserve (sv, numtype);
c2988b20
NC
2128 }
2129 }
28e5dec8 2130#endif /* NV_PRESERVES_UV */
25da4f38 2131 }
af359546
NC
2132 }
2133 else {
675c862f 2134 if (isGV_with_GP(sv))
a0933d07 2135 return glob_2number((GV *)sv);
180488f8 2136
af359546
NC
2137 if (!(SvFLAGS(sv) & SVs_PADTMP)) {
2138 if (!PL_localizing && ckWARN(WARN_UNINITIALIZED))
2139 report_uninit(sv);
2140 }
25da4f38
IZ
2141 if (SvTYPE(sv) < SVt_IV)
2142 /* Typically the caller expects that sv_any is not NULL now. */
2143 sv_upgrade(sv, SVt_IV);
af359546
NC
2144 /* Return 0 from the caller. */
2145 return TRUE;
2146 }
2147 return FALSE;
2148}
2149
2150/*
2151=for apidoc sv_2iv_flags
2152
2153Return the integer value of an SV, doing any necessary string
2154conversion. If flags includes SV_GMAGIC, does an mg_get() first.
2155Normally used via the C<SvIV(sv)> and C<SvIVx(sv)> macros.
2156
2157=cut
2158*/
2159
2160IV
2161Perl_sv_2iv_flags(pTHX_ register SV *sv, I32 flags)
2162{
97aff369 2163 dVAR;
af359546 2164 if (!sv)
a0d0e21e 2165 return 0;
af359546
NC
2166 if (SvGMAGICAL(sv)) {
2167 if (flags & SV_GMAGIC)
2168 mg_get(sv);
2169 if (SvIOKp(sv))
2170 return SvIVX(sv);
2171 if (SvNOKp(sv)) {
2172 return I_V(SvNVX(sv));
2173 }
71c558c3
NC
2174 if (SvPOKp(sv) && SvLEN(sv)) {
2175 UV value;
2176 const int numtype
2177 = grok_number(SvPVX_const(sv), SvCUR(sv), &value);
2178
2179 if ((numtype & (IS_NUMBER_IN_UV | IS_NUMBER_NOT_INT))
2180 == IS_NUMBER_IN_UV) {
2181 /* It's definitely an integer */
2182 if (numtype & IS_NUMBER_NEG) {
2183 if (value < (UV)IV_MIN)
2184 return -(IV)value;
2185 } else {
2186 if (value < (UV)IV_MAX)
2187 return (IV)value;
2188 }
2189 }
2190 if (!numtype) {
2191 if (ckWARN(WARN_NUMERIC))
2192 not_a_number(sv);
2193 }
2194 return I_V(Atof(SvPVX_const(sv)));
2195 }
1c7ff15e
NC
2196 if (SvROK(sv)) {
2197 goto return_rok;
af359546 2198 }
1c7ff15e
NC
2199 assert(SvTYPE(sv) >= SVt_PVMG);
2200 /* This falls through to the report_uninit inside S_sv_2iuv_common. */
4cb1ec55 2201 } else if (SvTHINKFIRST(sv)) {
af359546 2202 if (SvROK(sv)) {
1c7ff15e 2203 return_rok:
af359546
NC
2204 if (SvAMAGIC(sv)) {
2205 SV * const tmpstr=AMG_CALLun(sv,numer);
2206 if (tmpstr && (!SvROK(tmpstr) || (SvRV(tmpstr) != SvRV(sv)))) {
2207 return SvIV(tmpstr);
2208 }
2209 }
2210 return PTR2IV(SvRV(sv));
2211 }
2212 if (SvIsCOW(sv)) {
2213 sv_force_normal_flags(sv, 0);
2214 }
2215 if (SvREADONLY(sv) && !SvOK(sv)) {
2216 if (ckWARN(WARN_UNINITIALIZED))
2217 report_uninit(sv);
2218 return 0;
2219 }
2220 }
2221 if (!SvIOKp(sv)) {
2222 if (S_sv_2iuv_common(aTHX_ sv))
2223 return 0;
79072805 2224 }
1d7c1841
GS
2225 DEBUG_c(PerlIO_printf(Perl_debug_log, "0x%"UVxf" 2iv(%"IVdf")\n",
2226 PTR2UV(sv),SvIVX(sv)));
25da4f38 2227 return SvIsUV(sv) ? (IV)SvUVX(sv) : SvIVX(sv);
79072805
LW
2228}
2229
645c22ef 2230/*
891f9566 2231=for apidoc sv_2uv_flags
645c22ef
DM
2232
2233Return the unsigned integer value of an SV, doing any necessary string
891f9566
YST
2234conversion. If flags includes SV_GMAGIC, does an mg_get() first.
2235Normally used via the C<SvUV(sv)> and C<SvUVx(sv)> macros.
645c22ef
DM
2236
2237=cut
2238*/
2239
ff68c719 2240UV
891f9566 2241Perl_sv_2uv_flags(pTHX_ register SV *sv, I32 flags)
ff68c719 2242{
97aff369 2243 dVAR;
ff68c719 2244 if (!sv)
2245 return 0;
2246 if (SvGMAGICAL(sv)) {
891f9566
YST
2247 if (flags & SV_GMAGIC)
2248 mg_get(sv);
ff68c719 2249 if (SvIOKp(sv))
2250 return SvUVX(sv);
2251 if (SvNOKp(sv))
2252 return U_V(SvNVX(sv));
71c558c3
NC
2253 if (SvPOKp(sv) && SvLEN(sv)) {
2254 UV value;
2255 const int numtype
2256 = grok_number(SvPVX_const(sv), SvCUR(sv), &value);
2257
2258 if ((numtype & (IS_NUMBER_IN_UV | IS_NUMBER_NOT_INT))
2259 == IS_NUMBER_IN_UV) {
2260 /* It's definitely an integer */
2261 if (!(numtype & IS_NUMBER_NEG))
2262 return value;
2263 }
2264 if (!numtype) {
2265 if (ckWARN(WARN_NUMERIC))
2266 not_a_number(sv);
2267 }
2268 return U_V(Atof(SvPVX_const(sv)));
2269 }
1c7ff15e
NC
2270 if (SvROK(sv)) {
2271 goto return_rok;
3fe9a6f1 2272 }
1c7ff15e
NC
2273 assert(SvTYPE(sv) >= SVt_PVMG);
2274 /* This falls through to the report_uninit inside S_sv_2iuv_common. */
4cb1ec55 2275 } else if (SvTHINKFIRST(sv)) {
ff68c719 2276 if (SvROK(sv)) {
1c7ff15e 2277 return_rok:
deb46114
NC
2278 if (SvAMAGIC(sv)) {
2279 SV *const tmpstr = AMG_CALLun(sv,numer);
2280 if (tmpstr && (!SvROK(tmpstr) || (SvRV(tmpstr) != SvRV(sv)))) {
2281 return SvUV(tmpstr);
2282 }
2283 }
2284 return PTR2UV(SvRV(sv));
ff68c719 2285 }
765f542d
NC
2286 if (SvIsCOW(sv)) {
2287 sv_force_normal_flags(sv, 0);
8a818333 2288 }
0336b60e 2289 if (SvREADONLY(sv) && !SvOK(sv)) {
0336b60e 2290 if (ckWARN(WARN_UNINITIALIZED))
29489e7c 2291 report_uninit(sv);
ff68c719 2292 return 0;
2293 }
2294 }
af359546
NC
2295 if (!SvIOKp(sv)) {
2296 if (S_sv_2iuv_common(aTHX_ sv))
2297 return 0;
ff68c719 2298 }
25da4f38 2299
1d7c1841
GS
2300 DEBUG_c(PerlIO_printf(Perl_debug_log, "0x%"UVxf" 2uv(%"UVuf")\n",
2301 PTR2UV(sv),SvUVX(sv)));
25da4f38 2302 return SvIsUV(sv) ? SvUVX(sv) : (UV)SvIVX(sv);
ff68c719 2303}
2304
645c22ef
DM
2305/*
2306=for apidoc sv_2nv
2307
2308Return the num value of an SV, doing any necessary string or integer
2309conversion, magic etc. Normally used via the C<SvNV(sv)> and C<SvNVx(sv)>
2310macros.
2311
2312=cut
2313*/
2314
65202027 2315NV
864dbfa3 2316Perl_sv_2nv(pTHX_ register SV *sv)
79072805 2317{
97aff369 2318 dVAR;
79072805
LW
2319 if (!sv)
2320 return 0.0;
8990e307 2321 if (SvGMAGICAL(sv)) {
463ee0b2
LW
2322 mg_get(sv);
2323 if (SvNOKp(sv))
2324 return SvNVX(sv);
0aa395f8 2325 if ((SvPOKp(sv) && SvLEN(sv)) && !SvIOKp(sv)) {
041457d9 2326 if (!SvIOKp(sv) && ckWARN(WARN_NUMERIC) &&
504618e9 2327 !grok_number(SvPVX_const(sv), SvCUR(sv), NULL))
a0d0e21e 2328 not_a_number(sv);
3f7c398e 2329 return Atof(SvPVX_const(sv));
a0d0e21e 2330 }
25da4f38 2331 if (SvIOKp(sv)) {
1c846c1f 2332 if (SvIsUV(sv))
65202027 2333 return (NV)SvUVX(sv);
25da4f38 2334 else
65202027 2335 return (NV)SvIVX(sv);
47a72cb8
NC
2336 }
2337 if (SvROK(sv)) {
2338 goto return_rok;
2339 }
2340 assert(SvTYPE(sv) >= SVt_PVMG);
2341 /* This falls through to the report_uninit near the end of the
2342 function. */
2343 } else if (SvTHINKFIRST(sv)) {
a0d0e21e 2344 if (SvROK(sv)) {
47a72cb8 2345 return_rok:
deb46114
NC
2346 if (SvAMAGIC(sv)) {
2347 SV *const tmpstr = AMG_CALLun(sv,numer);
2348 if (tmpstr && (!SvROK(tmpstr) || (SvRV(tmpstr) != SvRV(sv)))) {
2349 return SvNV(tmpstr);
2350 }
2351 }
2352 return PTR2NV(SvRV(sv));
a0d0e21e 2353 }
765f542d
NC
2354 if (SvIsCOW(sv)) {
2355 sv_force_normal_flags(sv, 0);
8a818333 2356 }
0336b60e 2357 if (SvREADONLY(sv) && !SvOK(sv)) {
599cee73 2358 if (ckWARN(WARN_UNINITIALIZED))
29489e7c 2359 report_uninit(sv);
ed6116ce
LW
2360 return 0.0;
2361 }
79072805
LW
2362 }
2363 if (SvTYPE(sv) < SVt_NV) {
7e25a7e9
NC
2364 /* The logic to use SVt_PVNV if necessary is in sv_upgrade. */
2365 sv_upgrade(sv, SVt_NV);
906f284f 2366#ifdef USE_LONG_DOUBLE
097ee67d 2367 DEBUG_c({
f93f4e46 2368 STORE_NUMERIC_LOCAL_SET_STANDARD();
1d7c1841
GS
2369 PerlIO_printf(Perl_debug_log,
2370 "0x%"UVxf" num(%" PERL_PRIgldbl ")\n",
2371 PTR2UV(sv), SvNVX(sv));
572bbb43
GS
2372 RESTORE_NUMERIC_LOCAL();
2373 });
65202027 2374#else
572bbb43 2375 DEBUG_c({
f93f4e46 2376 STORE_NUMERIC_LOCAL_SET_STANDARD();
1779d84d 2377 PerlIO_printf(Perl_debug_log, "0x%"UVxf" num(%"NVgf")\n",
1d7c1841 2378 PTR2UV(sv), SvNVX(sv));
097ee67d
JH
2379 RESTORE_NUMERIC_LOCAL();
2380 });
572bbb43 2381#endif
79072805
LW
2382 }
2383 else if (SvTYPE(sv) < SVt_PVNV)
2384 sv_upgrade(sv, SVt_PVNV);
59d8ce62
NC
2385 if (SvNOKp(sv)) {
2386 return SvNVX(sv);
61604483 2387 }
59d8ce62 2388 if (SvIOKp(sv)) {
9d6ce603 2389 SvNV_set(sv, SvIsUV(sv) ? (NV)SvUVX(sv) : (NV)SvIVX(sv));
28e5dec8
JH
2390#ifdef NV_PRESERVES_UV
2391 SvNOK_on(sv);
2392#else
2393 /* Only set the public NV OK flag if this NV preserves the IV */
2394 /* Check it's not 0xFFFFFFFFFFFFFFFF */
2395 if (SvIsUV(sv) ? ((SvUVX(sv) != UV_MAX)&&(SvUVX(sv) == U_V(SvNVX(sv))))
2396 : (SvIVX(sv) == I_V(SvNVX(sv))))
2397 SvNOK_on(sv);
2398 else
2399 SvNOKp_on(sv);
2400#endif
93a17b20 2401 }
748a9306 2402 else if (SvPOKp(sv) && SvLEN(sv)) {
c2988b20 2403 UV value;
3f7c398e 2404 const int numtype = grok_number(SvPVX_const(sv), SvCUR(sv), &value);
041457d9 2405 if (!SvIOKp(sv) && !numtype && ckWARN(WARN_NUMERIC))
a0d0e21e 2406 not_a_number(sv);
28e5dec8 2407#ifdef NV_PRESERVES_UV
c2988b20
NC
2408 if ((numtype & (IS_NUMBER_IN_UV | IS_NUMBER_NOT_INT))
2409 == IS_NUMBER_IN_UV) {
5e045b90 2410 /* It's definitely an integer */
9d6ce603 2411 SvNV_set(sv, (numtype & IS_NUMBER_NEG) ? -(NV)value : (NV)value);
c2988b20 2412 } else
3f7c398e 2413 SvNV_set(sv, Atof(SvPVX_const(sv)));
28e5dec8
JH
2414 SvNOK_on(sv);
2415#else
3f7c398e 2416 SvNV_set(sv, Atof(SvPVX_const(sv)));
28e5dec8
JH
2417 /* Only set the public NV OK flag if this NV preserves the value in
2418 the PV at least as well as an IV/UV would.
2419 Not sure how to do this 100% reliably. */
2420 /* if that shift count is out of range then Configure's test is
2421 wonky. We shouldn't be in here with NV_PRESERVES_UV_BITS ==
2422 UV_BITS */
2423 if (((UV)1 << NV_PRESERVES_UV_BITS) >
c2988b20 2424 U_V(SvNVX(sv) > 0 ? SvNVX(sv) : -SvNVX(sv))) {
28e5dec8 2425 SvNOK_on(sv); /* Definitely small enough to preserve all bits */
c2988b20
NC
2426 } else if (!(numtype & IS_NUMBER_IN_UV)) {
2427 /* Can't use strtol etc to convert this string, so don't try.
2428 sv_2iv and sv_2uv will use the NV to convert, not the PV. */
2429 SvNOK_on(sv);
2430 } else {
2431 /* value has been set. It may not be precise. */
2432 if ((numtype & IS_NUMBER_NEG) && (value > (UV)IV_MIN)) {
2433 /* 2s complement assumption for (UV)IV_MIN */
2434 SvNOK_on(sv); /* Integer is too negative. */
2435 } else {
2436 SvNOKp_on(sv);
2437 SvIOKp_on(sv);
6fa402ec 2438
c2988b20 2439 if (numtype & IS_NUMBER_NEG) {
45977657 2440 SvIV_set(sv, -(IV)value);
c2988b20 2441 } else if (value <= (UV)IV_MAX) {
45977657 2442 SvIV_set(sv, (IV)value);
c2988b20 2443 } else {
607fa7f2 2444 SvUV_set(sv, value);
c2988b20
NC
2445 SvIsUV_on(sv);
2446 }
2447
2448 if (numtype & IS_NUMBER_NOT_INT) {
2449 /* I believe that even if the original PV had decimals,
2450 they are lost beyond the limit of the FP precision.
2451 However, neither is canonical, so both only get p
2452 flags. NWC, 2000/11/25 */
2453 /* Both already have p flags, so do nothing */
2454 } else {
66a1b24b 2455 const NV nv = SvNVX(sv);
c2988b20
NC
2456 if (SvNVX(sv) < (NV)IV_MAX + 0.5) {
2457 if (SvIVX(sv) == I_V(nv)) {
2458 SvNOK_on(sv);
c2988b20 2459 } else {
c2988b20
NC
2460 /* It had no "." so it must be integer. */
2461 }
00b6aa41 2462 SvIOK_on(sv);
c2988b20
NC
2463 } else {
2464 /* between IV_MAX and NV(UV_MAX).
2465 Could be slightly > UV_MAX */
6fa402ec 2466
c2988b20
NC
2467 if (numtype & IS_NUMBER_NOT_INT) {
2468 /* UV and NV both imprecise. */
2469 } else {
66a1b24b 2470 const UV nv_as_uv = U_V(nv);
c2988b20
NC
2471
2472 if (value == nv_as_uv && SvUVX(sv) != UV_MAX) {
2473 SvNOK_on(sv);
c2988b20 2474 }
00b6aa41 2475 SvIOK_on(sv);
c2988b20
NC
2476 }
2477 }
2478 }
2479 }
2480 }
28e5dec8 2481#endif /* NV_PRESERVES_UV */
93a17b20 2482 }
79072805 2483 else {
f7877b28 2484 if (isGV_with_GP(sv)) {
19f6321d 2485 glob_2number((GV *)sv);
180488f8
NC
2486 return 0.0;
2487 }
2488
041457d9 2489 if (!PL_localizing && !(SvFLAGS(sv) & SVs_PADTMP) && ckWARN(WARN_UNINITIALIZED))
29489e7c 2490 report_uninit(sv);
7e25a7e9
NC
2491 assert (SvTYPE(sv) >= SVt_NV);
2492 /* Typically the caller expects that sv_any is not NULL now. */
2493 /* XXX Ilya implies that this is a bug in callers that assume this
2494 and ideally should be fixed. */
a0d0e21e 2495 return 0.0;
79072805 2496 }
572bbb43 2497#if defined(USE_LONG_DOUBLE)
097ee67d 2498 DEBUG_c({
f93f4e46 2499 STORE_NUMERIC_LOCAL_SET_STANDARD();
1d7c1841
GS
2500 PerlIO_printf(Perl_debug_log, "0x%"UVxf" 2nv(%" PERL_PRIgldbl ")\n",
2501 PTR2UV(sv), SvNVX(sv));
572bbb43
GS
2502 RESTORE_NUMERIC_LOCAL();
2503 });
65202027 2504#else
572bbb43 2505 DEBUG_c({
f93f4e46 2506 STORE_NUMERIC_LOCAL_SET_STANDARD();
1779d84d 2507 PerlIO_printf(Perl_debug_log, "0x%"UVxf" 1nv(%"NVgf")\n",
1d7c1841 2508 PTR2UV(sv), SvNVX(sv));
097ee67d
JH
2509 RESTORE_NUMERIC_LOCAL();
2510 });
572bbb43 2511#endif
463ee0b2 2512 return SvNVX(sv);
79072805
LW
2513}
2514
645c22ef
DM
2515/* uiv_2buf(): private routine for use by sv_2pv_flags(): print an IV or
2516 * UV as a string towards the end of buf, and return pointers to start and
2517 * end of it.
2518 *
2519 * We assume that buf is at least TYPE_CHARS(UV) long.
2520 */
2521
864dbfa3 2522static char *
aec46f14 2523S_uiv_2buf(char *buf, IV iv, UV uv, int is_uv, char **peob)
25da4f38 2524{
25da4f38 2525 char *ptr = buf + TYPE_CHARS(UV);
823a54a3 2526 char * const ebuf = ptr;
25da4f38 2527 int sign;
25da4f38
IZ
2528
2529 if (is_uv)
2530 sign = 0;
2531 else if (iv >= 0) {
2532 uv = iv;
2533 sign = 0;
2534 } else {
2535 uv = -iv;
2536 sign = 1;
2537 }
2538 do {
eb160463 2539 *--ptr = '0' + (char)(uv % 10);
25da4f38
IZ
2540 } while (uv /= 10);
2541 if (sign)
2542 *--ptr = '-';
2543 *peob = ebuf;
2544 return ptr;
2545}
2546
9af30d34
NC
2547/* stringify_regexp(): private routine for use by sv_2pv_flags(): converts
2548 * a regexp to its stringified form.
2549 */
2550
2551static char *
2552S_stringify_regexp(pTHX_ SV *sv, MAGIC *mg, STRLEN *lp) {
97aff369 2553 dVAR;
00b6aa41 2554 const regexp * const re = (regexp *)mg->mg_obj;
9af30d34
NC
2555
2556 if (!mg->mg_ptr) {
2557 const char *fptr = "msix";
2558 char reflags[6];
2559 char ch;
2560 int left = 0;
2561 int right = 4;
00b6aa41 2562 bool need_newline = 0;
9af30d34
NC
2563 U16 reganch = (U16)((re->reganch & PMf_COMPILETIME) >> 12);
2564
2565 while((ch = *fptr++)) {
2566 if(reganch & 1) {
2567 reflags[left++] = ch;
2568 }
2569 else {
2570 reflags[right--] = ch;
2571 }
2572 reganch >>= 1;
2573 }
2574 if(left != 4) {
2575 reflags[left] = '-';
2576 left = 5;
2577 }
2578
2579 mg->mg_len = re->prelen + 4 + left;
2580 /*
2581 * If /x was used, we have to worry about a regex ending with a
2582 * comment later being embedded within another regex. If so, we don't
2583 * want this regex's "commentization" to leak out to the right part of
2584 * the enclosing regex, we must cap it with a newline.
2585 *
2586 * So, if /x was used, we scan backwards from the end of the regex. If
2587 * we find a '#' before we find a newline, we need to add a newline
2588 * ourself. If we find a '\n' first (or if we don't find '#' or '\n'),
2589 * we don't need to add anything. -jfriedl
2590 */
2591 if (PMf_EXTENDED & re->reganch) {
2592 const char *endptr = re->precomp + re->prelen;
2593 while (endptr >= re->precomp) {
2594 const char c = *(endptr--);
2595 if (c == '\n')
2596 break; /* don't need another */
2597 if (c == '#') {
2598 /* we end while in a comment, so we need a newline */
2599 mg->mg_len++; /* save space for it */
2600 need_newline = 1; /* note to add it */
2601 break;
2602 }
2603 }
2604 }
2605
2606 Newx(mg->mg_ptr, mg->mg_len + 1 + left, char);
2607 mg->mg_ptr[0] = '(';
2608 mg->mg_ptr[1] = '?';
2609 Copy(reflags, mg->mg_ptr+2, left, char);
2610 *(mg->mg_ptr+left+2) = ':';
2611 Copy(re->precomp, mg->mg_ptr+3+left, re->prelen, char);
2612 if (need_newline)
2613 mg->mg_ptr[mg->mg_len - 2] = '\n';
2614 mg->mg_ptr[mg->mg_len - 1] = ')';
2615 mg->mg_ptr[mg->mg_len] = 0;
2616 }
2617 PL_reginterp_cnt += re->program[0].next_off;
2618
2619 if (re->reganch & ROPT_UTF8)
2620 SvUTF8_on(sv);
2621 else
2622 SvUTF8_off(sv);
2623 if (lp)
2624 *lp = mg->mg_len;
2625 return mg->mg_ptr;
2626}
2627
645c22ef
DM
2628/*
2629=for apidoc sv_2pv_flags
2630
ff276b08 2631Returns a pointer to the string value of an SV, and sets *lp to its length.
645c22ef
DM
2632If flags includes SV_GMAGIC, does an mg_get() first. Coerces sv to a string
2633if necessary.
2634Normally invoked via the C<SvPV_flags> macro. C<sv_2pv()> and C<sv_2pv_nomg>
2635usually end up here too.
2636
2637=cut
2638*/
2639
8d6d96c1
HS
2640char *
2641Perl_sv_2pv_flags(pTHX_ register SV *sv, STRLEN *lp, I32 flags)
2642{
97aff369 2643 dVAR;
79072805 2644 register char *s;
79072805 2645
463ee0b2 2646 if (!sv) {
cdb061a3
NC
2647 if (lp)
2648 *lp = 0;
73d840c0 2649 return (char *)"";
463ee0b2 2650 }
8990e307 2651 if (SvGMAGICAL(sv)) {
8d6d96c1
HS
2652 if (flags & SV_GMAGIC)
2653 mg_get(sv);
463ee0b2 2654 if (SvPOKp(sv)) {
cdb061a3
NC
2655 if (lp)
2656 *lp = SvCUR(sv);
10516c54
NC
2657 if (flags & SV_MUTABLE_RETURN)
2658 return SvPVX_mutable(sv);
4d84ee25
NC
2659 if (flags & SV_CONST_RETURN)
2660 return (char *)SvPVX_const(sv);
463ee0b2
LW
2661 return SvPVX(sv);
2662 }
75dfc8ec
NC
2663 if (SvIOKp(sv) || SvNOKp(sv)) {
2664 char tbuf[64]; /* Must fit sprintf/Gconvert of longest IV/NV */
75dfc8ec
NC
2665 STRLEN len;
2666
2667 if (SvIOKp(sv)) {
e80fed9d 2668 len = SvIsUV(sv)
d9fad198
JH
2669 ? my_snprintf(tbuf, sizeof(tbuf), "%"UVuf, (UV)SvUVX(sv))
2670 : my_snprintf(tbuf, sizeof(tbuf), "%"IVdf, (IV)SvIVX(sv));
75dfc8ec 2671 } else {
e8ada2d0
NC
2672 Gconvert(SvNVX(sv), NV_DIG, 0, tbuf);
2673 len = strlen(tbuf);
75dfc8ec 2674 }
b5b886f0
NC
2675 assert(!SvROK(sv));
2676 {
75dfc8ec
NC
2677 dVAR;
2678
2679#ifdef FIXNEGATIVEZERO
e8ada2d0
NC
2680 if (len == 2 && tbuf[0] == '-' && tbuf[1] == '0') {
2681 tbuf[0] = '0';
2682 tbuf[1] = 0;
75dfc8ec
NC
2683 len = 1;
2684 }
2685#endif
2686 SvUPGRADE(sv, SVt_PV);
2687 if (lp)
2688 *lp = len;
2689 s = SvGROW_mutable(sv, len + 1);
2690 SvCUR_set(sv, len);
2691 SvPOKp_on(sv);
10edeb5d 2692 return (char*)memcpy(s, tbuf, len + 1);
75dfc8ec 2693 }
463ee0b2 2694 }
1c7ff15e
NC
2695 if (SvROK(sv)) {
2696 goto return_rok;
2697 }
2698 assert(SvTYPE(sv) >= SVt_PVMG);
2699 /* This falls through to the report_uninit near the end of the
2700 function. */
2701 } else if (SvTHINKFIRST(sv)) {
ed6116ce 2702 if (SvROK(sv)) {
1c7ff15e 2703 return_rok:
deb46114
NC
2704 if (SvAMAGIC(sv)) {
2705 SV *const tmpstr = AMG_CALLun(sv,string);
2706 if (tmpstr && (!SvROK(tmpstr) || (SvRV(tmpstr) != SvRV(sv)))) {
2707 /* Unwrap this: */
2708 /* char *pv = lp ? SvPV(tmpstr, *lp) : SvPV_nolen(tmpstr);
2709 */
2710
2711 char *pv;
2712 if ((SvFLAGS(tmpstr) & (SVf_POK)) == SVf_POK) {
2713 if (flags & SV_CONST_RETURN) {
2714 pv = (char *) SvPVX_const(tmpstr);
2715 } else {
2716 pv = (flags & SV_MUTABLE_RETURN)
2717 ? SvPVX_mutable(tmpstr) : SvPVX(tmpstr);
2718 }
2719 if (lp)
2720 *lp = SvCUR(tmpstr);
50adf7d2 2721 } else {
deb46114 2722 pv = sv_2pv_flags(tmpstr, lp, flags);
50adf7d2 2723 }
deb46114
NC
2724 if (SvUTF8(tmpstr))
2725 SvUTF8_on(sv);
2726 else
2727 SvUTF8_off(sv);
2728 return pv;
50adf7d2 2729 }
deb46114
NC
2730 }
2731 {
75dfc8ec 2732 SV *tsv;
f9277f47 2733 MAGIC *mg;
d8eae41e
NC
2734 const SV *const referent = (SV*)SvRV(sv);
2735
2736 if (!referent) {
396482e1 2737 tsv = sv_2mortal(newSVpvs("NULLREF"));
042dae7a
NC
2738 } else if (SvTYPE(referent) == SVt_PVMG
2739 && ((SvFLAGS(referent) &
2740 (SVs_OBJECT|SVf_OK|SVs_GMG|SVs_SMG|SVs_RMG))
2741 == (SVs_OBJECT|SVs_SMG))
2742 && (mg = mg_find(referent, PERL_MAGIC_qr))) {
c445ea15 2743 return stringify_regexp(sv, mg, lp);
d8eae41e
NC
2744 } else {
2745 const char *const typestr = sv_reftype(referent, 0);
2746
2747 tsv = sv_newmortal();
2748 if (SvOBJECT(referent)) {
2749 const char *const name = HvNAME_get(SvSTASH(referent));
2750 Perl_sv_setpvf(aTHX_ tsv, "%s=%s(0x%"UVxf")",
2751 name ? name : "__ANON__" , typestr,
2752 PTR2UV(referent));
2753 }
2754 else
2755 Perl_sv_setpvf(aTHX_ tsv, "%s(0x%"UVxf")", typestr,
2756 PTR2UV(referent));
c080367d 2757 }
042dae7a
NC
2758 if (lp)
2759 *lp = SvCUR(tsv);
2760 return SvPVX(tsv);
463ee0b2 2761 }
79072805 2762 }
0336b60e 2763 if (SvREADONLY(sv) && !SvOK(sv)) {
0336b60e 2764 if (ckWARN(WARN_UNINITIALIZED))
29489e7c 2765 report_uninit(sv);
cdb061a3
NC
2766 if (lp)
2767 *lp = 0;
73d840c0 2768 return (char *)"";
79072805 2769 }
79072805 2770 }
28e5dec8
JH
2771 if (SvIOK(sv) || ((SvIOKp(sv) && !SvNOKp(sv)))) {
2772 /* I'm assuming that if both IV and NV are equally valid then
2773 converting the IV is going to be more efficient */
e1ec3a88
AL
2774 const U32 isIOK = SvIOK(sv);
2775 const U32 isUIOK = SvIsUV(sv);
28e5dec8
JH
2776 char buf[TYPE_CHARS(UV)];
2777 char *ebuf, *ptr;
2778
2779 if (SvTYPE(sv) < SVt_PVIV)
2780 sv_upgrade(sv, SVt_PVIV);
4ea1d550 2781 ptr = uiv_2buf(buf, SvIVX(sv), SvUVX(sv), isUIOK, &ebuf);
5902b6a9
NC
2782 /* inlined from sv_setpvn */
2783 SvGROW_mutable(sv, (STRLEN)(ebuf - ptr + 1));
4d84ee25 2784 Move(ptr,SvPVX_mutable(sv),ebuf - ptr,char);
28e5dec8
JH
2785 SvCUR_set(sv, ebuf - ptr);
2786 s = SvEND(sv);
2787 *s = '\0';
2788 if (isIOK)
2789 SvIOK_on(sv);
2790 else
2791 SvIOKp_on(sv);
2792 if (isUIOK)
2793 SvIsUV_on(sv);
2794 }
2795 else if (SvNOKp(sv)) {
c81271c3 2796 const int olderrno = errno;
79072805
LW
2797 if (SvTYPE(sv) < SVt_PVNV)
2798 sv_upgrade(sv, SVt_PVNV);
1c846c1f 2799 /* The +20 is pure guesswork. Configure test needed. --jhi */
5902b6a9 2800 s = SvGROW_mutable(sv, NV_DIG + 20);
c81271c3 2801 /* some Xenix systems wipe out errno here */
79072805 2802#ifdef apollo
463ee0b2 2803 if (SvNVX(sv) == 0.0)
d1307786 2804 my_strlcpy(s, "0", SvLEN(sv));
79072805
LW
2805 else
2806#endif /*apollo*/
bbce6d69 2807 {
2d4389e4 2808 Gconvert(SvNVX(sv), NV_DIG, 0, s);
bbce6d69 2809 }
79072805 2810 errno = olderrno;
a0d0e21e
LW
2811#ifdef FIXNEGATIVEZERO
2812 if (*s == '-' && s[1] == '0' && !s[2])
d1307786 2813 my_strlcpy(s, "0", SvLEN(s));
a0d0e21e 2814#endif
79072805
LW
2815 while (*s) s++;
2816#ifdef hcx
2817 if (s[-1] == '.')
46fc3d4c 2818 *--s = '\0';
79072805
LW
2819#endif
2820 }
79072805 2821 else {
675c862f 2822 if (isGV_with_GP(sv))
19f6321d 2823 return glob_2pv((GV *)sv, lp);
180488f8 2824
041457d9 2825 if (!PL_localizing && !(SvFLAGS(sv) & SVs_PADTMP) && ckWARN(WARN_UNINITIALIZED))
29489e7c 2826 report_uninit(sv);
cdb061a3 2827 if (lp)
00b6aa41 2828 *lp = 0;
25da4f38
IZ
2829 if (SvTYPE(sv) < SVt_PV)
2830 /* Typically the caller expects that sv_any is not NULL now. */
2831 sv_upgrade(sv, SVt_PV);
73d840c0 2832 return (char *)"";
79072805 2833 }
cdb061a3 2834 {
823a54a3 2835 const STRLEN len = s - SvPVX_const(sv);
cdb061a3
NC
2836 if (lp)
2837 *lp = len;
2838 SvCUR_set(sv, len);
2839 }
79072805 2840 SvPOK_on(sv);
1d7c1841 2841 DEBUG_c(PerlIO_printf(Perl_debug_log, "0x%"UVxf" 2pv(%s)\n",
3f7c398e 2842 PTR2UV(sv),SvPVX_const(sv)));
4d84ee25
NC
2843 if (flags & SV_CONST_RETURN)
2844 return (char *)SvPVX_const(sv);
10516c54
NC
2845 if (flags & SV_MUTABLE_RETURN)
2846 return SvPVX_mutable(sv);
463ee0b2
LW
2847 return SvPVX(sv);
2848}
2849
645c22ef 2850/*
6050d10e
JP
2851=for apidoc sv_copypv
2852
2853Copies a stringified representation of the source SV into the
2854destination SV. Automatically performs any necessary mg_get and
54f0641b 2855coercion of numeric values into strings. Guaranteed to preserve
6050d10e 2856UTF-8 flag even from overloaded objects. Similar in nature to
54f0641b
NIS
2857sv_2pv[_flags] but operates directly on an SV instead of just the
2858string. Mostly uses sv_2pv_flags to do its work, except when that
6050d10e
JP
2859would lose the UTF-8'ness of the PV.
2860
2861=cut
2862*/
2863
2864void
2865Perl_sv_copypv(pTHX_ SV *dsv, register SV *ssv)
2866{
446eaa42 2867 STRLEN len;
53c1dcc0 2868 const char * const s = SvPV_const(ssv,len);
cb50f42d 2869 sv_setpvn(dsv,s,len);
446eaa42 2870 if (SvUTF8(ssv))
cb50f42d 2871 SvUTF8_on(dsv);
446eaa42 2872 else
cb50f42d 2873 SvUTF8_off(dsv);
6050d10e
JP
2874}
2875
2876/*
645c22ef
DM
2877=for apidoc sv_2pvbyte
2878
2879Return a pointer to the byte-encoded representation of the SV, and set *lp
1e54db1a 2880to its length. May cause the SV to be downgraded from UTF-8 as a
645c22ef
DM
2881side-effect.
2882
2883Usually accessed via the C<SvPVbyte> macro.
2884
2885=cut
2886*/
2887
7340a771
GS
2888char *
2889Perl_sv_2pvbyte(pTHX_ register SV *sv, STRLEN *lp)
2890{
0875d2fe 2891 sv_utf8_downgrade(sv,0);
97972285 2892 return lp ? SvPV(sv,*lp) : SvPV_nolen(sv);
7340a771
GS
2893}
2894
645c22ef 2895/*
035cbb0e
RGS
2896=for apidoc sv_2pvutf8
2897
2898Return a pointer to the UTF-8-encoded representation of the SV, and set *lp
2899to its length. May cause the SV to be upgraded to UTF-8 as a side-effect.
2900
2901Usually accessed via the C<SvPVutf8> macro.
2902
2903=cut
2904*/
645c22ef 2905
7340a771
GS
2906char *
2907Perl_sv_2pvutf8(pTHX_ register SV *sv, STRLEN *lp)
2908{
035cbb0e
RGS
2909 sv_utf8_upgrade(sv);
2910 return lp ? SvPV(sv,*lp) : SvPV_nolen(sv);
7340a771 2911}
1c846c1f 2912
7ee2227d 2913
645c22ef
DM
2914/*
2915=for apidoc sv_2bool
2916
2917This function is only called on magical items, and is only used by
8cf8f3d1 2918sv_true() or its macro equivalent.
645c22ef
DM
2919
2920=cut
2921*/
2922
463ee0b2 2923bool
864dbfa3 2924Perl_sv_2bool(pTHX_ register SV *sv)
463ee0b2 2925{
97aff369 2926 dVAR;
5b295bef 2927 SvGETMAGIC(sv);
463ee0b2 2928
a0d0e21e
LW
2929 if (!SvOK(sv))
2930 return 0;
2931 if (SvROK(sv)) {
fabdb6c0
AL
2932 if (SvAMAGIC(sv)) {
2933 SV * const tmpsv = AMG_CALLun(sv,bool_);
2934 if (tmpsv && (!SvROK(tmpsv) || (SvRV(tmpsv) != SvRV(sv))))
2935 return (bool)SvTRUE(tmpsv);
2936 }
2937 return SvRV(sv) != 0;
a0d0e21e 2938 }
463ee0b2 2939 if (SvPOKp(sv)) {
53c1dcc0
AL
2940 register XPV* const Xpvtmp = (XPV*)SvANY(sv);
2941 if (Xpvtmp &&
339049b0 2942 (*sv->sv_u.svu_pv > '0' ||
11343788 2943 Xpvtmp->xpv_cur > 1 ||
339049b0 2944 (Xpvtmp->xpv_cur && *sv->sv_u.svu_pv != '0')))
463ee0b2
LW
2945 return 1;
2946 else
2947 return 0;
2948 }
2949 else {
2950 if (SvIOKp(sv))
2951 return SvIVX(sv) != 0;
2952 else {
2953 if (SvNOKp(sv))
2954 return SvNVX(sv) != 0.0;
180488f8 2955 else {
f7877b28 2956 if (isGV_with_GP(sv))
180488f8
NC
2957 return TRUE;
2958 else
2959 return FALSE;
2960 }
463ee0b2
LW
2961 }
2962 }
79072805
LW
2963}
2964
c461cf8f
JH
2965/*
2966=for apidoc sv_utf8_upgrade
2967
78ea37eb 2968Converts the PV of an SV to its UTF-8-encoded form.
645c22ef 2969Forces the SV to string form if it is not already.
4411f3b6
NIS
2970Always sets the SvUTF8 flag to avoid future validity checks even
2971if all the bytes have hibit clear.
c461cf8f 2972
13a6c0e0
JH
2973This is not as a general purpose byte encoding to Unicode interface:
2974use the Encode extension for that.
2975
8d6d96c1
HS
2976=for apidoc sv_utf8_upgrade_flags
2977
78ea37eb 2978Converts the PV of an SV to its UTF-8-encoded form.
645c22ef 2979Forces the SV to string form if it is not already.
8d6d96c1
HS
2980Always sets the SvUTF8 flag to avoid future validity checks even
2981if all the bytes have hibit clear. If C<flags> has C<SV_GMAGIC> bit set,
2982will C<mg_get> on C<sv> if appropriate, else not. C<sv_utf8_upgrade> and
2983C<sv_utf8_upgrade_nomg> are implemented in terms of this function.
2984
13a6c0e0
JH
2985This is not as a general purpose byte encoding to Unicode interface:
2986use the Encode extension for that.
2987
8d6d96c1
HS
2988=cut
2989*/
2990
2991STRLEN
2992Perl_sv_utf8_upgrade_flags(pTHX_ register SV *sv, I32 flags)
2993{
97aff369 2994 dVAR;
808c356f
RGS
2995 if (sv == &PL_sv_undef)
2996 return 0;
e0e62c2a
NIS
2997 if (!SvPOK(sv)) {
2998 STRLEN len = 0;
d52b7888
NC
2999 if (SvREADONLY(sv) && (SvPOKp(sv) || SvIOKp(sv) || SvNOKp(sv))) {
3000 (void) sv_2pv_flags(sv,&len, flags);
3001 if (SvUTF8(sv))
3002 return len;
3003 } else {
3004 (void) SvPV_force(sv,len);
3005 }
e0e62c2a 3006 }
4411f3b6 3007
f5cee72b 3008 if (SvUTF8(sv)) {
5fec3b1d 3009 return SvCUR(sv);
f5cee72b 3010 }
5fec3b1d 3011
765f542d
NC
3012 if (SvIsCOW(sv)) {
3013 sv_force_normal_flags(sv, 0);
db42d148
NIS
3014 }
3015
88632417 3016 if (PL_encoding && !(flags & SV_UTF8_NO_ENCODING))
799ef3cb 3017 sv_recode_to_utf8(sv, PL_encoding);
9f4817db 3018 else { /* Assume Latin-1/EBCDIC */
c4e7c712
NC
3019 /* This function could be much more efficient if we
3020 * had a FLAG in SVs to signal if there are any hibit
3021 * chars in the PV. Given that there isn't such a flag
3022 * make the loop as fast as possible. */
00b6aa41 3023 const U8 * const s = (U8 *) SvPVX_const(sv);
c4420975 3024 const U8 * const e = (U8 *) SvEND(sv);
93524f2b 3025 const U8 *t = s;
c4e7c712
NC
3026
3027 while (t < e) {
53c1dcc0 3028 const U8 ch = *t++;
00b6aa41
AL
3029 /* Check for hi bit */
3030 if (!NATIVE_IS_INVARIANT(ch)) {
3031 STRLEN len = SvCUR(sv) + 1; /* Plus the \0 */
3032 U8 * const recoded = bytes_to_utf8((U8*)s, &len);
3033
3034 SvPV_free(sv); /* No longer using what was there before. */
3035 SvPV_set(sv, (char*)recoded);
3036 SvCUR_set(sv, len - 1);
3037 SvLEN_set(sv, len); /* No longer know the real size. */
c4e7c712 3038 break;
00b6aa41 3039 }
c4e7c712
NC
3040 }
3041 /* Mark as UTF-8 even if no hibit - saves scanning loop */
3042 SvUTF8_on(sv);
560a288e 3043 }
4411f3b6 3044 return SvCUR(sv);
560a288e
GS
3045}
3046
c461cf8f
JH
3047/*
3048=for apidoc sv_utf8_downgrade
3049
78ea37eb
TS
3050Attempts to convert the PV of an SV from characters to bytes.
3051If the PV contains a character beyond byte, this conversion will fail;
3052in this case, either returns false or, if C<fail_ok> is not
c461cf8f
JH
3053true, croaks.
3054
13a6c0e0
JH
3055This is not as a general purpose Unicode to byte encoding interface:
3056use the Encode extension for that.
3057
c461cf8f
JH
3058=cut
3059*/
3060
560a288e
GS
3061bool
3062Perl_sv_utf8_downgrade(pTHX_ register SV* sv, bool fail_ok)
3063{
97aff369 3064 dVAR;
78ea37eb 3065 if (SvPOKp(sv) && SvUTF8(sv)) {
fa301091 3066 if (SvCUR(sv)) {
03cfe0ae 3067 U8 *s;
652088fc 3068 STRLEN len;
fa301091 3069
765f542d
NC
3070 if (SvIsCOW(sv)) {
3071 sv_force_normal_flags(sv, 0);
3072 }
03cfe0ae
NIS
3073 s = (U8 *) SvPV(sv, len);
3074 if (!utf8_to_bytes(s, &len)) {
fa301091
JH
3075 if (fail_ok)
3076 return FALSE;
3077 else {
3078 if (PL_op)
3079 Perl_croak(aTHX_ "Wide character in %s",
53e06cf0 3080 OP_DESC(PL_op));
fa301091
JH
3081 else
3082 Perl_croak(aTHX_ "Wide character");
3083 }
4b3603a4 3084 }
b162af07 3085 SvCUR_set(sv, len);
67e989fb 3086 }
560a288e 3087 }
ffebcc3e 3088 SvUTF8_off(sv);
560a288e
GS
3089 return TRUE;
3090}
3091
c461cf8f
JH
3092/*
3093=for apidoc sv_utf8_encode
3094
78ea37eb
TS
3095Converts the PV of an SV to UTF-8, but then turns the C<SvUTF8>
3096flag off so that it looks like octets again.
c461cf8f
JH
3097
3098=cut
3099*/
3100
560a288e
GS
3101void
3102Perl_sv_utf8_encode(pTHX_ register SV *sv)
3103{
4c94c214
NC
3104 if (SvIsCOW(sv)) {
3105 sv_force_normal_flags(sv, 0);
3106 }
3107 if (SvREADONLY(sv)) {
3108 Perl_croak(aTHX_ PL_no_modify);
3109 }
a5f5288a 3110 (void) sv_utf8_upgrade(sv);
560a288e
GS
3111 SvUTF8_off(sv);
3112}
3113
4411f3b6
NIS
3114/*
3115=for apidoc sv_utf8_decode
3116
78ea37eb
TS
3117If the PV of the SV is an octet sequence in UTF-8
3118and contains a multiple-byte character, the C<SvUTF8> flag is turned on
3119so that it looks like a character. If the PV contains only single-byte
3120characters, the C<SvUTF8> flag stays being off.
3121Scans PV for validity and returns false if the PV is invalid UTF-8.
4411f3b6
NIS
3122
3123=cut
3124*/
3125
560a288e
GS
3126bool
3127Perl_sv_utf8_decode(pTHX_ register SV *sv)
3128{
78ea37eb 3129 if (SvPOKp(sv)) {
93524f2b
NC
3130 const U8 *c;
3131 const U8 *e;
9cbac4c7 3132
645c22ef
DM
3133 /* The octets may have got themselves encoded - get them back as
3134 * bytes
3135 */
3136 if (!sv_utf8_downgrade(sv, TRUE))
560a288e
GS
3137 return FALSE;
3138
3139 /* it is actually just a matter of turning the utf8 flag on, but
3140 * we want to make sure everything inside is valid utf8 first.
3141 */
93524f2b 3142 c = (const U8 *) SvPVX_const(sv);
63cd0674 3143 if (!is_utf8_string(c, SvCUR(sv)+1))
67e989fb 3144 return FALSE;
93524f2b 3145 e = (const U8 *) SvEND(sv);
511c2ff0 3146 while (c < e) {
b64e5050 3147 const U8 ch = *c++;
c4d5f83a 3148 if (!UTF8_IS_INVARIANT(ch)) {
67e989fb
JH
3149 SvUTF8_on(sv);
3150 break;
3151 }
560a288e 3152 }
560a288e
GS
3153 }
3154 return TRUE;
3155}
3156
954c1994
GS
3157/*
3158=for apidoc sv_setsv
3159
645c22ef
DM
3160Copies the contents of the source SV C<ssv> into the destination SV
3161C<dsv>. The source SV may be destroyed if it is mortal, so don't use this
3162function if the source SV needs to be reused. Does not handle 'set' magic.
3163Loosely speaking, it performs a copy-by-value, obliterating any previous
3164content of the destination.
3165
3166You probably want to use one of the assortment of wrappers, such as
3167C<SvSetSV>, C<SvSetSV_nosteal>, C<SvSetMagicSV> and
3168C<SvSetMagicSV_nosteal>.
3169
8d6d96c1
HS
3170=for apidoc sv_setsv_flags
3171
645c22ef
DM
3172Copies the contents of the source SV C<ssv> into the destination SV
3173C<dsv>. The source SV may be destroyed if it is mortal, so don't use this
3174function if the source SV needs to be reused. Does not handle 'set' magic.
3175Loosely speaking, it performs a copy-by-value, obliterating any previous
3176content of the destination.
3177If the C<flags> parameter has the C<SV_GMAGIC> bit set, will C<mg_get> on
5fcdf167
NC
3178C<ssv> if appropriate, else not. If the C<flags> parameter has the
3179C<NOSTEAL> bit set then the buffers of temps will not be stolen. <sv_setsv>
3180and C<sv_setsv_nomg> are implemented in terms of this function.
645c22ef
DM
3181
3182You probably want to use one of the assortment of wrappers, such as
3183C<SvSetSV>, C<SvSetSV_nosteal>, C<SvSetMagicSV> and
3184C<SvSetMagicSV_nosteal>.
3185
3186This is the primary function for copying scalars, and most other
3187copy-ish functions and macros use this underneath.
8d6d96c1
HS
3188
3189=cut
3190*/
3191
5d0301b7 3192static void
2eb42952 3193S_glob_assign_glob(pTHX_ SV *dstr, SV *sstr, const int dtype)
5d0301b7
NC
3194{
3195 if (dtype != SVt_PVGV) {
3196 const char * const name = GvNAME(sstr);
3197 const STRLEN len = GvNAMELEN(sstr);
3198 /* don't upgrade SVt_PVLV: it can hold a glob */
f7877b28
NC
3199 if (dtype != SVt_PVLV) {
3200 if (dtype >= SVt_PV) {
3201 SvPV_free(dstr);
3202 SvPV_set(dstr, 0);
3203 SvLEN_set(dstr, 0);
3204 SvCUR_set(dstr, 0);
3205 }
5d0301b7 3206 sv_upgrade(dstr, SVt_PVGV);
dedf8e73
NC
3207 (void)SvOK_off(dstr);
3208 SvSCREAM_on(dstr);
f7877b28 3209 }
5d0301b7
NC
3210 GvSTASH(dstr) = GvSTASH(sstr);
3211 if (GvSTASH(dstr))
3212 Perl_sv_add_backref(aTHX_ (SV*)GvSTASH(dstr), dstr);
ae8cc45f 3213 gv_name_set((GV *)dstr, name, len, GV_ADD);
5d0301b7
NC
3214 SvFAKE_on(dstr); /* can coerce to non-glob */
3215 }
3216
3217#ifdef GV_UNIQUE_CHECK
3218 if (GvUNIQUE((GV*)dstr)) {
3219 Perl_croak(aTHX_ PL_no_modify);
3220 }
3221#endif
3222
f7877b28
NC
3223 gp_free((GV*)dstr);
3224 SvSCREAM_off(dstr);
5d0301b7 3225 (void)SvOK_off(dstr);
f7877b28 3226 SvSCREAM_on(dstr);
dedf8e73 3227 GvINTRO_off(dstr); /* one-shot flag */
5d0301b7
NC
3228 GvGP(dstr) = gp_ref(GvGP(sstr));
3229 if (SvTAINTED(sstr))
3230 SvTAINT(dstr);
3231 if (GvIMPORTED(dstr) != GVf_IMPORTED
3232 && CopSTASH_ne(PL_curcop, GvSTASH(dstr)))
3233 {
3234 GvIMPORTED_on(dstr);
3235 }
3236 GvMULTI_on(dstr);
3237 return;
3238}
3239
b8473700 3240static void
2eb42952 3241S_glob_assign_ref(pTHX_ SV *dstr, SV *sstr) {
b8473700
NC
3242 SV * const sref = SvREFCNT_inc(SvRV(sstr));
3243 SV *dref = NULL;
3244 const int intro = GvINTRO(dstr);
2440974c 3245 SV **location;
3386d083 3246 U8 import_flag = 0;
27242d61
NC
3247 const U32 stype = SvTYPE(sref);
3248
b8473700
NC
3249
3250#ifdef GV_UNIQUE_CHECK
3251 if (GvUNIQUE((GV*)dstr)) {
3252 Perl_croak(aTHX_ PL_no_modify);
3253 }
3254#endif
3255
3256 if (intro) {
3257 GvINTRO_off(dstr); /* one-shot flag */
3258 GvLINE(dstr) = CopLINE(PL_curcop);
3259 GvEGV(dstr) = (GV*)dstr;
3260 }
3261 GvMULTI_on(dstr);
27242d61 3262 switch (stype) {
b8473700 3263 case SVt_PVCV:
27242d61
NC
3264 location = (SV **) &GvCV(dstr);
3265 import_flag = GVf_IMPORTED_CV;
3266 goto common;
3267 case SVt_PVHV:
3268 location = (SV **) &GvHV(dstr);
3269 import_flag = GVf_IMPORTED_HV;
3270 goto common;
3271 case SVt_PVAV:
3272 location = (SV **) &GvAV(dstr);
3273 import_flag = GVf_IMPORTED_AV;
3274 goto common;
3275 case SVt_PVIO:
3276 location = (SV **) &GvIOp(dstr);
3277 goto common;
3278 case SVt_PVFM:
3279 location = (SV **) &GvFORM(dstr);
3280 default:
3281 location = &GvSV(dstr);
3282 import_flag = GVf_IMPORTED_SV;
3283 common:
b8473700 3284 if (intro) {
27242d61
NC
3285 if (stype == SVt_PVCV) {
3286 if (GvCVGEN(dstr) && GvCV(dstr) != (CV*)sref) {
3287 SvREFCNT_dec(GvCV(dstr));
3288 GvCV(dstr) = NULL;
3289 GvCVGEN(dstr) = 0; /* Switch off cacheness. */
3290 PL_sub_generation++;
3291 }
b8473700 3292 }
27242d61 3293 SAVEGENERICSV(*location);
b8473700
NC
3294 }
3295 else
27242d61
NC
3296 dref = *location;
3297 if (stype == SVt_PVCV && *location != sref) {
3298 CV* const cv = (CV*)*location;
b8473700
NC
3299 if (cv) {
3300 if (!GvCVGEN((GV*)dstr) &&
3301 (CvROOT(cv) || CvXSUB(cv)))
3302 {
3303 /* Redefining a sub - warning is mandatory if
3304 it was a const and its value changed. */
3305 if (CvCONST(cv) && CvCONST((CV*)sref)
3306 && cv_const_sv(cv) == cv_const_sv((CV*)sref)) {
6f207bd3 3307 NOOP;
b8473700
NC
3308 /* They are 2 constant subroutines generated from
3309 the same constant. This probably means that
3310 they are really the "same" proxy subroutine
3311 instantiated in 2 places. Most likely this is
3312 when a constant is exported twice. Don't warn.
3313 */
3314 }
3315 else if (ckWARN(WARN_REDEFINE)
3316 || (CvCONST(cv)
3317 && (!CvCONST((CV*)sref)
3318 || sv_cmp(cv_const_sv(cv),
3319 cv_const_sv((CV*)sref))))) {
3320 Perl_warner(aTHX_ packWARN(WARN_REDEFINE),
10edeb5d
JH
3321 (const char *)
3322 (CvCONST(cv)
3323 ? "Constant subroutine %s::%s redefined"
3324 : "Subroutine %s::%s redefined"),
b8473700
NC
3325 HvNAME_get(GvSTASH((GV*)dstr)),
3326 GvENAME((GV*)dstr));
3327 }
3328 }
3329 if (!intro)
cbf82dd0
NC
3330 cv_ckproto_len(cv, (GV*)dstr,
3331 SvPOK(sref) ? SvPVX_const(sref) : NULL,
3332 SvPOK(sref) ? SvCUR(sref) : 0);
b8473700 3333 }
b8473700
NC
3334 GvCVGEN(dstr) = 0; /* Switch off cacheness. */
3335 GvASSUMECV_on(dstr);
3336 PL_sub_generation++;
3337 }
2440974c 3338 *location = sref;
3386d083
NC
3339 if (import_flag && !(GvFLAGS(dstr) & import_flag)
3340 && CopSTASH_ne(PL_curcop, GvSTASH(dstr))) {
3341 GvFLAGS(dstr) |= import_flag;
b8473700
NC
3342 }
3343 break;
3344 }
b37c2d43 3345 SvREFCNT_dec(dref);
b8473700
NC
3346 if (SvTAINTED(sstr))
3347 SvTAINT(dstr);
3348 return;
3349}
3350
8d6d96c1
HS
3351void
3352Perl_sv_setsv_flags(pTHX_ SV *dstr, register SV *sstr, I32 flags)
3353{
97aff369 3354 dVAR;
8990e307
LW
3355 register U32 sflags;
3356 register int dtype;
42d0e0b7 3357 register svtype stype;
463ee0b2 3358
79072805
LW
3359 if (sstr == dstr)
3360 return;
765f542d 3361 SV_CHECK_THINKFIRST_COW_DROP(dstr);
79072805 3362 if (!sstr)
3280af22 3363 sstr = &PL_sv_undef;
8990e307
LW
3364 stype = SvTYPE(sstr);
3365 dtype = SvTYPE(dstr);
79072805 3366
a0d0e21e 3367 SvAMAGIC_off(dstr);
7a5fa8a2 3368 if ( SvVOK(dstr) )
ece467f9
JP
3369 {
3370 /* need to nuke the magic */
3371 mg_free(dstr);
3372 SvRMAGICAL_off(dstr);
3373 }
9e7bc3e8 3374
463ee0b2 3375 /* There's a lot of redundancy below but we're going for speed here */
79072805 3376
8990e307 3377 switch (stype) {
79072805 3378 case SVt_NULL:
aece5585 3379 undef_sstr:
20408e3c
GS
3380 if (dtype != SVt_PVGV) {
3381 (void)SvOK_off(dstr);
3382 return;
3383 }
3384 break;
463ee0b2 3385 case SVt_IV:
aece5585
GA
3386 if (SvIOK(sstr)) {
3387 switch (dtype) {
3388 case SVt_NULL:
8990e307 3389 sv_upgrade(dstr, SVt_IV);
aece5585
GA
3390 break;
3391 case SVt_NV:
aece5585
GA
3392 case SVt_RV:
3393 case SVt_PV:
a0d0e21e 3394 sv_upgrade(dstr, SVt_PVIV);
aece5585
GA
3395 break;
3396 }
3397 (void)SvIOK_only(dstr);
45977657 3398 SvIV_set(dstr, SvIVX(sstr));
25da4f38
IZ
3399 if (SvIsUV(sstr))
3400 SvIsUV_on(dstr);
37c25af0
NC
3401 /* SvTAINTED can only be true if the SV has taint magic, which in
3402 turn means that the SV type is PVMG (or greater). This is the
3403 case statement for SVt_IV, so this cannot be true (whatever gcov
3404 may say). */
3405 assert(!SvTAINTED(sstr));
aece5585 3406 return;
8990e307 3407 }
aece5585
GA
3408 goto undef_sstr;
3409
463ee0b2 3410 case SVt_NV:
aece5585
GA
3411 if (SvNOK(sstr)) {
3412 switch (dtype) {
3413 case SVt_NULL:
3414 case SVt_IV:
8990e307 3415 sv_upgrade(dstr, SVt_NV);
aece5585
GA
3416 break;
3417 case SVt_RV:
3418 case SVt_PV:
3419 case SVt_PVIV:
a0d0e21e 3420 sv_upgrade(dstr, SVt_PVNV);
aece5585
GA
3421 break;
3422 }
9d6ce603 3423 SvNV_set(dstr, SvNVX(sstr));
aece5585 3424 (void)SvNOK_only(dstr);
37c25af0
NC
3425 /* SvTAINTED can only be true if the SV has taint magic, which in
3426 turn means that the SV type is PVMG (or greater). This is the
3427 case statement for SVt_NV, so this cannot be true (whatever gcov
3428 may say). */
3429 assert(!SvTAINTED(sstr));
aece5585 3430 return;
8990e307 3431 }
aece5585
GA
3432 goto undef_sstr;
3433
ed6116ce 3434 case SVt_RV:
8990e307 3435 if (dtype < SVt_RV)
ed6116ce 3436 sv_upgrade(dstr, SVt_RV);
ed6116ce 3437 break;
fc36a67e 3438 case SVt_PVFM:
f8c7b90f 3439#ifdef PERL_OLD_COPY_ON_WRITE
d89fc664
NC
3440 if ((SvFLAGS(sstr) & CAN_COW_MASK) == CAN_COW_FLAGS) {
3441 if (dtype < SVt_PVIV)
3442 sv_upgrade(dstr, SVt_PVIV);
3443 break;
3444 }
3445 /* Fall through */
3446#endif
3447 case SVt_PV:
8990e307 3448 if (dtype < SVt_PV)
463ee0b2 3449 sv_upgrade(dstr, SVt_PV);
463ee0b2
LW
3450 break;
3451 case SVt_PVIV:
8990e307 3452 if (dtype < SVt_PVIV)
463ee0b2 3453 sv_upgrade(dstr, SVt_PVIV);
463ee0b2
LW
3454 break;
3455 case SVt_PVNV:
8990e307 3456 if (dtype < SVt_PVNV)
463ee0b2 3457 sv_upgrade(dstr, SVt_PVNV);
463ee0b2 3458 break;
489f7bfe 3459 default:
a3b680e6
AL
3460 {
3461 const char * const type = sv_reftype(sstr,0);
533c011a 3462 if (PL_op)
a3b680e6 3463 Perl_croak(aTHX_ "Bizarre copy of %s in %s", type, OP_NAME(PL_op));
4633a7c4 3464 else
a3b680e6
AL
3465 Perl_croak(aTHX_ "Bizarre copy of %s", type);
3466 }
4633a7c4
LW
3467 break;
3468
79072805 3469 case SVt_PVGV:
8990e307 3470 if (dtype <= SVt_PVGV) {
d4c19fe8 3471 glob_assign_glob(dstr, sstr, dtype);
b8c701c1 3472 return;
79072805 3473 }
5f66b61c 3474 /*FALLTHROUGH*/
79072805 3475
489f7bfe
NC
3476 case SVt_PVMG:
3477 case SVt_PVLV:
3478 case SVt_PVBM:
8d6d96c1 3479 if (SvGMAGICAL(sstr) && (flags & SV_GMAGIC)) {
973f89ab 3480 mg_get(sstr);
1d9c78c6 3481 if (SvTYPE(sstr) != stype) {
973f89ab 3482 stype = SvTYPE(sstr);
b8c701c1 3483 if (stype == SVt_PVGV && dtype <= SVt_PVGV) {
d4c19fe8 3484 glob_assign_glob(dstr, sstr, dtype);
b8c701c1
NC
3485 return;
3486 }
973f89ab
CS
3487 }
3488 }
ded42b9f 3489 if (stype == SVt_PVLV)
862a34c6 3490 SvUPGRADE(dstr, SVt_PVNV);
ded42b9f 3491 else
42d0e0b7 3492 SvUPGRADE(dstr, (svtype)stype);
79072805
LW
3493 }
3494
ff920335
NC
3495 /* dstr may have been upgraded. */
3496 dtype = SvTYPE(dstr);
8990e307
LW
3497 sflags = SvFLAGS(sstr);
3498
3499 if (sflags & SVf_ROK) {
acaa9288
NC
3500 if (dtype == SVt_PVGV &&
3501 SvROK(sstr) && SvTYPE(SvRV(sstr)) == SVt_PVGV) {
3502 sstr = SvRV(sstr);
3503 if (sstr == dstr) {
3504 if (GvIMPORTED(dstr) != GVf_IMPORTED
3505 && CopSTASH_ne(PL_curcop, GvSTASH(dstr)))
3506 {
3507 GvIMPORTED_on(dstr);
3508 }
3509 GvMULTI_on(dstr);
3510 return;
3511 }
d4c19fe8 3512 glob_assign_glob(dstr, sstr, dtype);
acaa9288
NC
3513 return;
3514 }
3515
8990e307 3516 if (dtype >= SVt_PV) {
b8c701c1 3517 if (dtype == SVt_PVGV) {
d4c19fe8 3518 glob_assign_ref(dstr, sstr);
b8c701c1
NC
3519 return;
3520 }
3f7c398e 3521 if (SvPVX_const(dstr)) {
8bd4d4c5 3522 SvPV_free(dstr);
b162af07
SP
3523 SvLEN_set(dstr, 0);
3524 SvCUR_set(dstr, 0);
a0d0e21e 3525 }
8990e307 3526 }
a0d0e21e 3527 (void)SvOK_off(dstr);
b162af07 3528 SvRV_set(dstr, SvREFCNT_inc(SvRV(sstr)));
dfd48732
NC
3529 SvFLAGS(dstr) |= sflags & (SVf_ROK|SVf_AMAGIC);
3530 assert(!(sflags & SVp_NOK));
3531 assert(!(sflags & SVp_IOK));
3532 assert(!(sflags & SVf_NOK));
3533 assert(!(sflags & SVf_IOK));
ed6116ce 3534 }
c0c44674
NC
3535 else if (dtype == SVt_PVGV) {
3536 if (!(sflags & SVf_OK)) {
3537 if (ckWARN(WARN_MISC))
3538 Perl_warner(aTHX_ packWARN(WARN_MISC),
3539 "Undefined value assigned to typeglob");
3540 }
3541 else {
3542 GV *gv = gv_fetchsv(sstr, GV_ADD, SVt_PVGV);
3543 if (dstr != (SV*)gv) {
3544 if (GvGP(dstr))
3545 gp_free((GV*)dstr);
3546 GvGP(dstr) = gp_ref(GvGP(gv));
3547 }
3548 }
3549 }
8990e307 3550 else if (sflags & SVp_POK) {
765f542d 3551 bool isSwipe = 0;
79072805
LW
3552
3553 /*
3554 * Check to see if we can just swipe the string. If so, it's a
3555 * possible small lose on short strings, but a big win on long ones.
3f7c398e
SP
3556 * It might even be a win on short strings if SvPVX_const(dstr)
3557 * has to be allocated and SvPVX_const(sstr) has to be freed.
79072805
LW
3558 */
3559
120fac95
NC
3560 /* Whichever path we take through the next code, we want this true,
3561 and doing it now facilitates the COW check. */
3562 (void)SvPOK_only(dstr);
3563
765f542d 3564 if (
b8f9541a
NC
3565 /* We're not already COW */
3566 ((sflags & (SVf_FAKE | SVf_READONLY)) != (SVf_FAKE | SVf_READONLY)
f8c7b90f 3567#ifndef PERL_OLD_COPY_ON_WRITE
b8f9541a
NC
3568 /* or we are, but dstr isn't a suitable target. */
3569 || (SvFLAGS(dstr) & CAN_COW_MASK) != CAN_COW_FLAGS
3570#endif
3571 )
765f542d 3572 &&
765f542d
NC
3573 !(isSwipe =
3574 (sflags & SVs_TEMP) && /* slated for free anyway? */
3575 !(sflags & SVf_OOK) && /* and not involved in OOK hack? */
5fcdf167
NC
3576 (!(flags & SV_NOSTEAL)) &&
3577 /* and we're allowed to steal temps */
765f542d
NC
3578 SvREFCNT(sstr) == 1 && /* and no other references to it? */
3579 SvLEN(sstr) && /* and really is a string */
645c22ef 3580 /* and won't be needed again, potentially */
765f542d 3581 !(PL_op && PL_op->op_type == OP_AASSIGN))
f8c7b90f 3582#ifdef PERL_OLD_COPY_ON_WRITE
765f542d 3583 && !((sflags & CAN_COW_MASK) == CAN_COW_FLAGS
120fac95 3584 && (SvFLAGS(dstr) & CAN_COW_MASK) == CAN_COW_FLAGS
765f542d
NC
3585 && SvTYPE(sstr) >= SVt_PVIV)
3586#endif
3587 ) {
3588 /* Failed the swipe test, and it's not a shared hash key either.
3589 Have to copy the string. */
3590 STRLEN len = SvCUR(sstr);
3591 SvGROW(dstr, len + 1); /* inlined from sv_setpvn */
3f7c398e 3592 Move(SvPVX_const(sstr),SvPVX(dstr),len,char);
765f542d
NC
3593 SvCUR_set(dstr, len);
3594 *SvEND(dstr) = '\0';
765f542d 3595 } else {
f8c7b90f 3596 /* If PERL_OLD_COPY_ON_WRITE is not defined, then isSwipe will always
765f542d 3597 be true in here. */
765f542d
NC
3598 /* Either it's a shared hash key, or it's suitable for
3599 copy-on-write or we can swipe the string. */
46187eeb 3600 if (DEBUG_C_TEST) {
ed252734 3601 PerlIO_printf(Perl_debug_log, "Copy on write: sstr --> dstr\n");
e419cbc5
NC
3602 sv_dump(sstr);
3603 sv_dump(dstr);
46187eeb 3604 }
f8c7b90f 3605#ifdef PERL_OLD_COPY_ON_WRITE
765f542d
NC
3606 if (!isSwipe) {
3607 /* I believe I should acquire a global SV mutex if
3608 it's a COW sv (not a shared hash key) to stop
3609 it going un copy-on-write.
3610 If the source SV has gone un copy on write between up there
3611 and down here, then (assert() that) it is of the correct
3612 form to make it copy on write again */
3613 if ((sflags & (SVf_FAKE | SVf_READONLY))
3614 != (SVf_FAKE | SVf_READONLY)) {
3615 SvREADONLY_on(sstr);
3616 SvFAKE_on(sstr);
3617 /* Make the source SV into a loop of 1.
3618 (about to become 2) */
a29f6d03 3619 SV_COW_NEXT_SV_SET(sstr, sstr);
765f542d
NC
3620 }
3621 }
3622#endif
3623 /* Initial code is common. */
94010e71
NC
3624 if (SvPVX_const(dstr)) { /* we know that dtype >= SVt_PV */
3625 SvPV_free(dstr);
79072805 3626 }
765f542d 3627
765f542d
NC
3628 if (!isSwipe) {
3629 /* making another shared SV. */
3630 STRLEN cur = SvCUR(sstr);
3631 STRLEN len = SvLEN(sstr);
f8c7b90f 3632#ifdef PERL_OLD_COPY_ON_WRITE
765f542d 3633 if (len) {
b8f9541a 3634 assert (SvTYPE(dstr) >= SVt_PVIV);
765f542d
NC
3635 /* SvIsCOW_normal */
3636 /* splice us in between source and next-after-source. */
a29f6d03
NC
3637 SV_COW_NEXT_SV_SET(dstr, SV_COW_NEXT_SV(sstr));
3638 SV_COW_NEXT_SV_SET(sstr, dstr);
940132f3 3639 SvPV_set(dstr, SvPVX_mutable(sstr));
a604c751
NC
3640 } else
3641#endif
3642 {
765f542d 3643 /* SvIsCOW_shared_hash */
46187eeb
NC
3644 DEBUG_C(PerlIO_printf(Perl_debug_log,
3645 "Copy on write: Sharing hash\n"));
b8f9541a 3646
bdd68bc3 3647 assert (SvTYPE(dstr) >= SVt_PV);
765f542d 3648 SvPV_set(dstr,
d1db91c6 3649 HEK_KEY(share_hek_hek(SvSHARED_HEK_FROM_PV(SvPVX_const(sstr)))));
bdd68bc3 3650 }
87a1ef3d
SP
3651 SvLEN_set(dstr, len);
3652 SvCUR_set(dstr, cur);
765f542d
NC
3653 SvREADONLY_on(dstr);
3654 SvFAKE_on(dstr);
3655 /* Relesase a global SV mutex. */
3656 }
3657 else
765f542d 3658 { /* Passes the swipe test. */
78d1e721 3659 SvPV_set(dstr, SvPVX_mutable(sstr));
765f542d
NC
3660 SvLEN_set(dstr, SvLEN(sstr));
3661 SvCUR_set(dstr, SvCUR(sstr));
3662
3663 SvTEMP_off(dstr);
3664 (void)SvOK_off(sstr); /* NOTE: nukes most SvFLAGS on sstr */
6136c704 3665 SvPV_set(sstr, NULL);
765f542d
NC
3666 SvLEN_set(sstr, 0);
3667 SvCUR_set(sstr, 0);
3668 SvTEMP_off(sstr);
3669 }
3670 }
8990e307 3671 if (sflags & SVp_NOK) {
9d6ce603 3672 SvNV_set(dstr, SvNVX(sstr));
79072805 3673 }
8990e307 3674 if (sflags & SVp_IOK) {
23525414
NC
3675 SvRELEASE_IVX(dstr);
3676 SvIV_set(dstr, SvIVX(sstr));
3677 /* Must do this otherwise some other overloaded use of 0x80000000
3678 gets confused. I guess SVpbm_VALID */
2b1c7e3e 3679 if (sflags & SVf_IVisUV)
25da4f38 3680 SvIsUV_on(dstr);
79072805 3681 }
dd2eae66
NC
3682 SvFLAGS(dstr) |= sflags & (SVf_IOK|SVp_IOK|SVf_NOK|SVp_NOK|SVf_UTF8
3683 |SVf_AMAGIC);
4f2da183 3684 {
b0a11fe1 3685 const MAGIC * const smg = SvVSTRING_mg(sstr);
4f2da183
NC
3686 if (smg) {
3687 sv_magic(dstr, NULL, PERL_MAGIC_vstring,
3688 smg->mg_ptr, smg->mg_len);
3689 SvRMAGICAL_on(dstr);
3690 }
7a5fa8a2 3691 }
79072805 3692 }
5d581361 3693 else if (sflags & (SVp_IOK|SVp_NOK)) {
c2468cc7 3694 (void)SvOK_off(dstr);
dd2eae66
NC
3695 SvFLAGS(dstr) |= sflags & (SVf_IOK|SVp_IOK|SVf_IVisUV|SVf_NOK|SVp_NOK
3696 |SVf_AMAGIC);
5d581361
NC
3697 if (sflags & SVp_IOK) {
3698 /* XXXX Do we want to set IsUV for IV(ROK)? Be extra safe... */
3699 SvIV_set(dstr, SvIVX(sstr));
3700 }
3332b3c1 3701 if (sflags & SVp_NOK) {
9d6ce603 3702 SvNV_set(dstr, SvNVX(sstr));
3332b3c1
JH
3703 }
3704 }
79072805 3705 else {
f7877b28 3706 if (isGV_with_GP(sstr)) {
180488f8
NC
3707 /* This stringification rule for globs is spread in 3 places.
3708 This feels bad. FIXME. */
3709 const U32 wasfake = sflags & SVf_FAKE;
3710
3711 /* FAKE globs can get coerced, so need to turn this off
3712 temporarily if it is on. */
3713 SvFAKE_off(sstr);
3714 gv_efullname3(dstr, (GV *)sstr, "*");
3715 SvFLAGS(sstr) |= wasfake;
dd2eae66 3716 SvFLAGS(dstr) |= sflags & SVf_AMAGIC;
180488f8 3717 }
20408e3c
GS
3718 else
3719 (void)SvOK_off(dstr);
a0d0e21e 3720 }
27c9684d
AP
3721 if (SvTAINTED(sstr))
3722 SvTAINT(dstr);
79072805
LW
3723}
3724
954c1994
GS
3725/*
3726=for apidoc sv_setsv_mg
3727
3728Like C<sv_setsv>, but also handles 'set' magic.
3729
3730=cut
3731*/
3732
79072805 3733void
864dbfa3 3734Perl_sv_setsv_mg(pTHX_ SV *dstr, register SV *sstr)
ef50df4b
GS
3735{
3736 sv_setsv(dstr,sstr);
3737 SvSETMAGIC(dstr);
3738}
3739
f8c7b90f 3740#ifdef PERL_OLD_COPY_ON_WRITE
ed252734
NC
3741SV *
3742Perl_sv_setsv_cow(pTHX_ SV *dstr, SV *sstr)
3743{
3744 STRLEN cur = SvCUR(sstr);
3745 STRLEN len = SvLEN(sstr);
3746 register char *new_pv;
3747
3748 if (DEBUG_C_TEST) {
3749 PerlIO_printf(Perl_debug_log, "Fast copy on write: %p -> %p\n",
3750 sstr, dstr);
3751 sv_dump(sstr);
3752 if (dstr)
3753 sv_dump(dstr);
3754 }
3755
3756 if (dstr) {
3757 if (SvTHINKFIRST(dstr))
3758 sv_force_normal_flags(dstr, SV_COW_DROP_PV);
3f7c398e
SP
3759 else if (SvPVX_const(dstr))
3760 Safefree(SvPVX_const(dstr));
ed252734
NC
3761 }
3762 else
3763 new_SV(dstr);
862a34c6 3764 SvUPGRADE(dstr, SVt_PVIV);
ed252734
NC
3765
3766 assert (SvPOK(sstr));
3767 assert (SvPOKp(sstr));
3768 assert (!SvIOK(sstr));
3769 assert (!SvIOKp(sstr));
3770 assert (!SvNOK(sstr));
3771 assert (!SvNOKp(sstr));
3772
3773 if (SvIsCOW(sstr)) {
3774
3775 if (SvLEN(sstr) == 0) {
3776 /* source is a COW shared hash key. */
ed252734
NC
3777 DEBUG_C(PerlIO_printf(Perl_debug_log,
3778 "Fast copy on write: Sharing hash\n"));
d1db91c6 3779 new_pv = HEK_KEY(share_hek_hek(SvSHARED_HEK_FROM_PV(SvPVX_const(sstr))));
ed252734
NC
3780 goto common_exit;
3781 }
3782 SV_COW_NEXT_SV_SET(dstr, SV_COW_NEXT_SV(sstr));
3783 } else {
3784 assert ((SvFLAGS(sstr) & CAN_COW_MASK) == CAN_COW_FLAGS);
862a34c6 3785 SvUPGRADE(sstr, SVt_PVIV);
ed252734
NC
3786 SvREADONLY_on(sstr);
3787 SvFAKE_on(sstr);
3788 DEBUG_C(PerlIO_printf(Perl_debug_log,
3789 "Fast copy on write: Converting sstr to COW\n"));
3790 SV_COW_NEXT_SV_SET(dstr, sstr);
3791 }
3792 SV_COW_NEXT_SV_SET(sstr, dstr);
940132f3 3793 new_pv = SvPVX_mutable(sstr);
ed252734
NC
3794
3795 common_exit:
3796 SvPV_set(dstr, new_pv);
3797 SvFLAGS(dstr) = (SVt_PVIV|SVf_POK|SVp_POK|SVf_FAKE|SVf_READONLY);
3798 if (SvUTF8(sstr))
3799 SvUTF8_on(dstr);
87a1ef3d
SP
3800 SvLEN_set(dstr, len);
3801 SvCUR_set(dstr, cur);
ed252734
NC
3802 if (DEBUG_C_TEST) {
3803 sv_dump(dstr);
3804 }
3805 return dstr;
3806}
3807#endif
3808
954c1994
GS
3809/*
3810=for apidoc sv_setpvn
3811
3812Copies a string into an SV. The C<len> parameter indicates the number of
9e09f5f2
MHM
3813bytes to be copied. If the C<ptr> argument is NULL the SV will become
3814undefined. Does not handle 'set' magic. See C<sv_setpvn_mg>.
954c1994
GS
3815
3816=cut
3817*/
3818
ef50df4b 3819void
864dbfa3 3820Perl_sv_setpvn(pTHX_ register SV *sv, register const char *ptr, register STRLEN len)
79072805 3821{
97aff369 3822 dVAR;
c6f8c383 3823 register char *dptr;
22c522df 3824
765f542d 3825 SV_CHECK_THINKFIRST_COW_DROP(sv);
463ee0b2 3826 if (!ptr) {
a0d0e21e 3827 (void)SvOK_off(sv);
463ee0b2
LW
3828 return;
3829 }
22c522df
JH
3830 else {
3831 /* len is STRLEN which is unsigned, need to copy to signed */
a3b680e6 3832 const IV iv = len;
9c5ffd7c
JH
3833 if (iv < 0)
3834 Perl_croak(aTHX_ "panic: sv_setpvn called with negative strlen");
22c522df 3835 }
862a34c6 3836 SvUPGRADE(sv, SVt_PV);
c6f8c383 3837
5902b6a9 3838 dptr = SvGROW(sv, len + 1);
c6f8c383
GA
3839 Move(ptr,dptr,len,char);
3840 dptr[len] = '\0';
79072805 3841 SvCUR_set(sv, len);
1aa99e6b 3842 (void)SvPOK_only_UTF8(sv); /* validate pointer */
463ee0b2 3843 SvTAINT(sv);
79072805
LW
3844}
3845
954c1994
GS
3846/*
3847=for apidoc sv_setpvn_mg
3848
3849Like C<sv_setpvn>, but also handles 'set' magic.
3850
3851=cut
3852*/
3853
79072805 3854void
864dbfa3 3855Perl_sv_setpvn_mg(pTHX_ register SV *sv, register const char *ptr, register STRLEN len)
ef50df4b
GS
3856{
3857 sv_setpvn(sv,ptr,len);
3858 SvSETMAGIC(sv);
3859}
3860
954c1994
GS
3861/*
3862=for apidoc sv_setpv
3863
3864Copies a string into an SV. The string must be null-terminated. Does not
3865handle 'set' magic. See C<sv_setpv_mg>.
3866
3867=cut
3868*/
3869
ef50df4b 3870void
864dbfa3 3871Perl_sv_setpv(pTHX_ register SV *sv, register const char *ptr)
79072805 3872{
97aff369 3873 dVAR;
79072805
LW
3874 register STRLEN len;
3875
765f542d 3876 SV_CHECK_THINKFIRST_COW_DROP(sv);
463ee0b2 3877 if (!ptr) {
a0d0e21e 3878 (void)SvOK_off(sv);
463ee0b2
LW
3879 return;
3880 }
79072805 3881 len = strlen(ptr);
862a34c6 3882 SvUPGRADE(sv, SVt_PV);
c6f8c383 3883
79072805 3884 SvGROW(sv, len + 1);
463ee0b2 3885 Move(ptr,SvPVX(sv),len+1,char);
79072805 3886 SvCUR_set(sv, len);
1aa99e6b 3887 (void)SvPOK_only_UTF8(sv); /* validate pointer */
463ee0b2
LW
3888 SvTAINT(sv);
3889}
3890
954c1994
GS
3891/*
3892=for apidoc sv_setpv_mg
3893
3894Like C<sv_setpv>, but also handles 'set' magic.
3895
3896=cut
3897*/
3898
463ee0b2 3899void
864dbfa3 3900Perl_sv_setpv_mg(pTHX_ register SV *sv, register const char *ptr)
ef50df4b
GS
3901{
3902 sv_setpv(sv,ptr);
3903 SvSETMAGIC(sv);
3904}
3905
954c1994 3906/*
47518d95 3907=for apidoc sv_usepvn_flags
954c1994 3908
794a0d33
JH
3909Tells an SV to use C<ptr> to find its string value. Normally the
3910string is stored inside the SV but sv_usepvn allows the SV to use an
3911outside string. The C<ptr> should point to memory that was allocated
c1c21316
NC
3912by C<malloc>. The string length, C<len>, must be supplied. By default
3913this function will realloc (i.e. move) the memory pointed to by C<ptr>,
794a0d33
JH
3914so that pointer should not be freed or used by the programmer after
3915giving it to sv_usepvn, and neither should any pointers from "behind"
c1c21316
NC
3916that pointer (e.g. ptr + 1) be used.
3917
3918If C<flags> & SV_SMAGIC is true, will call SvSETMAGIC. If C<flags> &
3919SV_HAS_TRAILING_NUL is true, then C<ptr[len]> must be NUL, and the realloc
cbf82dd0 3920will be skipped. (i.e. the buffer is actually at least 1 byte longer than
c1c21316 3921C<len>, and already meets the requirements for storing in C<SvPVX>)
954c1994
GS
3922
3923=cut
3924*/
3925
ef50df4b 3926void
47518d95 3927Perl_sv_usepvn_flags(pTHX_ SV *sv, char *ptr, STRLEN len, U32 flags)
463ee0b2 3928{
97aff369 3929 dVAR;
1936d2a7 3930 STRLEN allocate;
765f542d 3931 SV_CHECK_THINKFIRST_COW_DROP(sv);
862a34c6 3932 SvUPGRADE(sv, SVt_PV);
463ee0b2 3933 if (!ptr) {
a0d0e21e 3934 (void)SvOK_off(sv);
47518d95
NC
3935 if (flags & SV_SMAGIC)
3936 SvSETMAGIC(sv);
463ee0b2
LW
3937 return;
3938 }
3f7c398e 3939 if (SvPVX_const(sv))
8bd4d4c5 3940 SvPV_free(sv);
1936d2a7 3941
0b7042f9 3942#ifdef DEBUGGING
2e90b4cd
NC
3943 if (flags & SV_HAS_TRAILING_NUL)
3944 assert(ptr[len] == '\0');
0b7042f9 3945#endif
2e90b4cd 3946
c1c21316 3947 allocate = (flags & SV_HAS_TRAILING_NUL)
8f01dc65 3948 ? len + 1: PERL_STRLEN_ROUNDUP(len + 1);
cbf82dd0
NC
3949 if (flags & SV_HAS_TRAILING_NUL) {
3950 /* It's long enough - do nothing.
3951 Specfically Perl_newCONSTSUB is relying on this. */
3952 } else {
69d25b4f 3953#ifdef DEBUGGING
69d25b4f 3954 /* Force a move to shake out bugs in callers. */
10edeb5d 3955 char *new_ptr = (char*)safemalloc(allocate);
69d25b4f
NC
3956 Copy(ptr, new_ptr, len, char);
3957 PoisonFree(ptr,len,char);
3958 Safefree(ptr);
3959 ptr = new_ptr;
69d25b4f 3960#else
10edeb5d 3961 ptr = (char*) saferealloc (ptr, allocate);
69d25b4f 3962#endif
cbf82dd0 3963 }
f880fe2f 3964 SvPV_set(sv, ptr);
463ee0b2 3965 SvCUR_set(sv, len);
1936d2a7 3966 SvLEN_set(sv, allocate);
c1c21316
NC
3967 if (!(flags & SV_HAS_TRAILING_NUL)) {
3968 *SvEND(sv) = '\0';
3969 }
1aa99e6b 3970 (void)SvPOK_only_UTF8(sv); /* validate pointer */
463ee0b2 3971 SvTAINT(sv);
47518d95
NC
3972 if (flags & SV_SMAGIC)
3973 SvSETMAGIC(sv);
ef50df4b
GS
3974}
3975
f8c7b90f 3976#ifdef PERL_OLD_COPY_ON_WRITE
765f542d
NC
3977/* Need to do this *after* making the SV normal, as we need the buffer
3978 pointer to remain valid until after we've copied it. If we let go too early,
3979 another thread could invalidate it by unsharing last of the same hash key
3980 (which it can do by means other than releasing copy-on-write Svs)
3981 or by changing the other copy-on-write SVs in the loop. */
3982STATIC void
bdd68bc3 3983S_sv_release_COW(pTHX_ register SV *sv, const char *pvx, STRLEN len, SV *after)
765f542d
NC
3984{
3985 if (len) { /* this SV was SvIsCOW_normal(sv) */
3986 /* we need to find the SV pointing to us. */
cf5629ad 3987 SV *current = SV_COW_NEXT_SV(after);
7a5fa8a2 3988
765f542d
NC
3989 if (current == sv) {
3990 /* The SV we point to points back to us (there were only two of us
3991 in the loop.)
3992 Hence other SV is no longer copy on write either. */
3993 SvFAKE_off(after);
3994 SvREADONLY_off(after);
3995 } else {
3996 /* We need to follow the pointers around the loop. */
3997 SV *next;
3998 while ((next = SV_COW_NEXT_SV(current)) != sv) {
3999 assert (next);
4000 current = next;
4001 /* don't loop forever if the structure is bust, and we have
4002 a pointer into a closed loop. */
4003 assert (current != after);
3f7c398e 4004 assert (SvPVX_const(current) == pvx);
765f542d
NC
4005 }
4006 /* Make the SV before us point to the SV after us. */
a29f6d03 4007 SV_COW_NEXT_SV_SET(current, after);
765f542d
NC
4008 }
4009 } else {
bdd68bc3 4010 unshare_hek(SvSHARED_HEK_FROM_PV(pvx));
765f542d
NC
4011 }
4012}
4013
4014int
4015Perl_sv_release_IVX(pTHX_ register SV *sv)
4016{
4017 if (SvIsCOW(sv))
4018 sv_force_normal_flags(sv, 0);
0c34ef67
MHM
4019 SvOOK_off(sv);
4020 return 0;
765f542d
NC
4021}
4022#endif
645c22ef
DM
4023/*
4024=for apidoc sv_force_normal_flags
4025
4026Undo various types of fakery on an SV: if the PV is a shared string, make
4027a private copy; if we're a ref, stop refing; if we're a glob, downgrade to
765f542d
NC
4028an xpvmg; if we're a copy-on-write scalar, this is the on-write time when
4029we do the copy, and is also used locally. If C<SV_COW_DROP_PV> is set
4030then a copy-on-write scalar drops its PV buffer (if any) and becomes
4031SvPOK_off rather than making a copy. (Used where this scalar is about to be
d3050d9d 4032set to some other value.) In addition, the C<flags> parameter gets passed to
765f542d
NC
4033C<sv_unref_flags()> when unrefing. C<sv_force_normal> calls this function
4034with flags set to 0.
645c22ef
DM
4035
4036=cut
4037*/
4038
6fc92669 4039void
840a7b70 4040Perl_sv_force_normal_flags(pTHX_ register SV *sv, U32 flags)
0f15f207 4041{
97aff369 4042 dVAR;
f8c7b90f 4043#ifdef PERL_OLD_COPY_ON_WRITE
765f542d
NC
4044 if (SvREADONLY(sv)) {
4045 /* At this point I believe I should acquire a global SV mutex. */
4046 if (SvFAKE(sv)) {
b64e5050 4047 const char * const pvx = SvPVX_const(sv);
a28509cc
AL
4048 const STRLEN len = SvLEN(sv);
4049 const STRLEN cur = SvCUR(sv);
a28509cc 4050 SV * const next = SV_COW_NEXT_SV(sv); /* next COW sv in the loop. */
46187eeb
NC
4051 if (DEBUG_C_TEST) {
4052 PerlIO_printf(Perl_debug_log,
4053 "Copy on write: Force normal %ld\n",
4054 (long) flags);
e419cbc5 4055 sv_dump(sv);
46187eeb 4056 }
765f542d
NC
4057 SvFAKE_off(sv);
4058 SvREADONLY_off(sv);
9f653bb5 4059 /* This SV doesn't own the buffer, so need to Newx() a new one: */
6136c704 4060 SvPV_set(sv, NULL);
87a1ef3d 4061 SvLEN_set(sv, 0);
765f542d
NC
4062 if (flags & SV_COW_DROP_PV) {
4063 /* OK, so we don't need to copy our buffer. */
4064 SvPOK_off(sv);
4065 } else {
4066 SvGROW(sv, cur + 1);
4067 Move(pvx,SvPVX(sv),cur,char);
87a1ef3d 4068 SvCUR_set(sv, cur);
765f542d
NC
4069 *SvEND(sv) = '\0';
4070 }
bdd68bc3 4071 sv_release_COW(sv, pvx, len, next);
46187eeb 4072 if (DEBUG_C_TEST) {
e419cbc5 4073 sv_dump(sv);
46187eeb 4074 }
765f542d 4075 }
923e4eb5 4076 else if (IN_PERL_RUNTIME)
765f542d
NC
4077 Perl_croak(aTHX_ PL_no_modify);
4078 /* At this point I believe that I can drop the global SV mutex. */
4079 }
4080#else
2213622d 4081 if (SvREADONLY(sv)) {
1c846c1f 4082 if (SvFAKE(sv)) {
b64e5050 4083 const char * const pvx = SvPVX_const(sv);
66a1b24b 4084 const STRLEN len = SvCUR(sv);
10bcdfd6
NC
4085 SvFAKE_off(sv);
4086 SvREADONLY_off(sv);
bd61b366 4087 SvPV_set(sv, NULL);
66a1b24b 4088 SvLEN_set(sv, 0);
1c846c1f 4089 SvGROW(sv, len + 1);
706aa1c9 4090 Move(pvx,SvPVX(sv),len,char);
1c846c1f 4091 *SvEND(sv) = '\0';
bdd68bc3 4092 unshare_hek(SvSHARED_HEK_FROM_PV(pvx));
1c846c1f 4093 }
923e4eb5 4094 else if (IN_PERL_RUNTIME)
cea2e8a9 4095 Perl_croak(aTHX_ PL_no_modify);
0f15f207 4096 }
765f542d 4097#endif
2213622d 4098 if (SvROK(sv))
840a7b70 4099 sv_unref_flags(sv, flags);
6fc92669
GS
4100 else if (SvFAKE(sv) && SvTYPE(sv) == SVt_PVGV)
4101 sv_unglob(sv);
0f15f207 4102}
1c846c1f 4103
645c22ef 4104/*
954c1994
GS
4105=for apidoc sv_chop
4106
1c846c1f 4107Efficient removal of characters from the beginning of the string buffer.
954c1994
GS
4108SvPOK(sv) must be true and the C<ptr> must be a pointer to somewhere inside
4109the string buffer. The C<ptr> becomes the first character of the adjusted
645c22ef 4110string. Uses the "OOK hack".
3f7c398e 4111Beware: after this function returns, C<ptr> and SvPVX_const(sv) may no longer
31869a79 4112refer to the same chunk of data.
954c1994
GS
4113
4114=cut
4115*/
4116
79072805 4117void
f54cb97a 4118Perl_sv_chop(pTHX_ register SV *sv, register const char *ptr)
79072805
LW
4119{
4120 register STRLEN delta;
a0d0e21e 4121 if (!ptr || !SvPOKp(sv))
79072805 4122 return;
3f7c398e 4123 delta = ptr - SvPVX_const(sv);
2213622d 4124 SV_CHECK_THINKFIRST(sv);
79072805
LW
4125 if (SvTYPE(sv) < SVt_PVIV)
4126 sv_upgrade(sv,SVt_PVIV);
4127
4128 if (!SvOOK(sv)) {
50483b2c 4129 if (!SvLEN(sv)) { /* make copy of shared string */
3f7c398e 4130 const char *pvx = SvPVX_const(sv);
a28509cc 4131 const STRLEN len = SvCUR(sv);
50483b2c 4132 SvGROW(sv, len + 1);
706aa1c9 4133 Move(pvx,SvPVX(sv),len,char);
50483b2c
JD
4134 *SvEND(sv) = '\0';
4135 }
45977657 4136 SvIV_set(sv, 0);
a4bfb290
AB
4137 /* Same SvOOK_on but SvOOK_on does a SvIOK_off
4138 and we do that anyway inside the SvNIOK_off
4139 */
7a5fa8a2 4140 SvFLAGS(sv) |= SVf_OOK;
79072805 4141 }
a4bfb290 4142 SvNIOK_off(sv);
b162af07
SP
4143 SvLEN_set(sv, SvLEN(sv) - delta);
4144 SvCUR_set(sv, SvCUR(sv) - delta);
f880fe2f 4145 SvPV_set(sv, SvPVX(sv) + delta);
45977657 4146 SvIV_set(sv, SvIVX(sv) + delta);
79072805
LW
4147}
4148
954c1994
GS
4149/*
4150=for apidoc sv_catpvn
4151
4152Concatenates the string onto the end of the string which is in the SV. The
1e54db1a
JH
4153C<len> indicates number of bytes to copy. If the SV has the UTF-8
4154status set, then the bytes appended should be valid UTF-8.
d5ce4a7c 4155Handles 'get' magic, but not 'set' magic. See C<sv_catpvn_mg>.
954c1994 4156
8d6d96c1
HS
4157=for apidoc sv_catpvn_flags
4158
4159Concatenates the string onto the end of the string which is in the SV. The
1e54db1a
JH
4160C<len> indicates number of bytes to copy. If the SV has the UTF-8
4161status set, then the bytes appended should be valid UTF-8.
8d6d96c1
HS
4162If C<flags> has C<SV_GMAGIC> bit set, will C<mg_get> on C<dsv> if
4163appropriate, else not. C<sv_catpvn> and C<sv_catpvn_nomg> are implemented
4164in terms of this function.
4165
4166=cut
4167*/
4168
4169void
4170Perl_sv_catpvn_flags(pTHX_ register SV *dsv, register const char *sstr, register STRLEN slen, I32 flags)
4171{
97aff369 4172 dVAR;
8d6d96c1 4173 STRLEN dlen;
fabdb6c0 4174 const char * const dstr = SvPV_force_flags(dsv, dlen, flags);
8d6d96c1 4175
8d6d96c1
HS
4176 SvGROW(dsv, dlen + slen + 1);
4177 if (sstr == dstr)
3f7c398e 4178 sstr = SvPVX_const(dsv);
8d6d96c1 4179 Move(sstr, SvPVX(dsv) + dlen, slen, char);
b162af07 4180 SvCUR_set(dsv, SvCUR(dsv) + slen);
8d6d96c1
HS
4181 *SvEND(dsv) = '\0';
4182 (void)SvPOK_only_UTF8(dsv); /* validate pointer */
4183 SvTAINT(dsv);
bddd5118
NC
4184 if (flags & SV_SMAGIC)
4185 SvSETMAGIC(dsv);
79072805
LW
4186}
4187
954c1994 4188/*
954c1994
GS
4189=for apidoc sv_catsv
4190
13e8c8e3
JH
4191Concatenates the string from SV C<ssv> onto the end of the string in
4192SV C<dsv>. Modifies C<dsv> but not C<ssv>. Handles 'get' magic, but
4193not 'set' magic. See C<sv_catsv_mg>.
954c1994 4194
8d6d96c1
HS
4195=for apidoc sv_catsv_flags
4196
4197Concatenates the string from SV C<ssv> onto the end of the string in
4198SV C<dsv>. Modifies C<dsv> but not C<ssv>. If C<flags> has C<SV_GMAGIC>
4199bit set, will C<mg_get> on the SVs if appropriate, else not. C<sv_catsv>
4200and C<sv_catsv_nomg> are implemented in terms of this function.
4201
4202=cut */
4203
ef50df4b 4204void
8d6d96c1 4205Perl_sv_catsv_flags(pTHX_ SV *dsv, register SV *ssv, I32 flags)
79072805 4206{
97aff369 4207 dVAR;
bddd5118 4208 if (ssv) {
00b6aa41
AL
4209 STRLEN slen;
4210 const char *spv = SvPV_const(ssv, slen);
4211 if (spv) {
bddd5118
NC
4212 /* sutf8 and dutf8 were type bool, but under USE_ITHREADS,
4213 gcc version 2.95.2 20000220 (Debian GNU/Linux) for
4214 Linux xxx 2.2.17 on sparc64 with gcc -O2, we erroneously
4215 get dutf8 = 0x20000000, (i.e. SVf_UTF8) even though
4216 dsv->sv_flags doesn't have that bit set.
4fd84b44 4217 Andy Dougherty 12 Oct 2001
bddd5118
NC
4218 */
4219 const I32 sutf8 = DO_UTF8(ssv);
4220 I32 dutf8;
13e8c8e3 4221
bddd5118
NC
4222 if (SvGMAGICAL(dsv) && (flags & SV_GMAGIC))
4223 mg_get(dsv);
4224 dutf8 = DO_UTF8(dsv);
8d6d96c1 4225
bddd5118
NC
4226 if (dutf8 != sutf8) {
4227 if (dutf8) {
4228 /* Not modifying source SV, so taking a temporary copy. */
00b6aa41 4229 SV* const csv = sv_2mortal(newSVpvn(spv, slen));
13e8c8e3 4230
bddd5118
NC
4231 sv_utf8_upgrade(csv);
4232 spv = SvPV_const(csv, slen);
4233 }
4234 else
4235 sv_utf8_upgrade_nomg(dsv);
13e8c8e3 4236 }
bddd5118 4237 sv_catpvn_nomg(dsv, spv, slen);
e84ff256 4238 }
560a288e 4239 }
bddd5118
NC
4240 if (flags & SV_SMAGIC)
4241 SvSETMAGIC(dsv);
79072805
LW
4242}
4243
954c1994 4244/*
954c1994
GS
4245=for apidoc sv_catpv
4246
4247Concatenates the string onto the end of the string which is in the SV.
1e54db1a
JH
4248If the SV has the UTF-8 status set, then the bytes appended should be
4249valid UTF-8. Handles 'get' magic, but not 'set' magic. See C<sv_catpv_mg>.
954c1994 4250
d5ce4a7c 4251=cut */
954c1994 4252
ef50df4b 4253void
0c981600 4254Perl_sv_catpv(pTHX_ register SV *sv, register const char *ptr)
79072805 4255{
97aff369 4256 dVAR;
79072805 4257 register STRLEN len;
463ee0b2 4258 STRLEN tlen;
748a9306 4259 char *junk;
79072805 4260
0c981600 4261 if (!ptr)
79072805 4262 return;
748a9306 4263 junk = SvPV_force(sv, tlen);
0c981600 4264 len = strlen(ptr);
463ee0b2 4265 SvGROW(sv, tlen + len + 1);
0c981600 4266 if (ptr == junk)
3f7c398e 4267 ptr = SvPVX_const(sv);
0c981600 4268 Move(ptr,SvPVX(sv)+tlen,len+1,char);
b162af07 4269 SvCUR_set(sv, SvCUR(sv) + len);
d41ff1b8 4270 (void)SvPOK_only_UTF8(sv); /* validate pointer */
463ee0b2 4271 SvTAINT(sv);
79072805
LW
4272}
4273
954c1994
GS
4274/*
4275=for apidoc sv_catpv_mg
4276
4277Like C<sv_catpv>, but also handles 'set' magic.
4278
4279=cut
4280*/
4281
ef50df4b 4282void
0c981600 4283Perl_sv_catpv_mg(pTHX_ register SV *sv, register const char *ptr)
ef50df4b 4284{
0c981600 4285 sv_catpv(sv,ptr);
ef50df4b
GS
4286 SvSETMAGIC(sv);
4287}
4288
645c22ef
DM
4289/*
4290=for apidoc newSV
4291
561b68a9
SH
4292Creates a new SV. A non-zero C<len> parameter indicates the number of
4293bytes of preallocated string space the SV should have. An extra byte for a
4294trailing NUL is also reserved. (SvPOK is not set for the SV even if string
4295space is allocated.) The reference count for the new SV is set to 1.
4296
4297In 5.9.3, newSV() replaces the older NEWSV() API, and drops the first
4298parameter, I<x>, a debug aid which allowed callers to identify themselves.
4299This aid has been superseded by a new build option, PERL_MEM_LOG (see
4300L<perlhack/PERL_MEM_LOG>). The older API is still there for use in XS
4301modules supporting older perls.
645c22ef
DM
4302
4303=cut
4304*/
4305
79072805 4306SV *
864dbfa3 4307Perl_newSV(pTHX_ STRLEN len)
79072805 4308{
97aff369 4309 dVAR;
79072805 4310 register SV *sv;
1c846c1f 4311
4561caa4 4312 new_SV(sv);
79072805
LW
4313 if (len) {
4314 sv_upgrade(sv, SVt_PV);
4315 SvGROW(sv, len + 1);
4316 }
4317 return sv;
4318}
954c1994 4319/*
92110913 4320=for apidoc sv_magicext
954c1994 4321
68795e93 4322Adds magic to an SV, upgrading it if necessary. Applies the
2d8d5d5a 4323supplied vtable and returns a pointer to the magic added.
92110913 4324
2d8d5d5a
SH
4325Note that C<sv_magicext> will allow things that C<sv_magic> will not.
4326In particular, you can add magic to SvREADONLY SVs, and add more than
4327one instance of the same 'how'.
645c22ef 4328
2d8d5d5a
SH
4329If C<namlen> is greater than zero then a C<savepvn> I<copy> of C<name> is
4330stored, if C<namlen> is zero then C<name> is stored as-is and - as another
4331special case - if C<(name && namlen == HEf_SVKEY)> then C<name> is assumed
4332to contain an C<SV*> and is stored as-is with its REFCNT incremented.
92110913 4333
2d8d5d5a 4334(This is now used as a subroutine by C<sv_magic>.)
954c1994
GS
4335
4336=cut
4337*/
92110913 4338MAGIC *
92e67595 4339Perl_sv_magicext(pTHX_ SV* sv, SV* obj, int how, MGVTBL *vtable,
92110913 4340 const char* name, I32 namlen)
79072805 4341{
97aff369 4342 dVAR;
79072805 4343 MAGIC* mg;
68795e93 4344
92110913 4345 if (SvTYPE(sv) < SVt_PVMG) {
862a34c6 4346 SvUPGRADE(sv, SVt_PVMG);
463ee0b2 4347 }
a02a5408 4348 Newxz(mg, 1, MAGIC);
79072805 4349 mg->mg_moremagic = SvMAGIC(sv);
b162af07 4350 SvMAGIC_set(sv, mg);
75f9d97a 4351
05f95b08
SB
4352 /* Sometimes a magic contains a reference loop, where the sv and
4353 object refer to each other. To prevent a reference loop that
4354 would prevent such objects being freed, we look for such loops
4355 and if we find one we avoid incrementing the object refcount.
87f0b213
JH
4356
4357 Note we cannot do this to avoid self-tie loops as intervening RV must
b5ccf5f2 4358 have its REFCNT incremented to keep it in existence.
87f0b213
JH
4359
4360 */
14befaf4
DM
4361 if (!obj || obj == sv ||
4362 how == PERL_MAGIC_arylen ||
4363 how == PERL_MAGIC_qr ||
8d2f4536 4364 how == PERL_MAGIC_symtab ||
75f9d97a
JH
4365 (SvTYPE(obj) == SVt_PVGV &&
4366 (GvSV(obj) == sv || GvHV(obj) == (HV*)sv || GvAV(obj) == (AV*)sv ||
4367 GvCV(obj) == (CV*)sv || GvIOp(obj) == (IO*)sv ||
2628be26 4368 GvFORM(obj) == (CV*)sv)))
75f9d97a 4369 {
8990e307 4370 mg->mg_obj = obj;
75f9d97a 4371 }
85e6fe83 4372 else {
b37c2d43 4373 mg->mg_obj = SvREFCNT_inc_simple(obj);
85e6fe83
LW
4374 mg->mg_flags |= MGf_REFCOUNTED;
4375 }
b5ccf5f2
YST
4376
4377 /* Normal self-ties simply pass a null object, and instead of
4378 using mg_obj directly, use the SvTIED_obj macro to produce a
4379 new RV as needed. For glob "self-ties", we are tieing the PVIO
4380 with an RV obj pointing to the glob containing the PVIO. In
4381 this case, to avoid a reference loop, we need to weaken the
4382 reference.
4383 */
4384
4385 if (how == PERL_MAGIC_tiedscalar && SvTYPE(sv) == SVt_PVIO &&
4386 obj && SvROK(obj) && GvIO(SvRV(obj)) == (IO*)sv)
4387 {
4388 sv_rvweaken(obj);
4389 }
4390
79072805 4391 mg->mg_type = how;
565764a8 4392 mg->mg_len = namlen;
9cbac4c7 4393 if (name) {
92110913 4394 if (namlen > 0)
1edc1566 4395 mg->mg_ptr = savepvn(name, namlen);
c6ee37c5 4396 else if (namlen == HEf_SVKEY)
b37c2d43 4397 mg->mg_ptr = (char*)SvREFCNT_inc_simple_NN((SV*)name);
68795e93 4398 else
92110913 4399 mg->mg_ptr = (char *) name;
9cbac4c7 4400 }
92110913 4401 mg->mg_virtual = vtable;
68795e93 4402
92110913
NIS
4403 mg_magical(sv);
4404 if (SvGMAGICAL(sv))
4405 SvFLAGS(sv) &= ~(SVf_IOK|SVf_NOK|SVf_POK);
4406 return mg;
4407}
4408
4409/*
4410=for apidoc sv_magic
1c846c1f 4411
92110913
NIS
4412Adds magic to an SV. First upgrades C<sv> to type C<SVt_PVMG> if necessary,
4413then adds a new magic item of type C<how> to the head of the magic list.
4414
2d8d5d5a
SH
4415See C<sv_magicext> (which C<sv_magic> now calls) for a description of the
4416handling of the C<name> and C<namlen> arguments.
4417
4509d3fb
SB
4418You need to use C<sv_magicext> to add magic to SvREADONLY SVs and also
4419to add more than one instance of the same 'how'.
4420
92110913
NIS
4421=cut
4422*/
4423
4424void
4425Perl_sv_magic(pTHX_ register SV *sv, SV *obj, int how, const char *name, I32 namlen)
68795e93 4426{
97aff369 4427 dVAR;
92e67595 4428 MGVTBL *vtable;
92110913 4429 MAGIC* mg;
92110913 4430
f8c7b90f 4431#ifdef PERL_OLD_COPY_ON_WRITE
765f542d
NC
4432 if (SvIsCOW(sv))
4433 sv_force_normal_flags(sv, 0);
4434#endif
92110913 4435 if (SvREADONLY(sv)) {
d8084ca5
DM
4436 if (
4437 /* its okay to attach magic to shared strings; the subsequent
4438 * upgrade to PVMG will unshare the string */
4439 !(SvFAKE(sv) && SvTYPE(sv) < SVt_PVMG)
4440
4441 && IN_PERL_RUNTIME
92110913
NIS
4442 && how != PERL_MAGIC_regex_global
4443 && how != PERL_MAGIC_bm
4444 && how != PERL_MAGIC_fm
4445 && how != PERL_MAGIC_sv
e6469971 4446 && how != PERL_MAGIC_backref
92110913
NIS
4447 )
4448 {
4449 Perl_croak(aTHX_ PL_no_modify);
4450 }
4451 }
4452 if (SvMAGICAL(sv) || (how == PERL_MAGIC_taint && SvTYPE(sv) >= SVt_PVMG)) {
4453 if (SvMAGIC(sv) && (mg = mg_find(sv, how))) {
68795e93
NIS
4454 /* sv_magic() refuses to add a magic of the same 'how' as an
4455 existing one
92110913 4456 */
2a509ed3 4457 if (how == PERL_MAGIC_taint) {
92110913 4458 mg->mg_len |= 1;
2a509ed3
NC
4459 /* Any scalar which already had taint magic on which someone
4460 (erroneously?) did SvIOK_on() or similar will now be
4461 incorrectly sporting public "OK" flags. */
4462 SvFLAGS(sv) &= ~(SVf_IOK|SVf_NOK|SVf_POK);
4463 }
92110913
NIS
4464 return;
4465 }
4466 }
68795e93 4467
79072805 4468 switch (how) {
14befaf4 4469 case PERL_MAGIC_sv:
92110913 4470 vtable = &PL_vtbl_sv;
79072805 4471 break;
14befaf4 4472 case PERL_MAGIC_overload:
92110913 4473 vtable = &PL_vtbl_amagic;
a0d0e21e 4474 break;
14befaf4 4475 case PERL_MAGIC_overload_elem:
92110913 4476 vtable = &PL_vtbl_amagicelem;
a0d0e21e 4477 break;
14befaf4 4478 case PERL_MAGIC_overload_table:
92110913 4479 vtable = &PL_vtbl_ovrld;
a0d0e21e 4480 break;
14befaf4 4481 case PERL_MAGIC_bm:
92110913 4482 vtable = &PL_vtbl_bm;
79072805 4483 break;
14befaf4 4484 case PERL_MAGIC_regdata:
92110913 4485 vtable = &PL_vtbl_regdata;
6cef1e77 4486 break;
14befaf4 4487 case PERL_MAGIC_regdatum:
92110913 4488 vtable = &PL_vtbl_regdatum;
6cef1e77 4489 break;
14befaf4 4490 case PERL_MAGIC_env:
92110913 4491 vtable = &PL_vtbl_env;
79072805 4492 break;
14befaf4 4493 case PERL_MAGIC_fm:
92110913 4494 vtable = &PL_vtbl_fm;
55497cff 4495 break;
14befaf4 4496 case PERL_MAGIC_envelem:
92110913 4497 vtable = &PL_vtbl_envelem;
79072805 4498 break;
14befaf4 4499 case PERL_MAGIC_regex_global:
92110913 4500 vtable = &PL_vtbl_mglob;
93a17b20 4501 break;
14befaf4 4502 case PERL_MAGIC_isa:
92110913 4503 vtable = &PL_vtbl_isa;
463ee0b2 4504 break;
14befaf4 4505 case PERL_MAGIC_isaelem:
92110913 4506 vtable = &PL_vtbl_isaelem;
463ee0b2 4507 break;
14befaf4 4508 case PERL_MAGIC_nkeys:
92110913 4509 vtable = &PL_vtbl_nkeys;
16660edb 4510 break;
14befaf4 4511 case PERL_MAGIC_dbfile:
aec46f14 4512 vtable = NULL;
93a17b20 4513 break;
14befaf4 4514 case PERL_MAGIC_dbline:
92110913 4515 vtable = &PL_vtbl_dbline;
79072805 4516 break;
36477c24 4517#ifdef USE_LOCALE_COLLATE
14befaf4 4518 case PERL_MAGIC_collxfrm:
92110913 4519 vtable = &PL_vtbl_collxfrm;
bbce6d69 4520 break;
36477c24 4521#endif /* USE_LOCALE_COLLATE */
14befaf4 4522 case PERL_MAGIC_tied:
92110913 4523 vtable = &PL_vtbl_pack;
463ee0b2 4524 break;
14befaf4
DM
4525 case PERL_MAGIC_tiedelem:
4526 case PERL_MAGIC_tiedscalar:
92110913 4527 vtable = &PL_vtbl_packelem;
463ee0b2 4528 break;
14befaf4 4529 case PERL_MAGIC_qr:
92110913 4530 vtable = &PL_vtbl_regexp;
c277df42 4531 break;
b3ca2e83
NC
4532 case PERL_MAGIC_hints:
4533 /* As this vtable is all NULL, we can reuse it. */
14befaf4 4534 case PERL_MAGIC_sig:
92110913 4535 vtable = &PL_vtbl_sig;
79072805 4536 break;
14befaf4 4537 case PERL_MAGIC_sigelem:
92110913 4538 vtable = &PL_vtbl_sigelem;
79072805 4539 break;
14befaf4 4540 case PERL_MAGIC_taint:
92110913 4541 vtable = &PL_vtbl_taint;
463ee0b2 4542 break;
14befaf4 4543 case PERL_MAGIC_uvar:
92110913 4544 vtable = &PL_vtbl_uvar;
79072805 4545 break;
14befaf4 4546 case PERL_MAGIC_vec:
92110913 4547 vtable = &PL_vtbl_vec;
79072805 4548 break;
a3874608 4549 case PERL_MAGIC_arylen_p:
bfcb3514 4550 case PERL_MAGIC_rhash:
8d2f4536 4551 case PERL_MAGIC_symtab:
ece467f9 4552 case PERL_MAGIC_vstring:
aec46f14 4553 vtable = NULL;
ece467f9 4554 break;
7e8c5dac
HS
4555 case PERL_MAGIC_utf8:
4556 vtable = &PL_vtbl_utf8;
4557 break;
14befaf4 4558 case PERL_MAGIC_substr:
92110913 4559 vtable = &PL_vtbl_substr;
79072805 4560 break;
14befaf4 4561 case PERL_MAGIC_defelem:
92110913 4562 vtable = &PL_vtbl_defelem;
5f05dabc 4563 break;
14befaf4 4564 case PERL_MAGIC_arylen:
92110913 4565 vtable = &PL_vtbl_arylen;
79072805 4566 break;
14befaf4 4567 case PERL_MAGIC_pos:
92110913 4568 vtable = &PL_vtbl_pos;
a0d0e21e 4569 break;
14befaf4 4570 case PERL_MAGIC_backref:
92110913 4571 vtable = &PL_vtbl_backref;
810b8aa5 4572 break;
b3ca2e83
NC
4573 case PERL_MAGIC_hintselem:
4574 vtable = &PL_vtbl_hintselem;
4575 break;
14befaf4
DM
4576 case PERL_MAGIC_ext:
4577 /* Reserved for use by extensions not perl internals. */
4633a7c4
LW
4578 /* Useful for attaching extension internal data to perl vars. */
4579 /* Note that multiple extensions may clash if magical scalars */
4580 /* etc holding private data from one are passed to another. */
aec46f14 4581 vtable = NULL;
a0d0e21e 4582 break;
79072805 4583 default:
14befaf4 4584 Perl_croak(aTHX_ "Don't know how to handle magic of type \\%o", how);
463ee0b2 4585 }
68795e93 4586
92110913 4587 /* Rest of work is done else where */
aec46f14 4588 mg = sv_magicext(sv,obj,how,vtable,name,namlen);
68795e93 4589
92110913
NIS
4590 switch (how) {
4591 case PERL_MAGIC_taint:
4592 mg->mg_len = 1;
4593 break;
4594 case PERL_MAGIC_ext:
4595 case PERL_MAGIC_dbfile:
4596 SvRMAGICAL_on(sv);
4597 break;
4598 }
463ee0b2
LW
4599}
4600
c461cf8f
JH
4601/*
4602=for apidoc sv_unmagic
4603
645c22ef 4604Removes all magic of type C<type> from an SV.
c461cf8f
JH
4605
4606=cut
4607*/
4608
463ee0b2 4609int
864dbfa3 4610Perl_sv_unmagic(pTHX_ SV *sv, int type)
463ee0b2
LW
4611{
4612 MAGIC* mg;
4613 MAGIC** mgp;
91bba347 4614 if (SvTYPE(sv) < SVt_PVMG || !SvMAGIC(sv))
463ee0b2 4615 return 0;
064cf529 4616 mgp = &(((XPVMG*) SvANY(sv))->xmg_u.xmg_magic);
463ee0b2
LW
4617 for (mg = *mgp; mg; mg = *mgp) {
4618 if (mg->mg_type == type) {
e1ec3a88 4619 const MGVTBL* const vtbl = mg->mg_virtual;
463ee0b2 4620 *mgp = mg->mg_moremagic;
1d7c1841 4621 if (vtbl && vtbl->svt_free)
fc0dc3b3 4622 CALL_FPTR(vtbl->svt_free)(aTHX_ sv, mg);
14befaf4 4623 if (mg->mg_ptr && mg->mg_type != PERL_MAGIC_regex_global) {
92110913 4624 if (mg->mg_len > 0)
1edc1566 4625 Safefree(mg->mg_ptr);
565764a8 4626 else if (mg->mg_len == HEf_SVKEY)
1edc1566 4627 SvREFCNT_dec((SV*)mg->mg_ptr);
d2923cdd 4628 else if (mg->mg_type == PERL_MAGIC_utf8)
7e8c5dac 4629 Safefree(mg->mg_ptr);
9cbac4c7 4630 }
a0d0e21e
LW
4631 if (mg->mg_flags & MGf_REFCOUNTED)
4632 SvREFCNT_dec(mg->mg_obj);
463ee0b2
LW
4633 Safefree(mg);
4634 }
4635 else
4636 mgp = &mg->mg_moremagic;
79072805 4637 }
91bba347 4638 if (!SvMAGIC(sv)) {
463ee0b2 4639 SvMAGICAL_off(sv);
c268c2a6 4640 SvFLAGS(sv) |= (SvFLAGS(sv) & (SVp_IOK|SVp_NOK|SVp_POK)) >> PRIVSHIFT;
86f55936 4641 SvMAGIC_set(sv, NULL);
463ee0b2
LW
4642 }
4643
4644 return 0;
79072805
LW
4645}
4646
c461cf8f
JH
4647/*
4648=for apidoc sv_rvweaken
4649
645c22ef
DM
4650Weaken a reference: set the C<SvWEAKREF> flag on this RV; give the
4651referred-to SV C<PERL_MAGIC_backref> magic if it hasn't already; and
4652push a back-reference to this RV onto the array of backreferences
1e73acc8
AS
4653associated with that magic. If the RV is magical, set magic will be
4654called after the RV is cleared.
c461cf8f
JH
4655
4656=cut
4657*/
4658
810b8aa5 4659SV *
864dbfa3 4660Perl_sv_rvweaken(pTHX_ SV *sv)
810b8aa5
GS
4661{
4662 SV *tsv;
4663 if (!SvOK(sv)) /* let undefs pass */
4664 return sv;
4665 if (!SvROK(sv))
cea2e8a9 4666 Perl_croak(aTHX_ "Can't weaken a nonreference");
810b8aa5 4667 else if (SvWEAKREF(sv)) {
810b8aa5 4668 if (ckWARN(WARN_MISC))
9014280d 4669 Perl_warner(aTHX_ packWARN(WARN_MISC), "Reference is already weak");
810b8aa5
GS
4670 return sv;
4671 }
4672 tsv = SvRV(sv);
e15faf7d 4673 Perl_sv_add_backref(aTHX_ tsv, sv);
810b8aa5 4674 SvWEAKREF_on(sv);
1c846c1f 4675 SvREFCNT_dec(tsv);
810b8aa5
GS
4676 return sv;
4677}
4678
645c22ef
DM
4679/* Give tsv backref magic if it hasn't already got it, then push a
4680 * back-reference to sv onto the array associated with the backref magic.
4681 */
4682
e15faf7d
NC
4683void
4684Perl_sv_add_backref(pTHX_ SV *tsv, SV *sv)
810b8aa5 4685{
97aff369 4686 dVAR;
810b8aa5 4687 AV *av;
86f55936
NC
4688
4689 if (SvTYPE(tsv) == SVt_PVHV) {
4690 AV **const avp = Perl_hv_backreferences_p(aTHX_ (HV*)tsv);
4691
4692 av = *avp;
4693 if (!av) {
4694 /* There is no AV in the offical place - try a fixup. */
4695 MAGIC *const mg = mg_find(tsv, PERL_MAGIC_backref);
4696
4697 if (mg) {
4698 /* Aha. They've got it stowed in magic. Bring it back. */
4699 av = (AV*)mg->mg_obj;
4700 /* Stop mg_free decreasing the refernce count. */
4701 mg->mg_obj = NULL;
4702 /* Stop mg_free even calling the destructor, given that
4703 there's no AV to free up. */
4704 mg->mg_virtual = 0;
4705 sv_unmagic(tsv, PERL_MAGIC_backref);
4706 } else {
4707 av = newAV();
4708 AvREAL_off(av);
b37c2d43 4709 SvREFCNT_inc_simple_void(av);
86f55936
NC
4710 }
4711 *avp = av;
4712 }
4713 } else {
4714 const MAGIC *const mg
4715 = SvMAGICAL(tsv) ? mg_find(tsv, PERL_MAGIC_backref) : NULL;
4716 if (mg)
4717 av = (AV*)mg->mg_obj;
4718 else {
4719 av = newAV();
4720 AvREAL_off(av);
4721 sv_magic(tsv, (SV*)av, PERL_MAGIC_backref, NULL, 0);
4722 /* av now has a refcnt of 2, which avoids it getting freed
4723 * before us during global cleanup. The extra ref is removed
4724 * by magic_killbackrefs() when tsv is being freed */
4725 }
810b8aa5 4726 }
d91d49e8 4727 if (AvFILLp(av) >= AvMAX(av)) {
d91d49e8
MM
4728 av_extend(av, AvFILLp(av)+1);
4729 }
4730 AvARRAY(av)[++AvFILLp(av)] = sv; /* av_push() */
810b8aa5
GS
4731}
4732
645c22ef
DM
4733/* delete a back-reference to ourselves from the backref magic associated
4734 * with the SV we point to.
4735 */
4736
1c846c1f 4737STATIC void
e15faf7d 4738S_sv_del_backref(pTHX_ SV *tsv, SV *sv)
810b8aa5 4739{
97aff369 4740 dVAR;
86f55936 4741 AV *av = NULL;
810b8aa5
GS
4742 SV **svp;
4743 I32 i;
86f55936
NC
4744
4745 if (SvTYPE(tsv) == SVt_PVHV && SvOOK(tsv)) {
4746 av = *Perl_hv_backreferences_p(aTHX_ (HV*)tsv);
5b285ea4
NC
4747 /* We mustn't attempt to "fix up" the hash here by moving the
4748 backreference array back to the hv_aux structure, as that is stored
4749 in the main HvARRAY(), and hfreentries assumes that no-one
4750 reallocates HvARRAY() while it is running. */
86f55936
NC
4751 }
4752 if (!av) {
4753 const MAGIC *const mg
4754 = SvMAGICAL(tsv) ? mg_find(tsv, PERL_MAGIC_backref) : NULL;
4755 if (mg)
4756 av = (AV *)mg->mg_obj;
4757 }
4758 if (!av) {
e15faf7d
NC
4759 if (PL_in_clean_all)
4760 return;
cea2e8a9 4761 Perl_croak(aTHX_ "panic: del_backref");
86f55936
NC
4762 }
4763
4764 if (SvIS_FREED(av))
4765 return;
4766
810b8aa5 4767 svp = AvARRAY(av);
6a76db8b
NC
4768 /* We shouldn't be in here more than once, but for paranoia reasons lets
4769 not assume this. */
4770 for (i = AvFILLp(av); i >= 0; i--) {
4771 if (svp[i] == sv) {
4772 const SSize_t fill = AvFILLp(av);
4773 if (i != fill) {
4774 /* We weren't the last entry.
4775 An unordered list has this property that you can take the
4776 last element off the end to fill the hole, and it's still
4777 an unordered list :-)
4778 */
4779 svp[i] = svp[fill];
4780 }
a0714e2c 4781 svp[fill] = NULL;
6a76db8b
NC
4782 AvFILLp(av) = fill - 1;
4783 }
4784 }
810b8aa5
GS
4785}
4786
86f55936
NC
4787int
4788Perl_sv_kill_backrefs(pTHX_ SV *sv, AV *av)
4789{
4790 SV **svp = AvARRAY(av);
4791
4792 PERL_UNUSED_ARG(sv);
4793
4794 /* Not sure why the av can get freed ahead of its sv, but somehow it does
4795 in ext/B/t/bytecode.t test 15 (involving print <DATA>) */
4796 if (svp && !SvIS_FREED(av)) {
4797 SV *const *const last = svp + AvFILLp(av);
4798
4799 while (svp <= last) {
4800 if (*svp) {
4801 SV *const referrer = *svp;
4802 if (SvWEAKREF(referrer)) {
4803 /* XXX Should we check that it hasn't changed? */
4804 SvRV_set(referrer, 0);
4805 SvOK_off(referrer);
4806 SvWEAKREF_off(referrer);
1e73acc8 4807 SvSETMAGIC(referrer);
86f55936
NC
4808 } else if (SvTYPE(referrer) == SVt_PVGV ||
4809 SvTYPE(referrer) == SVt_PVLV) {
4810 /* You lookin' at me? */
4811 assert(GvSTASH(referrer));
4812 assert(GvSTASH(referrer) == (HV*)sv);
4813 GvSTASH(referrer) = 0;
4814 } else {
4815 Perl_croak(aTHX_
4816 "panic: magic_killbackrefs (flags=%"UVxf")",
4817 (UV)SvFLAGS(referrer));
4818 }
4819
a0714e2c 4820 *svp = NULL;
86f55936
NC
4821 }
4822 svp++;
4823 }
4824 }
4825 SvREFCNT_dec(av); /* remove extra count added by sv_add_backref() */
4826 return 0;
4827}
4828
954c1994
GS
4829/*
4830=for apidoc sv_insert
4831
4832Inserts a string at the specified offset/length within the SV. Similar to
4833the Perl substr() function.
4834
4835=cut
4836*/
4837
79072805 4838void
e1ec3a88 4839Perl_sv_insert(pTHX_ SV *bigstr, STRLEN offset, STRLEN len, const char *little, STRLEN littlelen)
79072805 4840{
97aff369 4841 dVAR;
79072805
LW
4842 register char *big;
4843 register char *mid;
4844 register char *midend;
4845 register char *bigend;
4846 register I32 i;
6ff81951 4847 STRLEN curlen;
1c846c1f 4848
79072805 4849
8990e307 4850 if (!bigstr)
cea2e8a9 4851 Perl_croak(aTHX_ "Can't modify non-existent substring");
6ff81951 4852 SvPV_force(bigstr, curlen);
60fa28ff 4853 (void)SvPOK_only_UTF8(bigstr);
6ff81951
GS
4854 if (offset + len > curlen) {
4855 SvGROW(bigstr, offset+len+1);
93524f2b 4856 Zero(SvPVX(bigstr)+curlen, offset+len-curlen, char);
6ff81951
GS
4857 SvCUR_set(bigstr, offset+len);
4858 }
79072805 4859
69b47968 4860 SvTAINT(bigstr);
79072805
LW
4861 i = littlelen - len;
4862 if (i > 0) { /* string might grow */
a0d0e21e 4863 big = SvGROW(bigstr, SvCUR(bigstr) + i + 1);
79072805
LW
4864 mid = big + offset + len;
4865 midend = bigend = big + SvCUR(bigstr);
4866 bigend += i;
4867 *bigend = '\0';
4868 while (midend > mid) /* shove everything down */
4869 *--bigend = *--midend;
4870 Move(little,big+offset,littlelen,char);
b162af07 4871 SvCUR_set(bigstr, SvCUR(bigstr) + i);
79072805
LW
4872 SvSETMAGIC(bigstr);
4873 return;
4874 }
4875 else if (i == 0) {
463ee0b2 4876 Move(little,SvPVX(bigstr)+offset,len,char);
79072805
LW
4877 SvSETMAGIC(bigstr);
4878 return;
4879 }
4880
463ee0b2 4881 big = SvPVX(bigstr);
79072805
LW
4882 mid = big + offset;
4883 midend = mid + len;
4884 bigend = big + SvCUR(bigstr);
4885
4886 if (midend > bigend)
cea2e8a9 4887 Perl_croak(aTHX_ "panic: sv_insert");
79072805
LW
4888
4889 if (mid - big > bigend - midend) { /* faster to shorten from end */
4890 if (littlelen) {
4891 Move(little, mid, littlelen,char);
4892 mid += littlelen;
4893 }
4894 i = bigend - midend;
4895 if (i > 0) {
4896 Move(midend, mid, i,char);
4897 mid += i;
4898 }
4899 *mid = '\0';
4900 SvCUR_set(bigstr, mid - big);
4901 }
155aba94 4902 else if ((i = mid - big)) { /* faster from front */
79072805
LW
4903 midend -= littlelen;
4904 mid = midend;
4905 sv_chop(bigstr,midend-i);
4906 big += i;
4907 while (i--)
4908 *--midend = *--big;
4909 if (littlelen)
4910 Move(little, mid, littlelen,char);
4911 }
4912 else if (littlelen) {
4913 midend -= littlelen;
4914 sv_chop(bigstr,midend);
4915 Move(little,midend,littlelen,char);
4916 }
4917 else {
4918 sv_chop(bigstr,midend);
4919 }
4920 SvSETMAGIC(bigstr);
4921}
4922
c461cf8f
JH
4923/*
4924=for apidoc sv_replace
4925
4926Make the first argument a copy of the second, then delete the original.
645c22ef
DM
4927The target SV physically takes over ownership of the body of the source SV
4928and inherits its flags; however, the target keeps any magic it owns,
4929and any magic in the source is discarded.
ff276b08 4930Note that this is a rather specialist SV copying operation; most of the
645c22ef 4931time you'll want to use C<sv_setsv> or one of its many macro front-ends.
c461cf8f
JH
4932
4933=cut
4934*/
79072805
LW
4935
4936void
864dbfa3 4937Perl_sv_replace(pTHX_ register SV *sv, register SV *nsv)
79072805 4938{
97aff369 4939 dVAR;
a3b680e6 4940 const U32 refcnt = SvREFCNT(sv);
765f542d 4941 SV_CHECK_THINKFIRST_COW_DROP(sv);
30e5c352 4942 if (SvREFCNT(nsv) != 1) {
7437becc 4943 Perl_croak(aTHX_ "panic: reference miscount on nsv in sv_replace() (%"
30e5c352
NC
4944 UVuf " != 1)", (UV) SvREFCNT(nsv));
4945 }
93a17b20 4946 if (SvMAGICAL(sv)) {
a0d0e21e
LW
4947 if (SvMAGICAL(nsv))
4948 mg_free(nsv);
4949 else
4950 sv_upgrade(nsv, SVt_PVMG);
b162af07 4951 SvMAGIC_set(nsv, SvMAGIC(sv));
a0d0e21e 4952 SvFLAGS(nsv) |= SvMAGICAL(sv);
93a17b20 4953 SvMAGICAL_off(sv);
b162af07 4954 SvMAGIC_set(sv, NULL);
93a17b20 4955 }
79072805
LW
4956 SvREFCNT(sv) = 0;
4957 sv_clear(sv);
477f5d66 4958 assert(!SvREFCNT(sv));
fd0854ff
DM
4959#ifdef DEBUG_LEAKING_SCALARS
4960 sv->sv_flags = nsv->sv_flags;
4961 sv->sv_any = nsv->sv_any;
4962 sv->sv_refcnt = nsv->sv_refcnt;
f34d0642 4963 sv->sv_u = nsv->sv_u;
fd0854ff 4964#else
79072805 4965 StructCopy(nsv,sv,SV);
fd0854ff 4966#endif
7b2c381c
NC
4967 /* Currently could join these into one piece of pointer arithmetic, but
4968 it would be unclear. */
4969 if(SvTYPE(sv) == SVt_IV)
4970 SvANY(sv)
339049b0 4971 = (XPVIV*)((char*)&(sv->sv_u.svu_iv) - STRUCT_OFFSET(XPVIV, xiv_iv));
7b2c381c 4972 else if (SvTYPE(sv) == SVt_RV) {
339049b0 4973 SvANY(sv) = &sv->sv_u.svu_rv;
7b2c381c
NC
4974 }
4975
fd0854ff 4976
f8c7b90f 4977#ifdef PERL_OLD_COPY_ON_WRITE
d3d0e6f1
NC
4978 if (SvIsCOW_normal(nsv)) {
4979 /* We need to follow the pointers around the loop to make the
4980 previous SV point to sv, rather than nsv. */
4981 SV *next;
4982 SV *current = nsv;
4983 while ((next = SV_COW_NEXT_SV(current)) != nsv) {
4984 assert(next);
4985 current = next;
3f7c398e 4986 assert(SvPVX_const(current) == SvPVX_const(nsv));
d3d0e6f1
NC
4987 }
4988 /* Make the SV before us point to the SV after us. */
4989 if (DEBUG_C_TEST) {
4990 PerlIO_printf(Perl_debug_log, "previous is\n");
4991 sv_dump(current);
a29f6d03
NC
4992 PerlIO_printf(Perl_debug_log,
4993 "move it from 0x%"UVxf" to 0x%"UVxf"\n",
d3d0e6f1
NC
4994 (UV) SV_COW_NEXT_SV(current), (UV) sv);
4995 }
a29f6d03 4996 SV_COW_NEXT_SV_SET(current, sv);
d3d0e6f1
NC
4997 }
4998#endif
79072805 4999 SvREFCNT(sv) = refcnt;
1edc1566 5000 SvFLAGS(nsv) |= SVTYPEMASK; /* Mark as freed */
39cf41c2 5001 SvREFCNT(nsv) = 0;
463ee0b2 5002 del_SV(nsv);
79072805
LW
5003}
5004
c461cf8f
JH
5005/*
5006=for apidoc sv_clear
5007
645c22ef
DM
5008Clear an SV: call any destructors, free up any memory used by the body,
5009and free the body itself. The SV's head is I<not> freed, although
5010its type is set to all 1's so that it won't inadvertently be assumed
5011to be live during global destruction etc.
5012This function should only be called when REFCNT is zero. Most of the time
5013you'll want to call C<sv_free()> (or its macro wrapper C<SvREFCNT_dec>)
5014instead.
c461cf8f
JH
5015
5016=cut
5017*/
5018
79072805 5019void
864dbfa3 5020Perl_sv_clear(pTHX_ register SV *sv)
79072805 5021{
27da23d5 5022 dVAR;
82bb6deb 5023 const U32 type = SvTYPE(sv);
8edfc514
NC
5024 const struct body_details *const sv_type_details
5025 = bodies_by_type + type;
82bb6deb 5026
79072805
LW
5027 assert(sv);
5028 assert(SvREFCNT(sv) == 0);
5029
d2a0f284
JC
5030 if (type <= SVt_IV) {
5031 /* See the comment in sv.h about the collusion between this early
5032 return and the overloading of the NULL and IV slots in the size
5033 table. */
82bb6deb 5034 return;
d2a0f284 5035 }
82bb6deb 5036
ed6116ce 5037 if (SvOBJECT(sv)) {
3280af22 5038 if (PL_defstash) { /* Still have a symbol table? */
39644a26 5039 dSP;
893645bd 5040 HV* stash;
d460ef45 5041 do {
b464bac0 5042 CV* destructor;
4e8e7886 5043 stash = SvSTASH(sv);
32251b26 5044 destructor = StashHANDLER(stash,DESTROY);
4e8e7886 5045 if (destructor) {
1b6737cc 5046 SV* const tmpref = newRV(sv);
5cc433a6 5047 SvREADONLY_on(tmpref); /* DESTROY() could be naughty */
4e8e7886 5048 ENTER;
e788e7d3 5049 PUSHSTACKi(PERLSI_DESTROY);
4e8e7886
GS
5050 EXTEND(SP, 2);
5051 PUSHMARK(SP);
5cc433a6 5052 PUSHs(tmpref);
4e8e7886 5053 PUTBACK;
44389ee9 5054 call_sv((SV*)destructor, G_DISCARD|G_EVAL|G_KEEPERR|G_VOID);
7a5fa8a2
NIS
5055
5056
d3acc0f7 5057 POPSTACK;
3095d977 5058 SPAGAIN;
4e8e7886 5059 LEAVE;
5cc433a6
AB
5060 if(SvREFCNT(tmpref) < 2) {
5061 /* tmpref is not kept alive! */
5062 SvREFCNT(sv)--;
b162af07 5063 SvRV_set(tmpref, NULL);
5cc433a6
AB
5064 SvROK_off(tmpref);
5065 }
5066 SvREFCNT_dec(tmpref);
4e8e7886
GS
5067 }
5068 } while (SvOBJECT(sv) && SvSTASH(sv) != stash);
8ebc5c01 5069
6f44e0a4
JP
5070
5071 if (SvREFCNT(sv)) {
5072 if (PL_in_clean_objs)
cea2e8a9 5073 Perl_croak(aTHX_ "DESTROY created new reference to dead object '%s'",
bfcb3514 5074 HvNAME_get(stash));
6f44e0a4
JP
5075 /* DESTROY gave object new lease on life */
5076 return;
5077 }
a0d0e21e 5078 }
4e8e7886 5079
a0d0e21e 5080 if (SvOBJECT(sv)) {
4e8e7886 5081 SvREFCNT_dec(SvSTASH(sv)); /* possibly of changed persuasion */
a0d0e21e 5082 SvOBJECT_off(sv); /* Curse the object. */
82bb6deb 5083 if (type != SVt_PVIO)
3280af22 5084 --PL_sv_objcount; /* XXX Might want something more general */
a0d0e21e 5085 }
463ee0b2 5086 }
82bb6deb 5087 if (type >= SVt_PVMG) {
885ffcb3
NC
5088 if ((type == SVt_PVMG || type == SVt_PVGV) && SvPAD_OUR(sv)) {
5089 SvREFCNT_dec(OURSTASH(sv));
e736a858 5090 } else if (SvMAGIC(sv))
524189f1 5091 mg_free(sv);
00b1698f 5092 if (type == SVt_PVMG && SvPAD_TYPED(sv))
524189f1
JH
5093 SvREFCNT_dec(SvSTASH(sv));
5094 }
82bb6deb 5095 switch (type) {
8990e307 5096 case SVt_PVIO:
df0bd2f4
GS
5097 if (IoIFP(sv) &&
5098 IoIFP(sv) != PerlIO_stdin() &&
5f05dabc 5099 IoIFP(sv) != PerlIO_stdout() &&
5100 IoIFP(sv) != PerlIO_stderr())
93578b34 5101 {
f2b5be74 5102 io_close((IO*)sv, FALSE);
93578b34 5103 }
1d7c1841 5104 if (IoDIRP(sv) && !(IoFLAGS(sv) & IOf_FAKE_DIRP))
1236053a 5105 PerlDir_close(IoDIRP(sv));
1d7c1841 5106 IoDIRP(sv) = (DIR*)NULL;
8990e307
LW
5107 Safefree(IoTOP_NAME(sv));
5108 Safefree(IoFMT_NAME(sv));
5109 Safefree(IoBOTTOM_NAME(sv));
82bb6deb 5110 goto freescalar;
79072805 5111 case SVt_PVBM:
a0d0e21e 5112 goto freescalar;
79072805 5113 case SVt_PVCV:
748a9306 5114 case SVt_PVFM:
85e6fe83 5115 cv_undef((CV*)sv);
a0d0e21e 5116 goto freescalar;
79072805 5117 case SVt_PVHV:
86f55936 5118 Perl_hv_kill_backrefs(aTHX_ (HV*)sv);
85e6fe83 5119 hv_undef((HV*)sv);
a0d0e21e 5120 break;
79072805 5121 case SVt_PVAV:
85e6fe83 5122 av_undef((AV*)sv);
a0d0e21e 5123 break;
02270b4e 5124 case SVt_PVLV:
dd28f7bb
DM
5125 if (LvTYPE(sv) == 'T') { /* for tie: return HE to pool */
5126 SvREFCNT_dec(HeKEY_sv((HE*)LvTARG(sv)));
5127 HeNEXT((HE*)LvTARG(sv)) = PL_hv_fetch_ent_mh;
5128 PL_hv_fetch_ent_mh = (HE*)LvTARG(sv);
5129 }
5130 else if (LvTYPE(sv) != 't') /* unless tie: unrefcnted fake SV** */
5131 SvREFCNT_dec(LvTARG(sv));
02270b4e 5132 goto freescalar;
a0d0e21e 5133 case SVt_PVGV:
1edc1566 5134 gp_free((GV*)sv);
acda4c6a
NC
5135 if (GvNAME_HEK(sv)) {
5136 unshare_hek(GvNAME_HEK(sv));
5137 }
893645bd
NC
5138 /* If we're in a stash, we don't own a reference to it. However it does
5139 have a back reference to us, which needs to be cleared. */
5140 if (GvSTASH(sv))
5141 sv_del_backref((SV*)GvSTASH(sv), sv);
79072805 5142 case SVt_PVMG:
79072805
LW
5143 case SVt_PVNV:
5144 case SVt_PVIV:
a0d0e21e 5145 freescalar:
5228ca4e
NC
5146 /* Don't bother with SvOOK_off(sv); as we're only going to free it. */
5147 if (SvOOK(sv)) {
93524f2b 5148 SvPV_set(sv, SvPVX_mutable(sv) - SvIVX(sv));
5228ca4e
NC
5149 /* Don't even bother with turning off the OOK flag. */
5150 }
79072805 5151 case SVt_PV:
a0d0e21e 5152 case SVt_RV:
810b8aa5 5153 if (SvROK(sv)) {
b37c2d43 5154 SV * const target = SvRV(sv);
810b8aa5 5155 if (SvWEAKREF(sv))
e15faf7d 5156 sv_del_backref(target, sv);
810b8aa5 5157 else
e15faf7d 5158 SvREFCNT_dec(target);
810b8aa5 5159 }
f8c7b90f 5160#ifdef PERL_OLD_COPY_ON_WRITE
3f7c398e 5161 else if (SvPVX_const(sv)) {
765f542d
NC
5162 if (SvIsCOW(sv)) {
5163 /* I believe I need to grab the global SV mutex here and
5164 then recheck the COW status. */
46187eeb
NC
5165 if (DEBUG_C_TEST) {
5166 PerlIO_printf(Perl_debug_log, "Copy on write: clear\n");
e419cbc5 5167 sv_dump(sv);
46187eeb 5168 }
bdd68bc3
NC
5169 sv_release_COW(sv, SvPVX_const(sv), SvLEN(sv),
5170 SV_COW_NEXT_SV(sv));
765f542d
NC
5171 /* And drop it here. */
5172 SvFAKE_off(sv);
5173 } else if (SvLEN(sv)) {
3f7c398e 5174 Safefree(SvPVX_const(sv));
765f542d
NC
5175 }
5176 }
5177#else
3f7c398e 5178 else if (SvPVX_const(sv) && SvLEN(sv))
94010e71 5179 Safefree(SvPVX_mutable(sv));
3f7c398e 5180 else if (SvPVX_const(sv) && SvREADONLY(sv) && SvFAKE(sv)) {
bdd68bc3 5181 unshare_hek(SvSHARED_HEK_FROM_PV(SvPVX_const(sv)));
1c846c1f
NIS
5182 SvFAKE_off(sv);
5183 }
765f542d 5184#endif
79072805
LW
5185 break;
5186 case SVt_NV:
79072805
LW
5187 break;
5188 }
5189
893645bd
NC
5190 SvFLAGS(sv) &= SVf_BREAK;
5191 SvFLAGS(sv) |= SVTYPEMASK;
5192
8edfc514 5193 if (sv_type_details->arena) {
b9502f15 5194 del_body(((char *)SvANY(sv) + sv_type_details->offset),
8edfc514
NC
5195 &PL_body_roots[type]);
5196 }
d2a0f284 5197 else if (sv_type_details->body_size) {
8edfc514
NC
5198 my_safefree(SvANY(sv));
5199 }
79072805
LW
5200}
5201
645c22ef
DM
5202/*
5203=for apidoc sv_newref
5204
5205Increment an SV's reference count. Use the C<SvREFCNT_inc()> wrapper
5206instead.
5207
5208=cut
5209*/
5210
79072805 5211SV *
864dbfa3 5212Perl_sv_newref(pTHX_ SV *sv)
79072805 5213{
96a5add6 5214 PERL_UNUSED_CONTEXT;
463ee0b2 5215 if (sv)
4db098f4 5216 (SvREFCNT(sv))++;
79072805
LW
5217 return sv;
5218}
5219
c461cf8f
JH
5220/*
5221=for apidoc sv_free
5222
645c22ef
DM
5223Decrement an SV's reference count, and if it drops to zero, call
5224C<sv_clear> to invoke destructors and free up any memory used by
5225the body; finally, deallocate the SV's head itself.
5226Normally called via a wrapper macro C<SvREFCNT_dec>.
c461cf8f
JH
5227
5228=cut
5229*/
5230
79072805 5231void
864dbfa3 5232Perl_sv_free(pTHX_ SV *sv)
79072805 5233{
27da23d5 5234 dVAR;
79072805
LW
5235 if (!sv)
5236 return;
a0d0e21e
LW
5237 if (SvREFCNT(sv) == 0) {
5238 if (SvFLAGS(sv) & SVf_BREAK)
645c22ef
DM
5239 /* this SV's refcnt has been artificially decremented to
5240 * trigger cleanup */
a0d0e21e 5241 return;
3280af22 5242 if (PL_in_clean_all) /* All is fair */
1edc1566 5243 return;
d689ffdd
JP
5244 if (SvREADONLY(sv) && SvIMMORTAL(sv)) {
5245 /* make sure SvREFCNT(sv)==0 happens very seldom */
5246 SvREFCNT(sv) = (~(U32)0)/2;
5247 return;
5248 }
41e4abd8 5249 if (ckWARN_d(WARN_INTERNAL)) {
d5dede04 5250 Perl_warner(aTHX_ packWARN(WARN_INTERNAL),
472d47bc
SB
5251 "Attempt to free unreferenced scalar: SV 0x%"UVxf
5252 pTHX__FORMAT, PTR2UV(sv) pTHX__VALUE);
41e4abd8
NC
5253#ifdef DEBUG_LEAKING_SCALARS_FORK_DUMP
5254 Perl_dump_sv_child(aTHX_ sv);
5255#endif
5256 }
79072805
LW
5257 return;
5258 }
4db098f4 5259 if (--(SvREFCNT(sv)) > 0)
8990e307 5260 return;
8c4d3c90
NC
5261 Perl_sv_free2(aTHX_ sv);
5262}
5263
5264void
5265Perl_sv_free2(pTHX_ SV *sv)
5266{
27da23d5 5267 dVAR;
463ee0b2
LW
5268#ifdef DEBUGGING
5269 if (SvTEMP(sv)) {
0453d815 5270 if (ckWARN_d(WARN_DEBUGGING))
9014280d 5271 Perl_warner(aTHX_ packWARN(WARN_DEBUGGING),
472d47bc
SB
5272 "Attempt to free temp prematurely: SV 0x%"UVxf
5273 pTHX__FORMAT, PTR2UV(sv) pTHX__VALUE);
79072805 5274 return;
79072805 5275 }
463ee0b2 5276#endif
d689ffdd
JP
5277 if (SvREADONLY(sv) && SvIMMORTAL(sv)) {
5278 /* make sure SvREFCNT(sv)==0 happens very seldom */
5279 SvREFCNT(sv) = (~(U32)0)/2;
5280 return;
5281 }
79072805 5282 sv_clear(sv);
477f5d66
CS
5283 if (! SvREFCNT(sv))
5284 del_SV(sv);
79072805
LW
5285}
5286
954c1994
GS
5287/*
5288=for apidoc sv_len
5289
645c22ef
DM
5290Returns the length of the string in the SV. Handles magic and type
5291coercion. See also C<SvCUR>, which gives raw access to the xpv_cur slot.
954c1994
GS
5292
5293=cut
5294*/
5295
79072805 5296STRLEN
864dbfa3 5297Perl_sv_len(pTHX_ register SV *sv)
79072805 5298{
463ee0b2 5299 STRLEN len;
79072805
LW
5300
5301 if (!sv)
5302 return 0;
5303
8990e307 5304 if (SvGMAGICAL(sv))
565764a8 5305 len = mg_length(sv);
8990e307 5306 else
4d84ee25 5307 (void)SvPV_const(sv, len);
463ee0b2 5308 return len;
79072805
LW
5309}
5310
c461cf8f
JH
5311/*
5312=for apidoc sv_len_utf8
5313
5314Returns the number of characters in the string in an SV, counting wide
1e54db1a 5315UTF-8 bytes as a single character. Handles magic and type coercion.
c461cf8f
JH
5316
5317=cut
5318*/
5319
7e8c5dac
HS
5320/*
5321 * The length is cached in PERL_UTF8_magic, in the mg_len field. Also the
9564a3bd
NC
5322 * mg_ptr is used, by sv_pos_u2b() and sv_pos_b2u() - see the comments below.
5323 * (Note that the mg_len is not the length of the mg_ptr field.
5324 * This allows the cache to store the character length of the string without
5325 * needing to malloc() extra storage to attach to the mg_ptr.)
7a5fa8a2 5326 *
7e8c5dac
HS
5327 */
5328
a0ed51b3 5329STRLEN
864dbfa3 5330Perl_sv_len_utf8(pTHX_ register SV *sv)
a0ed51b3 5331{
a0ed51b3
LW
5332 if (!sv)
5333 return 0;
5334
a0ed51b3 5335 if (SvGMAGICAL(sv))
b76347f2 5336 return mg_length(sv);
a0ed51b3 5337 else
b76347f2 5338 {
26346457 5339 STRLEN len;
e62f0680 5340 const U8 *s = (U8*)SvPV_const(sv, len);
7e8c5dac 5341
26346457
NC
5342 if (PL_utf8cache) {
5343 STRLEN ulen;
5344 MAGIC *mg = SvMAGICAL(sv) ? mg_find(sv, PERL_MAGIC_utf8) : 0;
5345
5346 if (mg && mg->mg_len != -1) {
5347 ulen = mg->mg_len;
5348 if (PL_utf8cache < 0) {
5349 const STRLEN real = Perl_utf8_length(aTHX_ s, s + len);
5350 if (real != ulen) {
5351 /* Need to turn the assertions off otherwise we may
5352 recurse infinitely while printing error messages.
5353 */
5354 SAVEI8(PL_utf8cache);
5355 PL_utf8cache = 0;
f5992bc4
RB
5356 Perl_croak(aTHX_ "panic: sv_len_utf8 cache %"UVuf
5357 " real %"UVuf" for %"SVf,
95b63a38 5358 (UV) ulen, (UV) real, (void*)sv);
26346457
NC
5359 }
5360 }
5361 }
5362 else {
5363 ulen = Perl_utf8_length(aTHX_ s, s + len);
5364 if (!SvREADONLY(sv)) {
5365 if (!mg) {
5366 mg = sv_magicext(sv, 0, PERL_MAGIC_utf8,
5367 &PL_vtbl_utf8, 0, 0);
5368 }
cb9e20bb 5369 assert(mg);
26346457 5370 mg->mg_len = ulen;
cb9e20bb 5371 }
cb9e20bb 5372 }
26346457 5373 return ulen;
7e8c5dac 5374 }
26346457 5375 return Perl_utf8_length(aTHX_ s, s + len);
7e8c5dac
HS
5376 }
5377}
5378
9564a3bd
NC
5379/* Walk forwards to find the byte corresponding to the passed in UTF-8
5380 offset. */
bdf30dd6 5381static STRLEN
721e86b6 5382S_sv_pos_u2b_forwards(const U8 *const start, const U8 *const send,
bdf30dd6
NC
5383 STRLEN uoffset)
5384{
5385 const U8 *s = start;
5386
5387 while (s < send && uoffset--)
5388 s += UTF8SKIP(s);
5389 if (s > send) {
5390 /* This is the existing behaviour. Possibly it should be a croak, as
5391 it's actually a bounds error */
5392 s = send;
5393 }
5394 return s - start;
5395}
5396
9564a3bd
NC
5397/* Given the length of the string in both bytes and UTF-8 characters, decide
5398 whether to walk forwards or backwards to find the byte corresponding to
5399 the passed in UTF-8 offset. */
c336ad0b 5400static STRLEN
721e86b6 5401S_sv_pos_u2b_midway(const U8 *const start, const U8 *send,
c336ad0b
NC
5402 STRLEN uoffset, STRLEN uend)
5403{
5404 STRLEN backw = uend - uoffset;
5405 if (uoffset < 2 * backw) {
25a8a4ef 5406 /* The assumption is that going forwards is twice the speed of going
c336ad0b
NC
5407 forward (that's where the 2 * backw comes from).
5408 (The real figure of course depends on the UTF-8 data.) */
721e86b6 5409 return sv_pos_u2b_forwards(start, send, uoffset);
c336ad0b
NC
5410 }
5411
5412 while (backw--) {
5413 send--;
5414 while (UTF8_IS_CONTINUATION(*send))
5415 send--;
5416 }
5417 return send - start;
5418}
5419
9564a3bd
NC
5420/* For the string representation of the given scalar, find the byte
5421 corresponding to the passed in UTF-8 offset. uoffset0 and boffset0
5422 give another position in the string, *before* the sought offset, which
5423 (which is always true, as 0, 0 is a valid pair of positions), which should
5424 help reduce the amount of linear searching.
5425 If *mgp is non-NULL, it should point to the UTF-8 cache magic, which
5426 will be used to reduce the amount of linear searching. The cache will be
5427 created if necessary, and the found value offered to it for update. */
28ccbf94
NC
5428static STRLEN
5429S_sv_pos_u2b_cached(pTHX_ SV *sv, MAGIC **mgp, const U8 *const start,
5430 const U8 *const send, STRLEN uoffset,
5431 STRLEN uoffset0, STRLEN boffset0) {
7087a21c 5432 STRLEN boffset = 0; /* Actually always set, but let's keep gcc happy. */
c336ad0b
NC
5433 bool found = FALSE;
5434
75c33c12
NC
5435 assert (uoffset >= uoffset0);
5436
c336ad0b 5437 if (SvMAGICAL(sv) && !SvREADONLY(sv) && PL_utf8cache
0905937d 5438 && (*mgp || (*mgp = mg_find(sv, PERL_MAGIC_utf8)))) {
d8b2e1f9
NC
5439 if ((*mgp)->mg_ptr) {
5440 STRLEN *cache = (STRLEN *) (*mgp)->mg_ptr;
5441 if (cache[0] == uoffset) {
5442 /* An exact match. */
5443 return cache[1];
5444 }
ab455f60
NC
5445 if (cache[2] == uoffset) {
5446 /* An exact match. */
5447 return cache[3];
5448 }
668af93f
NC
5449
5450 if (cache[0] < uoffset) {
d8b2e1f9
NC
5451 /* The cache already knows part of the way. */
5452 if (cache[0] > uoffset0) {
5453 /* The cache knows more than the passed in pair */
5454 uoffset0 = cache[0];
5455 boffset0 = cache[1];
5456 }
5457 if ((*mgp)->mg_len != -1) {
5458 /* And we know the end too. */
5459 boffset = boffset0
721e86b6 5460 + sv_pos_u2b_midway(start + boffset0, send,
d8b2e1f9
NC
5461 uoffset - uoffset0,
5462 (*mgp)->mg_len - uoffset0);
5463 } else {
5464 boffset = boffset0
721e86b6 5465 + sv_pos_u2b_forwards(start + boffset0,
d8b2e1f9
NC
5466 send, uoffset - uoffset0);
5467 }
dd7c5fd3
NC
5468 }
5469 else if (cache[2] < uoffset) {
5470 /* We're between the two cache entries. */
5471 if (cache[2] > uoffset0) {
5472 /* and the cache knows more than the passed in pair */
5473 uoffset0 = cache[2];
5474 boffset0 = cache[3];
5475 }
5476
668af93f 5477 boffset = boffset0
721e86b6 5478 + sv_pos_u2b_midway(start + boffset0,
668af93f
NC
5479 start + cache[1],
5480 uoffset - uoffset0,
5481 cache[0] - uoffset0);
dd7c5fd3
NC
5482 } else {
5483 boffset = boffset0
721e86b6 5484 + sv_pos_u2b_midway(start + boffset0,
dd7c5fd3
NC
5485 start + cache[3],
5486 uoffset - uoffset0,
5487 cache[2] - uoffset0);
d8b2e1f9 5488 }
668af93f 5489 found = TRUE;
d8b2e1f9
NC
5490 }
5491 else if ((*mgp)->mg_len != -1) {
75c33c12
NC
5492 /* If we can take advantage of a passed in offset, do so. */
5493 /* In fact, offset0 is either 0, or less than offset, so don't
5494 need to worry about the other possibility. */
5495 boffset = boffset0
721e86b6 5496 + sv_pos_u2b_midway(start + boffset0, send,
75c33c12
NC
5497 uoffset - uoffset0,
5498 (*mgp)->mg_len - uoffset0);
c336ad0b
NC
5499 found = TRUE;
5500 }
28ccbf94 5501 }
c336ad0b
NC
5502
5503 if (!found || PL_utf8cache < 0) {
75c33c12 5504 const STRLEN real_boffset
721e86b6 5505 = boffset0 + sv_pos_u2b_forwards(start + boffset0,
75c33c12
NC
5506 send, uoffset - uoffset0);
5507
c336ad0b
NC
5508 if (found && PL_utf8cache < 0) {
5509 if (real_boffset != boffset) {
5510 /* Need to turn the assertions off otherwise we may recurse
5511 infinitely while printing error messages. */
5512 SAVEI8(PL_utf8cache);
5513 PL_utf8cache = 0;
f5992bc4
RB
5514 Perl_croak(aTHX_ "panic: sv_pos_u2b_cache cache %"UVuf
5515 " real %"UVuf" for %"SVf,
95b63a38 5516 (UV) boffset, (UV) real_boffset, (void*)sv);
c336ad0b
NC
5517 }
5518 }
5519 boffset = real_boffset;
28ccbf94 5520 }
0905937d 5521
ab455f60 5522 S_utf8_mg_pos_cache_update(aTHX_ sv, mgp, boffset, uoffset, send - start);
28ccbf94
NC
5523 return boffset;
5524}
5525
9564a3bd
NC
5526
5527/*
5528=for apidoc sv_pos_u2b
5529
5530Converts the value pointed to by offsetp from a count of UTF-8 chars from
5531the start of the string, to a count of the equivalent number of bytes; if
5532lenp is non-zero, it does the same to lenp, but this time starting from
5533the offset, rather than from the start of the string. Handles magic and
5534type coercion.
5535
5536=cut
5537*/
5538
5539/*
5540 * sv_pos_u2b() uses, like sv_pos_b2u(), the mg_ptr of the potential
5541 * PERL_UTF8_magic of the sv to store the mapping between UTF-8 and
5542 * byte offsets. See also the comments of S_utf8_mg_pos_cache_update().
5543 *
5544 */
5545
a0ed51b3 5546void
864dbfa3 5547Perl_sv_pos_u2b(pTHX_ register SV *sv, I32* offsetp, I32* lenp)
a0ed51b3 5548{
245d4a47 5549 const U8 *start;
a0ed51b3
LW
5550 STRLEN len;
5551
5552 if (!sv)
5553 return;
5554
245d4a47 5555 start = (U8*)SvPV_const(sv, len);
7e8c5dac 5556 if (len) {
bdf30dd6
NC
5557 STRLEN uoffset = (STRLEN) *offsetp;
5558 const U8 * const send = start + len;
0905937d 5559 MAGIC *mg = NULL;
721e86b6 5560 const STRLEN boffset = sv_pos_u2b_cached(sv, &mg, start, send,
28ccbf94 5561 uoffset, 0, 0);
bdf30dd6
NC
5562
5563 *offsetp = (I32) boffset;
5564
5565 if (lenp) {
28ccbf94 5566 /* Convert the relative offset to absolute. */
721e86b6
AL
5567 const STRLEN uoffset2 = uoffset + (STRLEN) *lenp;
5568 const STRLEN boffset2
5569 = sv_pos_u2b_cached(sv, &mg, start, send, uoffset2,
28ccbf94 5570 uoffset, boffset) - boffset;
bdf30dd6 5571
28ccbf94 5572 *lenp = boffset2;
bdf30dd6 5573 }
7e8c5dac
HS
5574 }
5575 else {
5576 *offsetp = 0;
5577 if (lenp)
5578 *lenp = 0;
a0ed51b3 5579 }
e23c8137 5580
a0ed51b3
LW
5581 return;
5582}
5583
9564a3bd
NC
5584/* Create and update the UTF8 magic offset cache, with the proffered utf8/
5585 byte length pairing. The (byte) length of the total SV is passed in too,
5586 as blen, because for some (more esoteric) SVs, the call to SvPV_const()
5587 may not have updated SvCUR, so we can't rely on reading it directly.
5588
5589 The proffered utf8/byte length pairing isn't used if the cache already has
5590 two pairs, and swapping either for the proffered pair would increase the
5591 RMS of the intervals between known byte offsets.
5592
5593 The cache itself consists of 4 STRLEN values
5594 0: larger UTF-8 offset
5595 1: corresponding byte offset
5596 2: smaller UTF-8 offset
5597 3: corresponding byte offset
5598
5599 Unused cache pairs have the value 0, 0.
5600 Keeping the cache "backwards" means that the invariant of
5601 cache[0] >= cache[2] is maintained even with empty slots, which means that
5602 the code that uses it doesn't need to worry if only 1 entry has actually
5603 been set to non-zero. It also makes the "position beyond the end of the
5604 cache" logic much simpler, as the first slot is always the one to start
5605 from.
645c22ef 5606*/
ec07b5e0 5607static void
ab455f60
NC
5608S_utf8_mg_pos_cache_update(pTHX_ SV *sv, MAGIC **mgp, STRLEN byte, STRLEN utf8,
5609 STRLEN blen)
ec07b5e0
NC
5610{
5611 STRLEN *cache;
5612 if (SvREADONLY(sv))
5613 return;
5614
5615 if (!*mgp) {
5616 *mgp = sv_magicext(sv, 0, PERL_MAGIC_utf8, (MGVTBL*)&PL_vtbl_utf8, 0,
5617 0);
5618 (*mgp)->mg_len = -1;
5619 }
5620 assert(*mgp);
5621
5622 if (!(cache = (STRLEN *)(*mgp)->mg_ptr)) {
5623 Newxz(cache, PERL_MAGIC_UTF8_CACHESIZE * 2, STRLEN);
5624 (*mgp)->mg_ptr = (char *) cache;
5625 }
5626 assert(cache);
5627
5628 if (PL_utf8cache < 0) {
ef816a78 5629 const U8 *start = (const U8 *) SvPVX_const(sv);
0905937d
NC
5630 const U8 *const end = start + byte;
5631 STRLEN realutf8 = 0;
5632
5633 while (start < end) {
5634 start += UTF8SKIP(start);
5635 realutf8++;
5636 }
5637
5638 /* Can't use S_sv_pos_b2u_forwards as it will scream warnings on
5639 surrogates. FIXME - is it inconsistent that b2u warns, but u2b
5640 doesn't? I don't know whether this difference was introduced with
5641 the caching code in 5.8.1. */
ec07b5e0
NC
5642
5643 if (realutf8 != utf8) {
5644 /* Need to turn the assertions off otherwise we may recurse
5645 infinitely while printing error messages. */
5646 SAVEI8(PL_utf8cache);
5647 PL_utf8cache = 0;
f5992bc4
RB
5648 Perl_croak(aTHX_ "panic: utf8_mg_pos_cache_update cache %"UVuf
5649 " real %"UVuf" for %"SVf, (UV) utf8, (UV) realutf8, (void*)sv);
ec07b5e0
NC
5650 }
5651 }
ab455f60
NC
5652
5653 /* Cache is held with the later position first, to simplify the code
5654 that deals with unbounded ends. */
5655
5656 ASSERT_UTF8_CACHE(cache);
5657 if (cache[1] == 0) {
5658 /* Cache is totally empty */
5659 cache[0] = utf8;
5660 cache[1] = byte;
5661 } else if (cache[3] == 0) {
5662 if (byte > cache[1]) {
5663 /* New one is larger, so goes first. */
5664 cache[2] = cache[0];
5665 cache[3] = cache[1];
5666 cache[0] = utf8;
5667 cache[1] = byte;
5668 } else {
5669 cache[2] = utf8;
5670 cache[3] = byte;
5671 }
5672 } else {
5673#define THREEWAY_SQUARE(a,b,c,d) \
5674 ((float)((d) - (c))) * ((float)((d) - (c))) \
5675 + ((float)((c) - (b))) * ((float)((c) - (b))) \
5676 + ((float)((b) - (a))) * ((float)((b) - (a)))
5677
5678 /* Cache has 2 slots in use, and we know three potential pairs.
5679 Keep the two that give the lowest RMS distance. Do the
5680 calcualation in bytes simply because we always know the byte
5681 length. squareroot has the same ordering as the positive value,
5682 so don't bother with the actual square root. */
5683 const float existing = THREEWAY_SQUARE(0, cache[3], cache[1], blen);
5684 if (byte > cache[1]) {
5685 /* New position is after the existing pair of pairs. */
5686 const float keep_earlier
5687 = THREEWAY_SQUARE(0, cache[3], byte, blen);
5688 const float keep_later
5689 = THREEWAY_SQUARE(0, cache[1], byte, blen);
5690
5691 if (keep_later < keep_earlier) {
5692 if (keep_later < existing) {
5693 cache[2] = cache[0];
5694 cache[3] = cache[1];
5695 cache[0] = utf8;
5696 cache[1] = byte;
5697 }
5698 }
5699 else {
5700 if (keep_earlier < existing) {
5701 cache[0] = utf8;
5702 cache[1] = byte;
5703 }
5704 }
5705 }
57d7fbf1
NC
5706 else if (byte > cache[3]) {
5707 /* New position is between the existing pair of pairs. */
5708 const float keep_earlier
5709 = THREEWAY_SQUARE(0, cache[3], byte, blen);
5710 const float keep_later
5711 = THREEWAY_SQUARE(0, byte, cache[1], blen);
5712
5713 if (keep_later < keep_earlier) {
5714 if (keep_later < existing) {
5715 cache[2] = utf8;
5716 cache[3] = byte;
5717 }
5718 }
5719 else {
5720 if (keep_earlier < existing) {
5721 cache[0] = utf8;
5722 cache[1] = byte;
5723 }
5724 }
5725 }
5726 else {
5727 /* New position is before the existing pair of pairs. */
5728 const float keep_earlier
5729 = THREEWAY_SQUARE(0, byte, cache[3], blen);
5730 const float keep_later
5731 = THREEWAY_SQUARE(0, byte, cache[1], blen);
5732
5733 if (keep_later < keep_earlier) {
5734 if (keep_later < existing) {
5735 cache[2] = utf8;
5736 cache[3] = byte;
5737 }
5738 }
5739 else {
5740 if (keep_earlier < existing) {
5741 cache[0] = cache[2];
5742 cache[1] = cache[3];
5743 cache[2] = utf8;
5744 cache[3] = byte;
5745 }
5746 }
5747 }
ab455f60 5748 }
0905937d 5749 ASSERT_UTF8_CACHE(cache);
ec07b5e0
NC
5750}
5751
5752/* If we don't know the character offset of the end of a region, our only
5753 option is to walk forwards to the target byte offset. */
5754static STRLEN
5755S_sv_pos_b2u_forwards(pTHX_ const U8 *s, const U8 *const target)
5756{
5757 STRLEN len = 0;
5758 while (s < target) {
5759 STRLEN n = 1;
5760
5761 /* Call utf8n_to_uvchr() to validate the sequence
5762 * (unless a simple non-UTF character) */
5763 if (!UTF8_IS_INVARIANT(*s))
5764 utf8n_to_uvchr(s, UTF8SKIP(s), &n, 0);
5765 if (n > 0) {
5766 s += n;
5767 len++;
5768 }
5769 else
5770 break;
5771 }
5772 return len;
5773}
5774
5775/* We already know all of the way, now we may be able to walk back. The same
25a8a4ef
NC
5776 assumption is made as in S_sv_pos_u2b_midway(), namely that walking
5777 backward is half the speed of walking forward. */
ec07b5e0
NC
5778static STRLEN
5779S_sv_pos_b2u_midway(pTHX_ const U8 *s, const U8 *const target, const U8 *end,
5780 STRLEN endu)
5781{
5782 const STRLEN forw = target - s;
5783 STRLEN backw = end - target;
5784
5785 if (forw < 2 * backw) {
5786 return S_sv_pos_b2u_forwards(aTHX_ s, target);
5787 }
5788
5789 while (end > target) {
5790 end--;
5791 while (UTF8_IS_CONTINUATION(*end)) {
5792 end--;
5793 }
5794 endu--;
5795 }
5796 return endu;
5797}
5798
9564a3bd
NC
5799/*
5800=for apidoc sv_pos_b2u
5801
5802Converts the value pointed to by offsetp from a count of bytes from the
5803start of the string, to a count of the equivalent number of UTF-8 chars.
5804Handles magic and type coercion.
5805
5806=cut
5807*/
5808
5809/*
5810 * sv_pos_b2u() uses, like sv_pos_u2b(), the mg_ptr of the potential
5811 * PERL_UTF8_magic of the sv to store the mapping between UTF-8 and
5812 * byte offsets.
5813 *
5814 */
a0ed51b3 5815void
7e8c5dac 5816Perl_sv_pos_b2u(pTHX_ register SV* sv, I32* offsetp)
a0ed51b3 5817{
83003860 5818 const U8* s;
ec07b5e0 5819 const STRLEN byte = *offsetp;
7087a21c 5820 STRLEN len = 0; /* Actually always set, but let's keep gcc happy. */
ab455f60 5821 STRLEN blen;
ec07b5e0
NC
5822 MAGIC* mg = NULL;
5823 const U8* send;
a922f900 5824 bool found = FALSE;
a0ed51b3
LW
5825
5826 if (!sv)
5827 return;
5828
ab455f60 5829 s = (const U8*)SvPV_const(sv, blen);
7e8c5dac 5830
ab455f60 5831 if (blen < byte)
ec07b5e0 5832 Perl_croak(aTHX_ "panic: sv_pos_b2u: bad byte offset");
7e8c5dac 5833
ec07b5e0 5834 send = s + byte;
a67d7df9 5835
ffca234a
NC
5836 if (SvMAGICAL(sv) && !SvREADONLY(sv) && PL_utf8cache
5837 && (mg = mg_find(sv, PERL_MAGIC_utf8))) {
5838 if (mg->mg_ptr) {
d4c19fe8 5839 STRLEN * const cache = (STRLEN *) mg->mg_ptr;
b9f984a5 5840 if (cache[1] == byte) {
ec07b5e0
NC
5841 /* An exact match. */
5842 *offsetp = cache[0];
ec07b5e0 5843 return;
7e8c5dac 5844 }
ab455f60
NC
5845 if (cache[3] == byte) {
5846 /* An exact match. */
5847 *offsetp = cache[2];
5848 return;
5849 }
668af93f
NC
5850
5851 if (cache[1] < byte) {
ec07b5e0 5852 /* We already know part of the way. */
b9f984a5
NC
5853 if (mg->mg_len != -1) {
5854 /* Actually, we know the end too. */
5855 len = cache[0]
5856 + S_sv_pos_b2u_midway(aTHX_ s + cache[1], send,
ab455f60 5857 s + blen, mg->mg_len - cache[0]);
b9f984a5
NC
5858 } else {
5859 len = cache[0]
5860 + S_sv_pos_b2u_forwards(aTHX_ s + cache[1], send);
5861 }
7e8c5dac 5862 }
9f985e4c
NC
5863 else if (cache[3] < byte) {
5864 /* We're between the two cached pairs, so we do the calculation
5865 offset by the byte/utf-8 positions for the earlier pair,
5866 then add the utf-8 characters from the string start to
5867 there. */
5868 len = S_sv_pos_b2u_midway(aTHX_ s + cache[3], send,
5869 s + cache[1], cache[0] - cache[2])
5870 + cache[2];
5871
5872 }
5873 else { /* cache[3] > byte */
5874 len = S_sv_pos_b2u_midway(aTHX_ s, send, s + cache[3],
5875 cache[2]);
7e8c5dac 5876
7e8c5dac 5877 }
ec07b5e0 5878 ASSERT_UTF8_CACHE(cache);
a922f900 5879 found = TRUE;
ffca234a 5880 } else if (mg->mg_len != -1) {
ab455f60 5881 len = S_sv_pos_b2u_midway(aTHX_ s, send, s + blen, mg->mg_len);
a922f900 5882 found = TRUE;
7e8c5dac 5883 }
a0ed51b3 5884 }
a922f900
NC
5885 if (!found || PL_utf8cache < 0) {
5886 const STRLEN real_len = S_sv_pos_b2u_forwards(aTHX_ s, send);
5887
5888 if (found && PL_utf8cache < 0) {
5889 if (len != real_len) {
5890 /* Need to turn the assertions off otherwise we may recurse
5891 infinitely while printing error messages. */
5892 SAVEI8(PL_utf8cache);
5893 PL_utf8cache = 0;
f5992bc4
RB
5894 Perl_croak(aTHX_ "panic: sv_pos_b2u cache %"UVuf
5895 " real %"UVuf" for %"SVf,
95b63a38 5896 (UV) len, (UV) real_len, (void*)sv);
a922f900
NC
5897 }
5898 }
5899 len = real_len;
ec07b5e0
NC
5900 }
5901 *offsetp = len;
5902
ab455f60 5903 S_utf8_mg_pos_cache_update(aTHX_ sv, &mg, byte, len, blen);
a0ed51b3
LW
5904}
5905
954c1994
GS
5906/*
5907=for apidoc sv_eq
5908
5909Returns a boolean indicating whether the strings in the two SVs are
645c22ef
DM
5910identical. Is UTF-8 and 'use bytes' aware, handles get magic, and will
5911coerce its args to strings if necessary.
954c1994
GS
5912
5913=cut
5914*/
5915
79072805 5916I32
e01b9e88 5917Perl_sv_eq(pTHX_ register SV *sv1, register SV *sv2)
79072805 5918{
97aff369 5919 dVAR;
e1ec3a88 5920 const char *pv1;
463ee0b2 5921 STRLEN cur1;
e1ec3a88 5922 const char *pv2;
463ee0b2 5923 STRLEN cur2;
e01b9e88 5924 I32 eq = 0;
bd61b366 5925 char *tpv = NULL;
a0714e2c 5926 SV* svrecode = NULL;
79072805 5927
e01b9e88 5928 if (!sv1) {
79072805
LW
5929 pv1 = "";
5930 cur1 = 0;
5931 }
ced497e2
YST
5932 else {
5933 /* if pv1 and pv2 are the same, second SvPV_const call may
5934 * invalidate pv1, so we may need to make a copy */
5935 if (sv1 == sv2 && (SvTHINKFIRST(sv1) || SvGMAGICAL(sv1))) {
5936 pv1 = SvPV_const(sv1, cur1);
5937 sv1 = sv_2mortal(newSVpvn(pv1, cur1));
5938 if (SvUTF8(sv2)) SvUTF8_on(sv1);
5939 }
4d84ee25 5940 pv1 = SvPV_const(sv1, cur1);
ced497e2 5941 }
79072805 5942
e01b9e88
SC
5943 if (!sv2){
5944 pv2 = "";
5945 cur2 = 0;
92d29cee 5946 }
e01b9e88 5947 else
4d84ee25 5948 pv2 = SvPV_const(sv2, cur2);
79072805 5949
cf48d248 5950 if (cur1 && cur2 && SvUTF8(sv1) != SvUTF8(sv2) && !IN_BYTES) {
799ef3cb
JH
5951 /* Differing utf8ness.
5952 * Do not UTF8size the comparands as a side-effect. */
5953 if (PL_encoding) {
5954 if (SvUTF8(sv1)) {
553e1bcc
AT
5955 svrecode = newSVpvn(pv2, cur2);
5956 sv_recode_to_utf8(svrecode, PL_encoding);
93524f2b 5957 pv2 = SvPV_const(svrecode, cur2);
799ef3cb
JH
5958 }
5959 else {
553e1bcc
AT
5960 svrecode = newSVpvn(pv1, cur1);
5961 sv_recode_to_utf8(svrecode, PL_encoding);
93524f2b 5962 pv1 = SvPV_const(svrecode, cur1);
799ef3cb
JH
5963 }
5964 /* Now both are in UTF-8. */
0a1bd7ac
DM
5965 if (cur1 != cur2) {
5966 SvREFCNT_dec(svrecode);
799ef3cb 5967 return FALSE;
0a1bd7ac 5968 }
799ef3cb
JH
5969 }
5970 else {
5971 bool is_utf8 = TRUE;
5972
5973 if (SvUTF8(sv1)) {
5974 /* sv1 is the UTF-8 one,
5975 * if is equal it must be downgrade-able */
9d4ba2ae 5976 char * const pv = (char*)bytes_from_utf8((const U8*)pv1,
799ef3cb
JH
5977 &cur1, &is_utf8);
5978 if (pv != pv1)
553e1bcc 5979 pv1 = tpv = pv;
799ef3cb
JH
5980 }
5981 else {
5982 /* sv2 is the UTF-8 one,
5983 * if is equal it must be downgrade-able */
9d4ba2ae 5984 char * const pv = (char *)bytes_from_utf8((const U8*)pv2,
799ef3cb
JH
5985 &cur2, &is_utf8);
5986 if (pv != pv2)
553e1bcc 5987 pv2 = tpv = pv;
799ef3cb
JH
5988 }
5989 if (is_utf8) {
5990 /* Downgrade not possible - cannot be eq */
bf694877 5991 assert (tpv == 0);
799ef3cb
JH
5992 return FALSE;
5993 }
5994 }
cf48d248
JH
5995 }
5996
5997 if (cur1 == cur2)
765f542d 5998 eq = (pv1 == pv2) || memEQ(pv1, pv2, cur1);
e01b9e88 5999
b37c2d43 6000 SvREFCNT_dec(svrecode);
553e1bcc
AT
6001 if (tpv)
6002 Safefree(tpv);
cf48d248 6003
e01b9e88 6004 return eq;
79072805
LW
6005}
6006
954c1994
GS
6007/*
6008=for apidoc sv_cmp
6009
6010Compares the strings in two SVs. Returns -1, 0, or 1 indicating whether the
6011string in C<sv1> is less than, equal to, or greater than the string in
645c22ef
DM
6012C<sv2>. Is UTF-8 and 'use bytes' aware, handles get magic, and will
6013coerce its args to strings if necessary. See also C<sv_cmp_locale>.
954c1994
GS
6014
6015=cut
6016*/
6017
79072805 6018I32
e01b9e88 6019Perl_sv_cmp(pTHX_ register SV *sv1, register SV *sv2)
79072805 6020{
97aff369 6021 dVAR;
560a288e 6022 STRLEN cur1, cur2;
e1ec3a88 6023 const char *pv1, *pv2;
bd61b366 6024 char *tpv = NULL;
cf48d248 6025 I32 cmp;
a0714e2c 6026 SV *svrecode = NULL;
560a288e 6027
e01b9e88
SC
6028 if (!sv1) {
6029 pv1 = "";
560a288e
GS
6030 cur1 = 0;
6031 }
e01b9e88 6032 else
4d84ee25 6033 pv1 = SvPV_const(sv1, cur1);
560a288e 6034
553e1bcc 6035 if (!sv2) {
e01b9e88 6036 pv2 = "";
560a288e
GS
6037 cur2 = 0;
6038 }
e01b9e88 6039 else
4d84ee25 6040 pv2 = SvPV_const(sv2, cur2);
79072805 6041
cf48d248 6042 if (cur1 && cur2 && SvUTF8(sv1) != SvUTF8(sv2) && !IN_BYTES) {
799ef3cb
JH
6043 /* Differing utf8ness.
6044 * Do not UTF8size the comparands as a side-effect. */
cf48d248 6045 if (SvUTF8(sv1)) {
799ef3cb 6046 if (PL_encoding) {
553e1bcc
AT
6047 svrecode = newSVpvn(pv2, cur2);
6048 sv_recode_to_utf8(svrecode, PL_encoding);
93524f2b 6049 pv2 = SvPV_const(svrecode, cur2);
799ef3cb
JH
6050 }
6051 else {
e1ec3a88 6052 pv2 = tpv = (char*)bytes_to_utf8((const U8*)pv2, &cur2);
799ef3cb 6053 }
cf48d248
JH
6054 }
6055 else {
799ef3cb 6056 if (PL_encoding) {
553e1bcc
AT
6057 svrecode = newSVpvn(pv1, cur1);
6058 sv_recode_to_utf8(svrecode, PL_encoding);
93524f2b 6059 pv1 = SvPV_const(svrecode, cur1);
799ef3cb
JH
6060 }
6061 else {
e1ec3a88 6062 pv1 = tpv = (char*)bytes_to_utf8((const U8*)pv1, &cur1);
799ef3cb 6063 }
cf48d248
JH
6064 }
6065 }
6066
e01b9e88 6067 if (!cur1) {
cf48d248 6068 cmp = cur2 ? -1 : 0;
e01b9e88 6069 } else if (!cur2) {
cf48d248
JH
6070 cmp = 1;
6071 } else {
e1ec3a88 6072 const I32 retval = memcmp((const void*)pv1, (const void*)pv2, cur1 < cur2 ? cur1 : cur2);
e01b9e88
SC
6073
6074 if (retval) {
cf48d248 6075 cmp = retval < 0 ? -1 : 1;
e01b9e88 6076 } else if (cur1 == cur2) {
cf48d248
JH
6077 cmp = 0;
6078 } else {
6079 cmp = cur1 < cur2 ? -1 : 1;
e01b9e88 6080 }
cf48d248 6081 }
16660edb 6082
b37c2d43 6083 SvREFCNT_dec(svrecode);
553e1bcc
AT
6084 if (tpv)
6085 Safefree(tpv);
cf48d248
JH
6086
6087 return cmp;
bbce6d69 6088}
16660edb 6089
c461cf8f
JH
6090/*
6091=for apidoc sv_cmp_locale
6092
645c22ef
DM
6093Compares the strings in two SVs in a locale-aware manner. Is UTF-8 and
6094'use bytes' aware, handles get magic, and will coerce its args to strings
6095if necessary. See also C<sv_cmp_locale>. See also C<sv_cmp>.
c461cf8f
JH
6096
6097=cut
6098*/
6099
bbce6d69 6100I32
864dbfa3 6101Perl_sv_cmp_locale(pTHX_ register SV *sv1, register SV *sv2)
bbce6d69 6102{
97aff369 6103 dVAR;
36477c24 6104#ifdef USE_LOCALE_COLLATE
16660edb 6105
bbce6d69 6106 char *pv1, *pv2;
6107 STRLEN len1, len2;
6108 I32 retval;
16660edb 6109
3280af22 6110 if (PL_collation_standard)
bbce6d69 6111 goto raw_compare;
16660edb 6112
bbce6d69 6113 len1 = 0;
8ac85365 6114 pv1 = sv1 ? sv_collxfrm(sv1, &len1) : (char *) NULL;
bbce6d69 6115 len2 = 0;
8ac85365 6116 pv2 = sv2 ? sv_collxfrm(sv2, &len2) : (char *) NULL;
16660edb 6117
bbce6d69 6118 if (!pv1 || !len1) {
6119 if (pv2 && len2)
6120 return -1;
6121 else
6122 goto raw_compare;
6123 }
6124 else {
6125 if (!pv2 || !len2)
6126 return 1;
6127 }
16660edb 6128
bbce6d69 6129 retval = memcmp((void*)pv1, (void*)pv2, len1 < len2 ? len1 : len2);
16660edb 6130
bbce6d69 6131 if (retval)
16660edb 6132 return retval < 0 ? -1 : 1;
6133
bbce6d69 6134 /*
6135 * When the result of collation is equality, that doesn't mean
6136 * that there are no differences -- some locales exclude some
6137 * characters from consideration. So to avoid false equalities,
6138 * we use the raw string as a tiebreaker.
6139 */
16660edb 6140
bbce6d69 6141 raw_compare:
5f66b61c 6142 /*FALLTHROUGH*/
16660edb 6143
36477c24 6144#endif /* USE_LOCALE_COLLATE */
16660edb 6145
bbce6d69 6146 return sv_cmp(sv1, sv2);
6147}
79072805 6148
645c22ef 6149
36477c24 6150#ifdef USE_LOCALE_COLLATE
645c22ef 6151
7a4c00b4 6152/*
645c22ef
DM
6153=for apidoc sv_collxfrm
6154
6155Add Collate Transform magic to an SV if it doesn't already have it.
6156
6157Any scalar variable may carry PERL_MAGIC_collxfrm magic that contains the
6158scalar data of the variable, but transformed to such a format that a normal
6159memory comparison can be used to compare the data according to the locale
6160settings.
6161
6162=cut
6163*/
6164
bbce6d69 6165char *
864dbfa3 6166Perl_sv_collxfrm(pTHX_ SV *sv, STRLEN *nxp)
bbce6d69 6167{
97aff369 6168 dVAR;
7a4c00b4 6169 MAGIC *mg;
16660edb 6170
14befaf4 6171 mg = SvMAGICAL(sv) ? mg_find(sv, PERL_MAGIC_collxfrm) : (MAGIC *) NULL;
3280af22 6172 if (!mg || !mg->mg_ptr || *(U32*)mg->mg_ptr != PL_collation_ix) {
93524f2b
NC
6173 const char *s;
6174 char *xf;
bbce6d69 6175 STRLEN len, xlen;
6176
7a4c00b4 6177 if (mg)
6178 Safefree(mg->mg_ptr);
93524f2b 6179 s = SvPV_const(sv, len);
bbce6d69 6180 if ((xf = mem_collxfrm(s, len, &xlen))) {
ff0cee69 6181 if (SvREADONLY(sv)) {
6182 SAVEFREEPV(xf);
6183 *nxp = xlen;
3280af22 6184 return xf + sizeof(PL_collation_ix);
ff0cee69 6185 }
7a4c00b4 6186 if (! mg) {
d83f0a82
NC
6187#ifdef PERL_OLD_COPY_ON_WRITE
6188 if (SvIsCOW(sv))
6189 sv_force_normal_flags(sv, 0);
6190#endif
6191 mg = sv_magicext(sv, 0, PERL_MAGIC_collxfrm, &PL_vtbl_collxfrm,
6192 0, 0);
7a4c00b4 6193 assert(mg);
bbce6d69 6194 }
7a4c00b4 6195 mg->mg_ptr = xf;
565764a8 6196 mg->mg_len = xlen;
7a4c00b4 6197 }
6198 else {
ff0cee69 6199 if (mg) {
6200 mg->mg_ptr = NULL;
565764a8 6201 mg->mg_len = -1;
ff0cee69 6202 }
bbce6d69 6203 }
6204 }
7a4c00b4 6205 if (mg && mg->mg_ptr) {
565764a8 6206 *nxp = mg->mg_len;
3280af22 6207 return mg->mg_ptr + sizeof(PL_collation_ix);
bbce6d69 6208 }
6209 else {
6210 *nxp = 0;
6211 return NULL;
16660edb 6212 }
79072805
LW
6213}
6214
36477c24 6215#endif /* USE_LOCALE_COLLATE */
bbce6d69 6216
c461cf8f
JH
6217/*
6218=for apidoc sv_gets
6219
6220Get a line from the filehandle and store it into the SV, optionally
6221appending to the currently-stored string.
6222
6223=cut
6224*/
6225
79072805 6226char *
864dbfa3 6227Perl_sv_gets(pTHX_ register SV *sv, register PerlIO *fp, I32 append)
79072805 6228{
97aff369 6229 dVAR;
e1ec3a88 6230 const char *rsptr;
c07a80fd 6231 STRLEN rslen;
6232 register STDCHAR rslast;
6233 register STDCHAR *bp;
6234 register I32 cnt;
9c5ffd7c 6235 I32 i = 0;
8bfdd7d9 6236 I32 rspara = 0;
c07a80fd 6237
bc44a8a2
NC
6238 if (SvTHINKFIRST(sv))
6239 sv_force_normal_flags(sv, append ? 0 : SV_COW_DROP_PV);
765f542d
NC
6240 /* XXX. If you make this PVIV, then copy on write can copy scalars read
6241 from <>.
6242 However, perlbench says it's slower, because the existing swipe code
6243 is faster than copy on write.
6244 Swings and roundabouts. */
862a34c6 6245 SvUPGRADE(sv, SVt_PV);
99491443 6246
ff68c719 6247 SvSCREAM_off(sv);
efd8b2ba
AE
6248
6249 if (append) {
6250 if (PerlIO_isutf8(fp)) {
6251 if (!SvUTF8(sv)) {
6252 sv_utf8_upgrade_nomg(sv);
6253 sv_pos_u2b(sv,&append,0);
6254 }
6255 } else if (SvUTF8(sv)) {
561b68a9 6256 SV * const tsv = newSV(0);
efd8b2ba
AE
6257 sv_gets(tsv, fp, 0);
6258 sv_utf8_upgrade_nomg(tsv);
6259 SvCUR_set(sv,append);
6260 sv_catsv(sv,tsv);
6261 sv_free(tsv);
6262 goto return_string_or_null;
6263 }
6264 }
6265
6266 SvPOK_only(sv);
6267 if (PerlIO_isutf8(fp))
6268 SvUTF8_on(sv);
c07a80fd 6269
923e4eb5 6270 if (IN_PERL_COMPILETIME) {
8bfdd7d9
HS
6271 /* we always read code in line mode */
6272 rsptr = "\n";
6273 rslen = 1;
6274 }
6275 else if (RsSNARF(PL_rs)) {
7a5fa8a2 6276 /* If it is a regular disk file use size from stat() as estimate
acbd132f
JH
6277 of amount we are going to read -- may result in mallocing
6278 more memory than we really need if the layers below reduce
6279 the size we read (e.g. CRLF or a gzip layer).
e468d35b 6280 */
e311fd51 6281 Stat_t st;
e468d35b 6282 if (!PerlLIO_fstat(PerlIO_fileno(fp), &st) && S_ISREG(st.st_mode)) {
f54cb97a 6283 const Off_t offset = PerlIO_tell(fp);
58f1856e 6284 if (offset != (Off_t) -1 && st.st_size + append > offset) {
e468d35b
NIS
6285 (void) SvGROW(sv, (STRLEN)((st.st_size - offset) + append + 1));
6286 }
6287 }
c07a80fd 6288 rsptr = NULL;
6289 rslen = 0;
6290 }
3280af22 6291 else if (RsRECORD(PL_rs)) {
e311fd51 6292 I32 bytesread;
5b2b9c68 6293 char *buffer;
acbd132f 6294 U32 recsize;
5b2b9c68
HM
6295
6296 /* Grab the size of the record we're getting */
acbd132f 6297 recsize = SvUV(SvRV(PL_rs)); /* RsRECORD() guarantees > 0. */
e311fd51 6298 buffer = SvGROW(sv, (STRLEN)(recsize + append + 1)) + append;
5b2b9c68
HM
6299 /* Go yank in */
6300#ifdef VMS
6301 /* VMS wants read instead of fread, because fread doesn't respect */
6302 /* RMS record boundaries. This is not necessarily a good thing to be */
e468d35b
NIS
6303 /* doing, but we've got no other real choice - except avoid stdio
6304 as implementation - perhaps write a :vms layer ?
6305 */
5b2b9c68
HM
6306 bytesread = PerlLIO_read(PerlIO_fileno(fp), buffer, recsize);
6307#else
6308 bytesread = PerlIO_read(fp, buffer, recsize);
6309#endif
27e6ca2d
AE
6310 if (bytesread < 0)
6311 bytesread = 0;
e311fd51 6312 SvCUR_set(sv, bytesread += append);
e670df4e 6313 buffer[bytesread] = '\0';
efd8b2ba 6314 goto return_string_or_null;
5b2b9c68 6315 }
3280af22 6316 else if (RsPARA(PL_rs)) {
c07a80fd 6317 rsptr = "\n\n";
6318 rslen = 2;
8bfdd7d9 6319 rspara = 1;
c07a80fd 6320 }
7d59b7e4
NIS
6321 else {
6322 /* Get $/ i.e. PL_rs into same encoding as stream wants */
6323 if (PerlIO_isutf8(fp)) {
6324 rsptr = SvPVutf8(PL_rs, rslen);
6325 }
6326 else {
6327 if (SvUTF8(PL_rs)) {
6328 if (!sv_utf8_downgrade(PL_rs, TRUE)) {
6329 Perl_croak(aTHX_ "Wide character in $/");
6330 }
6331 }
93524f2b 6332 rsptr = SvPV_const(PL_rs, rslen);
7d59b7e4
NIS
6333 }
6334 }
6335
c07a80fd 6336 rslast = rslen ? rsptr[rslen - 1] : '\0';
6337
8bfdd7d9 6338 if (rspara) { /* have to do this both before and after */
79072805 6339 do { /* to make sure file boundaries work right */
760ac839 6340 if (PerlIO_eof(fp))
a0d0e21e 6341 return 0;
760ac839 6342 i = PerlIO_getc(fp);
79072805 6343 if (i != '\n') {
a0d0e21e
LW
6344 if (i == -1)
6345 return 0;
760ac839 6346 PerlIO_ungetc(fp,i);
79072805
LW
6347 break;
6348 }
6349 } while (i != EOF);
6350 }
c07a80fd 6351
760ac839
LW
6352 /* See if we know enough about I/O mechanism to cheat it ! */
6353
6354 /* This used to be #ifdef test - it is made run-time test for ease
1c846c1f 6355 of abstracting out stdio interface. One call should be cheap
760ac839
LW
6356 enough here - and may even be a macro allowing compile
6357 time optimization.
6358 */
6359
6360 if (PerlIO_fast_gets(fp)) {
6361
6362 /*
6363 * We're going to steal some values from the stdio struct
6364 * and put EVERYTHING in the innermost loop into registers.
6365 */
6366 register STDCHAR *ptr;
6367 STRLEN bpx;
6368 I32 shortbuffered;
6369
16660edb 6370#if defined(VMS) && defined(PERLIO_IS_STDIO)
6371 /* An ungetc()d char is handled separately from the regular
6372 * buffer, so we getc() it back out and stuff it in the buffer.
6373 */
6374 i = PerlIO_getc(fp);
6375 if (i == EOF) return 0;
6376 *(--((*fp)->_ptr)) = (unsigned char) i;
6377 (*fp)->_cnt++;
6378#endif
c07a80fd 6379
c2960299 6380 /* Here is some breathtakingly efficient cheating */
c07a80fd 6381
a20bf0c3 6382 cnt = PerlIO_get_cnt(fp); /* get count into register */
e468d35b 6383 /* make sure we have the room */
7a5fa8a2 6384 if ((I32)(SvLEN(sv) - append) <= cnt + 1) {
e468d35b 6385 /* Not room for all of it
7a5fa8a2 6386 if we are looking for a separator and room for some
e468d35b
NIS
6387 */
6388 if (rslen && cnt > 80 && (I32)SvLEN(sv) > append) {
7a5fa8a2 6389 /* just process what we have room for */
79072805
LW
6390 shortbuffered = cnt - SvLEN(sv) + append + 1;
6391 cnt -= shortbuffered;
6392 }
6393 else {
6394 shortbuffered = 0;
bbce6d69 6395 /* remember that cnt can be negative */
eb160463 6396 SvGROW(sv, (STRLEN)(append + (cnt <= 0 ? 2 : (cnt + 1))));
79072805
LW
6397 }
6398 }
7a5fa8a2 6399 else
79072805 6400 shortbuffered = 0;
3f7c398e 6401 bp = (STDCHAR*)SvPVX_const(sv) + append; /* move these two too to registers */
a20bf0c3 6402 ptr = (STDCHAR*)PerlIO_get_ptr(fp);
16660edb 6403 DEBUG_P(PerlIO_printf(Perl_debug_log,
1d7c1841 6404 "Screamer: entering, ptr=%"UVuf", cnt=%ld\n",PTR2UV(ptr),(long)cnt));
16660edb 6405 DEBUG_P(PerlIO_printf(Perl_debug_log,
ba7abf9d 6406 "Screamer: entering: PerlIO * thinks ptr=%"UVuf", cnt=%ld, base=%"UVuf"\n",
1c846c1f 6407 PTR2UV(PerlIO_get_ptr(fp)), (long)PerlIO_get_cnt(fp),
1d7c1841 6408 PTR2UV(PerlIO_has_base(fp) ? PerlIO_get_base(fp) : 0)));
79072805
LW
6409 for (;;) {
6410 screamer:
93a17b20 6411 if (cnt > 0) {
c07a80fd 6412 if (rslen) {
760ac839
LW
6413 while (cnt > 0) { /* this | eat */
6414 cnt--;
c07a80fd 6415 if ((*bp++ = *ptr++) == rslast) /* really | dust */
6416 goto thats_all_folks; /* screams | sed :-) */
6417 }
6418 }
6419 else {
1c846c1f
NIS
6420 Copy(ptr, bp, cnt, char); /* this | eat */
6421 bp += cnt; /* screams | dust */
c07a80fd 6422 ptr += cnt; /* louder | sed :-) */
a5f75d66 6423 cnt = 0;
93a17b20 6424 }
79072805
LW
6425 }
6426
748a9306 6427 if (shortbuffered) { /* oh well, must extend */
79072805
LW
6428 cnt = shortbuffered;
6429 shortbuffered = 0;
3f7c398e 6430 bpx = bp - (STDCHAR*)SvPVX_const(sv); /* box up before relocation */
79072805
LW
6431 SvCUR_set(sv, bpx);
6432 SvGROW(sv, SvLEN(sv) + append + cnt + 2);
3f7c398e 6433 bp = (STDCHAR*)SvPVX_const(sv) + bpx; /* unbox after relocation */
79072805
LW
6434 continue;
6435 }
6436
16660edb 6437 DEBUG_P(PerlIO_printf(Perl_debug_log,
1d7c1841
GS
6438 "Screamer: going to getc, ptr=%"UVuf", cnt=%ld\n",
6439 PTR2UV(ptr),(long)cnt));
cc00df79 6440 PerlIO_set_ptrcnt(fp, (STDCHAR*)ptr, cnt); /* deregisterize cnt and ptr */
ba7abf9d 6441#if 0
16660edb 6442 DEBUG_P(PerlIO_printf(Perl_debug_log,
1d7c1841 6443 "Screamer: pre: FILE * thinks ptr=%"UVuf", cnt=%ld, base=%"UVuf"\n",
1c846c1f 6444 PTR2UV(PerlIO_get_ptr(fp)), (long)PerlIO_get_cnt(fp),
1d7c1841 6445 PTR2UV(PerlIO_has_base (fp) ? PerlIO_get_base(fp) : 0)));
ba7abf9d 6446#endif
1c846c1f 6447 /* This used to call 'filbuf' in stdio form, but as that behaves like
774d564b 6448 getc when cnt <= 0 we use PerlIO_getc here to avoid introducing
6449 another abstraction. */
760ac839 6450 i = PerlIO_getc(fp); /* get more characters */
ba7abf9d 6451#if 0
16660edb 6452 DEBUG_P(PerlIO_printf(Perl_debug_log,
1d7c1841 6453 "Screamer: post: FILE * thinks ptr=%"UVuf", cnt=%ld, base=%"UVuf"\n",
1c846c1f 6454 PTR2UV(PerlIO_get_ptr(fp)), (long)PerlIO_get_cnt(fp),
1d7c1841 6455 PTR2UV(PerlIO_has_base (fp) ? PerlIO_get_base(fp) : 0)));
ba7abf9d 6456#endif
a20bf0c3
JH
6457 cnt = PerlIO_get_cnt(fp);
6458 ptr = (STDCHAR*)PerlIO_get_ptr(fp); /* reregisterize cnt and ptr */
16660edb 6459 DEBUG_P(PerlIO_printf(Perl_debug_log,
1d7c1841 6460 "Screamer: after getc, ptr=%"UVuf", cnt=%ld\n",PTR2UV(ptr),(long)cnt));
79072805 6461
748a9306
LW
6462 if (i == EOF) /* all done for ever? */
6463 goto thats_really_all_folks;
6464
3f7c398e 6465 bpx = bp - (STDCHAR*)SvPVX_const(sv); /* box up before relocation */
79072805
LW
6466 SvCUR_set(sv, bpx);
6467 SvGROW(sv, bpx + cnt + 2);
3f7c398e 6468 bp = (STDCHAR*)SvPVX_const(sv) + bpx; /* unbox after relocation */
c07a80fd 6469
eb160463 6470 *bp++ = (STDCHAR)i; /* store character from PerlIO_getc */
79072805 6471
c07a80fd 6472 if (rslen && (STDCHAR)i == rslast) /* all done for now? */
79072805 6473 goto thats_all_folks;
79072805
LW
6474 }
6475
6476thats_all_folks:
3f7c398e 6477 if ((rslen > 1 && (STRLEN)(bp - (STDCHAR*)SvPVX_const(sv)) < rslen) ||
36477c24 6478 memNE((char*)bp - rslen, rsptr, rslen))
760ac839 6479 goto screamer; /* go back to the fray */
79072805
LW
6480thats_really_all_folks:
6481 if (shortbuffered)
6482 cnt += shortbuffered;
16660edb 6483 DEBUG_P(PerlIO_printf(Perl_debug_log,
1d7c1841 6484 "Screamer: quitting, ptr=%"UVuf", cnt=%ld\n",PTR2UV(ptr),(long)cnt));
cc00df79 6485 PerlIO_set_ptrcnt(fp, (STDCHAR*)ptr, cnt); /* put these back or we're in trouble */
16660edb 6486 DEBUG_P(PerlIO_printf(Perl_debug_log,
1d7c1841 6487 "Screamer: end: FILE * thinks ptr=%"UVuf", cnt=%ld, base=%"UVuf"\n",
1c846c1f 6488 PTR2UV(PerlIO_get_ptr(fp)), (long)PerlIO_get_cnt(fp),
1d7c1841 6489 PTR2UV(PerlIO_has_base (fp) ? PerlIO_get_base(fp) : 0)));
79072805 6490 *bp = '\0';
3f7c398e 6491 SvCUR_set(sv, bp - (STDCHAR*)SvPVX_const(sv)); /* set length */
16660edb 6492 DEBUG_P(PerlIO_printf(Perl_debug_log,
fb73857a 6493 "Screamer: done, len=%ld, string=|%.*s|\n",
3f7c398e 6494 (long)SvCUR(sv),(int)SvCUR(sv),SvPVX_const(sv)));
760ac839
LW
6495 }
6496 else
79072805 6497 {
6edd2cd5 6498 /*The big, slow, and stupid way. */
27da23d5 6499#ifdef USE_HEAP_INSTEAD_OF_STACK /* Even slower way. */
cbbf8932 6500 STDCHAR *buf = NULL;
a02a5408 6501 Newx(buf, 8192, STDCHAR);
6edd2cd5 6502 assert(buf);
4d2c4e07 6503#else
6edd2cd5 6504 STDCHAR buf[8192];
4d2c4e07 6505#endif
79072805 6506
760ac839 6507screamer2:
c07a80fd 6508 if (rslen) {
00b6aa41 6509 register const STDCHAR * const bpe = buf + sizeof(buf);
760ac839 6510 bp = buf;
eb160463 6511 while ((i = PerlIO_getc(fp)) != EOF && (*bp++ = (STDCHAR)i) != rslast && bp < bpe)
760ac839
LW
6512 ; /* keep reading */
6513 cnt = bp - buf;
c07a80fd 6514 }
6515 else {
760ac839 6516 cnt = PerlIO_read(fp,(char*)buf, sizeof(buf));
16660edb 6517 /* Accomodate broken VAXC compiler, which applies U8 cast to
6518 * both args of ?: operator, causing EOF to change into 255
6519 */
37be0adf 6520 if (cnt > 0)
cbe9e203
JH
6521 i = (U8)buf[cnt - 1];
6522 else
37be0adf 6523 i = EOF;
c07a80fd 6524 }
79072805 6525
cbe9e203
JH
6526 if (cnt < 0)
6527 cnt = 0; /* we do need to re-set the sv even when cnt <= 0 */
6528 if (append)
6529 sv_catpvn(sv, (char *) buf, cnt);
6530 else
6531 sv_setpvn(sv, (char *) buf, cnt);
c07a80fd 6532
6533 if (i != EOF && /* joy */
6534 (!rslen ||
6535 SvCUR(sv) < rslen ||
3f7c398e 6536 memNE(SvPVX_const(sv) + SvCUR(sv) - rslen, rsptr, rslen)))
79072805
LW
6537 {
6538 append = -1;
63e4d877
CS
6539 /*
6540 * If we're reading from a TTY and we get a short read,
6541 * indicating that the user hit his EOF character, we need
6542 * to notice it now, because if we try to read from the TTY
6543 * again, the EOF condition will disappear.
6544 *
6545 * The comparison of cnt to sizeof(buf) is an optimization
6546 * that prevents unnecessary calls to feof().
6547 *
6548 * - jik 9/25/96
6549 */
bb7a0f54 6550 if (!(cnt < (I32)sizeof(buf) && PerlIO_eof(fp)))
63e4d877 6551 goto screamer2;
79072805 6552 }
6edd2cd5 6553
27da23d5 6554#ifdef USE_HEAP_INSTEAD_OF_STACK
6edd2cd5
JH
6555 Safefree(buf);
6556#endif
79072805
LW
6557 }
6558
8bfdd7d9 6559 if (rspara) { /* have to do this both before and after */
c07a80fd 6560 while (i != EOF) { /* to make sure file boundaries work right */
760ac839 6561 i = PerlIO_getc(fp);
79072805 6562 if (i != '\n') {
760ac839 6563 PerlIO_ungetc(fp,i);
79072805
LW
6564 break;
6565 }
6566 }
6567 }
c07a80fd 6568
efd8b2ba 6569return_string_or_null:
bd61b366 6570 return (SvCUR(sv) - append) ? SvPVX(sv) : NULL;
79072805
LW
6571}
6572
954c1994
GS
6573/*
6574=for apidoc sv_inc
6575
645c22ef
DM
6576Auto-increment of the value in the SV, doing string to numeric conversion
6577if necessary. Handles 'get' magic.
954c1994
GS
6578
6579=cut
6580*/
6581
79072805 6582void
864dbfa3 6583Perl_sv_inc(pTHX_ register SV *sv)
79072805 6584{
97aff369 6585 dVAR;
79072805 6586 register char *d;
463ee0b2 6587 int flags;
79072805
LW
6588
6589 if (!sv)
6590 return;
5b295bef 6591 SvGETMAGIC(sv);
ed6116ce 6592 if (SvTHINKFIRST(sv)) {
765f542d
NC
6593 if (SvIsCOW(sv))
6594 sv_force_normal_flags(sv, 0);
0f15f207 6595 if (SvREADONLY(sv)) {
923e4eb5 6596 if (IN_PERL_RUNTIME)
cea2e8a9 6597 Perl_croak(aTHX_ PL_no_modify);
0f15f207 6598 }
a0d0e21e 6599 if (SvROK(sv)) {
b5be31e9 6600 IV i;
9e7bc3e8
JD
6601 if (SvAMAGIC(sv) && AMG_CALLun(sv,inc))
6602 return;
56431972 6603 i = PTR2IV(SvRV(sv));
b5be31e9
SM
6604 sv_unref(sv);
6605 sv_setiv(sv, i);
a0d0e21e 6606 }
ed6116ce 6607 }
8990e307 6608 flags = SvFLAGS(sv);
28e5dec8
JH
6609 if ((flags & (SVp_NOK|SVp_IOK)) == SVp_NOK) {
6610 /* It's (privately or publicly) a float, but not tested as an
6611 integer, so test it to see. */
d460ef45 6612 (void) SvIV(sv);
28e5dec8
JH
6613 flags = SvFLAGS(sv);
6614 }
6615 if ((flags & SVf_IOK) || ((flags & (SVp_IOK | SVp_NOK)) == SVp_IOK)) {
6616 /* It's publicly an integer, or privately an integer-not-float */
59d8ce62 6617#ifdef PERL_PRESERVE_IVUV
28e5dec8 6618 oops_its_int:
59d8ce62 6619#endif
25da4f38
IZ
6620 if (SvIsUV(sv)) {
6621 if (SvUVX(sv) == UV_MAX)
a1e868e7 6622 sv_setnv(sv, UV_MAX_P1);
25da4f38
IZ
6623 else
6624 (void)SvIOK_only_UV(sv);
607fa7f2 6625 SvUV_set(sv, SvUVX(sv) + 1);
25da4f38
IZ
6626 } else {
6627 if (SvIVX(sv) == IV_MAX)
28e5dec8 6628 sv_setuv(sv, (UV)IV_MAX + 1);
25da4f38
IZ
6629 else {
6630 (void)SvIOK_only(sv);
45977657 6631 SvIV_set(sv, SvIVX(sv) + 1);
1c846c1f 6632 }
55497cff 6633 }
79072805
LW
6634 return;
6635 }
28e5dec8
JH
6636 if (flags & SVp_NOK) {
6637 (void)SvNOK_only(sv);
9d6ce603 6638 SvNV_set(sv, SvNVX(sv) + 1.0);
28e5dec8
JH
6639 return;
6640 }
6641
3f7c398e 6642 if (!(flags & SVp_POK) || !*SvPVX_const(sv)) {
28e5dec8 6643 if ((flags & SVTYPEMASK) < SVt_PVIV)
f5282e15 6644 sv_upgrade(sv, ((flags & SVTYPEMASK) > SVt_IV ? SVt_PVIV : SVt_IV));
28e5dec8 6645 (void)SvIOK_only(sv);
45977657 6646 SvIV_set(sv, 1);
79072805
LW
6647 return;
6648 }
463ee0b2 6649 d = SvPVX(sv);
79072805
LW
6650 while (isALPHA(*d)) d++;
6651 while (isDIGIT(*d)) d++;
6652 if (*d) {
28e5dec8 6653#ifdef PERL_PRESERVE_IVUV
d1be9408 6654 /* Got to punt this as an integer if needs be, but we don't issue
28e5dec8
JH
6655 warnings. Probably ought to make the sv_iv_please() that does
6656 the conversion if possible, and silently. */
504618e9 6657 const int numtype = grok_number(SvPVX_const(sv), SvCUR(sv), NULL);
28e5dec8
JH
6658 if (numtype && !(numtype & IS_NUMBER_INFINITY)) {
6659 /* Need to try really hard to see if it's an integer.
6660 9.22337203685478e+18 is an integer.
6661 but "9.22337203685478e+18" + 0 is UV=9223372036854779904
6662 so $a="9.22337203685478e+18"; $a+0; $a++
6663 needs to be the same as $a="9.22337203685478e+18"; $a++
6664 or we go insane. */
d460ef45 6665
28e5dec8
JH
6666 (void) sv_2iv(sv);
6667 if (SvIOK(sv))
6668 goto oops_its_int;
6669
6670 /* sv_2iv *should* have made this an NV */
6671 if (flags & SVp_NOK) {
6672 (void)SvNOK_only(sv);
9d6ce603 6673 SvNV_set(sv, SvNVX(sv) + 1.0);
28e5dec8
JH
6674 return;
6675 }
6676 /* I don't think we can get here. Maybe I should assert this
6677 And if we do get here I suspect that sv_setnv will croak. NWC
6678 Fall through. */
6679#if defined(USE_LONG_DOUBLE)
6680 DEBUG_c(PerlIO_printf(Perl_debug_log,"sv_inc punt failed to convert '%s' to IOK or NOKp, UV=0x%"UVxf" NV=%"PERL_PRIgldbl"\n",
3f7c398e 6681 SvPVX_const(sv), SvIVX(sv), SvNVX(sv)));
28e5dec8 6682#else
1779d84d 6683 DEBUG_c(PerlIO_printf(Perl_debug_log,"sv_inc punt failed to convert '%s' to IOK or NOKp, UV=0x%"UVxf" NV=%"NVgf"\n",
3f7c398e 6684 SvPVX_const(sv), SvIVX(sv), SvNVX(sv)));
28e5dec8
JH
6685#endif
6686 }
6687#endif /* PERL_PRESERVE_IVUV */
3f7c398e 6688 sv_setnv(sv,Atof(SvPVX_const(sv)) + 1.0);
79072805
LW
6689 return;
6690 }
6691 d--;
3f7c398e 6692 while (d >= SvPVX_const(sv)) {
79072805
LW
6693 if (isDIGIT(*d)) {
6694 if (++*d <= '9')
6695 return;
6696 *(d--) = '0';
6697 }
6698 else {
9d116dd7
JH
6699#ifdef EBCDIC
6700 /* MKS: The original code here died if letters weren't consecutive.
6701 * at least it didn't have to worry about non-C locales. The
6702 * new code assumes that ('z'-'a')==('Z'-'A'), letters are
1c846c1f 6703 * arranged in order (although not consecutively) and that only
9d116dd7
JH
6704 * [A-Za-z] are accepted by isALPHA in the C locale.
6705 */
6706 if (*d != 'z' && *d != 'Z') {
6707 do { ++*d; } while (!isALPHA(*d));
6708 return;
6709 }
6710 *(d--) -= 'z' - 'a';
6711#else
79072805
LW
6712 ++*d;
6713 if (isALPHA(*d))
6714 return;
6715 *(d--) -= 'z' - 'a' + 1;
9d116dd7 6716#endif
79072805
LW
6717 }
6718 }
6719 /* oh,oh, the number grew */
6720 SvGROW(sv, SvCUR(sv) + 2);
b162af07 6721 SvCUR_set(sv, SvCUR(sv) + 1);
3f7c398e 6722 for (d = SvPVX(sv) + SvCUR(sv); d > SvPVX_const(sv); d--)
79072805
LW
6723 *d = d[-1];
6724 if (isDIGIT(d[1]))
6725 *d = '1';
6726 else
6727 *d = d[1];
6728}
6729
954c1994
GS
6730/*
6731=for apidoc sv_dec
6732
645c22ef
DM
6733Auto-decrement of the value in the SV, doing string to numeric conversion
6734if necessary. Handles 'get' magic.
954c1994
GS
6735
6736=cut
6737*/
6738
79072805 6739void
864dbfa3 6740Perl_sv_dec(pTHX_ register SV *sv)
79072805 6741{
97aff369 6742 dVAR;
463ee0b2
LW
6743 int flags;
6744
79072805
LW
6745 if (!sv)
6746 return;
5b295bef 6747 SvGETMAGIC(sv);
ed6116ce 6748 if (SvTHINKFIRST(sv)) {
765f542d
NC
6749 if (SvIsCOW(sv))
6750 sv_force_normal_flags(sv, 0);
0f15f207 6751 if (SvREADONLY(sv)) {
923e4eb5 6752 if (IN_PERL_RUNTIME)
cea2e8a9 6753 Perl_croak(aTHX_ PL_no_modify);
0f15f207 6754 }
a0d0e21e 6755 if (SvROK(sv)) {
b5be31e9 6756 IV i;
9e7bc3e8
JD
6757 if (SvAMAGIC(sv) && AMG_CALLun(sv,dec))
6758 return;
56431972 6759 i = PTR2IV(SvRV(sv));
b5be31e9
SM
6760 sv_unref(sv);
6761 sv_setiv(sv, i);
a0d0e21e 6762 }
ed6116ce 6763 }
28e5dec8
JH
6764 /* Unlike sv_inc we don't have to worry about string-never-numbers
6765 and keeping them magic. But we mustn't warn on punting */
8990e307 6766 flags = SvFLAGS(sv);
28e5dec8
JH
6767 if ((flags & SVf_IOK) || ((flags & (SVp_IOK | SVp_NOK)) == SVp_IOK)) {
6768 /* It's publicly an integer, or privately an integer-not-float */
59d8ce62 6769#ifdef PERL_PRESERVE_IVUV
28e5dec8 6770 oops_its_int:
59d8ce62 6771#endif
25da4f38
IZ
6772 if (SvIsUV(sv)) {
6773 if (SvUVX(sv) == 0) {
6774 (void)SvIOK_only(sv);
45977657 6775 SvIV_set(sv, -1);
25da4f38
IZ
6776 }
6777 else {
6778 (void)SvIOK_only_UV(sv);
f4eee32f 6779 SvUV_set(sv, SvUVX(sv) - 1);
1c846c1f 6780 }
25da4f38
IZ
6781 } else {
6782 if (SvIVX(sv) == IV_MIN)
65202027 6783 sv_setnv(sv, (NV)IV_MIN - 1.0);
25da4f38
IZ
6784 else {
6785 (void)SvIOK_only(sv);
45977657 6786 SvIV_set(sv, SvIVX(sv) - 1);
1c846c1f 6787 }
55497cff 6788 }
6789 return;
6790 }
28e5dec8 6791 if (flags & SVp_NOK) {
9d6ce603 6792 SvNV_set(sv, SvNVX(sv) - 1.0);
28e5dec8
JH
6793 (void)SvNOK_only(sv);
6794 return;
6795 }
8990e307 6796 if (!(flags & SVp_POK)) {
ef088171
NC
6797 if ((flags & SVTYPEMASK) < SVt_PVIV)
6798 sv_upgrade(sv, ((flags & SVTYPEMASK) > SVt_IV) ? SVt_PVIV : SVt_IV);
6799 SvIV_set(sv, -1);
6800 (void)SvIOK_only(sv);
79072805
LW
6801 return;
6802 }
28e5dec8
JH
6803#ifdef PERL_PRESERVE_IVUV
6804 {
504618e9 6805 const int numtype = grok_number(SvPVX_const(sv), SvCUR(sv), NULL);
28e5dec8
JH
6806 if (numtype && !(numtype & IS_NUMBER_INFINITY)) {
6807 /* Need to try really hard to see if it's an integer.
6808 9.22337203685478e+18 is an integer.
6809 but "9.22337203685478e+18" + 0 is UV=9223372036854779904
6810 so $a="9.22337203685478e+18"; $a+0; $a--
6811 needs to be the same as $a="9.22337203685478e+18"; $a--
6812 or we go insane. */
d460ef45 6813
28e5dec8
JH
6814 (void) sv_2iv(sv);
6815 if (SvIOK(sv))
6816 goto oops_its_int;
6817
6818 /* sv_2iv *should* have made this an NV */
6819 if (flags & SVp_NOK) {
6820 (void)SvNOK_only(sv);
9d6ce603 6821 SvNV_set(sv, SvNVX(sv) - 1.0);
28e5dec8
JH
6822 return;
6823 }
6824 /* I don't think we can get here. Maybe I should assert this
6825 And if we do get here I suspect that sv_setnv will croak. NWC
6826 Fall through. */
6827#if defined(USE_LONG_DOUBLE)
6828 DEBUG_c(PerlIO_printf(Perl_debug_log,"sv_dec punt failed to convert '%s' to IOK or NOKp, UV=0x%"UVxf" NV=%"PERL_PRIgldbl"\n",
3f7c398e 6829 SvPVX_const(sv), SvIVX(sv), SvNVX(sv)));
28e5dec8 6830#else
1779d84d 6831 DEBUG_c(PerlIO_printf(Perl_debug_log,"sv_dec punt failed to convert '%s' to IOK or NOKp, UV=0x%"UVxf" NV=%"NVgf"\n",
3f7c398e 6832 SvPVX_const(sv), SvIVX(sv), SvNVX(sv)));
28e5dec8
JH
6833#endif
6834 }
6835 }
6836#endif /* PERL_PRESERVE_IVUV */
3f7c398e 6837 sv_setnv(sv,Atof(SvPVX_const(sv)) - 1.0); /* punt */
79072805
LW
6838}
6839
954c1994
GS
6840/*
6841=for apidoc sv_mortalcopy
6842
645c22ef 6843Creates a new SV which is a copy of the original SV (using C<sv_setsv>).
d4236ebc
DM
6844The new SV is marked as mortal. It will be destroyed "soon", either by an
6845explicit call to FREETMPS, or by an implicit call at places such as
6846statement boundaries. See also C<sv_newmortal> and C<sv_2mortal>.
954c1994
GS
6847
6848=cut
6849*/
6850
79072805
LW
6851/* Make a string that will exist for the duration of the expression
6852 * evaluation. Actually, it may have to last longer than that, but
6853 * hopefully we won't free it until it has been assigned to a
6854 * permanent location. */
6855
6856SV *
864dbfa3 6857Perl_sv_mortalcopy(pTHX_ SV *oldstr)
79072805 6858{
97aff369 6859 dVAR;
463ee0b2 6860 register SV *sv;
b881518d 6861
4561caa4 6862 new_SV(sv);
79072805 6863 sv_setsv(sv,oldstr);
677b06e3
GS
6864 EXTEND_MORTAL(1);
6865 PL_tmps_stack[++PL_tmps_ix] = sv;
8990e307
LW
6866 SvTEMP_on(sv);
6867 return sv;
6868}
6869
954c1994
GS
6870/*
6871=for apidoc sv_newmortal
6872
645c22ef 6873Creates a new null SV which is mortal. The reference count of the SV is
d4236ebc
DM
6874set to 1. It will be destroyed "soon", either by an explicit call to
6875FREETMPS, or by an implicit call at places such as statement boundaries.
6876See also C<sv_mortalcopy> and C<sv_2mortal>.
954c1994
GS
6877
6878=cut
6879*/
6880
8990e307 6881SV *
864dbfa3 6882Perl_sv_newmortal(pTHX)
8990e307 6883{
97aff369 6884 dVAR;
8990e307
LW
6885 register SV *sv;
6886
4561caa4 6887 new_SV(sv);
8990e307 6888 SvFLAGS(sv) = SVs_TEMP;
677b06e3
GS
6889 EXTEND_MORTAL(1);
6890 PL_tmps_stack[++PL_tmps_ix] = sv;
79072805
LW
6891 return sv;
6892}
6893
954c1994
GS
6894/*
6895=for apidoc sv_2mortal
6896
d4236ebc
DM
6897Marks an existing SV as mortal. The SV will be destroyed "soon", either
6898by an explicit call to FREETMPS, or by an implicit call at places such as
37d2ac18
NC
6899statement boundaries. SvTEMP() is turned on which means that the SV's
6900string buffer can be "stolen" if this SV is copied. See also C<sv_newmortal>
6901and C<sv_mortalcopy>.
954c1994
GS
6902
6903=cut
6904*/
6905
79072805 6906SV *
864dbfa3 6907Perl_sv_2mortal(pTHX_ register SV *sv)
79072805 6908{
27da23d5 6909 dVAR;
79072805 6910 if (!sv)
7a5b473e 6911 return NULL;
d689ffdd 6912 if (SvREADONLY(sv) && SvIMMORTAL(sv))
11162842 6913 return sv;
677b06e3
GS
6914 EXTEND_MORTAL(1);
6915 PL_tmps_stack[++PL_tmps_ix] = sv;
8990e307 6916 SvTEMP_on(sv);
79072805
LW
6917 return sv;
6918}
6919
954c1994
GS
6920/*
6921=for apidoc newSVpv
6922
6923Creates a new SV and copies a string into it. The reference count for the
6924SV is set to 1. If C<len> is zero, Perl will compute the length using
6925strlen(). For efficiency, consider using C<newSVpvn> instead.
6926
6927=cut
6928*/
6929
79072805 6930SV *
864dbfa3 6931Perl_newSVpv(pTHX_ const char *s, STRLEN len)
79072805 6932{
97aff369 6933 dVAR;
463ee0b2 6934 register SV *sv;
79072805 6935
4561caa4 6936 new_SV(sv);
ddfa59c7 6937 sv_setpvn(sv, s, len || s == NULL ? len : strlen(s));
79072805
LW
6938 return sv;
6939}
6940
954c1994
GS
6941/*
6942=for apidoc newSVpvn
6943
6944Creates a new SV and copies a string into it. The reference count for the
1c846c1f 6945SV is set to 1. Note that if C<len> is zero, Perl will create a zero length
954c1994 6946string. You are responsible for ensuring that the source string is at least
9e09f5f2 6947C<len> bytes long. If the C<s> argument is NULL the new SV will be undefined.
954c1994
GS
6948
6949=cut
6950*/
6951
9da1e3b5 6952SV *
864dbfa3 6953Perl_newSVpvn(pTHX_ const char *s, STRLEN len)
9da1e3b5 6954{
97aff369 6955 dVAR;
9da1e3b5
MUN
6956 register SV *sv;
6957
6958 new_SV(sv);
9da1e3b5
MUN
6959 sv_setpvn(sv,s,len);
6960 return sv;
6961}
6962
bd08039b
NC
6963
6964/*
926f8064 6965=for apidoc newSVhek
bd08039b
NC
6966
6967Creates a new SV from the hash key structure. It will generate scalars that
5aaec2b4
NC
6968point to the shared string table where possible. Returns a new (undefined)
6969SV if the hek is NULL.
bd08039b
NC
6970
6971=cut
6972*/
6973
6974SV *
c1b02ed8 6975Perl_newSVhek(pTHX_ const HEK *hek)
bd08039b 6976{
97aff369 6977 dVAR;
5aaec2b4
NC
6978 if (!hek) {
6979 SV *sv;
6980
6981 new_SV(sv);
6982 return sv;
6983 }
6984
bd08039b
NC
6985 if (HEK_LEN(hek) == HEf_SVKEY) {
6986 return newSVsv(*(SV**)HEK_KEY(hek));
6987 } else {
6988 const int flags = HEK_FLAGS(hek);
6989 if (flags & HVhek_WASUTF8) {
6990 /* Trouble :-)
6991 Andreas would like keys he put in as utf8 to come back as utf8
6992 */
6993 STRLEN utf8_len = HEK_LEN(hek);
b64e5050
AL
6994 const U8 *as_utf8 = bytes_to_utf8 ((U8*)HEK_KEY(hek), &utf8_len);
6995 SV * const sv = newSVpvn ((const char*)as_utf8, utf8_len);
bd08039b
NC
6996
6997 SvUTF8_on (sv);
6998 Safefree (as_utf8); /* bytes_to_utf8() allocates a new string */
6999 return sv;
45e34800 7000 } else if (flags & (HVhek_REHASH|HVhek_UNSHARED)) {
bd08039b
NC
7001 /* We don't have a pointer to the hv, so we have to replicate the
7002 flag into every HEK. This hv is using custom a hasing
7003 algorithm. Hence we can't return a shared string scalar, as
7004 that would contain the (wrong) hash value, and might get passed
45e34800
NC
7005 into an hv routine with a regular hash.
7006 Similarly, a hash that isn't using shared hash keys has to have
7007 the flag in every key so that we know not to try to call
7008 share_hek_kek on it. */
bd08039b 7009
b64e5050 7010 SV * const sv = newSVpvn (HEK_KEY(hek), HEK_LEN(hek));
bd08039b
NC
7011 if (HEK_UTF8(hek))
7012 SvUTF8_on (sv);
7013 return sv;
7014 }
7015 /* This will be overwhelminly the most common case. */
409dfe77
NC
7016 {
7017 /* Inline most of newSVpvn_share(), because share_hek_hek() is far
7018 more efficient than sharepvn(). */
7019 SV *sv;
7020
7021 new_SV(sv);
7022 sv_upgrade(sv, SVt_PV);
7023 SvPV_set(sv, (char *)HEK_KEY(share_hek_hek(hek)));
7024 SvCUR_set(sv, HEK_LEN(hek));
7025 SvLEN_set(sv, 0);
7026 SvREADONLY_on(sv);
7027 SvFAKE_on(sv);
7028 SvPOK_on(sv);
7029 if (HEK_UTF8(hek))
7030 SvUTF8_on(sv);
7031 return sv;
7032 }
bd08039b
NC
7033 }
7034}
7035
1c846c1f
NIS
7036/*
7037=for apidoc newSVpvn_share
7038
3f7c398e 7039Creates a new SV with its SvPVX_const pointing to a shared string in the string
645c22ef
DM
7040table. If the string does not already exist in the table, it is created
7041first. Turns on READONLY and FAKE. The string's hash is stored in the UV
7042slot of the SV; if the C<hash> parameter is non-zero, that value is used;
7043otherwise the hash is computed. The idea here is that as the string table
3f7c398e 7044is used for shared hash keys these strings will have SvPVX_const == HeKEY and
645c22ef 7045hash lookup will avoid string compare.
1c846c1f
NIS
7046
7047=cut
7048*/
7049
7050SV *
c3654f1a 7051Perl_newSVpvn_share(pTHX_ const char *src, I32 len, U32 hash)
1c846c1f 7052{
97aff369 7053 dVAR;
1c846c1f 7054 register SV *sv;
c3654f1a 7055 bool is_utf8 = FALSE;
a51caccf
NC
7056 const char *const orig_src = src;
7057
c3654f1a 7058 if (len < 0) {
77caf834 7059 STRLEN tmplen = -len;
c3654f1a 7060 is_utf8 = TRUE;
75a54232 7061 /* See the note in hv.c:hv_fetch() --jhi */
e1ec3a88 7062 src = (char*)bytes_from_utf8((const U8*)src, &tmplen, &is_utf8);
75a54232
JH
7063 len = tmplen;
7064 }
1c846c1f 7065 if (!hash)
5afd6d42 7066 PERL_HASH(hash, src, len);
1c846c1f 7067 new_SV(sv);
bdd68bc3 7068 sv_upgrade(sv, SVt_PV);
f880fe2f 7069 SvPV_set(sv, sharepvn(src, is_utf8?-len:len, hash));
b162af07 7070 SvCUR_set(sv, len);
b162af07 7071 SvLEN_set(sv, 0);
1c846c1f
NIS
7072 SvREADONLY_on(sv);
7073 SvFAKE_on(sv);
7074 SvPOK_on(sv);
c3654f1a
IH
7075 if (is_utf8)
7076 SvUTF8_on(sv);
a51caccf
NC
7077 if (src != orig_src)
7078 Safefree(src);
1c846c1f
NIS
7079 return sv;
7080}
7081
645c22ef 7082
cea2e8a9 7083#if defined(PERL_IMPLICIT_CONTEXT)
645c22ef
DM
7084
7085/* pTHX_ magic can't cope with varargs, so this is a no-context
7086 * version of the main function, (which may itself be aliased to us).
7087 * Don't access this version directly.
7088 */
7089
46fc3d4c 7090SV *
cea2e8a9 7091Perl_newSVpvf_nocontext(const char* pat, ...)
46fc3d4c 7092{
cea2e8a9 7093 dTHX;
46fc3d4c 7094 register SV *sv;
7095 va_list args;
46fc3d4c 7096 va_start(args, pat);
c5be433b 7097 sv = vnewSVpvf(pat, &args);
46fc3d4c 7098 va_end(args);
7099 return sv;
7100}
cea2e8a9 7101#endif
46fc3d4c 7102
954c1994
GS
7103/*
7104=for apidoc newSVpvf
7105
645c22ef 7106Creates a new SV and initializes it with the string formatted like
954c1994
GS
7107C<sprintf>.
7108
7109=cut
7110*/
7111
cea2e8a9
GS
7112SV *
7113Perl_newSVpvf(pTHX_ const char* pat, ...)
7114{
7115 register SV *sv;
7116 va_list args;
cea2e8a9 7117 va_start(args, pat);
c5be433b 7118 sv = vnewSVpvf(pat, &args);
cea2e8a9
GS
7119 va_end(args);
7120 return sv;
7121}
46fc3d4c 7122
645c22ef
DM
7123/* backend for newSVpvf() and newSVpvf_nocontext() */
7124
79072805 7125SV *
c5be433b
GS
7126Perl_vnewSVpvf(pTHX_ const char* pat, va_list* args)
7127{
97aff369 7128 dVAR;
c5be433b
GS
7129 register SV *sv;
7130 new_SV(sv);
4608196e 7131 sv_vsetpvfn(sv, pat, strlen(pat), args, NULL, 0, NULL);
c5be433b
GS
7132 return sv;
7133}
7134
954c1994
GS
7135/*
7136=for apidoc newSVnv
7137
7138Creates a new SV and copies a floating point value into it.
7139The reference count for the SV is set to 1.
7140
7141=cut
7142*/
7143
c5be433b 7144SV *
65202027 7145Perl_newSVnv(pTHX_ NV n)
79072805 7146{
97aff369 7147 dVAR;
463ee0b2 7148 register SV *sv;
79072805 7149
4561caa4 7150 new_SV(sv);
79072805
LW
7151 sv_setnv(sv,n);
7152 return sv;
7153}
7154
954c1994
GS
7155/*
7156=for apidoc newSViv
7157
7158Creates a new SV and copies an integer into it. The reference count for the
7159SV is set to 1.
7160
7161=cut
7162*/
7163
79072805 7164SV *
864dbfa3 7165Perl_newSViv(pTHX_ IV i)
79072805 7166{
97aff369 7167 dVAR;
463ee0b2 7168 register SV *sv;
79072805 7169
4561caa4 7170 new_SV(sv);
79072805
LW
7171 sv_setiv(sv,i);
7172 return sv;
7173}
7174
954c1994 7175/*
1a3327fb
JH
7176=for apidoc newSVuv
7177
7178Creates a new SV and copies an unsigned integer into it.
7179The reference count for the SV is set to 1.
7180
7181=cut
7182*/
7183
7184SV *
7185Perl_newSVuv(pTHX_ UV u)
7186{
97aff369 7187 dVAR;
1a3327fb
JH
7188 register SV *sv;
7189
7190 new_SV(sv);
7191 sv_setuv(sv,u);
7192 return sv;
7193}
7194
7195/*
954c1994
GS
7196=for apidoc newRV_noinc
7197
7198Creates an RV wrapper for an SV. The reference count for the original
7199SV is B<not> incremented.
7200
7201=cut
7202*/
7203
2304df62 7204SV *
864dbfa3 7205Perl_newRV_noinc(pTHX_ SV *tmpRef)
2304df62 7206{
97aff369 7207 dVAR;
2304df62
AD
7208 register SV *sv;
7209
4561caa4 7210 new_SV(sv);
2304df62 7211 sv_upgrade(sv, SVt_RV);
76e3520e 7212 SvTEMP_off(tmpRef);
b162af07 7213 SvRV_set(sv, tmpRef);
2304df62 7214 SvROK_on(sv);
2304df62
AD
7215 return sv;
7216}
7217
ff276b08 7218/* newRV_inc is the official function name to use now.
645c22ef
DM
7219 * newRV_inc is in fact #defined to newRV in sv.h
7220 */
7221
5f05dabc 7222SV *
7f466ec7 7223Perl_newRV(pTHX_ SV *sv)
5f05dabc 7224{
97aff369 7225 dVAR;
7f466ec7 7226 return newRV_noinc(SvREFCNT_inc_simple_NN(sv));
5f05dabc 7227}
5f05dabc 7228
954c1994
GS
7229/*
7230=for apidoc newSVsv
7231
7232Creates a new SV which is an exact duplicate of the original SV.
645c22ef 7233(Uses C<sv_setsv>).
954c1994
GS
7234
7235=cut
7236*/
7237
79072805 7238SV *
864dbfa3 7239Perl_newSVsv(pTHX_ register SV *old)
79072805 7240{
97aff369 7241 dVAR;
463ee0b2 7242 register SV *sv;
79072805
LW
7243
7244 if (!old)
7a5b473e 7245 return NULL;
8990e307 7246 if (SvTYPE(old) == SVTYPEMASK) {
0453d815 7247 if (ckWARN_d(WARN_INTERNAL))
9014280d 7248 Perl_warner(aTHX_ packWARN(WARN_INTERNAL), "semi-panic: attempt to dup freed string");
a0714e2c 7249 return NULL;
79072805 7250 }
4561caa4 7251 new_SV(sv);
e90aabeb
NC
7252 /* SV_GMAGIC is the default for sv_setv()
7253 SV_NOSTEAL prevents TEMP buffers being, well, stolen, and saves games
7254 with SvTEMP_off and SvTEMP_on round a call to sv_setsv. */
7255 sv_setsv_flags(sv, old, SV_GMAGIC | SV_NOSTEAL);
463ee0b2 7256 return sv;
79072805
LW
7257}
7258
645c22ef
DM
7259/*
7260=for apidoc sv_reset
7261
7262Underlying implementation for the C<reset> Perl function.
7263Note that the perl-level function is vaguely deprecated.
7264
7265=cut
7266*/
7267
79072805 7268void
e1ec3a88 7269Perl_sv_reset(pTHX_ register const char *s, HV *stash)
79072805 7270{
27da23d5 7271 dVAR;
4802d5d7 7272 char todo[PERL_UCHAR_MAX+1];
79072805 7273
49d8d3a1
MB
7274 if (!stash)
7275 return;
7276
79072805 7277 if (!*s) { /* reset ?? searches */
aec46f14 7278 MAGIC * const mg = mg_find((SV *)stash, PERL_MAGIC_symtab);
8d2f4536
NC
7279 if (mg) {
7280 PMOP *pm = (PMOP *) mg->mg_obj;
7281 while (pm) {
7282 pm->op_pmdynflags &= ~PMdf_USED;
7283 pm = pm->op_pmnext;
7284 }
79072805
LW
7285 }
7286 return;
7287 }
7288
7289 /* reset variables */
7290
7291 if (!HvARRAY(stash))
7292 return;
463ee0b2
LW
7293
7294 Zero(todo, 256, char);
79072805 7295 while (*s) {
b464bac0
AL
7296 I32 max;
7297 I32 i = (unsigned char)*s;
79072805
LW
7298 if (s[1] == '-') {
7299 s += 2;
7300 }
4802d5d7 7301 max = (unsigned char)*s++;
79072805 7302 for ( ; i <= max; i++) {
463ee0b2
LW
7303 todo[i] = 1;
7304 }
a0d0e21e 7305 for (i = 0; i <= (I32) HvMAX(stash); i++) {
b464bac0 7306 HE *entry;
79072805 7307 for (entry = HvARRAY(stash)[i];
9e35f4b3
GS
7308 entry;
7309 entry = HeNEXT(entry))
7310 {
b464bac0
AL
7311 register GV *gv;
7312 register SV *sv;
7313
1edc1566 7314 if (!todo[(U8)*HeKEY(entry)])
463ee0b2 7315 continue;
1edc1566 7316 gv = (GV*)HeVAL(entry);
79072805 7317 sv = GvSV(gv);
e203899d
NC
7318 if (sv) {
7319 if (SvTHINKFIRST(sv)) {
7320 if (!SvREADONLY(sv) && SvROK(sv))
7321 sv_unref(sv);
7322 /* XXX Is this continue a bug? Why should THINKFIRST
7323 exempt us from resetting arrays and hashes? */
7324 continue;
7325 }
7326 SvOK_off(sv);
7327 if (SvTYPE(sv) >= SVt_PV) {
7328 SvCUR_set(sv, 0);
bd61b366 7329 if (SvPVX_const(sv) != NULL)
e203899d
NC
7330 *SvPVX(sv) = '\0';
7331 SvTAINT(sv);
7332 }
79072805
LW
7333 }
7334 if (GvAV(gv)) {
7335 av_clear(GvAV(gv));
7336 }
bfcb3514 7337 if (GvHV(gv) && !HvNAME_get(GvHV(gv))) {
b0269e46
AB
7338#if defined(VMS)
7339 Perl_die(aTHX_ "Can't reset %%ENV on this system");
7340#else /* ! VMS */
463ee0b2 7341 hv_clear(GvHV(gv));
b0269e46
AB
7342# if defined(USE_ENVIRON_ARRAY)
7343 if (gv == PL_envgv)
7344 my_clearenv();
7345# endif /* USE_ENVIRON_ARRAY */
7346#endif /* VMS */
79072805
LW
7347 }
7348 }
7349 }
7350 }
7351}
7352
645c22ef
DM
7353/*
7354=for apidoc sv_2io
7355
7356Using various gambits, try to get an IO from an SV: the IO slot if its a
7357GV; or the recursive result if we're an RV; or the IO slot of the symbol
7358named after the PV if we're a string.
7359
7360=cut
7361*/
7362
46fc3d4c 7363IO*
864dbfa3 7364Perl_sv_2io(pTHX_ SV *sv)
46fc3d4c 7365{
7366 IO* io;
7367 GV* gv;
7368
7369 switch (SvTYPE(sv)) {
7370 case SVt_PVIO:
7371 io = (IO*)sv;
7372 break;
7373 case SVt_PVGV:
7374 gv = (GV*)sv;
7375 io = GvIO(gv);
7376 if (!io)
cea2e8a9 7377 Perl_croak(aTHX_ "Bad filehandle: %s", GvNAME(gv));
46fc3d4c 7378 break;
7379 default:
7380 if (!SvOK(sv))
cea2e8a9 7381 Perl_croak(aTHX_ PL_no_usym, "filehandle");
46fc3d4c 7382 if (SvROK(sv))
7383 return sv_2io(SvRV(sv));
f776e3cd 7384 gv = gv_fetchsv(sv, 0, SVt_PVIO);
46fc3d4c 7385 if (gv)
7386 io = GvIO(gv);
7387 else
7388 io = 0;
7389 if (!io)
95b63a38 7390 Perl_croak(aTHX_ "Bad filehandle: %"SVf, (void*)sv);
46fc3d4c 7391 break;
7392 }
7393 return io;
7394}
7395
645c22ef
DM
7396/*
7397=for apidoc sv_2cv
7398
7399Using various gambits, try to get a CV from an SV; in addition, try if
7400possible to set C<*st> and C<*gvp> to the stash and GV associated with it.
f2c0649b 7401The flags in C<lref> are passed to sv_fetchsv.
645c22ef
DM
7402
7403=cut
7404*/
7405
79072805 7406CV *
864dbfa3 7407Perl_sv_2cv(pTHX_ SV *sv, HV **st, GV **gvp, I32 lref)
79072805 7408{
27da23d5 7409 dVAR;
a0714e2c 7410 GV *gv = NULL;
601f1833 7411 CV *cv = NULL;
79072805 7412
85dec29a
NC
7413 if (!sv) {
7414 *st = NULL;
7415 *gvp = NULL;
7416 return NULL;
7417 }
79072805 7418 switch (SvTYPE(sv)) {
79072805
LW
7419 case SVt_PVCV:
7420 *st = CvSTASH(sv);
a0714e2c 7421 *gvp = NULL;
79072805
LW
7422 return (CV*)sv;
7423 case SVt_PVHV:
7424 case SVt_PVAV:
ef58ba18 7425 *st = NULL;
a0714e2c 7426 *gvp = NULL;
601f1833 7427 return NULL;
8990e307
LW
7428 case SVt_PVGV:
7429 gv = (GV*)sv;
a0d0e21e 7430 *gvp = gv;
8990e307
LW
7431 *st = GvESTASH(gv);
7432 goto fix_gv;
7433
79072805 7434 default:
5b295bef 7435 SvGETMAGIC(sv);
a0d0e21e 7436 if (SvROK(sv)) {
823a54a3 7437 SV * const *sp = &sv; /* Used in tryAMAGICunDEREF macro. */
f5284f61
IZ
7438 tryAMAGICunDEREF(to_cv);
7439
62f274bf
GS
7440 sv = SvRV(sv);
7441 if (SvTYPE(sv) == SVt_PVCV) {
7442 cv = (CV*)sv;
a0714e2c 7443 *gvp = NULL;
62f274bf
GS
7444 *st = CvSTASH(cv);
7445 return cv;
7446 }
7447 else if(isGV(sv))
7448 gv = (GV*)sv;
7449 else
cea2e8a9 7450 Perl_croak(aTHX_ "Not a subroutine reference");
a0d0e21e 7451 }
62f274bf 7452 else if (isGV(sv))
79072805
LW
7453 gv = (GV*)sv;
7454 else
7a5fd60d 7455 gv = gv_fetchsv(sv, lref, SVt_PVCV);
79072805 7456 *gvp = gv;
ef58ba18
NC
7457 if (!gv) {
7458 *st = NULL;
601f1833 7459 return NULL;
ef58ba18 7460 }
e26df76a
NC
7461 /* Some flags to gv_fetchsv mean don't really create the GV */
7462 if (SvTYPE(gv) != SVt_PVGV) {
7463 *st = NULL;
7464 return NULL;
7465 }
79072805 7466 *st = GvESTASH(gv);
8990e307 7467 fix_gv:
8ebc5c01 7468 if (lref && !GvCVu(gv)) {
4633a7c4 7469 SV *tmpsv;
748a9306 7470 ENTER;
561b68a9 7471 tmpsv = newSV(0);
bd61b366 7472 gv_efullname3(tmpsv, gv, NULL);
f6ec51f7
GS
7473 /* XXX this is probably not what they think they're getting.
7474 * It has the same effect as "sub name;", i.e. just a forward
7475 * declaration! */
774d564b 7476 newSUB(start_subparse(FALSE, 0),
4633a7c4 7477 newSVOP(OP_CONST, 0, tmpsv),
5f66b61c 7478 NULL, NULL);
748a9306 7479 LEAVE;
8ebc5c01 7480 if (!GvCVu(gv))
35c1215d 7481 Perl_croak(aTHX_ "Unable to create sub named \"%"SVf"\"",
95b63a38 7482 (void*)sv);
8990e307 7483 }
8ebc5c01 7484 return GvCVu(gv);
79072805
LW
7485 }
7486}
7487
c461cf8f
JH
7488/*
7489=for apidoc sv_true
7490
7491Returns true if the SV has a true value by Perl's rules.
645c22ef
DM
7492Use the C<SvTRUE> macro instead, which may call C<sv_true()> or may
7493instead use an in-line version.
c461cf8f
JH
7494
7495=cut
7496*/
7497
79072805 7498I32
864dbfa3 7499Perl_sv_true(pTHX_ register SV *sv)
79072805 7500{
8990e307
LW
7501 if (!sv)
7502 return 0;
79072805 7503 if (SvPOK(sv)) {
823a54a3
AL
7504 register const XPV* const tXpv = (XPV*)SvANY(sv);
7505 if (tXpv &&
c2f1de04 7506 (tXpv->xpv_cur > 1 ||
339049b0 7507 (tXpv->xpv_cur && *sv->sv_u.svu_pv != '0')))
79072805
LW
7508 return 1;
7509 else
7510 return 0;
7511 }
7512 else {
7513 if (SvIOK(sv))
463ee0b2 7514 return SvIVX(sv) != 0;
79072805
LW
7515 else {
7516 if (SvNOK(sv))
463ee0b2 7517 return SvNVX(sv) != 0.0;
79072805 7518 else
463ee0b2 7519 return sv_2bool(sv);
79072805
LW
7520 }
7521 }
7522}
79072805 7523
645c22ef 7524/*
c461cf8f
JH
7525=for apidoc sv_pvn_force
7526
7527Get a sensible string out of the SV somehow.
645c22ef
DM
7528A private implementation of the C<SvPV_force> macro for compilers which
7529can't cope with complex macro expressions. Always use the macro instead.
c461cf8f 7530
8d6d96c1
HS
7531=for apidoc sv_pvn_force_flags
7532
7533Get a sensible string out of the SV somehow.
7534If C<flags> has C<SV_GMAGIC> bit set, will C<mg_get> on C<sv> if
7535appropriate, else not. C<sv_pvn_force> and C<sv_pvn_force_nomg> are
7536implemented in terms of this function.
645c22ef
DM
7537You normally want to use the various wrapper macros instead: see
7538C<SvPV_force> and C<SvPV_force_nomg>
8d6d96c1
HS
7539
7540=cut
7541*/
7542
7543char *
7544Perl_sv_pvn_force_flags(pTHX_ SV *sv, STRLEN *lp, I32 flags)
7545{
97aff369 7546 dVAR;
6fc92669 7547 if (SvTHINKFIRST(sv) && !SvROK(sv))
765f542d 7548 sv_force_normal_flags(sv, 0);
1c846c1f 7549
a0d0e21e 7550 if (SvPOK(sv)) {
13c5b33c
NC
7551 if (lp)
7552 *lp = SvCUR(sv);
a0d0e21e
LW
7553 }
7554 else {
a3b680e6 7555 char *s;
13c5b33c
NC
7556 STRLEN len;
7557
4d84ee25 7558 if (SvREADONLY(sv) && !(flags & SV_MUTABLE_RETURN)) {
b64e5050 7559 const char * const ref = sv_reftype(sv,0);
4d84ee25
NC
7560 if (PL_op)
7561 Perl_croak(aTHX_ "Can't coerce readonly %s to string in %s",
b64e5050 7562 ref, OP_NAME(PL_op));
4d84ee25 7563 else
b64e5050 7564 Perl_croak(aTHX_ "Can't coerce readonly %s to string", ref);
4d84ee25 7565 }
b64e5050 7566 if (SvTYPE(sv) > SVt_PVLV && SvTYPE(sv) != SVt_PVFM)
cea2e8a9 7567 Perl_croak(aTHX_ "Can't coerce %s to string in %s", sv_reftype(sv,0),
53e06cf0 7568 OP_NAME(PL_op));
b64e5050 7569 s = sv_2pv_flags(sv, &len, flags);
13c5b33c
NC
7570 if (lp)
7571 *lp = len;
7572
3f7c398e 7573 if (s != SvPVX_const(sv)) { /* Almost, but not quite, sv_setpvn() */
a0d0e21e
LW
7574 if (SvROK(sv))
7575 sv_unref(sv);
862a34c6 7576 SvUPGRADE(sv, SVt_PV); /* Never FALSE */
a0d0e21e 7577 SvGROW(sv, len + 1);
706aa1c9 7578 Move(s,SvPVX(sv),len,char);
a0d0e21e
LW
7579 SvCUR_set(sv, len);
7580 *SvEND(sv) = '\0';
7581 }
7582 if (!SvPOK(sv)) {
7583 SvPOK_on(sv); /* validate pointer */
7584 SvTAINT(sv);
1d7c1841 7585 DEBUG_c(PerlIO_printf(Perl_debug_log, "0x%"UVxf" 2pv(%s)\n",
3f7c398e 7586 PTR2UV(sv),SvPVX_const(sv)));
a0d0e21e
LW
7587 }
7588 }
4d84ee25 7589 return SvPVX_mutable(sv);
a0d0e21e
LW
7590}
7591
645c22ef 7592/*
645c22ef
DM
7593=for apidoc sv_pvbyten_force
7594
0feed65a 7595The backend for the C<SvPVbytex_force> macro. Always use the macro instead.
645c22ef
DM
7596
7597=cut
7598*/
7599
7340a771
GS
7600char *
7601Perl_sv_pvbyten_force(pTHX_ SV *sv, STRLEN *lp)
7602{
46ec2f14 7603 sv_pvn_force(sv,lp);
ffebcc3e 7604 sv_utf8_downgrade(sv,0);
46ec2f14
TS
7605 *lp = SvCUR(sv);
7606 return SvPVX(sv);
7340a771
GS
7607}
7608
645c22ef 7609/*
c461cf8f
JH
7610=for apidoc sv_pvutf8n_force
7611
0feed65a 7612The backend for the C<SvPVutf8x_force> macro. Always use the macro instead.
c461cf8f
JH
7613
7614=cut
7615*/
7616
7340a771
GS
7617char *
7618Perl_sv_pvutf8n_force(pTHX_ SV *sv, STRLEN *lp)
7619{
46ec2f14 7620 sv_pvn_force(sv,lp);
560a288e 7621 sv_utf8_upgrade(sv);
46ec2f14
TS
7622 *lp = SvCUR(sv);
7623 return SvPVX(sv);
7340a771
GS
7624}
7625
c461cf8f
JH
7626/*
7627=for apidoc sv_reftype
7628
7629Returns a string describing what the SV is a reference to.
7630
7631=cut
7632*/
7633
1cb0ed9b 7634char *
bfed75c6 7635Perl_sv_reftype(pTHX_ const SV *sv, int ob)
a0d0e21e 7636{
07409e01
NC
7637 /* The fact that I don't need to downcast to char * everywhere, only in ?:
7638 inside return suggests a const propagation bug in g++. */
c86bf373 7639 if (ob && SvOBJECT(sv)) {
1b6737cc 7640 char * const name = HvNAME_get(SvSTASH(sv));
07409e01 7641 return name ? name : (char *) "__ANON__";
c86bf373 7642 }
a0d0e21e
LW
7643 else {
7644 switch (SvTYPE(sv)) {
7645 case SVt_NULL:
7646 case SVt_IV:
7647 case SVt_NV:
7648 case SVt_RV:
7649 case SVt_PV:
7650 case SVt_PVIV:
7651 case SVt_PVNV:
7652 case SVt_PVMG:
7653 case SVt_PVBM:
1cb0ed9b 7654 if (SvVOK(sv))
439cb1c4 7655 return "VSTRING";
a0d0e21e
LW
7656 if (SvROK(sv))
7657 return "REF";
7658 else
7659 return "SCALAR";
1cb0ed9b 7660
07409e01 7661 case SVt_PVLV: return (char *) (SvROK(sv) ? "REF"
be65207d
DM
7662 /* tied lvalues should appear to be
7663 * scalars for backwards compatitbility */
7664 : (LvTYPE(sv) == 't' || LvTYPE(sv) == 'T')
07409e01 7665 ? "SCALAR" : "LVALUE");
a0d0e21e
LW
7666 case SVt_PVAV: return "ARRAY";
7667 case SVt_PVHV: return "HASH";
7668 case SVt_PVCV: return "CODE";
7669 case SVt_PVGV: return "GLOB";
1d2dff63 7670 case SVt_PVFM: return "FORMAT";
27f9d8f3 7671 case SVt_PVIO: return "IO";
a0d0e21e
LW
7672 default: return "UNKNOWN";
7673 }
7674 }
7675}
7676
954c1994
GS
7677/*
7678=for apidoc sv_isobject
7679
7680Returns a boolean indicating whether the SV is an RV pointing to a blessed
7681object. If the SV is not an RV, or if the object is not blessed, then this
7682will return false.
7683
7684=cut
7685*/
7686
463ee0b2 7687int
864dbfa3 7688Perl_sv_isobject(pTHX_ SV *sv)
85e6fe83 7689{
68dc0745 7690 if (!sv)
7691 return 0;
5b295bef 7692 SvGETMAGIC(sv);
85e6fe83
LW
7693 if (!SvROK(sv))
7694 return 0;
7695 sv = (SV*)SvRV(sv);
7696 if (!SvOBJECT(sv))
7697 return 0;
7698 return 1;
7699}
7700
954c1994
GS
7701/*
7702=for apidoc sv_isa
7703
7704Returns a boolean indicating whether the SV is blessed into the specified
7705class. This does not check for subtypes; use C<sv_derived_from> to verify
7706an inheritance relationship.
7707
7708=cut
7709*/
7710
85e6fe83 7711int
864dbfa3 7712Perl_sv_isa(pTHX_ SV *sv, const char *name)
463ee0b2 7713{
bfcb3514 7714 const char *hvname;
68dc0745 7715 if (!sv)
7716 return 0;
5b295bef 7717 SvGETMAGIC(sv);
ed6116ce 7718 if (!SvROK(sv))
463ee0b2 7719 return 0;
ed6116ce
LW
7720 sv = (SV*)SvRV(sv);
7721 if (!SvOBJECT(sv))
463ee0b2 7722 return 0;
bfcb3514
NC
7723 hvname = HvNAME_get(SvSTASH(sv));
7724 if (!hvname)
e27ad1f2 7725 return 0;
463ee0b2 7726
bfcb3514 7727 return strEQ(hvname, name);
463ee0b2
LW
7728}
7729
954c1994
GS
7730/*
7731=for apidoc newSVrv
7732
7733Creates a new SV for the RV, C<rv>, to point to. If C<rv> is not an RV then
7734it will be upgraded to one. If C<classname> is non-null then the new SV will
7735be blessed in the specified package. The new SV is returned and its
7736reference count is 1.
7737
7738=cut
7739*/
7740
463ee0b2 7741SV*
864dbfa3 7742Perl_newSVrv(pTHX_ SV *rv, const char *classname)
463ee0b2 7743{
97aff369 7744 dVAR;
463ee0b2
LW
7745 SV *sv;
7746
4561caa4 7747 new_SV(sv);
51cf62d8 7748
765f542d 7749 SV_CHECK_THINKFIRST_COW_DROP(rv);
51cf62d8 7750 SvAMAGIC_off(rv);
51cf62d8 7751
0199fce9 7752 if (SvTYPE(rv) >= SVt_PVMG) {
a3b680e6 7753 const U32 refcnt = SvREFCNT(rv);
0199fce9
JD
7754 SvREFCNT(rv) = 0;
7755 sv_clear(rv);
7756 SvFLAGS(rv) = 0;
7757 SvREFCNT(rv) = refcnt;
0199fce9 7758
dc5494d2
NC
7759 sv_upgrade(rv, SVt_RV);
7760 } else if (SvROK(rv)) {
7761 SvREFCNT_dec(SvRV(rv));
7762 } else if (SvTYPE(rv) < SVt_RV)
0199fce9
JD
7763 sv_upgrade(rv, SVt_RV);
7764 else if (SvTYPE(rv) > SVt_RV) {
8bd4d4c5 7765 SvPV_free(rv);
0199fce9
JD
7766 SvCUR_set(rv, 0);
7767 SvLEN_set(rv, 0);
7768 }
51cf62d8 7769
0c34ef67 7770 SvOK_off(rv);
b162af07 7771 SvRV_set(rv, sv);
ed6116ce 7772 SvROK_on(rv);
463ee0b2 7773
a0d0e21e 7774 if (classname) {
1b6737cc 7775 HV* const stash = gv_stashpv(classname, TRUE);
a0d0e21e
LW
7776 (void)sv_bless(rv, stash);
7777 }
7778 return sv;
7779}
7780
954c1994
GS
7781/*
7782=for apidoc sv_setref_pv
7783
7784Copies a pointer into a new SV, optionally blessing the SV. The C<rv>
7785argument will be upgraded to an RV. That RV will be modified to point to
7786the new SV. If the C<pv> argument is NULL then C<PL_sv_undef> will be placed
7787into the SV. The C<classname> argument indicates the package for the
bd61b366 7788blessing. Set C<classname> to C<NULL> to avoid the blessing. The new SV
d34c2299 7789will have a reference count of 1, and the RV will be returned.
954c1994
GS
7790
7791Do not use with other Perl types such as HV, AV, SV, CV, because those
7792objects will become corrupted by the pointer copy process.
7793
7794Note that C<sv_setref_pvn> copies the string while this copies the pointer.
7795
7796=cut
7797*/
7798
a0d0e21e 7799SV*
864dbfa3 7800Perl_sv_setref_pv(pTHX_ SV *rv, const char *classname, void *pv)
a0d0e21e 7801{
97aff369 7802 dVAR;
189b2af5 7803 if (!pv) {
3280af22 7804 sv_setsv(rv, &PL_sv_undef);
189b2af5
GS
7805 SvSETMAGIC(rv);
7806 }
a0d0e21e 7807 else
56431972 7808 sv_setiv(newSVrv(rv,classname), PTR2IV(pv));
a0d0e21e
LW
7809 return rv;
7810}
7811
954c1994
GS
7812/*
7813=for apidoc sv_setref_iv
7814
7815Copies an integer into a new SV, optionally blessing the SV. The C<rv>
7816argument will be upgraded to an RV. That RV will be modified to point to
7817the new SV. The C<classname> argument indicates the package for the
bd61b366 7818blessing. Set C<classname> to C<NULL> to avoid the blessing. The new SV
d34c2299 7819will have a reference count of 1, and the RV will be returned.
954c1994
GS
7820
7821=cut
7822*/
7823
a0d0e21e 7824SV*
864dbfa3 7825Perl_sv_setref_iv(pTHX_ SV *rv, const char *classname, IV iv)
a0d0e21e
LW
7826{
7827 sv_setiv(newSVrv(rv,classname), iv);
7828 return rv;
7829}
7830
954c1994 7831/*
e1c57cef
JH
7832=for apidoc sv_setref_uv
7833
7834Copies an unsigned integer into a new SV, optionally blessing the SV. The C<rv>
7835argument will be upgraded to an RV. That RV will be modified to point to
7836the new SV. The C<classname> argument indicates the package for the
bd61b366 7837blessing. Set C<classname> to C<NULL> to avoid the blessing. The new SV
d34c2299 7838will have a reference count of 1, and the RV will be returned.
e1c57cef
JH
7839
7840=cut
7841*/
7842
7843SV*
7844Perl_sv_setref_uv(pTHX_ SV *rv, const char *classname, UV uv)
7845{
7846 sv_setuv(newSVrv(rv,classname), uv);
7847 return rv;
7848}
7849
7850/*
954c1994
GS
7851=for apidoc sv_setref_nv
7852
7853Copies a double into a new SV, optionally blessing the SV. The C<rv>
7854argument will be upgraded to an RV. That RV will be modified to point to
7855the new SV. The C<classname> argument indicates the package for the
bd61b366 7856blessing. Set C<classname> to C<NULL> to avoid the blessing. The new SV
d34c2299 7857will have a reference count of 1, and the RV will be returned.
954c1994
GS
7858
7859=cut
7860*/
7861
a0d0e21e 7862SV*
65202027 7863Perl_sv_setref_nv(pTHX_ SV *rv, const char *classname, NV nv)
a0d0e21e
LW
7864{
7865 sv_setnv(newSVrv(rv,classname), nv);
7866 return rv;
7867}
463ee0b2 7868
954c1994
GS
7869/*
7870=for apidoc sv_setref_pvn
7871
7872Copies a string into a new SV, optionally blessing the SV. The length of the
7873string must be specified with C<n>. The C<rv> argument will be upgraded to
7874an RV. That RV will be modified to point to the new SV. The C<classname>
7875argument indicates the package for the blessing. Set C<classname> to
bd61b366 7876C<NULL> to avoid the blessing. The new SV will have a reference count
d34c2299 7877of 1, and the RV will be returned.
954c1994
GS
7878
7879Note that C<sv_setref_pv> copies the pointer while this copies the string.
7880
7881=cut
7882*/
7883
a0d0e21e 7884SV*
1b6737cc 7885Perl_sv_setref_pvn(pTHX_ SV *rv, const char *classname, const char *pv, STRLEN n)
a0d0e21e
LW
7886{
7887 sv_setpvn(newSVrv(rv,classname), pv, n);
463ee0b2
LW
7888 return rv;
7889}
7890
954c1994
GS
7891/*
7892=for apidoc sv_bless
7893
7894Blesses an SV into a specified package. The SV must be an RV. The package
7895must be designated by its stash (see C<gv_stashpv()>). The reference count
7896of the SV is unaffected.
7897
7898=cut
7899*/
7900
a0d0e21e 7901SV*
864dbfa3 7902Perl_sv_bless(pTHX_ SV *sv, HV *stash)
a0d0e21e 7903{
97aff369 7904 dVAR;
76e3520e 7905 SV *tmpRef;
a0d0e21e 7906 if (!SvROK(sv))
cea2e8a9 7907 Perl_croak(aTHX_ "Can't bless non-reference value");
76e3520e
GS
7908 tmpRef = SvRV(sv);
7909 if (SvFLAGS(tmpRef) & (SVs_OBJECT|SVf_READONLY)) {
7910 if (SvREADONLY(tmpRef))
cea2e8a9 7911 Perl_croak(aTHX_ PL_no_modify);
76e3520e
GS
7912 if (SvOBJECT(tmpRef)) {
7913 if (SvTYPE(tmpRef) != SVt_PVIO)
3280af22 7914 --PL_sv_objcount;
76e3520e 7915 SvREFCNT_dec(SvSTASH(tmpRef));
2e3febc6 7916 }
a0d0e21e 7917 }
76e3520e
GS
7918 SvOBJECT_on(tmpRef);
7919 if (SvTYPE(tmpRef) != SVt_PVIO)
3280af22 7920 ++PL_sv_objcount;
862a34c6 7921 SvUPGRADE(tmpRef, SVt_PVMG);
b37c2d43 7922 SvSTASH_set(tmpRef, (HV*)SvREFCNT_inc_simple(stash));
a0d0e21e 7923
2e3febc6
CS
7924 if (Gv_AMG(stash))
7925 SvAMAGIC_on(sv);
7926 else
7927 SvAMAGIC_off(sv);
a0d0e21e 7928
1edbfb88
AB
7929 if(SvSMAGICAL(tmpRef))
7930 if(mg_find(tmpRef, PERL_MAGIC_ext) || mg_find(tmpRef, PERL_MAGIC_uvar))
7931 mg_set(tmpRef);
7932
7933
ecdeb87c 7934
a0d0e21e
LW
7935 return sv;
7936}
7937
645c22ef 7938/* Downgrades a PVGV to a PVMG.
645c22ef
DM
7939 */
7940
76e3520e 7941STATIC void
cea2e8a9 7942S_sv_unglob(pTHX_ SV *sv)
a0d0e21e 7943{
97aff369 7944 dVAR;
850fabdf 7945 void *xpvmg;
b37c2d43 7946 SV * const temp = sv_newmortal();
850fabdf 7947
a0d0e21e
LW
7948 assert(SvTYPE(sv) == SVt_PVGV);
7949 SvFAKE_off(sv);
180488f8
NC
7950 gv_efullname3(temp, (GV *) sv, "*");
7951
f7877b28 7952 if (GvGP(sv)) {
1edc1566 7953 gp_free((GV*)sv);
f7877b28 7954 }
e826b3c7 7955 if (GvSTASH(sv)) {
e15faf7d 7956 sv_del_backref((SV*)GvSTASH(sv), sv);
5c284bb0 7957 GvSTASH(sv) = NULL;
e826b3c7 7958 }
a5f75d66 7959 GvMULTI_off(sv);
acda4c6a
NC
7960 if (GvNAME_HEK(sv)) {
7961 unshare_hek(GvNAME_HEK(sv));
7962 }
dedf8e73 7963 SvSCREAM_off(sv);
850fabdf
GS
7964
7965 /* need to keep SvANY(sv) in the right arena */
7966 xpvmg = new_XPVMG();
7967 StructCopy(SvANY(sv), xpvmg, XPVMG);
7968 del_XPVGV(SvANY(sv));
7969 SvANY(sv) = xpvmg;
7970
a0d0e21e
LW
7971 SvFLAGS(sv) &= ~SVTYPEMASK;
7972 SvFLAGS(sv) |= SVt_PVMG;
180488f8
NC
7973
7974 /* Intentionally not calling any local SET magic, as this isn't so much a
7975 set operation as merely an internal storage change. */
7976 sv_setsv_flags(sv, temp, 0);
a0d0e21e
LW
7977}
7978
954c1994 7979/*
840a7b70 7980=for apidoc sv_unref_flags
954c1994
GS
7981
7982Unsets the RV status of the SV, and decrements the reference count of
7983whatever was being referenced by the RV. This can almost be thought of
840a7b70
IZ
7984as a reversal of C<newSVrv>. The C<cflags> argument can contain
7985C<SV_IMMEDIATE_UNREF> to force the reference count to be decremented
7986(otherwise the decrementing is conditional on the reference count being
7987different from one or the reference being a readonly SV).
7889fe52 7988See C<SvROK_off>.
954c1994
GS
7989
7990=cut
7991*/
7992
ed6116ce 7993void
e15faf7d 7994Perl_sv_unref_flags(pTHX_ SV *ref, U32 flags)
ed6116ce 7995{
b64e5050 7996 SV* const target = SvRV(ref);
810b8aa5 7997
e15faf7d
NC
7998 if (SvWEAKREF(ref)) {
7999 sv_del_backref(target, ref);
8000 SvWEAKREF_off(ref);
8001 SvRV_set(ref, NULL);
810b8aa5
GS
8002 return;
8003 }
e15faf7d
NC
8004 SvRV_set(ref, NULL);
8005 SvROK_off(ref);
8006 /* You can't have a || SvREADONLY(target) here, as $a = $$a, where $a was
04ca4930 8007 assigned to as BEGIN {$a = \"Foo"} will fail. */
e15faf7d
NC
8008 if (SvREFCNT(target) != 1 || (flags & SV_IMMEDIATE_UNREF))
8009 SvREFCNT_dec(target);
840a7b70 8010 else /* XXX Hack, but hard to make $a=$a->[1] work otherwise */
e15faf7d 8011 sv_2mortal(target); /* Schedule for freeing later */
ed6116ce 8012}
8990e307 8013
840a7b70 8014/*
645c22ef
DM
8015=for apidoc sv_untaint
8016
8017Untaint an SV. Use C<SvTAINTED_off> instead.
8018=cut
8019*/
8020
bbce6d69 8021void
864dbfa3 8022Perl_sv_untaint(pTHX_ SV *sv)
bbce6d69 8023{
13f57bf8 8024 if (SvTYPE(sv) >= SVt_PVMG && SvMAGIC(sv)) {
b64e5050 8025 MAGIC * const mg = mg_find(sv, PERL_MAGIC_taint);
36477c24 8026 if (mg)
565764a8 8027 mg->mg_len &= ~1;
36477c24 8028 }
bbce6d69 8029}
8030
645c22ef
DM
8031/*
8032=for apidoc sv_tainted
8033
8034Test an SV for taintedness. Use C<SvTAINTED> instead.
8035=cut
8036*/
8037
bbce6d69 8038bool
864dbfa3 8039Perl_sv_tainted(pTHX_ SV *sv)
bbce6d69 8040{
13f57bf8 8041 if (SvTYPE(sv) >= SVt_PVMG && SvMAGIC(sv)) {
823a54a3 8042 const MAGIC * const mg = mg_find(sv, PERL_MAGIC_taint);
2ddb8a4f 8043 if (mg && (mg->mg_len & 1) )
36477c24 8044 return TRUE;
8045 }
8046 return FALSE;
bbce6d69 8047}
8048
09540bc3
JH
8049/*
8050=for apidoc sv_setpviv
8051
8052Copies an integer into the given SV, also updating its string value.
8053Does not handle 'set' magic. See C<sv_setpviv_mg>.
8054
8055=cut
8056*/
8057
8058void
8059Perl_sv_setpviv(pTHX_ SV *sv, IV iv)
8060{
8061 char buf[TYPE_CHARS(UV)];
8062 char *ebuf;
b64e5050 8063 char * const ptr = uiv_2buf(buf, iv, 0, 0, &ebuf);
09540bc3
JH
8064
8065 sv_setpvn(sv, ptr, ebuf - ptr);
8066}
8067
8068/*
8069=for apidoc sv_setpviv_mg
8070
8071Like C<sv_setpviv>, but also handles 'set' magic.
8072
8073=cut
8074*/
8075
8076void
8077Perl_sv_setpviv_mg(pTHX_ SV *sv, IV iv)
8078{
df7eb254 8079 sv_setpviv(sv, iv);
09540bc3
JH
8080 SvSETMAGIC(sv);
8081}
8082
cea2e8a9 8083#if defined(PERL_IMPLICIT_CONTEXT)
645c22ef
DM
8084
8085/* pTHX_ magic can't cope with varargs, so this is a no-context
8086 * version of the main function, (which may itself be aliased to us).
8087 * Don't access this version directly.
8088 */
8089
cea2e8a9
GS
8090void
8091Perl_sv_setpvf_nocontext(SV *sv, const char* pat, ...)
8092{
8093 dTHX;
8094 va_list args;
8095 va_start(args, pat);
c5be433b 8096 sv_vsetpvf(sv, pat, &args);
cea2e8a9
GS
8097 va_end(args);
8098}
8099
645c22ef
DM
8100/* pTHX_ magic can't cope with varargs, so this is a no-context
8101 * version of the main function, (which may itself be aliased to us).
8102 * Don't access this version directly.
8103 */
cea2e8a9
GS
8104
8105void
8106Perl_sv_setpvf_mg_nocontext(SV *sv, const char* pat, ...)
8107{
8108 dTHX;
8109 va_list args;
8110 va_start(args, pat);
c5be433b 8111 sv_vsetpvf_mg(sv, pat, &args);
cea2e8a9 8112 va_end(args);
cea2e8a9
GS
8113}
8114#endif
8115
954c1994
GS
8116/*
8117=for apidoc sv_setpvf
8118
bffc3d17
SH
8119Works like C<sv_catpvf> but copies the text into the SV instead of
8120appending it. Does not handle 'set' magic. See C<sv_setpvf_mg>.
954c1994
GS
8121
8122=cut
8123*/
8124
46fc3d4c 8125void
864dbfa3 8126Perl_sv_setpvf(pTHX_ SV *sv, const char* pat, ...)
46fc3d4c 8127{
8128 va_list args;
46fc3d4c 8129 va_start(args, pat);
c5be433b 8130 sv_vsetpvf(sv, pat, &args);
46fc3d4c 8131 va_end(args);
8132}
8133
bffc3d17
SH
8134/*
8135=for apidoc sv_vsetpvf
8136
8137Works like C<sv_vcatpvf> but copies the text into the SV instead of
8138appending it. Does not handle 'set' magic. See C<sv_vsetpvf_mg>.
8139
8140Usually used via its frontend C<sv_setpvf>.
8141
8142=cut
8143*/
645c22ef 8144
c5be433b
GS
8145void
8146Perl_sv_vsetpvf(pTHX_ SV *sv, const char* pat, va_list* args)
8147{
4608196e 8148 sv_vsetpvfn(sv, pat, strlen(pat), args, NULL, 0, NULL);
c5be433b 8149}
ef50df4b 8150
954c1994
GS
8151/*
8152=for apidoc sv_setpvf_mg
8153
8154Like C<sv_setpvf>, but also handles 'set' magic.
8155
8156=cut
8157*/
8158
ef50df4b 8159void
864dbfa3 8160Perl_sv_setpvf_mg(pTHX_ SV *sv, const char* pat, ...)
ef50df4b
GS
8161{
8162 va_list args;
ef50df4b 8163 va_start(args, pat);
c5be433b 8164 sv_vsetpvf_mg(sv, pat, &args);
ef50df4b 8165 va_end(args);
c5be433b
GS
8166}
8167
bffc3d17
SH
8168/*
8169=for apidoc sv_vsetpvf_mg
8170
8171Like C<sv_vsetpvf>, but also handles 'set' magic.
8172
8173Usually used via its frontend C<sv_setpvf_mg>.
8174
8175=cut
8176*/
645c22ef 8177
c5be433b
GS
8178void
8179Perl_sv_vsetpvf_mg(pTHX_ SV *sv, const char* pat, va_list* args)
8180{
4608196e 8181 sv_vsetpvfn(sv, pat, strlen(pat), args, NULL, 0, NULL);
ef50df4b
GS
8182 SvSETMAGIC(sv);
8183}
8184
cea2e8a9 8185#if defined(PERL_IMPLICIT_CONTEXT)
645c22ef
DM
8186
8187/* pTHX_ magic can't cope with varargs, so this is a no-context
8188 * version of the main function, (which may itself be aliased to us).
8189 * Don't access this version directly.
8190 */
8191
cea2e8a9
GS
8192void
8193Perl_sv_catpvf_nocontext(SV *sv, const char* pat, ...)
8194{
8195 dTHX;
8196 va_list args;
8197 va_start(args, pat);
c5be433b 8198 sv_vcatpvf(sv, pat, &args);
cea2e8a9
GS
8199 va_end(args);
8200}
8201
645c22ef
DM
8202/* pTHX_ magic can't cope with varargs, so this is a no-context
8203 * version of the main function, (which may itself be aliased to us).
8204 * Don't access this version directly.
8205 */
8206
cea2e8a9
GS
8207void
8208Perl_sv_catpvf_mg_nocontext(SV *sv, const char* pat, ...)
8209{
8210 dTHX;
8211 va_list args;
8212 va_start(args, pat);
c5be433b 8213 sv_vcatpvf_mg(sv, pat, &args);
cea2e8a9 8214 va_end(args);
cea2e8a9
GS
8215}
8216#endif
8217
954c1994
GS
8218/*
8219=for apidoc sv_catpvf
8220
d5ce4a7c
GA
8221Processes its arguments like C<sprintf> and appends the formatted
8222output to an SV. If the appended data contains "wide" characters
8223(including, but not limited to, SVs with a UTF-8 PV formatted with %s,
8224and characters >255 formatted with %c), the original SV might get
bffc3d17 8225upgraded to UTF-8. Handles 'get' magic, but not 'set' magic. See
cdd94ca7
NC
8226C<sv_catpvf_mg>. If the original SV was UTF-8, the pattern should be
8227valid UTF-8; if the original SV was bytes, the pattern should be too.
954c1994 8228
d5ce4a7c 8229=cut */
954c1994 8230
46fc3d4c 8231void
864dbfa3 8232Perl_sv_catpvf(pTHX_ SV *sv, const char* pat, ...)
46fc3d4c 8233{
8234 va_list args;
46fc3d4c 8235 va_start(args, pat);
c5be433b 8236 sv_vcatpvf(sv, pat, &args);
46fc3d4c 8237 va_end(args);
8238}
8239
bffc3d17
SH
8240/*
8241=for apidoc sv_vcatpvf
8242
8243Processes its arguments like C<vsprintf> and appends the formatted output
8244to an SV. Does not handle 'set' magic. See C<sv_vcatpvf_mg>.
8245
8246Usually used via its frontend C<sv_catpvf>.
8247
8248=cut
8249*/
645c22ef 8250
ef50df4b 8251void
c5be433b
GS
8252Perl_sv_vcatpvf(pTHX_ SV *sv, const char* pat, va_list* args)
8253{
4608196e 8254 sv_vcatpvfn(sv, pat, strlen(pat), args, NULL, 0, NULL);
c5be433b
GS
8255}
8256
954c1994
GS
8257/*
8258=for apidoc sv_catpvf_mg
8259
8260Like C<sv_catpvf>, but also handles 'set' magic.
8261
8262=cut
8263*/
8264
c5be433b 8265void
864dbfa3 8266Perl_sv_catpvf_mg(pTHX_ SV *sv, const char* pat, ...)
ef50df4b
GS
8267{
8268 va_list args;
ef50df4b 8269 va_start(args, pat);
c5be433b 8270 sv_vcatpvf_mg(sv, pat, &args);
ef50df4b 8271 va_end(args);
c5be433b
GS
8272}
8273
bffc3d17
SH
8274/*
8275=for apidoc sv_vcatpvf_mg
8276
8277Like C<sv_vcatpvf>, but also handles 'set' magic.
8278
8279Usually used via its frontend C<sv_catpvf_mg>.
8280
8281=cut
8282*/
645c22ef 8283
c5be433b
GS
8284void
8285Perl_sv_vcatpvf_mg(pTHX_ SV *sv, const char* pat, va_list* args)
8286{
4608196e 8287 sv_vcatpvfn(sv, pat, strlen(pat), args, NULL, 0, NULL);
ef50df4b
GS
8288 SvSETMAGIC(sv);
8289}
8290
954c1994
GS
8291/*
8292=for apidoc sv_vsetpvfn
8293
bffc3d17 8294Works like C<sv_vcatpvfn> but copies the text into the SV instead of
954c1994
GS
8295appending it.
8296
bffc3d17 8297Usually used via one of its frontends C<sv_vsetpvf> and C<sv_vsetpvf_mg>.
645c22ef 8298
954c1994
GS
8299=cut
8300*/
8301
46fc3d4c 8302void
7d5ea4e7 8303Perl_sv_vsetpvfn(pTHX_ SV *sv, const char *pat, STRLEN patlen, va_list *args, SV **svargs, I32 svmax, bool *maybe_tainted)
46fc3d4c 8304{
8305 sv_setpvn(sv, "", 0);
7d5ea4e7 8306 sv_vcatpvfn(sv, pat, patlen, args, svargs, svmax, maybe_tainted);
46fc3d4c 8307}
8308
2d00ba3b 8309STATIC I32
9dd79c3f 8310S_expect_number(pTHX_ char** pattern)
211dfcf1 8311{
97aff369 8312 dVAR;
211dfcf1
HS
8313 I32 var = 0;
8314 switch (**pattern) {
8315 case '1': case '2': case '3':
8316 case '4': case '5': case '6':
8317 case '7': case '8': case '9':
2fba7546
GA
8318 var = *(*pattern)++ - '0';
8319 while (isDIGIT(**pattern)) {
5f66b61c 8320 const I32 tmp = var * 10 + (*(*pattern)++ - '0');
2fba7546
GA
8321 if (tmp < var)
8322 Perl_croak(aTHX_ "Integer overflow in format string for %s", (PL_op ? OP_NAME(PL_op) : "sv_vcatpvfn"));
8323 var = tmp;
8324 }
211dfcf1
HS
8325 }
8326 return var;
8327}
211dfcf1 8328
c445ea15
AL
8329STATIC char *
8330S_F0convert(NV nv, char *endbuf, STRLEN *len)
4151a5fe 8331{
a3b680e6 8332 const int neg = nv < 0;
4151a5fe 8333 UV uv;
4151a5fe
IZ
8334
8335 if (neg)
8336 nv = -nv;
8337 if (nv < UV_MAX) {
b464bac0 8338 char *p = endbuf;
4151a5fe 8339 nv += 0.5;
028f8eaa 8340 uv = (UV)nv;
4151a5fe
IZ
8341 if (uv & 1 && uv == nv)
8342 uv--; /* Round to even */
8343 do {
a3b680e6 8344 const unsigned dig = uv % 10;
4151a5fe
IZ
8345 *--p = '0' + dig;
8346 } while (uv /= 10);
8347 if (neg)
8348 *--p = '-';
8349 *len = endbuf - p;
8350 return p;
8351 }
bd61b366 8352 return NULL;
4151a5fe
IZ
8353}
8354
8355
954c1994
GS
8356/*
8357=for apidoc sv_vcatpvfn
8358
8359Processes its arguments like C<vsprintf> and appends the formatted output
8360to an SV. Uses an array of SVs if the C style variable argument list is
8361missing (NULL). When running with taint checks enabled, indicates via
8362C<maybe_tainted> if results are untrustworthy (often due to the use of
8363locales).
8364
bffc3d17 8365Usually used via one of its frontends C<sv_vcatpvf> and C<sv_vcatpvf_mg>.
645c22ef 8366
954c1994
GS
8367=cut
8368*/
8369
8896765a
RB
8370
8371#define VECTORIZE_ARGS vecsv = va_arg(*args, SV*);\
8372 vecstr = (U8*)SvPV_const(vecsv,veclen);\
8373 vec_utf8 = DO_UTF8(vecsv);
8374
1ef29b0e
RGS
8375/* XXX maybe_tainted is never assigned to, so the doc above is lying. */
8376
46fc3d4c 8377void
7d5ea4e7 8378Perl_sv_vcatpvfn(pTHX_ SV *sv, const char *pat, STRLEN patlen, va_list *args, SV **svargs, I32 svmax, bool *maybe_tainted)
46fc3d4c 8379{
97aff369 8380 dVAR;
46fc3d4c 8381 char *p;
8382 char *q;
a3b680e6 8383 const char *patend;
fc36a67e 8384 STRLEN origlen;
46fc3d4c 8385 I32 svix = 0;
27da23d5 8386 static const char nullstr[] = "(null)";
a0714e2c 8387 SV *argsv = NULL;
b464bac0
AL
8388 bool has_utf8 = DO_UTF8(sv); /* has the result utf8? */
8389 const bool pat_utf8 = has_utf8; /* the pattern is in utf8? */
a0714e2c 8390 SV *nsv = NULL;
4151a5fe
IZ
8391 /* Times 4: a decimal digit takes more than 3 binary digits.
8392 * NV_DIG: mantissa takes than many decimal digits.
8393 * Plus 32: Playing safe. */
8394 char ebuf[IV_DIG * 4 + NV_DIG + 32];
8395 /* large enough for "%#.#f" --chip */
8396 /* what about long double NVs? --jhi */
db79b45b 8397
53c1dcc0
AL
8398 PERL_UNUSED_ARG(maybe_tainted);
8399
46fc3d4c 8400 /* no matter what, this is a string now */
fc36a67e 8401 (void)SvPV_force(sv, origlen);
46fc3d4c 8402
8896765a 8403 /* special-case "", "%s", and "%-p" (SVf - see below) */
46fc3d4c 8404 if (patlen == 0)
8405 return;
0dbb1585 8406 if (patlen == 2 && pat[0] == '%' && pat[1] == 's') {
2d03de9c
AL
8407 if (args) {
8408 const char * const s = va_arg(*args, char*);
8409 sv_catpv(sv, s ? s : nullstr);
8410 }
8411 else if (svix < svmax) {
8412 sv_catsv(sv, *svargs);
2d03de9c
AL
8413 }
8414 return;
0dbb1585 8415 }
8896765a
RB
8416 if (args && patlen == 3 && pat[0] == '%' &&
8417 pat[1] == '-' && pat[2] == 'p') {
8418 argsv = va_arg(*args, SV*);
8419 sv_catsv(sv, argsv);
8896765a 8420 return;
46fc3d4c 8421 }
8422
1d917b39 8423#ifndef USE_LONG_DOUBLE
4151a5fe 8424 /* special-case "%.<number>[gf]" */
7af36d83 8425 if ( !args && patlen <= 5 && pat[0] == '%' && pat[1] == '.'
4151a5fe
IZ
8426 && (pat[patlen-1] == 'g' || pat[patlen-1] == 'f') ) {
8427 unsigned digits = 0;
8428 const char *pp;
8429
8430 pp = pat + 2;
8431 while (*pp >= '0' && *pp <= '9')
8432 digits = 10 * digits + (*pp++ - '0');
028f8eaa 8433 if (pp - pat == (int)patlen - 1) {
4151a5fe
IZ
8434 NV nv;
8435
7af36d83 8436 if (svix < svmax)
4151a5fe
IZ
8437 nv = SvNV(*svargs);
8438 else
8439 return;
8440 if (*pp == 'g') {
2873255c
NC
8441 /* Add check for digits != 0 because it seems that some
8442 gconverts are buggy in this case, and we don't yet have
8443 a Configure test for this. */
8444 if (digits && digits < sizeof(ebuf) - NV_DIG - 10) {
8445 /* 0, point, slack */
2e59c212 8446 Gconvert(nv, (int)digits, 0, ebuf);
4151a5fe
IZ
8447 sv_catpv(sv, ebuf);
8448 if (*ebuf) /* May return an empty string for digits==0 */
8449 return;
8450 }
8451 } else if (!digits) {
8452 STRLEN l;
8453
8454 if ((p = F0convert(nv, ebuf + sizeof ebuf, &l))) {
8455 sv_catpvn(sv, p, l);
8456 return;
8457 }
8458 }
8459 }
8460 }
1d917b39 8461#endif /* !USE_LONG_DOUBLE */
4151a5fe 8462
2cf2cfc6 8463 if (!args && svix < svmax && DO_UTF8(*svargs))
205f51d8 8464 has_utf8 = TRUE;
2cf2cfc6 8465
46fc3d4c 8466 patend = (char*)pat + patlen;
8467 for (p = (char*)pat; p < patend; p = q) {
8468 bool alt = FALSE;
8469 bool left = FALSE;
b22c7a20 8470 bool vectorize = FALSE;
211dfcf1 8471 bool vectorarg = FALSE;
2cf2cfc6 8472 bool vec_utf8 = FALSE;
46fc3d4c 8473 char fill = ' ';
8474 char plus = 0;
8475 char intsize = 0;
8476 STRLEN width = 0;
fc36a67e 8477 STRLEN zeros = 0;
46fc3d4c 8478 bool has_precis = FALSE;
8479 STRLEN precis = 0;
c445ea15 8480 const I32 osvix = svix;
2cf2cfc6 8481 bool is_utf8 = FALSE; /* is this item utf8? */
20f6aaab
AS
8482#ifdef HAS_LDBL_SPRINTF_BUG
8483 /* This is to try to fix a bug with irix/nonstop-ux/powerux and
205f51d8 8484 with sfio - Allen <allens@cpan.org> */
20f6aaab
AS
8485 bool fix_ldbl_sprintf_bug = FALSE;
8486#endif
205f51d8 8487
46fc3d4c 8488 char esignbuf[4];
89ebb4a3 8489 U8 utf8buf[UTF8_MAXBYTES+1];
46fc3d4c 8490 STRLEN esignlen = 0;
8491
bd61b366 8492 const char *eptr = NULL;
fc36a67e 8493 STRLEN elen = 0;
a0714e2c 8494 SV *vecsv = NULL;
4608196e 8495 const U8 *vecstr = NULL;
b22c7a20 8496 STRLEN veclen = 0;
934abaf1 8497 char c = 0;
46fc3d4c 8498 int i;
9c5ffd7c 8499 unsigned base = 0;
8c8eb53c
RB
8500 IV iv = 0;
8501 UV uv = 0;
9e5b023a
JH
8502 /* we need a long double target in case HAS_LONG_DOUBLE but
8503 not USE_LONG_DOUBLE
8504 */
35fff930 8505#if defined(HAS_LONG_DOUBLE) && LONG_DOUBLESIZE > DOUBLESIZE
9e5b023a
JH
8506 long double nv;
8507#else
65202027 8508 NV nv;
9e5b023a 8509#endif
46fc3d4c 8510 STRLEN have;
8511 STRLEN need;
8512 STRLEN gap;
7af36d83 8513 const char *dotstr = ".";
b22c7a20 8514 STRLEN dotstrlen = 1;
211dfcf1 8515 I32 efix = 0; /* explicit format parameter index */
eb3fce90 8516 I32 ewix = 0; /* explicit width index */
211dfcf1
HS
8517 I32 epix = 0; /* explicit precision index */
8518 I32 evix = 0; /* explicit vector index */
eb3fce90 8519 bool asterisk = FALSE;
46fc3d4c 8520
211dfcf1 8521 /* echo everything up to the next format specification */
46fc3d4c 8522 for (q = p; q < patend && *q != '%'; ++q) ;
8523 if (q > p) {
db79b45b
JH
8524 if (has_utf8 && !pat_utf8)
8525 sv_catpvn_utf8_upgrade(sv, p, q - p, nsv);
8526 else
8527 sv_catpvn(sv, p, q - p);
46fc3d4c 8528 p = q;
8529 }
8530 if (q++ >= patend)
8531 break;
8532
211dfcf1
HS
8533/*
8534 We allow format specification elements in this order:
8535 \d+\$ explicit format parameter index
8536 [-+ 0#]+ flags
a472f209 8537 v|\*(\d+\$)?v vector with optional (optionally specified) arg
f3583277 8538 0 flag (as above): repeated to allow "v02"
211dfcf1
HS
8539 \d+|\*(\d+\$)? width using optional (optionally specified) arg
8540 \.(\d*|\*(\d+\$)?) precision using optional (optionally specified) arg
8541 [hlqLV] size
8896765a
RB
8542 [%bcdefginopsuxDFOUX] format (mandatory)
8543*/
8544
8545 if (args) {
8546/*
8547 As of perl5.9.3, printf format checking is on by default.
8548 Internally, perl uses %p formats to provide an escape to
8549 some extended formatting. This block deals with those
8550 extensions: if it does not match, (char*)q is reset and
8551 the normal format processing code is used.
8552
8553 Currently defined extensions are:
8554 %p include pointer address (standard)
8555 %-p (SVf) include an SV (previously %_)
8556 %-<num>p include an SV with precision <num>
8557 %1p (VDf) include a v-string (as %vd)
8558 %<num>p reserved for future extensions
8559
8560 Robin Barker 2005-07-14
211dfcf1 8561*/
8896765a
RB
8562 char* r = q;
8563 bool sv = FALSE;
8564 STRLEN n = 0;
8565 if (*q == '-')
8566 sv = *q++;
c445ea15 8567 n = expect_number(&q);
8896765a
RB
8568 if (*q++ == 'p') {
8569 if (sv) { /* SVf */
8570 if (n) {
8571 precis = n;
8572 has_precis = TRUE;
8573 }
8574 argsv = va_arg(*args, SV*);
8575 eptr = SvPVx_const(argsv, elen);
8576 if (DO_UTF8(argsv))
8577 is_utf8 = TRUE;
8578 goto string;
8579 }
8580#if vdNUMBER
8581 else if (n == vdNUMBER) { /* VDf */
8582 vectorize = TRUE;
8583 VECTORIZE_ARGS
8584 goto format_vd;
8585 }
8586#endif
8587 else if (n) {
8588 if (ckWARN_d(WARN_INTERNAL))
8589 Perl_warner(aTHX_ packWARN(WARN_INTERNAL),
8590 "internal %%<num>p might conflict with future printf extensions");
8591 }
8592 }
8593 q = r;
8594 }
8595
c445ea15 8596 if ( (width = expect_number(&q)) ) {
211dfcf1
HS
8597 if (*q == '$') {
8598 ++q;
8599 efix = width;
8600 } else {
8601 goto gotwidth;
8602 }
8603 }
8604
fc36a67e 8605 /* FLAGS */
8606
46fc3d4c 8607 while (*q) {
8608 switch (*q) {
8609 case ' ':
8610 case '+':
8611 plus = *q++;
8612 continue;
8613
8614 case '-':
8615 left = TRUE;
8616 q++;
8617 continue;
8618
8619 case '0':
8620 fill = *q++;
8621 continue;
8622
8623 case '#':
8624 alt = TRUE;
8625 q++;
8626 continue;
8627
fc36a67e 8628 default:
8629 break;
8630 }
8631 break;
8632 }
46fc3d4c 8633
211dfcf1 8634 tryasterisk:
eb3fce90 8635 if (*q == '*') {
211dfcf1 8636 q++;
c445ea15 8637 if ( (ewix = expect_number(&q)) )
211dfcf1
HS
8638 if (*q++ != '$')
8639 goto unknown;
eb3fce90 8640 asterisk = TRUE;
211dfcf1
HS
8641 }
8642 if (*q == 'v') {
eb3fce90 8643 q++;
211dfcf1
HS
8644 if (vectorize)
8645 goto unknown;
9cbac4c7 8646 if ((vectorarg = asterisk)) {
211dfcf1
HS
8647 evix = ewix;
8648 ewix = 0;
8649 asterisk = FALSE;
8650 }
8651 vectorize = TRUE;
8652 goto tryasterisk;
eb3fce90
JH
8653 }
8654
211dfcf1 8655 if (!asterisk)
858a90f9 8656 {
7a5fa8a2 8657 if( *q == '0' )
f3583277 8658 fill = *q++;
c445ea15 8659 width = expect_number(&q);
858a90f9 8660 }
211dfcf1
HS
8661
8662 if (vectorize) {
8663 if (vectorarg) {
8664 if (args)
8665 vecsv = va_arg(*args, SV*);
7ad96abb
NC
8666 else if (evix) {
8667 vecsv = (evix > 0 && evix <= svmax)
8668 ? svargs[evix-1] : &PL_sv_undef;
8669 } else {
8670 vecsv = svix < svmax ? svargs[svix++] : &PL_sv_undef;
8671 }
245d4a47 8672 dotstr = SvPV_const(vecsv, dotstrlen);
640283f5
NC
8673 /* Keep the DO_UTF8 test *after* the SvPV call, else things go
8674 bad with tied or overloaded values that return UTF8. */
211dfcf1 8675 if (DO_UTF8(vecsv))
2cf2cfc6 8676 is_utf8 = TRUE;
640283f5
NC
8677 else if (has_utf8) {
8678 vecsv = sv_mortalcopy(vecsv);
8679 sv_utf8_upgrade(vecsv);
8680 dotstr = SvPV_const(vecsv, dotstrlen);
8681 is_utf8 = TRUE;
8682 }
211dfcf1
HS
8683 }
8684 if (args) {
8896765a 8685 VECTORIZE_ARGS
eb3fce90 8686 }
7ad96abb 8687 else if (efix ? (efix > 0 && efix <= svmax) : svix < svmax) {
211dfcf1 8688 vecsv = svargs[efix ? efix-1 : svix++];
245d4a47 8689 vecstr = (U8*)SvPV_const(vecsv,veclen);
2cf2cfc6 8690 vec_utf8 = DO_UTF8(vecsv);
96b8f7ce
JP
8691
8692 /* if this is a version object, we need to convert
8693 * back into v-string notation and then let the
8694 * vectorize happen normally
d7aa5382 8695 */
96b8f7ce
JP
8696 if (sv_derived_from(vecsv, "version")) {
8697 char *version = savesvpv(vecsv);
34ba6322
SP
8698 if ( hv_exists((HV*)SvRV(vecsv), "alpha", 5 ) ) {
8699 Perl_warner(aTHX_ packWARN(WARN_INTERNAL),
8700 "vector argument not supported with alpha versions");
8701 goto unknown;
8702 }
96b8f7ce
JP
8703 vecsv = sv_newmortal();
8704 /* scan_vstring is expected to be called during
8705 * tokenization, so we need to fake up the end
8706 * of the buffer for it
8707 */
8708 PL_bufend = version + veclen;
8709 scan_vstring(version, vecsv);
8710 vecstr = (U8*)SvPV_const(vecsv, veclen);
8711 vec_utf8 = DO_UTF8(vecsv);
8712 Safefree(version);
d7aa5382 8713 }
211dfcf1
HS
8714 }
8715 else {
8716 vecstr = (U8*)"";
8717 veclen = 0;
8718 }
eb3fce90 8719 }
fc36a67e 8720
eb3fce90 8721 if (asterisk) {
fc36a67e 8722 if (args)
8723 i = va_arg(*args, int);
8724 else
eb3fce90
JH
8725 i = (ewix ? ewix <= svmax : svix < svmax) ?
8726 SvIVx(svargs[ewix ? ewix-1 : svix++]) : 0;
fc36a67e 8727 left |= (i < 0);
8728 width = (i < 0) ? -i : i;
fc36a67e 8729 }
211dfcf1 8730 gotwidth:
fc36a67e 8731
8732 /* PRECISION */
46fc3d4c 8733
fc36a67e 8734 if (*q == '.') {
8735 q++;
8736 if (*q == '*') {
211dfcf1 8737 q++;
c445ea15 8738 if ( ((epix = expect_number(&q))) && (*q++ != '$') )
7b8dd722
HS
8739 goto unknown;
8740 /* XXX: todo, support specified precision parameter */
8741 if (epix)
211dfcf1 8742 goto unknown;
46fc3d4c 8743 if (args)
8744 i = va_arg(*args, int);
8745 else
eb3fce90
JH
8746 i = (ewix ? ewix <= svmax : svix < svmax)
8747 ? SvIVx(svargs[ewix ? ewix-1 : svix++]) : 0;
fc36a67e 8748 precis = (i < 0) ? 0 : i;
fc36a67e 8749 }
8750 else {
8751 precis = 0;
8752 while (isDIGIT(*q))
8753 precis = precis * 10 + (*q++ - '0');
8754 }
8755 has_precis = TRUE;
8756 }
46fc3d4c 8757
fc36a67e 8758 /* SIZE */
46fc3d4c 8759
fc36a67e 8760 switch (*q) {
c623ac67
GS
8761#ifdef WIN32
8762 case 'I': /* Ix, I32x, and I64x */
8763# ifdef WIN64
8764 if (q[1] == '6' && q[2] == '4') {
8765 q += 3;
8766 intsize = 'q';
8767 break;
8768 }
8769# endif
8770 if (q[1] == '3' && q[2] == '2') {
8771 q += 3;
8772 break;
8773 }
8774# ifdef WIN64
8775 intsize = 'q';
8776# endif
8777 q++;
8778 break;
8779#endif
9e5b023a 8780#if defined(HAS_QUAD) || defined(HAS_LONG_DOUBLE)
6f9bb7fd 8781 case 'L': /* Ld */
5f66b61c 8782 /*FALLTHROUGH*/
e5c81feb 8783#ifdef HAS_QUAD
6f9bb7fd 8784 case 'q': /* qd */
9e5b023a 8785#endif
6f9bb7fd
GS
8786 intsize = 'q';
8787 q++;
8788 break;
8789#endif
fc36a67e 8790 case 'l':
9e5b023a 8791#if defined(HAS_QUAD) || defined(HAS_LONG_DOUBLE)
205f51d8 8792 if (*(q + 1) == 'l') { /* lld, llf */
fc36a67e 8793 intsize = 'q';
8794 q += 2;
46fc3d4c 8795 break;
cf2093f6 8796 }
fc36a67e 8797#endif
5f66b61c 8798 /*FALLTHROUGH*/
fc36a67e 8799 case 'h':
5f66b61c 8800 /*FALLTHROUGH*/
fc36a67e 8801 case 'V':
8802 intsize = *q++;
46fc3d4c 8803 break;
8804 }
8805
fc36a67e 8806 /* CONVERSION */
8807
211dfcf1
HS
8808 if (*q == '%') {
8809 eptr = q++;
8810 elen = 1;
26372e71
GA
8811 if (vectorize) {
8812 c = '%';
8813 goto unknown;
8814 }
211dfcf1
HS
8815 goto string;
8816 }
8817
26372e71 8818 if (!vectorize && !args) {
86c51f8b
NC
8819 if (efix) {
8820 const I32 i = efix-1;
8821 argsv = (i >= 0 && i < svmax) ? svargs[i] : &PL_sv_undef;
8822 } else {
8823 argsv = (svix >= 0 && svix < svmax)
8824 ? svargs[svix++] : &PL_sv_undef;
8825 }
863811b2 8826 }
211dfcf1 8827
46fc3d4c 8828 switch (c = *q++) {
8829
8830 /* STRINGS */
8831
46fc3d4c 8832 case 'c':
26372e71
GA
8833 if (vectorize)
8834 goto unknown;
8835 uv = (args) ? va_arg(*args, int) : SvIVx(argsv);
1bd104fb
JH
8836 if ((uv > 255 ||
8837 (!UNI_IS_INVARIANT(uv) && SvUTF8(sv)))
0064a8a9 8838 && !IN_BYTES) {
dfe13c55 8839 eptr = (char*)utf8buf;
9041c2e3 8840 elen = uvchr_to_utf8((U8*)eptr, uv) - utf8buf;
2cf2cfc6 8841 is_utf8 = TRUE;
7e2040f0
GS
8842 }
8843 else {
8844 c = (char)uv;
8845 eptr = &c;
8846 elen = 1;
a0ed51b3 8847 }
46fc3d4c 8848 goto string;
8849
46fc3d4c 8850 case 's':
26372e71
GA
8851 if (vectorize)
8852 goto unknown;
8853 if (args) {
fc36a67e 8854 eptr = va_arg(*args, char*);
c635e13b 8855 if (eptr)
1d7c1841
GS
8856#ifdef MACOS_TRADITIONAL
8857 /* On MacOS, %#s format is used for Pascal strings */
8858 if (alt)
8859 elen = *eptr++;
8860 else
8861#endif
c635e13b 8862 elen = strlen(eptr);
8863 else {
27da23d5 8864 eptr = (char *)nullstr;
c635e13b 8865 elen = sizeof nullstr - 1;
8866 }
46fc3d4c 8867 }
211dfcf1 8868 else {
4d84ee25 8869 eptr = SvPVx_const(argsv, elen);
7e2040f0 8870 if (DO_UTF8(argsv)) {
a0ed51b3
LW
8871 if (has_precis && precis < elen) {
8872 I32 p = precis;
7e2040f0 8873 sv_pos_u2b(argsv, &p, 0); /* sticks at end */
a0ed51b3
LW
8874 precis = p;
8875 }
8876 if (width) { /* fudge width (can't fudge elen) */
7e2040f0 8877 width += elen - sv_len_utf8(argsv);
a0ed51b3 8878 }
2cf2cfc6 8879 is_utf8 = TRUE;
a0ed51b3
LW
8880 }
8881 }
fc36a67e 8882
46fc3d4c 8883 string:
8884 if (has_precis && elen > precis)
8885 elen = precis;
8886 break;
8887
8888 /* INTEGERS */
8889
fc36a67e 8890 case 'p':
be75b157 8891 if (alt || vectorize)
c2e66d9e 8892 goto unknown;
211dfcf1 8893 uv = PTR2UV(args ? va_arg(*args, void*) : argsv);
fc36a67e 8894 base = 16;
8895 goto integer;
8896
46fc3d4c 8897 case 'D':
29fe7a80 8898#ifdef IV_IS_QUAD
22f3ae8c 8899 intsize = 'q';
29fe7a80 8900#else
46fc3d4c 8901 intsize = 'l';
29fe7a80 8902#endif
5f66b61c 8903 /*FALLTHROUGH*/
46fc3d4c 8904 case 'd':
8905 case 'i':
8896765a
RB
8906#if vdNUMBER
8907 format_vd:
8908#endif
b22c7a20 8909 if (vectorize) {
ba210ebe 8910 STRLEN ulen;
211dfcf1
HS
8911 if (!veclen)
8912 continue;
2cf2cfc6
A
8913 if (vec_utf8)
8914 uv = utf8n_to_uvchr(vecstr, veclen, &ulen,
8915 UTF8_ALLOW_ANYUV);
b22c7a20 8916 else {
e83d50c9 8917 uv = *vecstr;
b22c7a20
GS
8918 ulen = 1;
8919 }
8920 vecstr += ulen;
8921 veclen -= ulen;
e83d50c9
JP
8922 if (plus)
8923 esignbuf[esignlen++] = plus;
b22c7a20
GS
8924 }
8925 else if (args) {
46fc3d4c 8926 switch (intsize) {
8927 case 'h': iv = (short)va_arg(*args, int); break;
46fc3d4c 8928 case 'l': iv = va_arg(*args, long); break;
fc36a67e 8929 case 'V': iv = va_arg(*args, IV); break;
b10c0dba 8930 default: iv = va_arg(*args, int); break;
cf2093f6
JH
8931#ifdef HAS_QUAD
8932 case 'q': iv = va_arg(*args, Quad_t); break;
8933#endif
46fc3d4c 8934 }
8935 }
8936 else {
b10c0dba 8937 IV tiv = SvIVx(argsv); /* work around GCC bug #13488 */
46fc3d4c 8938 switch (intsize) {
b10c0dba
MHM
8939 case 'h': iv = (short)tiv; break;
8940 case 'l': iv = (long)tiv; break;
8941 case 'V':
8942 default: iv = tiv; break;
cf2093f6 8943#ifdef HAS_QUAD
b10c0dba 8944 case 'q': iv = (Quad_t)tiv; break;
cf2093f6 8945#endif
46fc3d4c 8946 }
8947 }
e83d50c9
JP
8948 if ( !vectorize ) /* we already set uv above */
8949 {
8950 if (iv >= 0) {
8951 uv = iv;
8952 if (plus)
8953 esignbuf[esignlen++] = plus;
8954 }
8955 else {
8956 uv = -iv;
8957 esignbuf[esignlen++] = '-';
8958 }
46fc3d4c 8959 }
8960 base = 10;
8961 goto integer;
8962
fc36a67e 8963 case 'U':
29fe7a80 8964#ifdef IV_IS_QUAD
22f3ae8c 8965 intsize = 'q';
29fe7a80 8966#else
fc36a67e 8967 intsize = 'l';
29fe7a80 8968#endif
5f66b61c 8969 /*FALLTHROUGH*/
fc36a67e 8970 case 'u':
8971 base = 10;
8972 goto uns_integer;
8973
4f19785b
WSI
8974 case 'b':
8975 base = 2;
8976 goto uns_integer;
8977
46fc3d4c 8978 case 'O':
29fe7a80 8979#ifdef IV_IS_QUAD
22f3ae8c 8980 intsize = 'q';
29fe7a80 8981#else
46fc3d4c 8982 intsize = 'l';
29fe7a80 8983#endif
5f66b61c 8984 /*FALLTHROUGH*/
46fc3d4c 8985 case 'o':
8986 base = 8;
8987 goto uns_integer;
8988
8989 case 'X':
46fc3d4c 8990 case 'x':
8991 base = 16;
46fc3d4c 8992
8993 uns_integer:
b22c7a20 8994 if (vectorize) {
ba210ebe 8995 STRLEN ulen;
b22c7a20 8996 vector:
211dfcf1
HS
8997 if (!veclen)
8998 continue;
2cf2cfc6
A
8999 if (vec_utf8)
9000 uv = utf8n_to_uvchr(vecstr, veclen, &ulen,
9001 UTF8_ALLOW_ANYUV);
b22c7a20 9002 else {
a05b299f 9003 uv = *vecstr;
b22c7a20
GS
9004 ulen = 1;
9005 }
9006 vecstr += ulen;
9007 veclen -= ulen;
9008 }
9009 else if (args) {
46fc3d4c 9010 switch (intsize) {
9011 case 'h': uv = (unsigned short)va_arg(*args, unsigned); break;
46fc3d4c 9012 case 'l': uv = va_arg(*args, unsigned long); break;
fc36a67e 9013 case 'V': uv = va_arg(*args, UV); break;
b10c0dba 9014 default: uv = va_arg(*args, unsigned); break;
cf2093f6 9015#ifdef HAS_QUAD
9e3321a5 9016 case 'q': uv = va_arg(*args, Uquad_t); break;
cf2093f6 9017#endif
46fc3d4c 9018 }
9019 }
9020 else {
b10c0dba 9021 UV tuv = SvUVx(argsv); /* work around GCC bug #13488 */
46fc3d4c 9022 switch (intsize) {
b10c0dba
MHM
9023 case 'h': uv = (unsigned short)tuv; break;
9024 case 'l': uv = (unsigned long)tuv; break;
9025 case 'V':
9026 default: uv = tuv; break;
cf2093f6 9027#ifdef HAS_QUAD
b10c0dba 9028 case 'q': uv = (Uquad_t)tuv; break;
cf2093f6 9029#endif
46fc3d4c 9030 }
9031 }
9032
9033 integer:
4d84ee25
NC
9034 {
9035 char *ptr = ebuf + sizeof ebuf;
1387f30c
DD
9036 bool tempalt = uv ? alt : FALSE; /* Vectors can't change alt */
9037 zeros = 0;
9038
4d84ee25
NC
9039 switch (base) {
9040 unsigned dig;
9041 case 16:
4d84ee25
NC
9042 p = (char*)((c == 'X')
9043 ? "0123456789ABCDEF" : "0123456789abcdef");
9044 do {
9045 dig = uv & 15;
9046 *--ptr = p[dig];
9047 } while (uv >>= 4);
1387f30c 9048 if (tempalt) {
4d84ee25
NC
9049 esignbuf[esignlen++] = '0';
9050 esignbuf[esignlen++] = c; /* 'x' or 'X' */
9051 }
9052 break;
9053 case 8:
9054 do {
9055 dig = uv & 7;
9056 *--ptr = '0' + dig;
9057 } while (uv >>= 3);
9058 if (alt && *ptr != '0')
9059 *--ptr = '0';
9060 break;
9061 case 2:
9062 do {
9063 dig = uv & 1;
9064 *--ptr = '0' + dig;
9065 } while (uv >>= 1);
1387f30c 9066 if (tempalt) {
4d84ee25
NC
9067 esignbuf[esignlen++] = '0';
9068 esignbuf[esignlen++] = 'b';
9069 }
9070 break;
9071 default: /* it had better be ten or less */
9072 do {
9073 dig = uv % base;
9074 *--ptr = '0' + dig;
9075 } while (uv /= base);
9076 break;
46fc3d4c 9077 }
4d84ee25
NC
9078 elen = (ebuf + sizeof ebuf) - ptr;
9079 eptr = ptr;
9080 if (has_precis) {
9081 if (precis > elen)
9082 zeros = precis - elen;
9083 else if (precis == 0 && elen == 1 && *eptr == '0')
9084 elen = 0;
eda88b6d 9085 }
c10ed8b9 9086 }
46fc3d4c 9087 break;
9088
9089 /* FLOATING POINT */
9090
fc36a67e 9091 case 'F':
9092 c = 'f'; /* maybe %F isn't supported here */
5f66b61c 9093 /*FALLTHROUGH*/
46fc3d4c 9094 case 'e': case 'E':
fc36a67e 9095 case 'f':
46fc3d4c 9096 case 'g': case 'G':
26372e71
GA
9097 if (vectorize)
9098 goto unknown;
46fc3d4c 9099
9100 /* This is evil, but floating point is even more evil */
9101
9e5b023a
JH
9102 /* for SV-style calling, we can only get NV
9103 for C-style calling, we assume %f is double;
9104 for simplicity we allow any of %Lf, %llf, %qf for long double
9105 */
9106 switch (intsize) {
9107 case 'V':
9108#if defined(USE_LONG_DOUBLE)
9109 intsize = 'q';
9110#endif
9111 break;
8a2e3f14 9112/* [perl #20339] - we should accept and ignore %lf rather than die */
00e17364 9113 case 'l':
5f66b61c 9114 /*FALLTHROUGH*/
9e5b023a
JH
9115 default:
9116#if defined(USE_LONG_DOUBLE)
9117 intsize = args ? 0 : 'q';
9118#endif
9119 break;
9120 case 'q':
9121#if defined(HAS_LONG_DOUBLE)
9122 break;
9123#else
5f66b61c 9124 /*FALLTHROUGH*/
9e5b023a
JH
9125#endif
9126 case 'h':
9e5b023a
JH
9127 goto unknown;
9128 }
9129
9130 /* now we need (long double) if intsize == 'q', else (double) */
26372e71 9131 nv = (args) ?
35fff930
JH
9132#if LONG_DOUBLESIZE > DOUBLESIZE
9133 intsize == 'q' ?
205f51d8
AS
9134 va_arg(*args, long double) :
9135 va_arg(*args, double)
35fff930 9136#else
205f51d8 9137 va_arg(*args, double)
35fff930 9138#endif
9e5b023a 9139 : SvNVx(argsv);
fc36a67e 9140
9141 need = 0;
9142 if (c != 'e' && c != 'E') {
9143 i = PERL_INT_MIN;
9e5b023a
JH
9144 /* FIXME: if HAS_LONG_DOUBLE but not USE_LONG_DOUBLE this
9145 will cast our (long double) to (double) */
73b309ea 9146 (void)Perl_frexp(nv, &i);
fc36a67e 9147 if (i == PERL_INT_MIN)
cea2e8a9 9148 Perl_die(aTHX_ "panic: frexp");
c635e13b 9149 if (i > 0)
fc36a67e 9150 need = BIT_DIGITS(i);
9151 }
9152 need += has_precis ? precis : 6; /* known default */
20f6aaab 9153
fc36a67e 9154 if (need < width)
9155 need = width;
9156
20f6aaab
AS
9157#ifdef HAS_LDBL_SPRINTF_BUG
9158 /* This is to try to fix a bug with irix/nonstop-ux/powerux and
205f51d8
AS
9159 with sfio - Allen <allens@cpan.org> */
9160
9161# ifdef DBL_MAX
9162# define MY_DBL_MAX DBL_MAX
9163# else /* XXX guessing! HUGE_VAL may be defined as infinity, so not using */
9164# if DOUBLESIZE >= 8
9165# define MY_DBL_MAX 1.7976931348623157E+308L
9166# else
9167# define MY_DBL_MAX 3.40282347E+38L
9168# endif
9169# endif
9170
9171# ifdef HAS_LDBL_SPRINTF_BUG_LESS1 /* only between -1L & 1L - Allen */
9172# define MY_DBL_MAX_BUG 1L
20f6aaab 9173# else
205f51d8 9174# define MY_DBL_MAX_BUG MY_DBL_MAX
20f6aaab 9175# endif
20f6aaab 9176
205f51d8
AS
9177# ifdef DBL_MIN
9178# define MY_DBL_MIN DBL_MIN
9179# else /* XXX guessing! -Allen */
9180# if DOUBLESIZE >= 8
9181# define MY_DBL_MIN 2.2250738585072014E-308L
9182# else
9183# define MY_DBL_MIN 1.17549435E-38L
9184# endif
9185# endif
20f6aaab 9186
205f51d8
AS
9187 if ((intsize == 'q') && (c == 'f') &&
9188 ((nv < MY_DBL_MAX_BUG) && (nv > -MY_DBL_MAX_BUG)) &&
9189 (need < DBL_DIG)) {
9190 /* it's going to be short enough that
9191 * long double precision is not needed */
9192
9193 if ((nv <= 0L) && (nv >= -0L))
9194 fix_ldbl_sprintf_bug = TRUE; /* 0 is 0 - easiest */
9195 else {
9196 /* would use Perl_fp_class as a double-check but not
9197 * functional on IRIX - see perl.h comments */
9198
9199 if ((nv >= MY_DBL_MIN) || (nv <= -MY_DBL_MIN)) {
9200 /* It's within the range that a double can represent */
9201#if defined(DBL_MAX) && !defined(DBL_MIN)
9202 if ((nv >= ((long double)1/DBL_MAX)) ||
9203 (nv <= (-(long double)1/DBL_MAX)))
20f6aaab 9204#endif
205f51d8 9205 fix_ldbl_sprintf_bug = TRUE;
20f6aaab 9206 }
205f51d8
AS
9207 }
9208 if (fix_ldbl_sprintf_bug == TRUE) {
9209 double temp;
9210
9211 intsize = 0;
9212 temp = (double)nv;
9213 nv = (NV)temp;
9214 }
20f6aaab 9215 }
205f51d8
AS
9216
9217# undef MY_DBL_MAX
9218# undef MY_DBL_MAX_BUG
9219# undef MY_DBL_MIN
9220
20f6aaab
AS
9221#endif /* HAS_LDBL_SPRINTF_BUG */
9222
46fc3d4c 9223 need += 20; /* fudge factor */
80252599
GS
9224 if (PL_efloatsize < need) {
9225 Safefree(PL_efloatbuf);
9226 PL_efloatsize = need + 20; /* more fudge */
a02a5408 9227 Newx(PL_efloatbuf, PL_efloatsize, char);
7d5ea4e7 9228 PL_efloatbuf[0] = '\0';
46fc3d4c 9229 }
9230
4151a5fe
IZ
9231 if ( !(width || left || plus || alt) && fill != '0'
9232 && has_precis && intsize != 'q' ) { /* Shortcuts */
2873255c
NC
9233 /* See earlier comment about buggy Gconvert when digits,
9234 aka precis is 0 */
9235 if ( c == 'g' && precis) {
2e59c212 9236 Gconvert((NV)nv, (int)precis, 0, PL_efloatbuf);
4150c189
NC
9237 /* May return an empty string for digits==0 */
9238 if (*PL_efloatbuf) {
9239 elen = strlen(PL_efloatbuf);
4151a5fe 9240 goto float_converted;
4150c189 9241 }
4151a5fe
IZ
9242 } else if ( c == 'f' && !precis) {
9243 if ((eptr = F0convert(nv, ebuf + sizeof ebuf, &elen)))
9244 break;
9245 }
9246 }
4d84ee25
NC
9247 {
9248 char *ptr = ebuf + sizeof ebuf;
9249 *--ptr = '\0';
9250 *--ptr = c;
9251 /* FIXME: what to do if HAS_LONG_DOUBLE but not PERL_PRIfldbl? */
9e5b023a 9252#if defined(HAS_LONG_DOUBLE) && defined(PERL_PRIfldbl)
4d84ee25
NC
9253 if (intsize == 'q') {
9254 /* Copy the one or more characters in a long double
9255 * format before the 'base' ([efgEFG]) character to
9256 * the format string. */
9257 static char const prifldbl[] = PERL_PRIfldbl;
9258 char const *p = prifldbl + sizeof(prifldbl) - 3;
9259 while (p >= prifldbl) { *--ptr = *p--; }
9260 }
65202027 9261#endif
4d84ee25
NC
9262 if (has_precis) {
9263 base = precis;
9264 do { *--ptr = '0' + (base % 10); } while (base /= 10);
9265 *--ptr = '.';
9266 }
9267 if (width) {
9268 base = width;
9269 do { *--ptr = '0' + (base % 10); } while (base /= 10);
9270 }
9271 if (fill == '0')
9272 *--ptr = fill;
9273 if (left)
9274 *--ptr = '-';
9275 if (plus)
9276 *--ptr = plus;
9277 if (alt)
9278 *--ptr = '#';
9279 *--ptr = '%';
9280
9281 /* No taint. Otherwise we are in the strange situation
9282 * where printf() taints but print($float) doesn't.
9283 * --jhi */
9e5b023a 9284#if defined(HAS_LONG_DOUBLE)
4150c189 9285 elen = ((intsize == 'q')
d9fad198
JH
9286 ? my_snprintf(PL_efloatbuf, PL_efloatsize, ptr, nv)
9287 : my_snprintf(PL_efloatbuf, PL_efloatsize, ptr, (double)nv));
9e5b023a 9288#else
4150c189 9289 elen = my_sprintf(PL_efloatbuf, ptr, nv);
9e5b023a 9290#endif
4d84ee25 9291 }
4151a5fe 9292 float_converted:
80252599 9293 eptr = PL_efloatbuf;
46fc3d4c 9294 break;
9295
fc36a67e 9296 /* SPECIAL */
9297
9298 case 'n':
26372e71
GA
9299 if (vectorize)
9300 goto unknown;
fc36a67e 9301 i = SvCUR(sv) - origlen;
26372e71 9302 if (args) {
c635e13b 9303 switch (intsize) {
9304 case 'h': *(va_arg(*args, short*)) = i; break;
9305 default: *(va_arg(*args, int*)) = i; break;
9306 case 'l': *(va_arg(*args, long*)) = i; break;
9307 case 'V': *(va_arg(*args, IV*)) = i; break;
cf2093f6
JH
9308#ifdef HAS_QUAD
9309 case 'q': *(va_arg(*args, Quad_t*)) = i; break;
9310#endif
c635e13b 9311 }
fc36a67e 9312 }
9dd79c3f 9313 else
211dfcf1 9314 sv_setuv_mg(argsv, (UV)i);
fc36a67e 9315 continue; /* not "break" */
9316
9317 /* UNKNOWN */
9318
46fc3d4c 9319 default:
fc36a67e 9320 unknown:
041457d9
DM
9321 if (!args
9322 && (PL_op->op_type == OP_PRTF || PL_op->op_type == OP_SPRINTF)
9323 && ckWARN(WARN_PRINTF))
9324 {
c4420975 9325 SV * const msg = sv_newmortal();
35c1215d
NC
9326 Perl_sv_setpvf(aTHX_ msg, "Invalid conversion in %sprintf: ",
9327 (PL_op->op_type == OP_PRTF) ? "" : "s");
0f4b6630 9328 if (c) {
0f4b6630 9329 if (isPRINT(c))
1c846c1f 9330 Perl_sv_catpvf(aTHX_ msg,
0f4b6630
JH
9331 "\"%%%c\"", c & 0xFF);
9332 else
9333 Perl_sv_catpvf(aTHX_ msg,
57def98f 9334 "\"%%\\%03"UVof"\"",
0f4b6630 9335 (UV)c & 0xFF);
0f4b6630 9336 } else
396482e1 9337 sv_catpvs(msg, "end of string");
95b63a38 9338 Perl_warner(aTHX_ packWARN(WARN_PRINTF), "%"SVf, (void*)msg); /* yes, this is reentrant */
c635e13b 9339 }
fb73857a 9340
9341 /* output mangled stuff ... */
9342 if (c == '\0')
9343 --q;
46fc3d4c 9344 eptr = p;
9345 elen = q - p;
fb73857a 9346
9347 /* ... right here, because formatting flags should not apply */
9348 SvGROW(sv, SvCUR(sv) + elen + 1);
9349 p = SvEND(sv);
4459522c 9350 Copy(eptr, p, elen, char);
fb73857a 9351 p += elen;
9352 *p = '\0';
3f7c398e 9353 SvCUR_set(sv, p - SvPVX_const(sv));
58e33a90 9354 svix = osvix;
fb73857a 9355 continue; /* not "break" */
46fc3d4c 9356 }
9357
cc61b222
TS
9358 if (is_utf8 != has_utf8) {
9359 if (is_utf8) {
9360 if (SvCUR(sv))
9361 sv_utf8_upgrade(sv);
9362 }
9363 else {
9364 const STRLEN old_elen = elen;
9365 SV * const nsv = sv_2mortal(newSVpvn(eptr, elen));
9366 sv_utf8_upgrade(nsv);
9367 eptr = SvPVX_const(nsv);
9368 elen = SvCUR(nsv);
9369
9370 if (width) { /* fudge width (can't fudge elen) */
9371 width += elen - old_elen;
9372 }
9373 is_utf8 = TRUE;
9374 }
9375 }
9376
6c94ec8b 9377 have = esignlen + zeros + elen;
ed2b91d2
GA
9378 if (have < zeros)
9379 Perl_croak_nocontext(PL_memory_wrap);
6c94ec8b 9380
46fc3d4c 9381 need = (have > width ? have : width);
9382 gap = need - have;
9383
d2641cbd
PC
9384 if (need >= (((STRLEN)~0) - SvCUR(sv) - dotstrlen - 1))
9385 Perl_croak_nocontext(PL_memory_wrap);
b22c7a20 9386 SvGROW(sv, SvCUR(sv) + need + dotstrlen + 1);
46fc3d4c 9387 p = SvEND(sv);
9388 if (esignlen && fill == '0') {
53c1dcc0 9389 int i;
eb160463 9390 for (i = 0; i < (int)esignlen; i++)
46fc3d4c 9391 *p++ = esignbuf[i];
9392 }
9393 if (gap && !left) {
9394 memset(p, fill, gap);
9395 p += gap;
9396 }
9397 if (esignlen && fill != '0') {
53c1dcc0 9398 int i;
eb160463 9399 for (i = 0; i < (int)esignlen; i++)
46fc3d4c 9400 *p++ = esignbuf[i];
9401 }
fc36a67e 9402 if (zeros) {
53c1dcc0 9403 int i;
fc36a67e 9404 for (i = zeros; i; i--)
9405 *p++ = '0';
9406 }
46fc3d4c 9407 if (elen) {
4459522c 9408 Copy(eptr, p, elen, char);
46fc3d4c 9409 p += elen;
9410 }
9411 if (gap && left) {
9412 memset(p, ' ', gap);
9413 p += gap;
9414 }
b22c7a20
GS
9415 if (vectorize) {
9416 if (veclen) {
4459522c 9417 Copy(dotstr, p, dotstrlen, char);
b22c7a20
GS
9418 p += dotstrlen;
9419 }
9420 else
9421 vectorize = FALSE; /* done iterating over vecstr */
9422 }
2cf2cfc6
A
9423 if (is_utf8)
9424 has_utf8 = TRUE;
9425 if (has_utf8)
7e2040f0 9426 SvUTF8_on(sv);
46fc3d4c 9427 *p = '\0';
3f7c398e 9428 SvCUR_set(sv, p - SvPVX_const(sv));
b22c7a20
GS
9429 if (vectorize) {
9430 esignlen = 0;
9431 goto vector;
9432 }
46fc3d4c 9433 }
9434}
51371543 9435
645c22ef
DM
9436/* =========================================================================
9437
9438=head1 Cloning an interpreter
9439
9440All the macros and functions in this section are for the private use of
9441the main function, perl_clone().
9442
9443The foo_dup() functions make an exact copy of an existing foo thinngy.
9444During the course of a cloning, a hash table is used to map old addresses
9445to new addresses. The table is created and manipulated with the
9446ptr_table_* functions.
9447
9448=cut
9449
9450============================================================================*/
9451
9452
1d7c1841
GS
9453#if defined(USE_ITHREADS)
9454
d4c19fe8 9455/* XXX Remove this so it doesn't have to go thru the macro and return for nothing */
1d7c1841
GS
9456#ifndef GpREFCNT_inc
9457# define GpREFCNT_inc(gp) ((gp) ? (++(gp)->gp_refcnt, (gp)) : (GP*)NULL)
9458#endif
9459
9460
a41cc44e
NC
9461/* Certain cases in Perl_ss_dup have been merged, by relying on the fact
9462 that currently av_dup and hv_dup are the same as sv_dup. If this changes,
9463 please unmerge ss_dup. */
d2d73c3e 9464#define sv_dup_inc(s,t) SvREFCNT_inc(sv_dup(s,t))
7f466ec7 9465#define sv_dup_inc_NN(s,t) SvREFCNT_inc_NN(sv_dup(s,t))
d2d73c3e
AB
9466#define av_dup(s,t) (AV*)sv_dup((SV*)s,t)
9467#define av_dup_inc(s,t) (AV*)SvREFCNT_inc(sv_dup((SV*)s,t))
9468#define hv_dup(s,t) (HV*)sv_dup((SV*)s,t)
9469#define hv_dup_inc(s,t) (HV*)SvREFCNT_inc(sv_dup((SV*)s,t))
9470#define cv_dup(s,t) (CV*)sv_dup((SV*)s,t)
9471#define cv_dup_inc(s,t) (CV*)SvREFCNT_inc(sv_dup((SV*)s,t))
9472#define io_dup(s,t) (IO*)sv_dup((SV*)s,t)
9473#define io_dup_inc(s,t) (IO*)SvREFCNT_inc(sv_dup((SV*)s,t))
9474#define gv_dup(s,t) (GV*)sv_dup((SV*)s,t)
9475#define gv_dup_inc(s,t) (GV*)SvREFCNT_inc(sv_dup((SV*)s,t))
6136c704
AL
9476#define SAVEPV(p) ((p) ? savepv(p) : NULL)
9477#define SAVEPVN(p,n) ((p) ? savepvn(p,n) : NULL)
8cf8f3d1 9478
d2d73c3e 9479
d2f185dc
AMS
9480/* Duplicate a regexp. Required reading: pregcomp() and pregfree() in
9481 regcomp.c. AMS 20010712 */
645c22ef 9482
1d7c1841 9483REGEXP *
53c1dcc0 9484Perl_re_dup(pTHX_ const REGEXP *r, CLONE_PARAMS *param)
1d7c1841 9485{
84da74a7 9486 return CALLREGDUPE(aTHX_ r,param);
1d7c1841
GS
9487}
9488
d2d73c3e 9489/* duplicate a file handle */
645c22ef 9490
1d7c1841 9491PerlIO *
a8fc9800 9492Perl_fp_dup(pTHX_ PerlIO *fp, char type, CLONE_PARAMS *param)
1d7c1841
GS
9493{
9494 PerlIO *ret;
53c1dcc0
AL
9495
9496 PERL_UNUSED_ARG(type);
73d840c0 9497
1d7c1841
GS
9498 if (!fp)
9499 return (PerlIO*)NULL;
9500
9501 /* look for it in the table first */
9502 ret = (PerlIO*)ptr_table_fetch(PL_ptr_table, fp);
9503 if (ret)
9504 return ret;
9505
9506 /* create anew and remember what it is */
ecdeb87c 9507 ret = PerlIO_fdupopen(aTHX_ fp, param, PERLIO_DUP_CLONE);
1d7c1841
GS
9508 ptr_table_store(PL_ptr_table, fp, ret);
9509 return ret;
9510}
9511
645c22ef
DM
9512/* duplicate a directory handle */
9513
1d7c1841
GS
9514DIR *
9515Perl_dirp_dup(pTHX_ DIR *dp)
9516{
96a5add6 9517 PERL_UNUSED_CONTEXT;
1d7c1841
GS
9518 if (!dp)
9519 return (DIR*)NULL;
9520 /* XXX TODO */
9521 return dp;
9522}
9523
ff276b08 9524/* duplicate a typeglob */
645c22ef 9525
1d7c1841 9526GP *
a8fc9800 9527Perl_gp_dup(pTHX_ GP *gp, CLONE_PARAMS* param)
1d7c1841
GS
9528{
9529 GP *ret;
b37c2d43 9530
1d7c1841
GS
9531 if (!gp)
9532 return (GP*)NULL;
9533 /* look for it in the table first */
9534 ret = (GP*)ptr_table_fetch(PL_ptr_table, gp);
9535 if (ret)
9536 return ret;
9537
9538 /* create anew and remember what it is */
a02a5408 9539 Newxz(ret, 1, GP);
1d7c1841
GS
9540 ptr_table_store(PL_ptr_table, gp, ret);
9541
9542 /* clone */
9543 ret->gp_refcnt = 0; /* must be before any other dups! */
d2d73c3e
AB
9544 ret->gp_sv = sv_dup_inc(gp->gp_sv, param);
9545 ret->gp_io = io_dup_inc(gp->gp_io, param);
9546 ret->gp_form = cv_dup_inc(gp->gp_form, param);
9547 ret->gp_av = av_dup_inc(gp->gp_av, param);
9548 ret->gp_hv = hv_dup_inc(gp->gp_hv, param);
9549 ret->gp_egv = gv_dup(gp->gp_egv, param);/* GvEGV is not refcounted */
9550 ret->gp_cv = cv_dup_inc(gp->gp_cv, param);
1d7c1841 9551 ret->gp_cvgen = gp->gp_cvgen;
1d7c1841 9552 ret->gp_line = gp->gp_line;
f4890806 9553 ret->gp_file_hek = hek_dup(gp->gp_file_hek, param);
1d7c1841
GS
9554 return ret;
9555}
9556
645c22ef
DM
9557/* duplicate a chain of magic */
9558
1d7c1841 9559MAGIC *
a8fc9800 9560Perl_mg_dup(pTHX_ MAGIC *mg, CLONE_PARAMS* param)
1d7c1841 9561{
cb359b41
JH
9562 MAGIC *mgprev = (MAGIC*)NULL;
9563 MAGIC *mgret;
1d7c1841
GS
9564 if (!mg)
9565 return (MAGIC*)NULL;
9566 /* look for it in the table first */
9567 mgret = (MAGIC*)ptr_table_fetch(PL_ptr_table, mg);
9568 if (mgret)
9569 return mgret;
9570
9571 for (; mg; mg = mg->mg_moremagic) {
9572 MAGIC *nmg;
a02a5408 9573 Newxz(nmg, 1, MAGIC);
cb359b41 9574 if (mgprev)
1d7c1841 9575 mgprev->mg_moremagic = nmg;
cb359b41
JH
9576 else
9577 mgret = nmg;
1d7c1841
GS
9578 nmg->mg_virtual = mg->mg_virtual; /* XXX copy dynamic vtable? */
9579 nmg->mg_private = mg->mg_private;
9580 nmg->mg_type = mg->mg_type;
9581 nmg->mg_flags = mg->mg_flags;
14befaf4 9582 if (mg->mg_type == PERL_MAGIC_qr) {
d2f185dc 9583 nmg->mg_obj = (SV*)re_dup((REGEXP*)mg->mg_obj, param);
1d7c1841 9584 }
05bd4103 9585 else if(mg->mg_type == PERL_MAGIC_backref) {
d7cbc7b5
NC
9586 /* The backref AV has its reference count deliberately bumped by
9587 1. */
9588 nmg->mg_obj = SvREFCNT_inc(av_dup_inc((AV*) mg->mg_obj, param));
05bd4103 9589 }
8d2f4536
NC
9590 else if (mg->mg_type == PERL_MAGIC_symtab) {
9591 nmg->mg_obj = mg->mg_obj;
9592 }
1d7c1841
GS
9593 else {
9594 nmg->mg_obj = (mg->mg_flags & MGf_REFCOUNTED)
d2d73c3e
AB
9595 ? sv_dup_inc(mg->mg_obj, param)
9596 : sv_dup(mg->mg_obj, param);
1d7c1841
GS
9597 }
9598 nmg->mg_len = mg->mg_len;
9599 nmg->mg_ptr = mg->mg_ptr; /* XXX random ptr? */
14befaf4 9600 if (mg->mg_ptr && mg->mg_type != PERL_MAGIC_regex_global) {
68795e93 9601 if (mg->mg_len > 0) {
1d7c1841 9602 nmg->mg_ptr = SAVEPVN(mg->mg_ptr, mg->mg_len);
14befaf4
DM
9603 if (mg->mg_type == PERL_MAGIC_overload_table &&
9604 AMT_AMAGIC((AMT*)mg->mg_ptr))
9605 {
c445ea15 9606 const AMT * const amtp = (AMT*)mg->mg_ptr;
0bcc34c2 9607 AMT * const namtp = (AMT*)nmg->mg_ptr;
1d7c1841
GS
9608 I32 i;
9609 for (i = 1; i < NofAMmeth; i++) {
d2d73c3e 9610 namtp->table[i] = cv_dup_inc(amtp->table[i], param);
1d7c1841
GS
9611 }
9612 }
9613 }
9614 else if (mg->mg_len == HEf_SVKEY)
d2d73c3e 9615 nmg->mg_ptr = (char*)sv_dup_inc((SV*)mg->mg_ptr, param);
1d7c1841 9616 }
68795e93
NIS
9617 if ((mg->mg_flags & MGf_DUP) && mg->mg_virtual && mg->mg_virtual->svt_dup) {
9618 CALL_FPTR(nmg->mg_virtual->svt_dup)(aTHX_ nmg, param);
9619 }
1d7c1841
GS
9620 mgprev = nmg;
9621 }
9622 return mgret;
9623}
9624
645c22ef
DM
9625/* create a new pointer-mapping table */
9626
1d7c1841
GS
9627PTR_TBL_t *
9628Perl_ptr_table_new(pTHX)
9629{
9630 PTR_TBL_t *tbl;
96a5add6
AL
9631 PERL_UNUSED_CONTEXT;
9632
a02a5408 9633 Newxz(tbl, 1, PTR_TBL_t);
1d7c1841
GS
9634 tbl->tbl_max = 511;
9635 tbl->tbl_items = 0;
a02a5408 9636 Newxz(tbl->tbl_ary, tbl->tbl_max + 1, PTR_TBL_ENT_t*);
1d7c1841
GS
9637 return tbl;
9638}
9639
7119fd33
NC
9640#define PTR_TABLE_HASH(ptr) \
9641 ((PTR2UV(ptr) >> 3) ^ (PTR2UV(ptr) >> (3 + 7)) ^ (PTR2UV(ptr) >> (3 + 17)))
134ca3d6 9642
93e68bfb
JC
9643/*
9644 we use the PTE_SVSLOT 'reservation' made above, both here (in the
9645 following define) and at call to new_body_inline made below in
9646 Perl_ptr_table_store()
9647 */
9648
9649#define del_pte(p) del_body_type(p, PTE_SVSLOT)
32e691d0 9650
645c22ef
DM
9651/* map an existing pointer using a table */
9652
7bf61b54 9653STATIC PTR_TBL_ENT_t *
b0e6ae5b 9654S_ptr_table_find(PTR_TBL_t *tbl, const void *sv) {
1d7c1841 9655 PTR_TBL_ENT_t *tblent;
4373e329 9656 const UV hash = PTR_TABLE_HASH(sv);
1d7c1841
GS
9657 assert(tbl);
9658 tblent = tbl->tbl_ary[hash & tbl->tbl_max];
9659 for (; tblent; tblent = tblent->next) {
9660 if (tblent->oldval == sv)
7bf61b54 9661 return tblent;
1d7c1841 9662 }
d4c19fe8 9663 return NULL;
7bf61b54
NC
9664}
9665
9666void *
9667Perl_ptr_table_fetch(pTHX_ PTR_TBL_t *tbl, const void *sv)
9668{
b0e6ae5b 9669 PTR_TBL_ENT_t const *const tblent = ptr_table_find(tbl, sv);
96a5add6 9670 PERL_UNUSED_CONTEXT;
d4c19fe8 9671 return tblent ? tblent->newval : NULL;
1d7c1841
GS
9672}
9673
645c22ef
DM
9674/* add a new entry to a pointer-mapping table */
9675
1d7c1841 9676void
44f8325f 9677Perl_ptr_table_store(pTHX_ PTR_TBL_t *tbl, const void *oldsv, void *newsv)
1d7c1841 9678{
0c9fdfe0 9679 PTR_TBL_ENT_t *tblent = ptr_table_find(tbl, oldsv);
96a5add6 9680 PERL_UNUSED_CONTEXT;
1d7c1841 9681
7bf61b54
NC
9682 if (tblent) {
9683 tblent->newval = newsv;
9684 } else {
9685 const UV entry = PTR_TABLE_HASH(oldsv) & tbl->tbl_max;
9686
d2a0f284
JC
9687 new_body_inline(tblent, PTE_SVSLOT);
9688
7bf61b54
NC
9689 tblent->oldval = oldsv;
9690 tblent->newval = newsv;
9691 tblent->next = tbl->tbl_ary[entry];
9692 tbl->tbl_ary[entry] = tblent;
9693 tbl->tbl_items++;
9694 if (tblent->next && tbl->tbl_items > tbl->tbl_max)
9695 ptr_table_split(tbl);
1d7c1841 9696 }
1d7c1841
GS
9697}
9698
645c22ef
DM
9699/* double the hash bucket size of an existing ptr table */
9700
1d7c1841
GS
9701void
9702Perl_ptr_table_split(pTHX_ PTR_TBL_t *tbl)
9703{
9704 PTR_TBL_ENT_t **ary = tbl->tbl_ary;
4373e329 9705 const UV oldsize = tbl->tbl_max + 1;
1d7c1841
GS
9706 UV newsize = oldsize * 2;
9707 UV i;
96a5add6 9708 PERL_UNUSED_CONTEXT;
1d7c1841
GS
9709
9710 Renew(ary, newsize, PTR_TBL_ENT_t*);
9711 Zero(&ary[oldsize], newsize-oldsize, PTR_TBL_ENT_t*);
9712 tbl->tbl_max = --newsize;
9713 tbl->tbl_ary = ary;
9714 for (i=0; i < oldsize; i++, ary++) {
9715 PTR_TBL_ENT_t **curentp, **entp, *ent;
9716 if (!*ary)
9717 continue;
9718 curentp = ary + oldsize;
9719 for (entp = ary, ent = *ary; ent; ent = *entp) {
134ca3d6 9720 if ((newsize & PTR_TABLE_HASH(ent->oldval)) != i) {
1d7c1841
GS
9721 *entp = ent->next;
9722 ent->next = *curentp;
9723 *curentp = ent;
9724 continue;
9725 }
9726 else
9727 entp = &ent->next;
9728 }
9729 }
9730}
9731
645c22ef
DM
9732/* remove all the entries from a ptr table */
9733
a0739874
DM
9734void
9735Perl_ptr_table_clear(pTHX_ PTR_TBL_t *tbl)
9736{
d5cefff9 9737 if (tbl && tbl->tbl_items) {
c445ea15 9738 register PTR_TBL_ENT_t * const * const array = tbl->tbl_ary;
d5cefff9 9739 UV riter = tbl->tbl_max;
a0739874 9740
d5cefff9
NC
9741 do {
9742 PTR_TBL_ENT_t *entry = array[riter];
ab1e7f95 9743
d5cefff9 9744 while (entry) {
00b6aa41 9745 PTR_TBL_ENT_t * const oentry = entry;
d5cefff9
NC
9746 entry = entry->next;
9747 del_pte(oentry);
9748 }
9749 } while (riter--);
a0739874 9750
d5cefff9
NC
9751 tbl->tbl_items = 0;
9752 }
a0739874
DM
9753}
9754
645c22ef
DM
9755/* clear and free a ptr table */
9756
a0739874
DM
9757void
9758Perl_ptr_table_free(pTHX_ PTR_TBL_t *tbl)
9759{
9760 if (!tbl) {
9761 return;
9762 }
9763 ptr_table_clear(tbl);
9764 Safefree(tbl->tbl_ary);
9765 Safefree(tbl);
9766}
9767
5bd07a3d 9768
83841fad 9769void
eb86f8b3 9770Perl_rvpv_dup(pTHX_ SV *dstr, const SV *sstr, CLONE_PARAMS* param)
83841fad
NIS
9771{
9772 if (SvROK(sstr)) {
b162af07
SP
9773 SvRV_set(dstr, SvWEAKREF(sstr)
9774 ? sv_dup(SvRV(sstr), param)
9775 : sv_dup_inc(SvRV(sstr), param));
f880fe2f 9776
83841fad 9777 }
3f7c398e 9778 else if (SvPVX_const(sstr)) {
83841fad
NIS
9779 /* Has something there */
9780 if (SvLEN(sstr)) {
68795e93 9781 /* Normal PV - clone whole allocated space */
3f7c398e 9782 SvPV_set(dstr, SAVEPVN(SvPVX_const(sstr), SvLEN(sstr)-1));
d3d0e6f1
NC
9783 if (SvREADONLY(sstr) && SvFAKE(sstr)) {
9784 /* Not that normal - actually sstr is copy on write.
9785 But we are a true, independant SV, so: */
9786 SvREADONLY_off(dstr);
9787 SvFAKE_off(dstr);
9788 }
68795e93 9789 }
83841fad
NIS
9790 else {
9791 /* Special case - not normally malloced for some reason */
f7877b28
NC
9792 if (isGV_with_GP(sstr)) {
9793 /* Don't need to do anything here. */
9794 }
9795 else if ((SvREADONLY(sstr) && SvFAKE(sstr))) {
ef10be65
NC
9796 /* A "shared" PV - clone it as "shared" PV */
9797 SvPV_set(dstr,
9798 HEK_KEY(hek_dup(SvSHARED_HEK_FROM_PV(SvPVX_const(sstr)),
9799 param)));
83841fad
NIS
9800 }
9801 else {
9802 /* Some other special case - random pointer */
f880fe2f 9803 SvPV_set(dstr, SvPVX(sstr));
d3d0e6f1 9804 }
83841fad
NIS
9805 }
9806 }
9807 else {
4608196e 9808 /* Copy the NULL */
f880fe2f 9809 if (SvTYPE(dstr) == SVt_RV)
b162af07 9810 SvRV_set(dstr, NULL);
f880fe2f 9811 else
6136c704 9812 SvPV_set(dstr, NULL);
83841fad
NIS
9813 }
9814}
9815
662fb8b2
NC
9816/* duplicate an SV of any type (including AV, HV etc) */
9817
1d7c1841 9818SV *
eb86f8b3 9819Perl_sv_dup(pTHX_ const SV *sstr, CLONE_PARAMS* param)
1d7c1841 9820{
27da23d5 9821 dVAR;
1d7c1841
GS
9822 SV *dstr;
9823
9824 if (!sstr || SvTYPE(sstr) == SVTYPEMASK)
6136c704 9825 return NULL;
1d7c1841
GS
9826 /* look for it in the table first */
9827 dstr = (SV*)ptr_table_fetch(PL_ptr_table, sstr);
9828 if (dstr)
9829 return dstr;
9830
0405e91e
AB
9831 if(param->flags & CLONEf_JOIN_IN) {
9832 /** We are joining here so we don't want do clone
9833 something that is bad **/
eb86f8b3
AL
9834 if (SvTYPE(sstr) == SVt_PVHV) {
9835 const char * const hvname = HvNAME_get(sstr);
9836 if (hvname)
9837 /** don't clone stashes if they already exist **/
9838 return (SV*)gv_stashpv(hvname,0);
0405e91e
AB
9839 }
9840 }
9841
1d7c1841
GS
9842 /* create anew and remember what it is */
9843 new_SV(dstr);
fd0854ff
DM
9844
9845#ifdef DEBUG_LEAKING_SCALARS
9846 dstr->sv_debug_optype = sstr->sv_debug_optype;
9847 dstr->sv_debug_line = sstr->sv_debug_line;
9848 dstr->sv_debug_inpad = sstr->sv_debug_inpad;
9849 dstr->sv_debug_cloned = 1;
fd0854ff 9850 dstr->sv_debug_file = savepv(sstr->sv_debug_file);
fd0854ff
DM
9851#endif
9852
1d7c1841
GS
9853 ptr_table_store(PL_ptr_table, sstr, dstr);
9854
9855 /* clone */
9856 SvFLAGS(dstr) = SvFLAGS(sstr);
9857 SvFLAGS(dstr) &= ~SVf_OOK; /* don't propagate OOK hack */
9858 SvREFCNT(dstr) = 0; /* must be before any other dups! */
9859
9860#ifdef DEBUGGING
3f7c398e 9861 if (SvANY(sstr) && PL_watch_pvx && SvPVX_const(sstr) == PL_watch_pvx)
1d7c1841 9862 PerlIO_printf(Perl_debug_log, "watch at %p hit, found string \"%s\"\n",
3f7c398e 9863 PL_watch_pvx, SvPVX_const(sstr));
1d7c1841
GS
9864#endif
9865
9660f481
DM
9866 /* don't clone objects whose class has asked us not to */
9867 if (SvOBJECT(sstr) && ! (SvFLAGS(SvSTASH(sstr)) & SVphv_CLONEABLE)) {
9868 SvFLAGS(dstr) &= ~SVTYPEMASK;
9869 SvOBJECT_off(dstr);
9870 return dstr;
9871 }
9872
1d7c1841
GS
9873 switch (SvTYPE(sstr)) {
9874 case SVt_NULL:
9875 SvANY(dstr) = NULL;
9876 break;
9877 case SVt_IV:
339049b0 9878 SvANY(dstr) = (XPVIV*)((char*)&(dstr->sv_u.svu_iv) - STRUCT_OFFSET(XPVIV, xiv_iv));
45977657 9879 SvIV_set(dstr, SvIVX(sstr));
1d7c1841
GS
9880 break;
9881 case SVt_NV:
9882 SvANY(dstr) = new_XNV();
9d6ce603 9883 SvNV_set(dstr, SvNVX(sstr));
1d7c1841
GS
9884 break;
9885 case SVt_RV:
339049b0 9886 SvANY(dstr) = &(dstr->sv_u.svu_rv);
83841fad 9887 Perl_rvpv_dup(aTHX_ dstr, sstr, param);
1d7c1841 9888 break;
662fb8b2
NC
9889 default:
9890 {
9891 /* These are all the types that need complex bodies allocating. */
662fb8b2 9892 void *new_body;
2bcc16b3
NC
9893 const svtype sv_type = SvTYPE(sstr);
9894 const struct body_details *const sv_type_details
9895 = bodies_by_type + sv_type;
662fb8b2 9896
93e68bfb 9897 switch (sv_type) {
662fb8b2 9898 default:
bb263b4e 9899 Perl_croak(aTHX_ "Bizarre SvTYPE [%" IVdf "]", (IV)SvTYPE(sstr));
662fb8b2
NC
9900 break;
9901
662fb8b2
NC
9902 case SVt_PVGV:
9903 if (GvUNIQUE((GV*)sstr)) {
6f207bd3 9904 NOOP; /* Do sharing here, and fall through */
662fb8b2 9905 }
c22188b4
NC
9906 case SVt_PVIO:
9907 case SVt_PVFM:
9908 case SVt_PVHV:
9909 case SVt_PVAV:
93e68bfb 9910 case SVt_PVBM:
662fb8b2 9911 case SVt_PVCV:
662fb8b2 9912 case SVt_PVLV:
662fb8b2 9913 case SVt_PVMG:
662fb8b2 9914 case SVt_PVNV:
662fb8b2 9915 case SVt_PVIV:
662fb8b2 9916 case SVt_PV:
d2a0f284 9917 assert(sv_type_details->body_size);
c22188b4 9918 if (sv_type_details->arena) {
d2a0f284 9919 new_body_inline(new_body, sv_type);
c22188b4 9920 new_body
b9502f15 9921 = (void*)((char*)new_body - sv_type_details->offset);
c22188b4
NC
9922 } else {
9923 new_body = new_NOARENA(sv_type_details);
9924 }
1d7c1841 9925 }
662fb8b2
NC
9926 assert(new_body);
9927 SvANY(dstr) = new_body;
9928
2bcc16b3 9929#ifndef PURIFY
b9502f15
NC
9930 Copy(((char*)SvANY(sstr)) + sv_type_details->offset,
9931 ((char*)SvANY(dstr)) + sv_type_details->offset,
f32993d6 9932 sv_type_details->copy, char);
2bcc16b3
NC
9933#else
9934 Copy(((char*)SvANY(sstr)),
9935 ((char*)SvANY(dstr)),
d2a0f284 9936 sv_type_details->body_size + sv_type_details->offset, char);
2bcc16b3 9937#endif
662fb8b2 9938
f7877b28
NC
9939 if (sv_type != SVt_PVAV && sv_type != SVt_PVHV
9940 && !isGV_with_GP(dstr))
662fb8b2
NC
9941 Perl_rvpv_dup(aTHX_ dstr, sstr, param);
9942
9943 /* The Copy above means that all the source (unduplicated) pointers
9944 are now in the destination. We can check the flags and the
9945 pointers in either, but it's possible that there's less cache
9946 missing by always going for the destination.
9947 FIXME - instrument and check that assumption */
f32993d6 9948 if (sv_type >= SVt_PVMG) {
885ffcb3
NC
9949 if ((sv_type == SVt_PVMG) && SvPAD_OUR(dstr)) {
9950 OURSTASH_set(dstr, hv_dup_inc(OURSTASH(dstr), param));
e736a858 9951 } else if (SvMAGIC(dstr))
662fb8b2
NC
9952 SvMAGIC_set(dstr, mg_dup(SvMAGIC(dstr), param));
9953 if (SvSTASH(dstr))
9954 SvSTASH_set(dstr, hv_dup_inc(SvSTASH(dstr), param));
1d7c1841 9955 }
662fb8b2 9956
f32993d6
NC
9957 /* The cast silences a GCC warning about unhandled types. */
9958 switch ((int)sv_type) {
662fb8b2
NC
9959 case SVt_PV:
9960 break;
9961 case SVt_PVIV:
9962 break;
9963 case SVt_PVNV:
9964 break;
9965 case SVt_PVMG:
9966 break;
9967 case SVt_PVBM:
9968 break;
9969 case SVt_PVLV:
9970 /* XXX LvTARGOFF sometimes holds PMOP* when DEBUGGING */
9971 if (LvTYPE(dstr) == 't') /* for tie: unrefcnted fake (SV**) */
9972 LvTARG(dstr) = dstr;
9973 else if (LvTYPE(dstr) == 'T') /* for tie: fake HE */
9974 LvTARG(dstr) = (SV*)he_dup((HE*)LvTARG(dstr), 0, param);
9975 else
9976 LvTARG(dstr) = sv_dup_inc(LvTARG(dstr), param);
9977 break;
9978 case SVt_PVGV:
acda4c6a
NC
9979 if (GvNAME_HEK(dstr))
9980 GvNAME_HEK(dstr) = hek_dup(GvNAME_HEK(dstr), param);
f5c1e807 9981
e15faf7d
NC
9982 /* Don't call sv_add_backref here as it's going to be created
9983 as part of the magic cloning of the symbol table. */
f7877b28
NC
9984 GvSTASH(dstr) = hv_dup(GvSTASH(dstr), param);
9985 if(isGV_with_GP(sstr)) {
9986 /* Danger Will Robinson - GvGP(dstr) isn't initialised
9987 at the point of this comment. */
9988 GvGP(dstr) = gp_dup(GvGP(sstr), param);
9989 (void)GpREFCNT_inc(GvGP(dstr));
9990 } else
9991 Perl_rvpv_dup(aTHX_ dstr, sstr, param);
662fb8b2
NC
9992 break;
9993 case SVt_PVIO:
9994 IoIFP(dstr) = fp_dup(IoIFP(dstr), IoTYPE(dstr), param);
9995 if (IoOFP(dstr) == IoIFP(sstr))
9996 IoOFP(dstr) = IoIFP(dstr);
9997 else
9998 IoOFP(dstr) = fp_dup(IoOFP(dstr), IoTYPE(dstr), param);
9999 /* PL_rsfp_filters entries have fake IoDIRP() */
662fb8b2
NC
10000 if(IoFLAGS(dstr) & IOf_FAKE_DIRP) {
10001 /* I have no idea why fake dirp (rsfps)
10002 should be treated differently but otherwise
10003 we end up with leaks -- sky*/
10004 IoTOP_GV(dstr) = gv_dup_inc(IoTOP_GV(dstr), param);
10005 IoFMT_GV(dstr) = gv_dup_inc(IoFMT_GV(dstr), param);
10006 IoBOTTOM_GV(dstr) = gv_dup_inc(IoBOTTOM_GV(dstr), param);
10007 } else {
10008 IoTOP_GV(dstr) = gv_dup(IoTOP_GV(dstr), param);
10009 IoFMT_GV(dstr) = gv_dup(IoFMT_GV(dstr), param);
10010 IoBOTTOM_GV(dstr) = gv_dup(IoBOTTOM_GV(dstr), param);
100ce7e1
NC
10011 if (IoDIRP(dstr)) {
10012 IoDIRP(dstr) = dirp_dup(IoDIRP(dstr));
10013 } else {
6f207bd3 10014 NOOP;
100ce7e1
NC
10015 /* IoDIRP(dstr) is already a copy of IoDIRP(sstr) */
10016 }
662fb8b2
NC
10017 }
10018 IoTOP_NAME(dstr) = SAVEPV(IoTOP_NAME(dstr));
10019 IoFMT_NAME(dstr) = SAVEPV(IoFMT_NAME(dstr));
10020 IoBOTTOM_NAME(dstr) = SAVEPV(IoBOTTOM_NAME(dstr));
10021 break;
10022 case SVt_PVAV:
10023 if (AvARRAY((AV*)sstr)) {
10024 SV **dst_ary, **src_ary;
10025 SSize_t items = AvFILLp((AV*)sstr) + 1;
10026
10027 src_ary = AvARRAY((AV*)sstr);
a02a5408 10028 Newxz(dst_ary, AvMAX((AV*)sstr)+1, SV*);
662fb8b2
NC
10029 ptr_table_store(PL_ptr_table, src_ary, dst_ary);
10030 SvPV_set(dstr, (char*)dst_ary);
10031 AvALLOC((AV*)dstr) = dst_ary;
10032 if (AvREAL((AV*)sstr)) {
10033 while (items-- > 0)
10034 *dst_ary++ = sv_dup_inc(*src_ary++, param);
10035 }
10036 else {
10037 while (items-- > 0)
10038 *dst_ary++ = sv_dup(*src_ary++, param);
10039 }
10040 items = AvMAX((AV*)sstr) - AvFILLp((AV*)sstr);
10041 while (items-- > 0) {
10042 *dst_ary++ = &PL_sv_undef;
10043 }
bfcb3514 10044 }
662fb8b2 10045 else {
bd61b366 10046 SvPV_set(dstr, NULL);
662fb8b2 10047 AvALLOC((AV*)dstr) = (SV**)NULL;
b79f7545 10048 }
662fb8b2
NC
10049 break;
10050 case SVt_PVHV:
7e265ef3
AL
10051 if (HvARRAY((HV*)sstr)) {
10052 STRLEN i = 0;
10053 const bool sharekeys = !!HvSHAREKEYS(sstr);
10054 XPVHV * const dxhv = (XPVHV*)SvANY(dstr);
10055 XPVHV * const sxhv = (XPVHV*)SvANY(sstr);
10056 char *darray;
10057 Newx(darray, PERL_HV_ARRAY_ALLOC_BYTES(dxhv->xhv_max+1)
10058 + (SvOOK(sstr) ? sizeof(struct xpvhv_aux) : 0),
10059 char);
10060 HvARRAY(dstr) = (HE**)darray;
10061 while (i <= sxhv->xhv_max) {
10062 const HE * const source = HvARRAY(sstr)[i];
10063 HvARRAY(dstr)[i] = source
10064 ? he_dup(source, sharekeys, param) : 0;
10065 ++i;
10066 }
10067 if (SvOOK(sstr)) {
10068 HEK *hvname;
10069 const struct xpvhv_aux * const saux = HvAUX(sstr);
10070 struct xpvhv_aux * const daux = HvAUX(dstr);
10071 /* This flag isn't copied. */
10072 /* SvOOK_on(hv) attacks the IV flags. */
10073 SvFLAGS(dstr) |= SVf_OOK;
10074
10075 hvname = saux->xhv_name;
10076 daux->xhv_name = hvname ? hek_dup(hvname, param) : hvname;
10077
10078 daux->xhv_riter = saux->xhv_riter;
10079 daux->xhv_eiter = saux->xhv_eiter
10080 ? he_dup(saux->xhv_eiter,
10081 (bool)!!HvSHAREKEYS(sstr), param) : 0;
10082 daux->xhv_backreferences =
10083 saux->xhv_backreferences
86f55936 10084 ? (AV*) SvREFCNT_inc(
7e265ef3 10085 sv_dup((SV*)saux->xhv_backreferences, param))
86f55936 10086 : 0;
7e265ef3
AL
10087 /* Record stashes for possible cloning in Perl_clone(). */
10088 if (hvname)
10089 av_push(param->stashes, dstr);
662fb8b2 10090 }
662fb8b2 10091 }
7e265ef3
AL
10092 else
10093 SvPV_set(dstr, NULL);
662fb8b2 10094 break;
662fb8b2 10095 case SVt_PVCV:
bb172083
NC
10096 if (!(param->flags & CLONEf_COPY_STACKS)) {
10097 CvDEPTH(dstr) = 0;
10098 }
10099 case SVt_PVFM:
662fb8b2
NC
10100 /* NOTE: not refcounted */
10101 CvSTASH(dstr) = hv_dup(CvSTASH(dstr), param);
10102 OP_REFCNT_LOCK;
d04ba589
NC
10103 if (!CvISXSUB(dstr))
10104 CvROOT(dstr) = OpREFCNT_inc(CvROOT(dstr));
662fb8b2 10105 OP_REFCNT_UNLOCK;
cfae286e 10106 if (CvCONST(dstr) && CvISXSUB(dstr)) {
662fb8b2
NC
10107 CvXSUBANY(dstr).any_ptr = GvUNIQUE(CvGV(dstr)) ?
10108 SvREFCNT_inc(CvXSUBANY(dstr).any_ptr) :
10109 sv_dup_inc((SV *)CvXSUBANY(dstr).any_ptr, param);
10110 }
10111 /* don't dup if copying back - CvGV isn't refcounted, so the
10112 * duped GV may never be freed. A bit of a hack! DAPM */
10113 CvGV(dstr) = (param->flags & CLONEf_JOIN_IN) ?
a0714e2c 10114 NULL : gv_dup(CvGV(dstr), param) ;
662fb8b2
NC
10115 PAD_DUP(CvPADLIST(dstr), CvPADLIST(sstr), param);
10116 CvOUTSIDE(dstr) =
10117 CvWEAKOUTSIDE(sstr)
10118 ? cv_dup( CvOUTSIDE(dstr), param)
10119 : cv_dup_inc(CvOUTSIDE(dstr), param);
aed2304a 10120 if (!CvISXSUB(dstr))
662fb8b2
NC
10121 CvFILE(dstr) = SAVEPV(CvFILE(dstr));
10122 break;
bfcb3514 10123 }
1d7c1841 10124 }
1d7c1841
GS
10125 }
10126
10127 if (SvOBJECT(dstr) && SvTYPE(dstr) != SVt_PVIO)
10128 ++PL_sv_objcount;
10129
10130 return dstr;
d2d73c3e 10131 }
1d7c1841 10132
645c22ef
DM
10133/* duplicate a context */
10134
1d7c1841 10135PERL_CONTEXT *
a8fc9800 10136Perl_cx_dup(pTHX_ PERL_CONTEXT *cxs, I32 ix, I32 max, CLONE_PARAMS* param)
1d7c1841
GS
10137{
10138 PERL_CONTEXT *ncxs;
10139
10140 if (!cxs)
10141 return (PERL_CONTEXT*)NULL;
10142
10143 /* look for it in the table first */
10144 ncxs = (PERL_CONTEXT*)ptr_table_fetch(PL_ptr_table, cxs);
10145 if (ncxs)
10146 return ncxs;
10147
10148 /* create anew and remember what it is */
a02a5408 10149 Newxz(ncxs, max + 1, PERL_CONTEXT);
1d7c1841
GS
10150 ptr_table_store(PL_ptr_table, cxs, ncxs);
10151
10152 while (ix >= 0) {
c445ea15
AL
10153 PERL_CONTEXT * const cx = &cxs[ix];
10154 PERL_CONTEXT * const ncx = &ncxs[ix];
1d7c1841
GS
10155 ncx->cx_type = cx->cx_type;
10156 if (CxTYPE(cx) == CXt_SUBST) {
10157 Perl_croak(aTHX_ "Cloning substitution context is unimplemented");
10158 }
10159 else {
10160 ncx->blk_oldsp = cx->blk_oldsp;
10161 ncx->blk_oldcop = cx->blk_oldcop;
1d7c1841
GS
10162 ncx->blk_oldmarksp = cx->blk_oldmarksp;
10163 ncx->blk_oldscopesp = cx->blk_oldscopesp;
10164 ncx->blk_oldpm = cx->blk_oldpm;
10165 ncx->blk_gimme = cx->blk_gimme;
10166 switch (CxTYPE(cx)) {
10167 case CXt_SUB:
10168 ncx->blk_sub.cv = (cx->blk_sub.olddepth == 0
d2d73c3e
AB
10169 ? cv_dup_inc(cx->blk_sub.cv, param)
10170 : cv_dup(cx->blk_sub.cv,param));
cc8d50a7 10171 ncx->blk_sub.argarray = (cx->blk_sub.hasargs
d2d73c3e 10172 ? av_dup_inc(cx->blk_sub.argarray, param)
7d49f689 10173 : NULL);
d2d73c3e 10174 ncx->blk_sub.savearray = av_dup_inc(cx->blk_sub.savearray, param);
1d7c1841 10175 ncx->blk_sub.olddepth = cx->blk_sub.olddepth;
cc8d50a7
NC
10176 ncx->blk_sub.hasargs = cx->blk_sub.hasargs;
10177 ncx->blk_sub.lval = cx->blk_sub.lval;
f39bc417 10178 ncx->blk_sub.retop = cx->blk_sub.retop;
d8d97e70
DM
10179 ncx->blk_sub.oldcomppad = (PAD*)ptr_table_fetch(PL_ptr_table,
10180 cx->blk_sub.oldcomppad);
1d7c1841
GS
10181 break;
10182 case CXt_EVAL:
10183 ncx->blk_eval.old_in_eval = cx->blk_eval.old_in_eval;
10184 ncx->blk_eval.old_op_type = cx->blk_eval.old_op_type;
b47cad08 10185 ncx->blk_eval.old_namesv = sv_dup_inc(cx->blk_eval.old_namesv, param);
1d7c1841 10186 ncx->blk_eval.old_eval_root = cx->blk_eval.old_eval_root;
d2d73c3e 10187 ncx->blk_eval.cur_text = sv_dup(cx->blk_eval.cur_text, param);
f39bc417 10188 ncx->blk_eval.retop = cx->blk_eval.retop;
1d7c1841
GS
10189 break;
10190 case CXt_LOOP:
10191 ncx->blk_loop.label = cx->blk_loop.label;
10192 ncx->blk_loop.resetsp = cx->blk_loop.resetsp;
022eaa24 10193 ncx->blk_loop.my_op = cx->blk_loop.my_op;
1d7c1841
GS
10194 ncx->blk_loop.iterdata = (CxPADLOOP(cx)
10195 ? cx->blk_loop.iterdata
d2d73c3e 10196 : gv_dup((GV*)cx->blk_loop.iterdata, param));
f3548bdc
DM
10197 ncx->blk_loop.oldcomppad
10198 = (PAD*)ptr_table_fetch(PL_ptr_table,
10199 cx->blk_loop.oldcomppad);
d2d73c3e
AB
10200 ncx->blk_loop.itersave = sv_dup_inc(cx->blk_loop.itersave, param);
10201 ncx->blk_loop.iterlval = sv_dup_inc(cx->blk_loop.iterlval, param);
10202 ncx->blk_loop.iterary = av_dup_inc(cx->blk_loop.iterary, param);
1d7c1841
GS
10203 ncx->blk_loop.iterix = cx->blk_loop.iterix;
10204 ncx->blk_loop.itermax = cx->blk_loop.itermax;
10205 break;
10206 case CXt_FORMAT:
d2d73c3e
AB
10207 ncx->blk_sub.cv = cv_dup(cx->blk_sub.cv, param);
10208 ncx->blk_sub.gv = gv_dup(cx->blk_sub.gv, param);
10209 ncx->blk_sub.dfoutgv = gv_dup_inc(cx->blk_sub.dfoutgv, param);
cc8d50a7 10210 ncx->blk_sub.hasargs = cx->blk_sub.hasargs;
f39bc417 10211 ncx->blk_sub.retop = cx->blk_sub.retop;
1d7c1841
GS
10212 break;
10213 case CXt_BLOCK:
10214 case CXt_NULL:
10215 break;
10216 }
10217 }
10218 --ix;
10219 }
10220 return ncxs;
10221}
10222
645c22ef
DM
10223/* duplicate a stack info structure */
10224
1d7c1841 10225PERL_SI *
a8fc9800 10226Perl_si_dup(pTHX_ PERL_SI *si, CLONE_PARAMS* param)
1d7c1841
GS
10227{
10228 PERL_SI *nsi;
10229
10230 if (!si)
10231 return (PERL_SI*)NULL;
10232
10233 /* look for it in the table first */
10234 nsi = (PERL_SI*)ptr_table_fetch(PL_ptr_table, si);
10235 if (nsi)
10236 return nsi;
10237
10238 /* create anew and remember what it is */
a02a5408 10239 Newxz(nsi, 1, PERL_SI);
1d7c1841
GS
10240 ptr_table_store(PL_ptr_table, si, nsi);
10241
d2d73c3e 10242 nsi->si_stack = av_dup_inc(si->si_stack, param);
1d7c1841
GS
10243 nsi->si_cxix = si->si_cxix;
10244 nsi->si_cxmax = si->si_cxmax;
d2d73c3e 10245 nsi->si_cxstack = cx_dup(si->si_cxstack, si->si_cxix, si->si_cxmax, param);
1d7c1841 10246 nsi->si_type = si->si_type;
d2d73c3e
AB
10247 nsi->si_prev = si_dup(si->si_prev, param);
10248 nsi->si_next = si_dup(si->si_next, param);
1d7c1841
GS
10249 nsi->si_markoff = si->si_markoff;
10250
10251 return nsi;
10252}
10253
10254#define POPINT(ss,ix) ((ss)[--(ix)].any_i32)
10255#define TOPINT(ss,ix) ((ss)[ix].any_i32)
10256#define POPLONG(ss,ix) ((ss)[--(ix)].any_long)
10257#define TOPLONG(ss,ix) ((ss)[ix].any_long)
10258#define POPIV(ss,ix) ((ss)[--(ix)].any_iv)
10259#define TOPIV(ss,ix) ((ss)[ix].any_iv)
38d8b13e
HS
10260#define POPBOOL(ss,ix) ((ss)[--(ix)].any_bool)
10261#define TOPBOOL(ss,ix) ((ss)[ix].any_bool)
1d7c1841
GS
10262#define POPPTR(ss,ix) ((ss)[--(ix)].any_ptr)
10263#define TOPPTR(ss,ix) ((ss)[ix].any_ptr)
10264#define POPDPTR(ss,ix) ((ss)[--(ix)].any_dptr)
10265#define TOPDPTR(ss,ix) ((ss)[ix].any_dptr)
10266#define POPDXPTR(ss,ix) ((ss)[--(ix)].any_dxptr)
10267#define TOPDXPTR(ss,ix) ((ss)[ix].any_dxptr)
10268
10269/* XXXXX todo */
10270#define pv_dup_inc(p) SAVEPV(p)
10271#define pv_dup(p) SAVEPV(p)
10272#define svp_dup_inc(p,pp) any_dup(p,pp)
10273
645c22ef
DM
10274/* map any object to the new equivent - either something in the
10275 * ptr table, or something in the interpreter structure
10276 */
10277
1d7c1841 10278void *
53c1dcc0 10279Perl_any_dup(pTHX_ void *v, const PerlInterpreter *proto_perl)
1d7c1841
GS
10280{
10281 void *ret;
10282
10283 if (!v)
10284 return (void*)NULL;
10285
10286 /* look for it in the table first */
10287 ret = ptr_table_fetch(PL_ptr_table, v);
10288 if (ret)
10289 return ret;
10290
10291 /* see if it is part of the interpreter structure */
10292 if (v >= (void*)proto_perl && v < (void*)(proto_perl+1))
acfe0abc 10293 ret = (void*)(((char*)aTHX) + (((char*)v) - (char*)proto_perl));
05ec9bb3 10294 else {
1d7c1841 10295 ret = v;
05ec9bb3 10296 }
1d7c1841
GS
10297
10298 return ret;
10299}
10300
645c22ef
DM
10301/* duplicate the save stack */
10302
1d7c1841 10303ANY *
a8fc9800 10304Perl_ss_dup(pTHX_ PerlInterpreter *proto_perl, CLONE_PARAMS* param)
1d7c1841 10305{
53c1dcc0
AL
10306 ANY * const ss = proto_perl->Tsavestack;
10307 const I32 max = proto_perl->Tsavestack_max;
10308 I32 ix = proto_perl->Tsavestack_ix;
1d7c1841
GS
10309 ANY *nss;
10310 SV *sv;
10311 GV *gv;
10312 AV *av;
10313 HV *hv;
10314 void* ptr;
10315 int intval;
10316 long longval;
10317 GP *gp;
10318 IV iv;
c4e33207 10319 char *c = NULL;
1d7c1841 10320 void (*dptr) (void*);
acfe0abc 10321 void (*dxptr) (pTHX_ void*);
1d7c1841 10322
a02a5408 10323 Newxz(nss, max, ANY);
1d7c1841
GS
10324
10325 while (ix > 0) {
b464bac0 10326 I32 i = POPINT(ss,ix);
1d7c1841
GS
10327 TOPINT(nss,ix) = i;
10328 switch (i) {
10329 case SAVEt_ITEM: /* normal string */
a41cc44e 10330 case SAVEt_SV: /* scalar reference */
1d7c1841 10331 sv = (SV*)POPPTR(ss,ix);
d2d73c3e 10332 TOPPTR(nss,ix) = sv_dup_inc(sv, param);
1d7c1841 10333 sv = (SV*)POPPTR(ss,ix);
d2d73c3e 10334 TOPPTR(nss,ix) = sv_dup_inc(sv, param);
1d7c1841 10335 break;
05ec9bb3
NIS
10336 case SAVEt_SHARED_PVREF: /* char* in shared space */
10337 c = (char*)POPPTR(ss,ix);
10338 TOPPTR(nss,ix) = savesharedpv(c);
10339 ptr = POPPTR(ss,ix);
10340 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10341 break;
1d7c1841
GS
10342 case SAVEt_GENERIC_SVREF: /* generic sv */
10343 case SAVEt_SVREF: /* scalar reference */
10344 sv = (SV*)POPPTR(ss,ix);
d2d73c3e 10345 TOPPTR(nss,ix) = sv_dup_inc(sv, param);
1d7c1841
GS
10346 ptr = POPPTR(ss,ix);
10347 TOPPTR(nss,ix) = svp_dup_inc((SV**)ptr, proto_perl);/* XXXXX */
10348 break;
a41cc44e 10349 case SAVEt_HV: /* hash reference */
1d7c1841 10350 case SAVEt_AV: /* array reference */
11b79775 10351 sv = (SV*) POPPTR(ss,ix);
337d28f5 10352 TOPPTR(nss,ix) = sv_dup_inc(sv, param);
1d7c1841 10353 gv = (GV*)POPPTR(ss,ix);
d2d73c3e 10354 TOPPTR(nss,ix) = gv_dup(gv, param);
1d7c1841
GS
10355 break;
10356 case SAVEt_INT: /* int reference */
10357 ptr = POPPTR(ss,ix);
10358 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10359 intval = (int)POPINT(ss,ix);
10360 TOPINT(nss,ix) = intval;
10361 break;
10362 case SAVEt_LONG: /* long reference */
10363 ptr = POPPTR(ss,ix);
10364 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10365 longval = (long)POPLONG(ss,ix);
10366 TOPLONG(nss,ix) = longval;
10367 break;
10368 case SAVEt_I32: /* I32 reference */
10369 case SAVEt_I16: /* I16 reference */
10370 case SAVEt_I8: /* I8 reference */
88effcc9 10371 case SAVEt_COP_ARYBASE: /* call CopARYBASE_set */
1d7c1841
GS
10372 ptr = POPPTR(ss,ix);
10373 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10374 i = POPINT(ss,ix);
10375 TOPINT(nss,ix) = i;
10376 break;
10377 case SAVEt_IV: /* IV reference */
10378 ptr = POPPTR(ss,ix);
10379 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10380 iv = POPIV(ss,ix);
10381 TOPIV(nss,ix) = iv;
10382 break;
a41cc44e
NC
10383 case SAVEt_HPTR: /* HV* reference */
10384 case SAVEt_APTR: /* AV* reference */
1d7c1841
GS
10385 case SAVEt_SPTR: /* SV* reference */
10386 ptr = POPPTR(ss,ix);
10387 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10388 sv = (SV*)POPPTR(ss,ix);
d2d73c3e 10389 TOPPTR(nss,ix) = sv_dup(sv, param);
1d7c1841
GS
10390 break;
10391 case SAVEt_VPTR: /* random* reference */
10392 ptr = POPPTR(ss,ix);
10393 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10394 ptr = POPPTR(ss,ix);
10395 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10396 break;
b03d03b0 10397 case SAVEt_GENERIC_PVREF: /* generic char* */
1d7c1841
GS
10398 case SAVEt_PPTR: /* char* reference */
10399 ptr = POPPTR(ss,ix);
10400 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10401 c = (char*)POPPTR(ss,ix);
10402 TOPPTR(nss,ix) = pv_dup(c);
10403 break;
1d7c1841
GS
10404 case SAVEt_NSTAB:
10405 gv = (GV*)POPPTR(ss,ix);
d2d73c3e 10406 TOPPTR(nss,ix) = gv_dup(gv, param);
1d7c1841
GS
10407 break;
10408 case SAVEt_GP: /* scalar reference */
10409 gp = (GP*)POPPTR(ss,ix);
d2d73c3e 10410 TOPPTR(nss,ix) = gp = gp_dup(gp, param);
1d7c1841
GS
10411 (void)GpREFCNT_inc(gp);
10412 gv = (GV*)POPPTR(ss,ix);
2ed3c8fc 10413 TOPPTR(nss,ix) = gv_dup_inc(gv, param);
1d7c1841
GS
10414 c = (char*)POPPTR(ss,ix);
10415 TOPPTR(nss,ix) = pv_dup(c);
10416 iv = POPIV(ss,ix);
10417 TOPIV(nss,ix) = iv;
10418 iv = POPIV(ss,ix);
10419 TOPIV(nss,ix) = iv;
10420 break;
10421 case SAVEt_FREESV:
26d9b02f 10422 case SAVEt_MORTALIZESV:
1d7c1841 10423 sv = (SV*)POPPTR(ss,ix);
d2d73c3e 10424 TOPPTR(nss,ix) = sv_dup_inc(sv, param);
1d7c1841
GS
10425 break;
10426 case SAVEt_FREEOP:
10427 ptr = POPPTR(ss,ix);
10428 if (ptr && (((OP*)ptr)->op_private & OPpREFCOUNTED)) {
10429 /* these are assumed to be refcounted properly */
53c1dcc0 10430 OP *o;
1d7c1841
GS
10431 switch (((OP*)ptr)->op_type) {
10432 case OP_LEAVESUB:
10433 case OP_LEAVESUBLV:
10434 case OP_LEAVEEVAL:
10435 case OP_LEAVE:
10436 case OP_SCOPE:
10437 case OP_LEAVEWRITE:
e977893f
GS
10438 TOPPTR(nss,ix) = ptr;
10439 o = (OP*)ptr;
10440 OpREFCNT_inc(o);
1d7c1841
GS
10441 break;
10442 default:
5f66b61c 10443 TOPPTR(nss,ix) = NULL;
1d7c1841
GS
10444 break;
10445 }
10446 }
10447 else
5f66b61c 10448 TOPPTR(nss,ix) = NULL;
1d7c1841
GS
10449 break;
10450 case SAVEt_FREEPV:
10451 c = (char*)POPPTR(ss,ix);
10452 TOPPTR(nss,ix) = pv_dup_inc(c);
10453 break;
10454 case SAVEt_CLEARSV:
10455 longval = POPLONG(ss,ix);
10456 TOPLONG(nss,ix) = longval;
10457 break;
10458 case SAVEt_DELETE:
10459 hv = (HV*)POPPTR(ss,ix);
d2d73c3e 10460 TOPPTR(nss,ix) = hv_dup_inc(hv, param);
1d7c1841
GS
10461 c = (char*)POPPTR(ss,ix);
10462 TOPPTR(nss,ix) = pv_dup_inc(c);
10463 i = POPINT(ss,ix);
10464 TOPINT(nss,ix) = i;
10465 break;
10466 case SAVEt_DESTRUCTOR:
10467 ptr = POPPTR(ss,ix);
10468 TOPPTR(nss,ix) = any_dup(ptr, proto_perl); /* XXX quite arbitrary */
10469 dptr = POPDPTR(ss,ix);
8141890a
JH
10470 TOPDPTR(nss,ix) = DPTR2FPTR(void (*)(void*),
10471 any_dup(FPTR2DPTR(void *, dptr),
10472 proto_perl));
1d7c1841
GS
10473 break;
10474 case SAVEt_DESTRUCTOR_X:
10475 ptr = POPPTR(ss,ix);
10476 TOPPTR(nss,ix) = any_dup(ptr, proto_perl); /* XXX quite arbitrary */
10477 dxptr = POPDXPTR(ss,ix);
8141890a
JH
10478 TOPDXPTR(nss,ix) = DPTR2FPTR(void (*)(pTHX_ void*),
10479 any_dup(FPTR2DPTR(void *, dxptr),
10480 proto_perl));
1d7c1841
GS
10481 break;
10482 case SAVEt_REGCONTEXT:
10483 case SAVEt_ALLOC:
10484 i = POPINT(ss,ix);
10485 TOPINT(nss,ix) = i;
10486 ix -= i;
10487 break;
10488 case SAVEt_STACK_POS: /* Position on Perl stack */
10489 i = POPINT(ss,ix);
10490 TOPINT(nss,ix) = i;
10491 break;
10492 case SAVEt_AELEM: /* array element */
10493 sv = (SV*)POPPTR(ss,ix);
d2d73c3e 10494 TOPPTR(nss,ix) = sv_dup_inc(sv, param);
1d7c1841
GS
10495 i = POPINT(ss,ix);
10496 TOPINT(nss,ix) = i;
10497 av = (AV*)POPPTR(ss,ix);
d2d73c3e 10498 TOPPTR(nss,ix) = av_dup_inc(av, param);
1d7c1841
GS
10499 break;
10500 case SAVEt_HELEM: /* hash element */
10501 sv = (SV*)POPPTR(ss,ix);
d2d73c3e 10502 TOPPTR(nss,ix) = sv_dup_inc(sv, param);
1d7c1841 10503 sv = (SV*)POPPTR(ss,ix);
d2d73c3e 10504 TOPPTR(nss,ix) = sv_dup_inc(sv, param);
1d7c1841 10505 hv = (HV*)POPPTR(ss,ix);
d2d73c3e 10506 TOPPTR(nss,ix) = hv_dup_inc(hv, param);
1d7c1841
GS
10507 break;
10508 case SAVEt_OP:
10509 ptr = POPPTR(ss,ix);
10510 TOPPTR(nss,ix) = ptr;
10511 break;
10512 case SAVEt_HINTS:
10513 i = POPINT(ss,ix);
10514 TOPINT(nss,ix) = i;
b3ca2e83 10515 ptr = POPPTR(ss,ix);
080ac856 10516 if (ptr) {
7b6dd8c3 10517 HINTS_REFCNT_LOCK;
080ac856 10518 ((struct refcounted_he *)ptr)->refcounted_he_refcnt++;
7b6dd8c3
NC
10519 HINTS_REFCNT_UNLOCK;
10520 }
cbb1fbea 10521 TOPPTR(nss,ix) = ptr;
a8f8b6a7
NC
10522 if (i & HINT_LOCALIZE_HH) {
10523 hv = (HV*)POPPTR(ss,ix);
10524 TOPPTR(nss,ix) = hv_dup_inc(hv, param);
10525 }
1d7c1841 10526 break;
c4410b1b
GS
10527 case SAVEt_COMPPAD:
10528 av = (AV*)POPPTR(ss,ix);
58ed4fbe 10529 TOPPTR(nss,ix) = av_dup(av, param);
c4410b1b 10530 break;
c3564e5c
GS
10531 case SAVEt_PADSV:
10532 longval = (long)POPLONG(ss,ix);
10533 TOPLONG(nss,ix) = longval;
10534 ptr = POPPTR(ss,ix);
10535 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
10536 sv = (SV*)POPPTR(ss,ix);
d2d73c3e 10537 TOPPTR(nss,ix) = sv_dup(sv, param);
c3564e5c 10538 break;
a1bb4754 10539 case SAVEt_BOOL:
38d8b13e 10540 ptr = POPPTR(ss,ix);
b9609c01 10541 TOPPTR(nss,ix) = any_dup(ptr, proto_perl);
38d8b13e 10542 longval = (long)POPBOOL(ss,ix);
b9609c01 10543 TOPBOOL(nss,ix) = (bool)longval;
a1bb4754 10544 break;
8bd2680e
MHM
10545 case SAVEt_SET_SVFLAGS:
10546 i = POPINT(ss,ix);
10547 TOPINT(nss,ix) = i;
10548 i = POPINT(ss,ix);
10549 TOPINT(nss,ix) = i;
10550 sv = (SV*)POPPTR(ss,ix);
10551 TOPPTR(nss,ix) = sv_dup(sv, param);
10552 break;
5bfb7d0e
NC
10553 case SAVEt_RE_STATE:
10554 {
10555 const struct re_save_state *const old_state
10556 = (struct re_save_state *)
10557 (ss + ix - SAVESTACK_ALLOC_FOR_RE_SAVE_STATE);
10558 struct re_save_state *const new_state
10559 = (struct re_save_state *)
10560 (nss + ix - SAVESTACK_ALLOC_FOR_RE_SAVE_STATE);
10561
10562 Copy(old_state, new_state, 1, struct re_save_state);
10563 ix -= SAVESTACK_ALLOC_FOR_RE_SAVE_STATE;
10564
10565 new_state->re_state_bostr
10566 = pv_dup(old_state->re_state_bostr);
10567 new_state->re_state_reginput
10568 = pv_dup(old_state->re_state_reginput);
5bfb7d0e
NC
10569 new_state->re_state_regeol
10570 = pv_dup(old_state->re_state_regeol);
10571 new_state->re_state_regstartp
11b79775 10572 = (I32*) any_dup(old_state->re_state_regstartp, proto_perl);
5bfb7d0e 10573 new_state->re_state_regendp
11b79775 10574 = (I32*) any_dup(old_state->re_state_regendp, proto_perl);
5bfb7d0e 10575 new_state->re_state_reglastparen
11b79775
DD
10576 = (U32*) any_dup(old_state->re_state_reglastparen,
10577 proto_perl);
5bfb7d0e 10578 new_state->re_state_reglastcloseparen
11b79775 10579 = (U32*)any_dup(old_state->re_state_reglastcloseparen,
5bfb7d0e 10580 proto_perl);
5bfb7d0e
NC
10581 /* XXX This just has to be broken. The old save_re_context
10582 code did SAVEGENERICPV(PL_reg_start_tmp);
10583 PL_reg_start_tmp is char **.
10584 Look above to what the dup code does for
10585 SAVEt_GENERIC_PVREF
10586 It can never have worked.
10587 So this is merely a faithful copy of the exiting bug: */
10588 new_state->re_state_reg_start_tmp
10589 = (char **) pv_dup((char *)
10590 old_state->re_state_reg_start_tmp);
10591 /* I assume that it only ever "worked" because no-one called
10592 (pseudo)fork while the regexp engine had re-entered itself.
10593 */
5bfb7d0e
NC
10594#ifdef PERL_OLD_COPY_ON_WRITE
10595 new_state->re_state_nrs
10596 = sv_dup(old_state->re_state_nrs, param);
10597#endif
10598 new_state->re_state_reg_magic
11b79775
DD
10599 = (MAGIC*) any_dup(old_state->re_state_reg_magic,
10600 proto_perl);
5bfb7d0e 10601 new_state->re_state_reg_oldcurpm
11b79775
DD
10602 = (PMOP*) any_dup(old_state->re_state_reg_oldcurpm,
10603 proto_perl);
5bfb7d0e 10604 new_state->re_state_reg_curpm
11b79775
DD
10605 = (PMOP*) any_dup(old_state->re_state_reg_curpm,
10606 proto_perl);
5bfb7d0e
NC
10607 new_state->re_state_reg_oldsaved
10608 = pv_dup(old_state->re_state_reg_oldsaved);
10609 new_state->re_state_reg_poscache
10610 = pv_dup(old_state->re_state_reg_poscache);
5bfb7d0e
NC
10611 new_state->re_state_reg_starttry
10612 = pv_dup(old_state->re_state_reg_starttry);
5bfb7d0e
NC
10613 break;
10614 }
68da3b2f
NC
10615 case SAVEt_COMPILE_WARNINGS:
10616 ptr = POPPTR(ss,ix);
10617 TOPPTR(nss,ix) = DUP_WARNINGS((STRLEN*)ptr);
7b6dd8c3 10618 break;
1d7c1841 10619 default:
ca05af4a 10620 Perl_croak(aTHX_ "panic: ss_dup inconsistency (%"IVdf")", (IV) i);
1d7c1841
GS
10621 }
10622 }
10623
bd81e77b
NC
10624 return nss;
10625}
10626
10627
10628/* if sv is a stash, call $class->CLONE_SKIP(), and set the SVphv_CLONEABLE
10629 * flag to the result. This is done for each stash before cloning starts,
10630 * so we know which stashes want their objects cloned */
10631
10632static void
10633do_mark_cloneable_stash(pTHX_ SV *sv)
10634{
10635 const HEK * const hvname = HvNAME_HEK((HV*)sv);
10636 if (hvname) {
10637 GV* const cloner = gv_fetchmethod_autoload((HV*)sv, "CLONE_SKIP", 0);
10638 SvFLAGS(sv) |= SVphv_CLONEABLE; /* clone objects by default */
10639 if (cloner && GvCV(cloner)) {
10640 dSP;
10641 UV status;
10642
10643 ENTER;
10644 SAVETMPS;
10645 PUSHMARK(SP);
10646 XPUSHs(sv_2mortal(newSVhek(hvname)));
10647 PUTBACK;
10648 call_sv((SV*)GvCV(cloner), G_SCALAR);
10649 SPAGAIN;
10650 status = POPu;
10651 PUTBACK;
10652 FREETMPS;
10653 LEAVE;
10654 if (status)
10655 SvFLAGS(sv) &= ~SVphv_CLONEABLE;
10656 }
10657 }
10658}
10659
10660
10661
10662/*
10663=for apidoc perl_clone
10664
10665Create and return a new interpreter by cloning the current one.
10666
10667perl_clone takes these flags as parameters:
10668
10669CLONEf_COPY_STACKS - is used to, well, copy the stacks also,
10670without it we only clone the data and zero the stacks,
10671with it we copy the stacks and the new perl interpreter is
10672ready to run at the exact same point as the previous one.
10673The pseudo-fork code uses COPY_STACKS while the
10674threads->new doesn't.
10675
10676CLONEf_KEEP_PTR_TABLE
10677perl_clone keeps a ptr_table with the pointer of the old
10678variable as a key and the new variable as a value,
10679this allows it to check if something has been cloned and not
10680clone it again but rather just use the value and increase the
10681refcount. If KEEP_PTR_TABLE is not set then perl_clone will kill
10682the ptr_table using the function
10683C<ptr_table_free(PL_ptr_table); PL_ptr_table = NULL;>,
10684reason to keep it around is if you want to dup some of your own
10685variable who are outside the graph perl scans, example of this
10686code is in threads.xs create
10687
10688CLONEf_CLONE_HOST
10689This is a win32 thing, it is ignored on unix, it tells perls
10690win32host code (which is c++) to clone itself, this is needed on
10691win32 if you want to run two threads at the same time,
10692if you just want to do some stuff in a separate perl interpreter
10693and then throw it away and return to the original one,
10694you don't need to do anything.
10695
10696=cut
10697*/
10698
10699/* XXX the above needs expanding by someone who actually understands it ! */
10700EXTERN_C PerlInterpreter *
10701perl_clone_host(PerlInterpreter* proto_perl, UV flags);
10702
10703PerlInterpreter *
10704perl_clone(PerlInterpreter *proto_perl, UV flags)
10705{
10706 dVAR;
10707#ifdef PERL_IMPLICIT_SYS
10708
10709 /* perlhost.h so we need to call into it
10710 to clone the host, CPerlHost should have a c interface, sky */
10711
10712 if (flags & CLONEf_CLONE_HOST) {
10713 return perl_clone_host(proto_perl,flags);
10714 }
10715 return perl_clone_using(proto_perl, flags,
10716 proto_perl->IMem,
10717 proto_perl->IMemShared,
10718 proto_perl->IMemParse,
10719 proto_perl->IEnv,
10720 proto_perl->IStdIO,
10721 proto_perl->ILIO,
10722 proto_perl->IDir,
10723 proto_perl->ISock,
10724 proto_perl->IProc);
10725}
10726
10727PerlInterpreter *
10728perl_clone_using(PerlInterpreter *proto_perl, UV flags,
10729 struct IPerlMem* ipM, struct IPerlMem* ipMS,
10730 struct IPerlMem* ipMP, struct IPerlEnv* ipE,
10731 struct IPerlStdIO* ipStd, struct IPerlLIO* ipLIO,
10732 struct IPerlDir* ipD, struct IPerlSock* ipS,
10733 struct IPerlProc* ipP)
10734{
10735 /* XXX many of the string copies here can be optimized if they're
10736 * constants; they need to be allocated as common memory and just
10737 * their pointers copied. */
10738
10739 IV i;
10740 CLONE_PARAMS clone_params;
5f66b61c 10741 CLONE_PARAMS* const param = &clone_params;
bd81e77b 10742
5f66b61c 10743 PerlInterpreter * const my_perl = (PerlInterpreter*)(*ipM->pMalloc)(ipM, sizeof(PerlInterpreter));
bd81e77b
NC
10744 /* for each stash, determine whether its objects should be cloned */
10745 S_visit(proto_perl, do_mark_cloneable_stash, SVt_PVHV, SVTYPEMASK);
10746 PERL_SET_THX(my_perl);
10747
10748# ifdef DEBUGGING
7e337ee0 10749 PoisonNew(my_perl, 1, PerlInterpreter);
5f66b61c
AL
10750 PL_op = NULL;
10751 PL_curcop = NULL;
bd81e77b
NC
10752 PL_markstack = 0;
10753 PL_scopestack = 0;
10754 PL_savestack = 0;
10755 PL_savestack_ix = 0;
10756 PL_savestack_max = -1;
10757 PL_sig_pending = 0;
10758 Zero(&PL_debug_pad, 1, struct perl_debug_pad);
10759# else /* !DEBUGGING */
10760 Zero(my_perl, 1, PerlInterpreter);
10761# endif /* DEBUGGING */
10762
10763 /* host pointers */
10764 PL_Mem = ipM;
10765 PL_MemShared = ipMS;
10766 PL_MemParse = ipMP;
10767 PL_Env = ipE;
10768 PL_StdIO = ipStd;
10769 PL_LIO = ipLIO;
10770 PL_Dir = ipD;
10771 PL_Sock = ipS;
10772 PL_Proc = ipP;
10773#else /* !PERL_IMPLICIT_SYS */
10774 IV i;
10775 CLONE_PARAMS clone_params;
10776 CLONE_PARAMS* param = &clone_params;
5f66b61c 10777 PerlInterpreter * const my_perl = (PerlInterpreter*)PerlMem_malloc(sizeof(PerlInterpreter));
bd81e77b
NC
10778 /* for each stash, determine whether its objects should be cloned */
10779 S_visit(proto_perl, do_mark_cloneable_stash, SVt_PVHV, SVTYPEMASK);
10780 PERL_SET_THX(my_perl);
10781
10782# ifdef DEBUGGING
7e337ee0 10783 PoisonNew(my_perl, 1, PerlInterpreter);
5f66b61c
AL
10784 PL_op = NULL;
10785 PL_curcop = NULL;
bd81e77b
NC
10786 PL_markstack = 0;
10787 PL_scopestack = 0;
10788 PL_savestack = 0;
10789 PL_savestack_ix = 0;
10790 PL_savestack_max = -1;
10791 PL_sig_pending = 0;
10792 Zero(&PL_debug_pad, 1, struct perl_debug_pad);
10793# else /* !DEBUGGING */
10794 Zero(my_perl, 1, PerlInterpreter);
10795# endif /* DEBUGGING */
10796#endif /* PERL_IMPLICIT_SYS */
10797 param->flags = flags;
10798 param->proto_perl = proto_perl;
10799
7cb608b5
NC
10800 INIT_TRACK_MEMPOOL(my_perl->Imemory_debug_header, my_perl);
10801
fdda85ca 10802 PL_body_arenas = NULL;
bd81e77b
NC
10803 Zero(&PL_body_roots, 1, PL_body_roots);
10804
10805 PL_nice_chunk = NULL;
10806 PL_nice_chunk_size = 0;
10807 PL_sv_count = 0;
10808 PL_sv_objcount = 0;
a0714e2c
SS
10809 PL_sv_root = NULL;
10810 PL_sv_arenaroot = NULL;
bd81e77b
NC
10811
10812 PL_debug = proto_perl->Idebug;
10813
10814 PL_hash_seed = proto_perl->Ihash_seed;
10815 PL_rehash_seed = proto_perl->Irehash_seed;
10816
10817#ifdef USE_REENTRANT_API
10818 /* XXX: things like -Dm will segfault here in perlio, but doing
10819 * PERL_SET_CONTEXT(proto_perl);
10820 * breaks too many other things
10821 */
10822 Perl_reentrant_init(aTHX);
10823#endif
10824
10825 /* create SV map for pointer relocation */
10826 PL_ptr_table = ptr_table_new();
10827
10828 /* initialize these special pointers as early as possible */
10829 SvANY(&PL_sv_undef) = NULL;
10830 SvREFCNT(&PL_sv_undef) = (~(U32)0)/2;
10831 SvFLAGS(&PL_sv_undef) = SVf_READONLY|SVt_NULL;
10832 ptr_table_store(PL_ptr_table, &proto_perl->Isv_undef, &PL_sv_undef);
10833
10834 SvANY(&PL_sv_no) = new_XPVNV();
10835 SvREFCNT(&PL_sv_no) = (~(U32)0)/2;
10836 SvFLAGS(&PL_sv_no) = SVp_IOK|SVf_IOK|SVp_NOK|SVf_NOK
10837 |SVp_POK|SVf_POK|SVf_READONLY|SVt_PVNV;
bb7a0f54 10838 SvPV_set(&PL_sv_no, savepvn(PL_No, 0));
bd81e77b
NC
10839 SvCUR_set(&PL_sv_no, 0);
10840 SvLEN_set(&PL_sv_no, 1);
10841 SvIV_set(&PL_sv_no, 0);
10842 SvNV_set(&PL_sv_no, 0);
10843 ptr_table_store(PL_ptr_table, &proto_perl->Isv_no, &PL_sv_no);
10844
10845 SvANY(&PL_sv_yes) = new_XPVNV();
10846 SvREFCNT(&PL_sv_yes) = (~(U32)0)/2;
10847 SvFLAGS(&PL_sv_yes) = SVp_IOK|SVf_IOK|SVp_NOK|SVf_NOK
10848 |SVp_POK|SVf_POK|SVf_READONLY|SVt_PVNV;
bb7a0f54 10849 SvPV_set(&PL_sv_yes, savepvn(PL_Yes, 1));
bd81e77b
NC
10850 SvCUR_set(&PL_sv_yes, 1);
10851 SvLEN_set(&PL_sv_yes, 2);
10852 SvIV_set(&PL_sv_yes, 1);
10853 SvNV_set(&PL_sv_yes, 1);
10854 ptr_table_store(PL_ptr_table, &proto_perl->Isv_yes, &PL_sv_yes);
10855
10856 /* create (a non-shared!) shared string table */
10857 PL_strtab = newHV();
10858 HvSHAREKEYS_off(PL_strtab);
10859 hv_ksplit(PL_strtab, HvTOTALKEYS(proto_perl->Istrtab));
10860 ptr_table_store(PL_ptr_table, proto_perl->Istrtab, PL_strtab);
10861
10862 PL_compiling = proto_perl->Icompiling;
10863
10864 /* These two PVs will be free'd special way so must set them same way op.c does */
10865 PL_compiling.cop_stashpv = savesharedpv(PL_compiling.cop_stashpv);
10866 ptr_table_store(PL_ptr_table, proto_perl->Icompiling.cop_stashpv, PL_compiling.cop_stashpv);
10867
10868 PL_compiling.cop_file = savesharedpv(PL_compiling.cop_file);
10869 ptr_table_store(PL_ptr_table, proto_perl->Icompiling.cop_file, PL_compiling.cop_file);
10870
10871 ptr_table_store(PL_ptr_table, &proto_perl->Icompiling, &PL_compiling);
72dc9ed5 10872 PL_compiling.cop_warnings = DUP_WARNINGS(PL_compiling.cop_warnings);
c28fe1ec 10873 if (PL_compiling.cop_hints_hash) {
cbb1fbea 10874 HINTS_REFCNT_LOCK;
c28fe1ec 10875 PL_compiling.cop_hints_hash->refcounted_he_refcnt++;
cbb1fbea
NC
10876 HINTS_REFCNT_UNLOCK;
10877 }
bd81e77b
NC
10878 PL_curcop = (COP*)any_dup(proto_perl->Tcurcop, proto_perl);
10879
10880 /* pseudo environmental stuff */
10881 PL_origargc = proto_perl->Iorigargc;
10882 PL_origargv = proto_perl->Iorigargv;
10883
10884 param->stashes = newAV(); /* Setup array of objects to call clone on */
10885
10886 /* Set tainting stuff before PerlIO_debug can possibly get called */
10887 PL_tainting = proto_perl->Itainting;
10888 PL_taint_warn = proto_perl->Itaint_warn;
10889
10890#ifdef PERLIO_LAYERS
10891 /* Clone PerlIO tables as soon as we can handle general xx_dup() */
10892 PerlIO_clone(aTHX_ proto_perl, param);
10893#endif
10894
10895 PL_envgv = gv_dup(proto_perl->Ienvgv, param);
10896 PL_incgv = gv_dup(proto_perl->Iincgv, param);
10897 PL_hintgv = gv_dup(proto_perl->Ihintgv, param);
10898 PL_origfilename = SAVEPV(proto_perl->Iorigfilename);
10899 PL_diehook = sv_dup_inc(proto_perl->Idiehook, param);
10900 PL_warnhook = sv_dup_inc(proto_perl->Iwarnhook, param);
10901
10902 /* switches */
10903 PL_minus_c = proto_perl->Iminus_c;
10904 PL_patchlevel = sv_dup_inc(proto_perl->Ipatchlevel, param);
10905 PL_localpatches = proto_perl->Ilocalpatches;
10906 PL_splitstr = proto_perl->Isplitstr;
10907 PL_preprocess = proto_perl->Ipreprocess;
10908 PL_minus_n = proto_perl->Iminus_n;
10909 PL_minus_p = proto_perl->Iminus_p;
10910 PL_minus_l = proto_perl->Iminus_l;
10911 PL_minus_a = proto_perl->Iminus_a;
bc9b29db 10912 PL_minus_E = proto_perl->Iminus_E;
bd81e77b
NC
10913 PL_minus_F = proto_perl->Iminus_F;
10914 PL_doswitches = proto_perl->Idoswitches;
10915 PL_dowarn = proto_perl->Idowarn;
10916 PL_doextract = proto_perl->Idoextract;
10917 PL_sawampersand = proto_perl->Isawampersand;
10918 PL_unsafe = proto_perl->Iunsafe;
10919 PL_inplace = SAVEPV(proto_perl->Iinplace);
10920 PL_e_script = sv_dup_inc(proto_perl->Ie_script, param);
10921 PL_perldb = proto_perl->Iperldb;
10922 PL_perl_destruct_level = proto_perl->Iperl_destruct_level;
10923 PL_exit_flags = proto_perl->Iexit_flags;
10924
10925 /* magical thingies */
10926 /* XXX time(&PL_basetime) when asked for? */
10927 PL_basetime = proto_perl->Ibasetime;
10928 PL_formfeed = sv_dup(proto_perl->Iformfeed, param);
10929
10930 PL_maxsysfd = proto_perl->Imaxsysfd;
bd81e77b
NC
10931 PL_statusvalue = proto_perl->Istatusvalue;
10932#ifdef VMS
10933 PL_statusvalue_vms = proto_perl->Istatusvalue_vms;
10934#else
10935 PL_statusvalue_posix = proto_perl->Istatusvalue_posix;
10936#endif
10937 PL_encoding = sv_dup(proto_perl->Iencoding, param);
10938
10939 sv_setpvn(PERL_DEBUG_PAD(0), "", 0); /* For regex debugging. */
10940 sv_setpvn(PERL_DEBUG_PAD(1), "", 0); /* ext/re needs these */
10941 sv_setpvn(PERL_DEBUG_PAD(2), "", 0); /* even without DEBUGGING. */
10942
84da74a7
YO
10943
10944 /* RE engine - function pointers -- must initilize these before
10945 re_dup() is called. dmq. */
10946 PL_regcompp = proto_perl->Tregcompp;
10947 PL_regexecp = proto_perl->Tregexecp;
10948 PL_regint_start = proto_perl->Tregint_start;
10949 PL_regint_string = proto_perl->Tregint_string;
10950 PL_regfree = proto_perl->Tregfree;
10951 PL_regdupe = proto_perl->Tregdupe;
10952
10953 Zero(&PL_reg_state, 1, struct re_save_state);
10954 PL_reginterp_cnt = 0;
10955 PL_regmatch_slab = NULL;
10956
bd81e77b
NC
10957 /* Clone the regex array */
10958 PL_regex_padav = newAV();
10959 {
10960 const I32 len = av_len((AV*)proto_perl->Iregex_padav);
7a5b473e 10961 SV* const * const regexen = AvARRAY((AV*)proto_perl->Iregex_padav);
bd81e77b 10962 IV i;
7f466ec7 10963 av_push(PL_regex_padav, sv_dup_inc_NN(regexen[0],param));
bd81e77b 10964 for(i = 1; i <= len; i++) {
7a5b473e
AL
10965 const SV * const regex = regexen[i];
10966 SV * const sv =
10967 SvREPADTMP(regex)
10968 ? sv_dup_inc(regex, param)
10969 : SvREFCNT_inc(
10970 newSViv(PTR2IV(re_dup(
10971 INT2PTR(REGEXP *, SvIVX(regex)), param))))
10972 ;
10973 av_push(PL_regex_padav, sv);
bd81e77b
NC
10974 }
10975 }
10976 PL_regex_pad = AvARRAY(PL_regex_padav);
10977
10978 /* shortcuts to various I/O objects */
10979 PL_stdingv = gv_dup(proto_perl->Istdingv, param);
10980 PL_stderrgv = gv_dup(proto_perl->Istderrgv, param);
10981 PL_defgv = gv_dup(proto_perl->Idefgv, param);
10982 PL_argvgv = gv_dup(proto_perl->Iargvgv, param);
10983 PL_argvoutgv = gv_dup(proto_perl->Iargvoutgv, param);
10984 PL_argvout_stack = av_dup_inc(proto_perl->Iargvout_stack, param);
1d7c1841 10985
bd81e77b
NC
10986 /* shortcuts to regexp stuff */
10987 PL_replgv = gv_dup(proto_perl->Ireplgv, param);
9660f481 10988
bd81e77b
NC
10989 /* shortcuts to misc objects */
10990 PL_errgv = gv_dup(proto_perl->Ierrgv, param);
9660f481 10991
bd81e77b
NC
10992 /* shortcuts to debugging objects */
10993 PL_DBgv = gv_dup(proto_perl->IDBgv, param);
10994 PL_DBline = gv_dup(proto_perl->IDBline, param);
10995 PL_DBsub = gv_dup(proto_perl->IDBsub, param);
10996 PL_DBsingle = sv_dup(proto_perl->IDBsingle, param);
10997 PL_DBtrace = sv_dup(proto_perl->IDBtrace, param);
10998 PL_DBsignal = sv_dup(proto_perl->IDBsignal, param);
10999 PL_DBassertion = sv_dup(proto_perl->IDBassertion, param);
11000 PL_lineary = av_dup(proto_perl->Ilineary, param);
11001 PL_dbargs = av_dup(proto_perl->Idbargs, param);
9660f481 11002
bd81e77b
NC
11003 /* symbol tables */
11004 PL_defstash = hv_dup_inc(proto_perl->Tdefstash, param);
11005 PL_curstash = hv_dup(proto_perl->Tcurstash, param);
11006 PL_debstash = hv_dup(proto_perl->Idebstash, param);
11007 PL_globalstash = hv_dup(proto_perl->Iglobalstash, param);
11008 PL_curstname = sv_dup_inc(proto_perl->Icurstname, param);
11009
11010 PL_beginav = av_dup_inc(proto_perl->Ibeginav, param);
11011 PL_beginav_save = av_dup_inc(proto_perl->Ibeginav_save, param);
11012 PL_checkav_save = av_dup_inc(proto_perl->Icheckav_save, param);
11013 PL_endav = av_dup_inc(proto_perl->Iendav, param);
11014 PL_checkav = av_dup_inc(proto_perl->Icheckav, param);
11015 PL_initav = av_dup_inc(proto_perl->Iinitav, param);
11016
11017 PL_sub_generation = proto_perl->Isub_generation;
11018
11019 /* funky return mechanisms */
11020 PL_forkprocess = proto_perl->Iforkprocess;
11021
11022 /* subprocess state */
11023 PL_fdpid = av_dup_inc(proto_perl->Ifdpid, param);
11024
11025 /* internal state */
11026 PL_maxo = proto_perl->Imaxo;
11027 if (proto_perl->Iop_mask)
11028 PL_op_mask = SAVEPVN(proto_perl->Iop_mask, PL_maxo);
11029 else
bd61b366 11030 PL_op_mask = NULL;
bd81e77b
NC
11031 /* PL_asserting = proto_perl->Iasserting; */
11032
11033 /* current interpreter roots */
11034 PL_main_cv = cv_dup_inc(proto_perl->Imain_cv, param);
11035 PL_main_root = OpREFCNT_inc(proto_perl->Imain_root);
11036 PL_main_start = proto_perl->Imain_start;
11037 PL_eval_root = proto_perl->Ieval_root;
11038 PL_eval_start = proto_perl->Ieval_start;
11039
11040 /* runtime control stuff */
11041 PL_curcopdb = (COP*)any_dup(proto_perl->Icurcopdb, proto_perl);
11042 PL_copline = proto_perl->Icopline;
11043
11044 PL_filemode = proto_perl->Ifilemode;
11045 PL_lastfd = proto_perl->Ilastfd;
11046 PL_oldname = proto_perl->Ioldname; /* XXX not quite right */
11047 PL_Argv = NULL;
bd61b366 11048 PL_Cmd = NULL;
bd81e77b
NC
11049 PL_gensym = proto_perl->Igensym;
11050 PL_preambled = proto_perl->Ipreambled;
11051 PL_preambleav = av_dup_inc(proto_perl->Ipreambleav, param);
11052 PL_laststatval = proto_perl->Ilaststatval;
11053 PL_laststype = proto_perl->Ilaststype;
a0714e2c 11054 PL_mess_sv = NULL;
bd81e77b
NC
11055
11056 PL_ors_sv = sv_dup_inc(proto_perl->Iors_sv, param);
11057
11058 /* interpreter atexit processing */
11059 PL_exitlistlen = proto_perl->Iexitlistlen;
11060 if (PL_exitlistlen) {
11061 Newx(PL_exitlist, PL_exitlistlen, PerlExitListEntry);
11062 Copy(proto_perl->Iexitlist, PL_exitlist, PL_exitlistlen, PerlExitListEntry);
9660f481 11063 }
bd81e77b
NC
11064 else
11065 PL_exitlist = (PerlExitListEntry*)NULL;
f16dd614
DM
11066
11067 PL_my_cxt_size = proto_perl->Imy_cxt_size;
4c901e72 11068 if (PL_my_cxt_size) {
f16dd614
DM
11069 Newx(PL_my_cxt_list, PL_my_cxt_size, void *);
11070 Copy(proto_perl->Imy_cxt_list, PL_my_cxt_list, PL_my_cxt_size, void *);
11071 }
11072 else
11073 PL_my_cxt_list = (void**)NULL;
bd81e77b
NC
11074 PL_modglobal = hv_dup_inc(proto_perl->Imodglobal, param);
11075 PL_custom_op_names = hv_dup_inc(proto_perl->Icustom_op_names,param);
11076 PL_custom_op_descs = hv_dup_inc(proto_perl->Icustom_op_descs,param);
11077
11078 PL_profiledata = NULL;
11079 PL_rsfp = fp_dup(proto_perl->Irsfp, '<', param);
11080 /* PL_rsfp_filters entries have fake IoDIRP() */
11081 PL_rsfp_filters = av_dup_inc(proto_perl->Irsfp_filters, param);
9660f481 11082
bd81e77b 11083 PL_compcv = cv_dup(proto_perl->Icompcv, param);
9660f481 11084
bd81e77b 11085 PAD_CLONE_VARS(proto_perl, param);
9660f481 11086
bd81e77b
NC
11087#ifdef HAVE_INTERP_INTERN
11088 sys_intern_dup(&proto_perl->Isys_intern, &PL_sys_intern);
11089#endif
645c22ef 11090
bd81e77b
NC
11091 /* more statics moved here */
11092 PL_generation = proto_perl->Igeneration;
11093 PL_DBcv = cv_dup(proto_perl->IDBcv, param);
645c22ef 11094
bd81e77b
NC
11095 PL_in_clean_objs = proto_perl->Iin_clean_objs;
11096 PL_in_clean_all = proto_perl->Iin_clean_all;
6a78b4db 11097
bd81e77b
NC
11098 PL_uid = proto_perl->Iuid;
11099 PL_euid = proto_perl->Ieuid;
11100 PL_gid = proto_perl->Igid;
11101 PL_egid = proto_perl->Iegid;
11102 PL_nomemok = proto_perl->Inomemok;
11103 PL_an = proto_perl->Ian;
11104 PL_evalseq = proto_perl->Ievalseq;
11105 PL_origenviron = proto_perl->Iorigenviron; /* XXX not quite right */
11106 PL_origalen = proto_perl->Iorigalen;
11107#ifdef PERL_USES_PL_PIDSTATUS
11108 PL_pidstatus = newHV(); /* XXX flag for cloning? */
11109#endif
11110 PL_osname = SAVEPV(proto_perl->Iosname);
11111 PL_sighandlerp = proto_perl->Isighandlerp;
6a78b4db 11112
bd81e77b 11113 PL_runops = proto_perl->Irunops;
6a78b4db 11114
bd81e77b 11115 Copy(proto_perl->Itokenbuf, PL_tokenbuf, 256, char);
6a78b4db 11116
bd81e77b
NC
11117#ifdef CSH
11118 PL_cshlen = proto_perl->Icshlen;
11119 PL_cshname = proto_perl->Icshname; /* XXX never deallocated */
11120#endif
645c22ef 11121
bd81e77b
NC
11122 PL_lex_state = proto_perl->Ilex_state;
11123 PL_lex_defer = proto_perl->Ilex_defer;
11124 PL_lex_expect = proto_perl->Ilex_expect;
11125 PL_lex_formbrack = proto_perl->Ilex_formbrack;
11126 PL_lex_dojoin = proto_perl->Ilex_dojoin;
11127 PL_lex_starts = proto_perl->Ilex_starts;
11128 PL_lex_stuff = sv_dup_inc(proto_perl->Ilex_stuff, param);
11129 PL_lex_repl = sv_dup_inc(proto_perl->Ilex_repl, param);
11130 PL_lex_op = proto_perl->Ilex_op;
11131 PL_lex_inpat = proto_perl->Ilex_inpat;
11132 PL_lex_inwhat = proto_perl->Ilex_inwhat;
11133 PL_lex_brackets = proto_perl->Ilex_brackets;
11134 i = (PL_lex_brackets < 120 ? 120 : PL_lex_brackets);
11135 PL_lex_brackstack = SAVEPVN(proto_perl->Ilex_brackstack,i);
11136 PL_lex_casemods = proto_perl->Ilex_casemods;
11137 i = (PL_lex_casemods < 12 ? 12 : PL_lex_casemods);
11138 PL_lex_casestack = SAVEPVN(proto_perl->Ilex_casestack,i);
645c22ef 11139
5db06880
NC
11140#ifdef PERL_MAD
11141 Copy(proto_perl->Inexttoke, PL_nexttoke, 5, NEXTTOKE);
11142 PL_lasttoke = proto_perl->Ilasttoke;
5336380d
NC
11143 PL_realtokenstart = proto_perl->Irealtokenstart;
11144 PL_faketokens = proto_perl->Ifaketokens;
11145 PL_thismad = proto_perl->Ithismad;
11146 PL_thistoken = proto_perl->Ithistoken;
11147 PL_thisopen = proto_perl->Ithisopen;
11148 PL_thisstuff = proto_perl->Ithisstuff;
11149 PL_thisclose = proto_perl->Ithisclose;
11150 PL_thiswhite = proto_perl->Ithiswhite;
11151 PL_nextwhite = proto_perl->Inextwhite;
11152 PL_skipwhite = proto_perl->Iskipwhite;
11153 PL_endwhite = proto_perl->Iendwhite;
11154 PL_curforce = proto_perl->Icurforce;
5db06880 11155#else
bd81e77b
NC
11156 Copy(proto_perl->Inextval, PL_nextval, 5, YYSTYPE);
11157 Copy(proto_perl->Inexttype, PL_nexttype, 5, I32);
11158 PL_nexttoke = proto_perl->Inexttoke;
5db06880 11159#endif
c43294b8 11160
bd81e77b
NC
11161 /* XXX This is probably masking the deeper issue of why
11162 * SvANY(proto_perl->Ilinestr) can be NULL at this point. For test case:
11163 * http://archive.develooper.com/perl5-porters%40perl.org/msg83298.html
11164 * (A little debugging with a watchpoint on it may help.)
11165 */
11166 if (SvANY(proto_perl->Ilinestr)) {
11167 PL_linestr = sv_dup_inc(proto_perl->Ilinestr, param);
11168 i = proto_perl->Ibufptr - SvPVX_const(proto_perl->Ilinestr);
11169 PL_bufptr = SvPVX(PL_linestr) + (i < 0 ? 0 : i);
11170 i = proto_perl->Ioldbufptr - SvPVX_const(proto_perl->Ilinestr);
11171 PL_oldbufptr = SvPVX(PL_linestr) + (i < 0 ? 0 : i);
11172 i = proto_perl->Ioldoldbufptr - SvPVX_const(proto_perl->Ilinestr);
11173 PL_oldoldbufptr = SvPVX(PL_linestr) + (i < 0 ? 0 : i);
11174 i = proto_perl->Ilinestart - SvPVX_const(proto_perl->Ilinestr);
11175 PL_linestart = SvPVX(PL_linestr) + (i < 0 ? 0 : i);
11176 }
11177 else {
561b68a9 11178 PL_linestr = newSV(79);
bd81e77b
NC
11179 sv_upgrade(PL_linestr,SVt_PVIV);
11180 sv_setpvn(PL_linestr,"",0);
11181 PL_bufptr = PL_oldbufptr = PL_oldoldbufptr = PL_linestart = SvPVX(PL_linestr);
11182 }
11183 PL_bufend = SvPVX(PL_linestr) + SvCUR(PL_linestr);
11184 PL_pending_ident = proto_perl->Ipending_ident;
11185 PL_sublex_info = proto_perl->Isublex_info; /* XXX not quite right */
11186
11187 PL_expect = proto_perl->Iexpect;
11188
11189 PL_multi_start = proto_perl->Imulti_start;
11190 PL_multi_end = proto_perl->Imulti_end;
11191 PL_multi_open = proto_perl->Imulti_open;
11192 PL_multi_close = proto_perl->Imulti_close;
11193
11194 PL_error_count = proto_perl->Ierror_count;
11195 PL_subline = proto_perl->Isubline;
11196 PL_subname = sv_dup_inc(proto_perl->Isubname, param);
c43294b8 11197
bd81e77b
NC
11198 /* XXX See comment on SvANY(proto_perl->Ilinestr) above */
11199 if (SvANY(proto_perl->Ilinestr)) {
11200 i = proto_perl->Ilast_uni - SvPVX_const(proto_perl->Ilinestr);
11201 PL_last_uni = SvPVX(PL_linestr) + (i < 0 ? 0 : i);
11202 i = proto_perl->Ilast_lop - SvPVX_const(proto_perl->Ilinestr);
11203 PL_last_lop = SvPVX(PL_linestr) + (i < 0 ? 0 : i);
11204 PL_last_lop_op = proto_perl->Ilast_lop_op;
11205 }
11206 else {
11207 PL_last_uni = SvPVX(PL_linestr);
11208 PL_last_lop = SvPVX(PL_linestr);
11209 PL_last_lop_op = 0;
11210 }
11211 PL_in_my = proto_perl->Iin_my;
11212 PL_in_my_stash = hv_dup(proto_perl->Iin_my_stash, param);
11213#ifdef FCRYPT
11214 PL_cryptseen = proto_perl->Icryptseen;
11215#endif
1d7c1841 11216
bd81e77b 11217 PL_hints = proto_perl->Ihints;
1d7c1841 11218
bd81e77b 11219 PL_amagic_generation = proto_perl->Iamagic_generation;
d2d73c3e 11220
bd81e77b
NC
11221#ifdef USE_LOCALE_COLLATE
11222 PL_collation_ix = proto_perl->Icollation_ix;
11223 PL_collation_name = SAVEPV(proto_perl->Icollation_name);
11224 PL_collation_standard = proto_perl->Icollation_standard;
11225 PL_collxfrm_base = proto_perl->Icollxfrm_base;
11226 PL_collxfrm_mult = proto_perl->Icollxfrm_mult;
11227#endif /* USE_LOCALE_COLLATE */
1d7c1841 11228
bd81e77b
NC
11229#ifdef USE_LOCALE_NUMERIC
11230 PL_numeric_name = SAVEPV(proto_perl->Inumeric_name);
11231 PL_numeric_standard = proto_perl->Inumeric_standard;
11232 PL_numeric_local = proto_perl->Inumeric_local;
11233 PL_numeric_radix_sv = sv_dup_inc(proto_perl->Inumeric_radix_sv, param);
11234#endif /* !USE_LOCALE_NUMERIC */
1d7c1841 11235
bd81e77b
NC
11236 /* utf8 character classes */
11237 PL_utf8_alnum = sv_dup_inc(proto_perl->Iutf8_alnum, param);
11238 PL_utf8_alnumc = sv_dup_inc(proto_perl->Iutf8_alnumc, param);
11239 PL_utf8_ascii = sv_dup_inc(proto_perl->Iutf8_ascii, param);
11240 PL_utf8_alpha = sv_dup_inc(proto_perl->Iutf8_alpha, param);
11241 PL_utf8_space = sv_dup_inc(proto_perl->Iutf8_space, param);
11242 PL_utf8_cntrl = sv_dup_inc(proto_perl->Iutf8_cntrl, param);
11243 PL_utf8_graph = sv_dup_inc(proto_perl->Iutf8_graph, param);
11244 PL_utf8_digit = sv_dup_inc(proto_perl->Iutf8_digit, param);
11245 PL_utf8_upper = sv_dup_inc(proto_perl->Iutf8_upper, param);
11246 PL_utf8_lower = sv_dup_inc(proto_perl->Iutf8_lower, param);
11247 PL_utf8_print = sv_dup_inc(proto_perl->Iutf8_print, param);
11248 PL_utf8_punct = sv_dup_inc(proto_perl->Iutf8_punct, param);
11249 PL_utf8_xdigit = sv_dup_inc(proto_perl->Iutf8_xdigit, param);
11250 PL_utf8_mark = sv_dup_inc(proto_perl->Iutf8_mark, param);
11251 PL_utf8_toupper = sv_dup_inc(proto_perl->Iutf8_toupper, param);
11252 PL_utf8_totitle = sv_dup_inc(proto_perl->Iutf8_totitle, param);
11253 PL_utf8_tolower = sv_dup_inc(proto_perl->Iutf8_tolower, param);
11254 PL_utf8_tofold = sv_dup_inc(proto_perl->Iutf8_tofold, param);
11255 PL_utf8_idstart = sv_dup_inc(proto_perl->Iutf8_idstart, param);
11256 PL_utf8_idcont = sv_dup_inc(proto_perl->Iutf8_idcont, param);
1d7c1841 11257
bd81e77b
NC
11258 /* Did the locale setup indicate UTF-8? */
11259 PL_utf8locale = proto_perl->Iutf8locale;
11260 /* Unicode features (see perlrun/-C) */
11261 PL_unicode = proto_perl->Iunicode;
1d7c1841 11262
bd81e77b
NC
11263 /* Pre-5.8 signals control */
11264 PL_signals = proto_perl->Isignals;
1d7c1841 11265
bd81e77b
NC
11266 /* times() ticks per second */
11267 PL_clocktick = proto_perl->Iclocktick;
1d7c1841 11268
bd81e77b
NC
11269 /* Recursion stopper for PerlIO_find_layer */
11270 PL_in_load_module = proto_perl->Iin_load_module;
8df990a8 11271
bd81e77b
NC
11272 /* sort() routine */
11273 PL_sort_RealCmp = proto_perl->Isort_RealCmp;
e5dd39fc 11274
bd81e77b
NC
11275 /* Not really needed/useful since the reenrant_retint is "volatile",
11276 * but do it for consistency's sake. */
11277 PL_reentrant_retint = proto_perl->Ireentrant_retint;
1d7c1841 11278
bd81e77b
NC
11279 /* Hooks to shared SVs and locks. */
11280 PL_sharehook = proto_perl->Isharehook;
11281 PL_lockhook = proto_perl->Ilockhook;
11282 PL_unlockhook = proto_perl->Iunlockhook;
11283 PL_threadhook = proto_perl->Ithreadhook;
1d7c1841 11284
bd81e77b
NC
11285 PL_runops_std = proto_perl->Irunops_std;
11286 PL_runops_dbg = proto_perl->Irunops_dbg;
1d7c1841 11287
bd81e77b
NC
11288#ifdef THREADS_HAVE_PIDS
11289 PL_ppid = proto_perl->Ippid;
11290#endif
1d7c1841 11291
bd81e77b 11292 /* swatch cache */
5c284bb0 11293 PL_last_swash_hv = NULL; /* reinits on demand */
bd81e77b
NC
11294 PL_last_swash_klen = 0;
11295 PL_last_swash_key[0]= '\0';
11296 PL_last_swash_tmps = (U8*)NULL;
11297 PL_last_swash_slen = 0;
1d7c1841 11298
bd81e77b
NC
11299 PL_glob_index = proto_perl->Iglob_index;
11300 PL_srand_called = proto_perl->Isrand_called;
11b79775 11301 PL_uudmap[(U32) 'M'] = 0; /* reinits on demand */
bd61b366 11302 PL_bitcount = NULL; /* reinits on demand */
05ec9bb3 11303
bd81e77b
NC
11304 if (proto_perl->Ipsig_pend) {
11305 Newxz(PL_psig_pend, SIG_SIZE, int);
11306 }
11307 else {
11308 PL_psig_pend = (int*)NULL;
11309 }
05ec9bb3 11310
bd81e77b
NC
11311 if (proto_perl->Ipsig_ptr) {
11312 Newxz(PL_psig_ptr, SIG_SIZE, SV*);
11313 Newxz(PL_psig_name, SIG_SIZE, SV*);
11314 for (i = 1; i < SIG_SIZE; i++) {
11315 PL_psig_ptr[i] = sv_dup_inc(proto_perl->Ipsig_ptr[i], param);
11316 PL_psig_name[i] = sv_dup_inc(proto_perl->Ipsig_name[i], param);
11317 }
11318 }
11319 else {
11320 PL_psig_ptr = (SV**)NULL;
11321 PL_psig_name = (SV**)NULL;
11322 }
05ec9bb3 11323
bd81e77b 11324 /* thrdvar.h stuff */
1d7c1841 11325
bd81e77b
NC
11326 if (flags & CLONEf_COPY_STACKS) {
11327 /* next allocation will be PL_tmps_stack[PL_tmps_ix+1] */
11328 PL_tmps_ix = proto_perl->Ttmps_ix;
11329 PL_tmps_max = proto_perl->Ttmps_max;
11330 PL_tmps_floor = proto_perl->Ttmps_floor;
11331 Newxz(PL_tmps_stack, PL_tmps_max, SV*);
11332 i = 0;
11333 while (i <= PL_tmps_ix) {
11334 PL_tmps_stack[i] = sv_dup_inc(proto_perl->Ttmps_stack[i], param);
11335 ++i;
11336 }
d2d73c3e 11337
bd81e77b
NC
11338 /* next PUSHMARK() sets *(PL_markstack_ptr+1) */
11339 i = proto_perl->Tmarkstack_max - proto_perl->Tmarkstack;
11340 Newxz(PL_markstack, i, I32);
11341 PL_markstack_max = PL_markstack + (proto_perl->Tmarkstack_max
11342 - proto_perl->Tmarkstack);
11343 PL_markstack_ptr = PL_markstack + (proto_perl->Tmarkstack_ptr
11344 - proto_perl->Tmarkstack);
11345 Copy(proto_perl->Tmarkstack, PL_markstack,
11346 PL_markstack_ptr - PL_markstack + 1, I32);
d2d73c3e 11347
bd81e77b
NC
11348 /* next push_scope()/ENTER sets PL_scopestack[PL_scopestack_ix]
11349 * NOTE: unlike the others! */
11350 PL_scopestack_ix = proto_perl->Tscopestack_ix;
11351 PL_scopestack_max = proto_perl->Tscopestack_max;
11352 Newxz(PL_scopestack, PL_scopestack_max, I32);
11353 Copy(proto_perl->Tscopestack, PL_scopestack, PL_scopestack_ix, I32);
d419787a 11354
bd81e77b
NC
11355 /* NOTE: si_dup() looks at PL_markstack */
11356 PL_curstackinfo = si_dup(proto_perl->Tcurstackinfo, param);
d2d73c3e 11357
bd81e77b
NC
11358 /* PL_curstack = PL_curstackinfo->si_stack; */
11359 PL_curstack = av_dup(proto_perl->Tcurstack, param);
11360 PL_mainstack = av_dup(proto_perl->Tmainstack, param);
1d7c1841 11361
bd81e77b
NC
11362 /* next PUSHs() etc. set *(PL_stack_sp+1) */
11363 PL_stack_base = AvARRAY(PL_curstack);
11364 PL_stack_sp = PL_stack_base + (proto_perl->Tstack_sp
11365 - proto_perl->Tstack_base);
11366 PL_stack_max = PL_stack_base + AvMAX(PL_curstack);
1d7c1841 11367
bd81e77b
NC
11368 /* next SSPUSHFOO() sets PL_savestack[PL_savestack_ix]
11369 * NOTE: unlike the others! */
11370 PL_savestack_ix = proto_perl->Tsavestack_ix;
11371 PL_savestack_max = proto_perl->Tsavestack_max;
11372 /*Newxz(PL_savestack, PL_savestack_max, ANY);*/
11373 PL_savestack = ss_dup(proto_perl, param);
11374 }
11375 else {
11376 init_stacks();
11377 ENTER; /* perl_destruct() wants to LEAVE; */
34394ecd
DM
11378
11379 /* although we're not duplicating the tmps stack, we should still
11380 * add entries for any SVs on the tmps stack that got cloned by a
11381 * non-refcount means (eg a temp in @_); otherwise they will be
11382 * orphaned
11383 */
11384 for (i = 0; i<= proto_perl->Ttmps_ix; i++) {
6136c704 11385 SV * const nsv = (SV*)ptr_table_fetch(PL_ptr_table,
34394ecd
DM
11386 proto_perl->Ttmps_stack[i]);
11387 if (nsv && !SvREFCNT(nsv)) {
11388 EXTEND_MORTAL(1);
b37c2d43 11389 PL_tmps_stack[++PL_tmps_ix] = SvREFCNT_inc_simple(nsv);
34394ecd
DM
11390 }
11391 }
bd81e77b 11392 }
1d7c1841 11393
bd81e77b
NC
11394 PL_start_env = proto_perl->Tstart_env; /* XXXXXX */
11395 PL_top_env = &PL_start_env;
1d7c1841 11396
bd81e77b 11397 PL_op = proto_perl->Top;
4a4c6fe3 11398
a0714e2c 11399 PL_Sv = NULL;
bd81e77b
NC
11400 PL_Xpv = (XPV*)NULL;
11401 PL_na = proto_perl->Tna;
1fcf4c12 11402
bd81e77b
NC
11403 PL_statbuf = proto_perl->Tstatbuf;
11404 PL_statcache = proto_perl->Tstatcache;
11405 PL_statgv = gv_dup(proto_perl->Tstatgv, param);
11406 PL_statname = sv_dup_inc(proto_perl->Tstatname, param);
11407#ifdef HAS_TIMES
11408 PL_timesbuf = proto_perl->Ttimesbuf;
11409#endif
1d7c1841 11410
bd81e77b
NC
11411 PL_tainted = proto_perl->Ttainted;
11412 PL_curpm = proto_perl->Tcurpm; /* XXX No PMOP ref count */
11413 PL_rs = sv_dup_inc(proto_perl->Trs, param);
11414 PL_last_in_gv = gv_dup(proto_perl->Tlast_in_gv, param);
11415 PL_ofs_sv = sv_dup_inc(proto_perl->Tofs_sv, param);
11416 PL_defoutgv = gv_dup_inc(proto_perl->Tdefoutgv, param);
11417 PL_chopset = proto_perl->Tchopset; /* XXX never deallocated */
11418 PL_toptarget = sv_dup_inc(proto_perl->Ttoptarget, param);
11419 PL_bodytarget = sv_dup_inc(proto_perl->Tbodytarget, param);
11420 PL_formtarget = sv_dup(proto_perl->Tformtarget, param);
1d7c1841 11421
bd81e77b
NC
11422 PL_restartop = proto_perl->Trestartop;
11423 PL_in_eval = proto_perl->Tin_eval;
11424 PL_delaymagic = proto_perl->Tdelaymagic;
11425 PL_dirty = proto_perl->Tdirty;
11426 PL_localizing = proto_perl->Tlocalizing;
1d7c1841 11427
bd81e77b 11428 PL_errors = sv_dup_inc(proto_perl->Terrors, param);
4608196e 11429 PL_hv_fetch_ent_mh = NULL;
bd81e77b 11430 PL_modcount = proto_perl->Tmodcount;
5f66b61c 11431 PL_lastgotoprobe = NULL;
bd81e77b 11432 PL_dumpindent = proto_perl->Tdumpindent;
1d7c1841 11433
bd81e77b
NC
11434 PL_sortcop = (OP*)any_dup(proto_perl->Tsortcop, proto_perl);
11435 PL_sortstash = hv_dup(proto_perl->Tsortstash, param);
11436 PL_firstgv = gv_dup(proto_perl->Tfirstgv, param);
11437 PL_secondgv = gv_dup(proto_perl->Tsecondgv, param);
bd61b366 11438 PL_efloatbuf = NULL; /* reinits on demand */
bd81e77b 11439 PL_efloatsize = 0; /* reinits on demand */
d2d73c3e 11440
bd81e77b 11441 /* regex stuff */
1d7c1841 11442
bd81e77b
NC
11443 PL_screamfirst = NULL;
11444 PL_screamnext = NULL;
11445 PL_maxscream = -1; /* reinits on demand */
a0714e2c 11446 PL_lastscream = NULL;
1d7c1841 11447
bd81e77b 11448 PL_watchaddr = NULL;
bd61b366 11449 PL_watchok = NULL;
1d7c1841 11450
bd81e77b 11451 PL_regdummy = proto_perl->Tregdummy;
bd81e77b
NC
11452 PL_colorset = 0; /* reinits PL_colors[] */
11453 /*PL_colors[6] = {0,0,0,0,0,0};*/
1d7c1841 11454
84da74a7 11455
1d7c1841 11456
bd81e77b
NC
11457 /* Pluggable optimizer */
11458 PL_peepp = proto_perl->Tpeepp;
1d7c1841 11459
bd81e77b 11460 PL_stashcache = newHV();
1d7c1841 11461
bd81e77b
NC
11462 if (!(flags & CLONEf_KEEP_PTR_TABLE)) {
11463 ptr_table_free(PL_ptr_table);
11464 PL_ptr_table = NULL;
11465 }
1d7c1841 11466
bd81e77b
NC
11467 /* Call the ->CLONE method, if it exists, for each of the stashes
11468 identified by sv_dup() above.
11469 */
11470 while(av_len(param->stashes) != -1) {
11471 HV* const stash = (HV*) av_shift(param->stashes);
11472 GV* const cloner = gv_fetchmethod_autoload(stash, "CLONE", 0);
11473 if (cloner && GvCV(cloner)) {
11474 dSP;
11475 ENTER;
11476 SAVETMPS;
11477 PUSHMARK(SP);
11478 XPUSHs(sv_2mortal(newSVhek(HvNAME_HEK(stash))));
11479 PUTBACK;
11480 call_sv((SV*)GvCV(cloner), G_DISCARD);
11481 FREETMPS;
11482 LEAVE;
11483 }
1d7c1841 11484 }
1d7c1841 11485
bd81e77b 11486 SvREFCNT_dec(param->stashes);
1d7c1841 11487
bd81e77b
NC
11488 /* orphaned? eg threads->new inside BEGIN or use */
11489 if (PL_compcv && ! SvREFCNT(PL_compcv)) {
b37c2d43 11490 SvREFCNT_inc_simple_void(PL_compcv);
bd81e77b
NC
11491 SAVEFREESV(PL_compcv);
11492 }
dd2155a4 11493
bd81e77b
NC
11494 return my_perl;
11495}
1d7c1841 11496
bd81e77b 11497#endif /* USE_ITHREADS */
1d7c1841 11498
bd81e77b
NC
11499/*
11500=head1 Unicode Support
1d7c1841 11501
bd81e77b 11502=for apidoc sv_recode_to_utf8
1d7c1841 11503
bd81e77b
NC
11504The encoding is assumed to be an Encode object, on entry the PV
11505of the sv is assumed to be octets in that encoding, and the sv
11506will be converted into Unicode (and UTF-8).
1d7c1841 11507
bd81e77b
NC
11508If the sv already is UTF-8 (or if it is not POK), or if the encoding
11509is not a reference, nothing is done to the sv. If the encoding is not
11510an C<Encode::XS> Encoding object, bad things will happen.
11511(See F<lib/encoding.pm> and L<Encode>).
1d7c1841 11512
bd81e77b 11513The PV of the sv is returned.
1d7c1841 11514
bd81e77b 11515=cut */
1d7c1841 11516
bd81e77b
NC
11517char *
11518Perl_sv_recode_to_utf8(pTHX_ SV *sv, SV *encoding)
11519{
11520 dVAR;
11521 if (SvPOK(sv) && !SvUTF8(sv) && !IN_BYTES && SvROK(encoding)) {
11522 SV *uni;
11523 STRLEN len;
11524 const char *s;
11525 dSP;
11526 ENTER;
11527 SAVETMPS;
11528 save_re_context();
11529 PUSHMARK(sp);
11530 EXTEND(SP, 3);
11531 XPUSHs(encoding);
11532 XPUSHs(sv);
11533/*
11534 NI-S 2002/07/09
11535 Passing sv_yes is wrong - it needs to be or'ed set of constants
11536 for Encode::XS, while UTf-8 decode (currently) assumes a true value means
11537 remove converted chars from source.
1d7c1841 11538
bd81e77b 11539 Both will default the value - let them.
1d7c1841 11540
bd81e77b
NC
11541 XPUSHs(&PL_sv_yes);
11542*/
11543 PUTBACK;
11544 call_method("decode", G_SCALAR);
11545 SPAGAIN;
11546 uni = POPs;
11547 PUTBACK;
11548 s = SvPV_const(uni, len);
11549 if (s != SvPVX_const(sv)) {
11550 SvGROW(sv, len + 1);
11551 Move(s, SvPVX(sv), len + 1, char);
11552 SvCUR_set(sv, len);
11553 }
11554 FREETMPS;
11555 LEAVE;
11556 SvUTF8_on(sv);
11557 return SvPVX(sv);
389edf32 11558 }
bd81e77b
NC
11559 return SvPOKp(sv) ? SvPVX(sv) : NULL;
11560}
1d7c1841 11561
bd81e77b
NC
11562/*
11563=for apidoc sv_cat_decode
1d7c1841 11564
bd81e77b
NC
11565The encoding is assumed to be an Encode object, the PV of the ssv is
11566assumed to be octets in that encoding and decoding the input starts
11567from the position which (PV + *offset) pointed to. The dsv will be
11568concatenated the decoded UTF-8 string from ssv. Decoding will terminate
11569when the string tstr appears in decoding output or the input ends on
11570the PV of the ssv. The value which the offset points will be modified
11571to the last input position on the ssv.
1d7c1841 11572
bd81e77b 11573Returns TRUE if the terminator was found, else returns FALSE.
1d7c1841 11574
bd81e77b
NC
11575=cut */
11576
11577bool
11578Perl_sv_cat_decode(pTHX_ SV *dsv, SV *encoding,
11579 SV *ssv, int *offset, char *tstr, int tlen)
11580{
11581 dVAR;
11582 bool ret = FALSE;
11583 if (SvPOK(ssv) && SvPOK(dsv) && SvROK(encoding) && offset) {
11584 SV *offsv;
11585 dSP;
11586 ENTER;
11587 SAVETMPS;
11588 save_re_context();
11589 PUSHMARK(sp);
11590 EXTEND(SP, 6);
11591 XPUSHs(encoding);
11592 XPUSHs(dsv);
11593 XPUSHs(ssv);
11594 XPUSHs(offsv = sv_2mortal(newSViv(*offset)));
11595 XPUSHs(sv_2mortal(newSVpvn(tstr, tlen)));
11596 PUTBACK;
11597 call_method("cat_decode", G_SCALAR);
11598 SPAGAIN;
11599 ret = SvTRUE(TOPs);
11600 *offset = SvIV(offsv);
11601 PUTBACK;
11602 FREETMPS;
11603 LEAVE;
389edf32 11604 }
bd81e77b
NC
11605 else
11606 Perl_croak(aTHX_ "Invalid argument to sv_cat_decode");
11607 return ret;
1d7c1841 11608
bd81e77b 11609}
1d7c1841 11610
bd81e77b
NC
11611/* ---------------------------------------------------------------------
11612 *
11613 * support functions for report_uninit()
11614 */
1d7c1841 11615
bd81e77b
NC
11616/* the maxiumum size of array or hash where we will scan looking
11617 * for the undefined element that triggered the warning */
1d7c1841 11618
bd81e77b 11619#define FUV_MAX_SEARCH_SIZE 1000
1d7c1841 11620
bd81e77b
NC
11621/* Look for an entry in the hash whose value has the same SV as val;
11622 * If so, return a mortal copy of the key. */
1d7c1841 11623
bd81e77b
NC
11624STATIC SV*
11625S_find_hash_subscript(pTHX_ HV *hv, SV* val)
11626{
11627 dVAR;
11628 register HE **array;
11629 I32 i;
6c3182a5 11630
bd81e77b
NC
11631 if (!hv || SvMAGICAL(hv) || !HvARRAY(hv) ||
11632 (HvTOTALKEYS(hv) > FUV_MAX_SEARCH_SIZE))
a0714e2c 11633 return NULL;
6c3182a5 11634
bd81e77b 11635 array = HvARRAY(hv);
6c3182a5 11636
bd81e77b
NC
11637 for (i=HvMAX(hv); i>0; i--) {
11638 register HE *entry;
11639 for (entry = array[i]; entry; entry = HeNEXT(entry)) {
11640 if (HeVAL(entry) != val)
11641 continue;
11642 if ( HeVAL(entry) == &PL_sv_undef ||
11643 HeVAL(entry) == &PL_sv_placeholder)
11644 continue;
11645 if (!HeKEY(entry))
a0714e2c 11646 return NULL;
bd81e77b
NC
11647 if (HeKLEN(entry) == HEf_SVKEY)
11648 return sv_mortalcopy(HeKEY_sv(entry));
11649 return sv_2mortal(newSVpvn(HeKEY(entry), HeKLEN(entry)));
11650 }
11651 }
a0714e2c 11652 return NULL;
bd81e77b 11653}
6c3182a5 11654
bd81e77b
NC
11655/* Look for an entry in the array whose value has the same SV as val;
11656 * If so, return the index, otherwise return -1. */
6c3182a5 11657
bd81e77b
NC
11658STATIC I32
11659S_find_array_subscript(pTHX_ AV *av, SV* val)
11660{
97aff369 11661 dVAR;
bd81e77b
NC
11662 if (!av || SvMAGICAL(av) || !AvARRAY(av) ||
11663 (AvFILLp(av) > FUV_MAX_SEARCH_SIZE))
11664 return -1;
57c6e6d2 11665
4a021917
AL
11666 if (val != &PL_sv_undef) {
11667 SV ** const svp = AvARRAY(av);
11668 I32 i;
11669
11670 for (i=AvFILLp(av); i>=0; i--)
11671 if (svp[i] == val)
11672 return i;
bd81e77b
NC
11673 }
11674 return -1;
11675}
15a5279a 11676
bd81e77b
NC
11677/* S_varname(): return the name of a variable, optionally with a subscript.
11678 * If gv is non-zero, use the name of that global, along with gvtype (one
11679 * of "$", "@", "%"); otherwise use the name of the lexical at pad offset
11680 * targ. Depending on the value of the subscript_type flag, return:
11681 */
bce260cd 11682
bd81e77b
NC
11683#define FUV_SUBSCRIPT_NONE 1 /* "@foo" */
11684#define FUV_SUBSCRIPT_ARRAY 2 /* "$foo[aindex]" */
11685#define FUV_SUBSCRIPT_HASH 3 /* "$foo{keyname}" */
11686#define FUV_SUBSCRIPT_WITHIN 4 /* "within @foo" */
bce260cd 11687
bd81e77b
NC
11688STATIC SV*
11689S_varname(pTHX_ GV *gv, const char gvtype, PADOFFSET targ,
11690 SV* keyname, I32 aindex, int subscript_type)
11691{
1d7c1841 11692
bd81e77b
NC
11693 SV * const name = sv_newmortal();
11694 if (gv) {
11695 char buffer[2];
11696 buffer[0] = gvtype;
11697 buffer[1] = 0;
1d7c1841 11698
bd81e77b 11699 /* as gv_fullname4(), but add literal '^' for $^FOO names */
66fe0623 11700
bd81e77b 11701 gv_fullname4(name, gv, buffer, 0);
1d7c1841 11702
bd81e77b
NC
11703 if ((unsigned int)SvPVX(name)[1] <= 26) {
11704 buffer[0] = '^';
11705 buffer[1] = SvPVX(name)[1] + 'A' - 1;
1d7c1841 11706
bd81e77b
NC
11707 /* Swap the 1 unprintable control character for the 2 byte pretty
11708 version - ie substr($name, 1, 1) = $buffer; */
11709 sv_insert(name, 1, 1, buffer, 2);
1d7c1841 11710 }
bd81e77b
NC
11711 }
11712 else {
11713 U32 unused;
11714 CV * const cv = find_runcv(&unused);
11715 SV *sv;
11716 AV *av;
1d7c1841 11717
bd81e77b 11718 if (!cv || !CvPADLIST(cv))
a0714e2c 11719 return NULL;
bd81e77b
NC
11720 av = (AV*)(*av_fetch(CvPADLIST(cv), 0, FALSE));
11721 sv = *av_fetch(av, targ, FALSE);
11722 /* SvLEN in a pad name is not to be trusted */
11723 sv_setpv(name, SvPV_nolen_const(sv));
11724 }
1d7c1841 11725
bd81e77b 11726 if (subscript_type == FUV_SUBSCRIPT_HASH) {
561b68a9 11727 SV * const sv = newSV(0);
bd81e77b
NC
11728 *SvPVX(name) = '$';
11729 Perl_sv_catpvf(aTHX_ name, "{%s}",
11730 pv_display(sv,SvPVX_const(keyname), SvCUR(keyname), 0, 32));
11731 SvREFCNT_dec(sv);
11732 }
11733 else if (subscript_type == FUV_SUBSCRIPT_ARRAY) {
11734 *SvPVX(name) = '$';
11735 Perl_sv_catpvf(aTHX_ name, "[%"IVdf"]", (IV)aindex);
11736 }
11737 else if (subscript_type == FUV_SUBSCRIPT_WITHIN)
89529cee 11738 Perl_sv_insert(aTHX_ name, 0, 0, STR_WITH_LEN("within "));
1d7c1841 11739
bd81e77b
NC
11740 return name;
11741}
1d7c1841 11742
1d7c1841 11743
bd81e77b
NC
11744/*
11745=for apidoc find_uninit_var
1d7c1841 11746
bd81e77b
NC
11747Find the name of the undefined variable (if any) that caused the operator o
11748to issue a "Use of uninitialized value" warning.
11749If match is true, only return a name if it's value matches uninit_sv.
11750So roughly speaking, if a unary operator (such as OP_COS) generates a
11751warning, then following the direct child of the op may yield an
11752OP_PADSV or OP_GV that gives the name of the undefined variable. On the
11753other hand, with OP_ADD there are two branches to follow, so we only print
11754the variable name if we get an exact match.
1d7c1841 11755
bd81e77b 11756The name is returned as a mortal SV.
1d7c1841 11757
bd81e77b
NC
11758Assumes that PL_op is the op that originally triggered the error, and that
11759PL_comppad/PL_curpad points to the currently executing pad.
1d7c1841 11760
bd81e77b
NC
11761=cut
11762*/
1d7c1841 11763
bd81e77b
NC
11764STATIC SV *
11765S_find_uninit_var(pTHX_ OP* obase, SV* uninit_sv, bool match)
11766{
11767 dVAR;
11768 SV *sv;
11769 AV *av;
11770 GV *gv;
11771 OP *o, *o2, *kid;
1d7c1841 11772
bd81e77b
NC
11773 if (!obase || (match && (!uninit_sv || uninit_sv == &PL_sv_undef ||
11774 uninit_sv == &PL_sv_placeholder)))
a0714e2c 11775 return NULL;
1d7c1841 11776
bd81e77b 11777 switch (obase->op_type) {
1d7c1841 11778
bd81e77b
NC
11779 case OP_RV2AV:
11780 case OP_RV2HV:
11781 case OP_PADAV:
11782 case OP_PADHV:
11783 {
11784 const bool pad = (obase->op_type == OP_PADAV || obase->op_type == OP_PADHV);
11785 const bool hash = (obase->op_type == OP_PADHV || obase->op_type == OP_RV2HV);
11786 I32 index = 0;
a0714e2c 11787 SV *keysv = NULL;
bd81e77b 11788 int subscript_type = FUV_SUBSCRIPT_WITHIN;
1d7c1841 11789
bd81e77b
NC
11790 if (pad) { /* @lex, %lex */
11791 sv = PAD_SVl(obase->op_targ);
a0714e2c 11792 gv = NULL;
bd81e77b
NC
11793 }
11794 else {
11795 if (cUNOPx(obase)->op_first->op_type == OP_GV) {
11796 /* @global, %global */
11797 gv = cGVOPx_gv(cUNOPx(obase)->op_first);
11798 if (!gv)
11799 break;
11800 sv = hash ? (SV*)GvHV(gv): (SV*)GvAV(gv);
11801 }
11802 else /* @{expr}, %{expr} */
11803 return find_uninit_var(cUNOPx(obase)->op_first,
11804 uninit_sv, match);
11805 }
1d7c1841 11806
bd81e77b
NC
11807 /* attempt to find a match within the aggregate */
11808 if (hash) {
d4c19fe8 11809 keysv = find_hash_subscript((HV*)sv, uninit_sv);
bd81e77b
NC
11810 if (keysv)
11811 subscript_type = FUV_SUBSCRIPT_HASH;
11812 }
11813 else {
e15d5972 11814 index = find_array_subscript((AV*)sv, uninit_sv);
bd81e77b
NC
11815 if (index >= 0)
11816 subscript_type = FUV_SUBSCRIPT_ARRAY;
11817 }
1d7c1841 11818
bd81e77b
NC
11819 if (match && subscript_type == FUV_SUBSCRIPT_WITHIN)
11820 break;
1d7c1841 11821
bd81e77b
NC
11822 return varname(gv, hash ? '%' : '@', obase->op_targ,
11823 keysv, index, subscript_type);
11824 }
1d7c1841 11825
bd81e77b
NC
11826 case OP_PADSV:
11827 if (match && PAD_SVl(obase->op_targ) != uninit_sv)
11828 break;
a0714e2c
SS
11829 return varname(NULL, '$', obase->op_targ,
11830 NULL, 0, FUV_SUBSCRIPT_NONE);
1d7c1841 11831
bd81e77b
NC
11832 case OP_GVSV:
11833 gv = cGVOPx_gv(obase);
11834 if (!gv || (match && GvSV(gv) != uninit_sv))
11835 break;
a0714e2c 11836 return varname(gv, '$', 0, NULL, 0, FUV_SUBSCRIPT_NONE);
1d7c1841 11837
bd81e77b
NC
11838 case OP_AELEMFAST:
11839 if (obase->op_flags & OPf_SPECIAL) { /* lexical array */
11840 if (match) {
11841 SV **svp;
11842 av = (AV*)PAD_SV(obase->op_targ);
11843 if (!av || SvRMAGICAL(av))
11844 break;
11845 svp = av_fetch(av, (I32)obase->op_private, FALSE);
11846 if (!svp || *svp != uninit_sv)
11847 break;
11848 }
a0714e2c
SS
11849 return varname(NULL, '$', obase->op_targ,
11850 NULL, (I32)obase->op_private, FUV_SUBSCRIPT_ARRAY);
bd81e77b
NC
11851 }
11852 else {
11853 gv = cGVOPx_gv(obase);
11854 if (!gv)
11855 break;
11856 if (match) {
11857 SV **svp;
11858 av = GvAV(gv);
11859 if (!av || SvRMAGICAL(av))
11860 break;
11861 svp = av_fetch(av, (I32)obase->op_private, FALSE);
11862 if (!svp || *svp != uninit_sv)
11863 break;
11864 }
11865 return varname(gv, '$', 0,
a0714e2c 11866 NULL, (I32)obase->op_private, FUV_SUBSCRIPT_ARRAY);
bd81e77b
NC
11867 }
11868 break;
1d7c1841 11869
bd81e77b
NC
11870 case OP_EXISTS:
11871 o = cUNOPx(obase)->op_first;
11872 if (!o || o->op_type != OP_NULL ||
11873 ! (o->op_targ == OP_AELEM || o->op_targ == OP_HELEM))
11874 break;
11875 return find_uninit_var(cBINOPo->op_last, uninit_sv, match);
a2efc822 11876
bd81e77b
NC
11877 case OP_AELEM:
11878 case OP_HELEM:
11879 if (PL_op == obase)
11880 /* $a[uninit_expr] or $h{uninit_expr} */
11881 return find_uninit_var(cBINOPx(obase)->op_last, uninit_sv, match);
081fc587 11882
a0714e2c 11883 gv = NULL;
bd81e77b
NC
11884 o = cBINOPx(obase)->op_first;
11885 kid = cBINOPx(obase)->op_last;
8cf8f3d1 11886
bd81e77b 11887 /* get the av or hv, and optionally the gv */
a0714e2c 11888 sv = NULL;
bd81e77b
NC
11889 if (o->op_type == OP_PADAV || o->op_type == OP_PADHV) {
11890 sv = PAD_SV(o->op_targ);
11891 }
11892 else if ((o->op_type == OP_RV2AV || o->op_type == OP_RV2HV)
11893 && cUNOPo->op_first->op_type == OP_GV)
11894 {
11895 gv = cGVOPx_gv(cUNOPo->op_first);
11896 if (!gv)
11897 break;
11898 sv = o->op_type == OP_RV2HV ? (SV*)GvHV(gv) : (SV*)GvAV(gv);
11899 }
11900 if (!sv)
11901 break;
11902
11903 if (kid && kid->op_type == OP_CONST && SvOK(cSVOPx_sv(kid))) {
11904 /* index is constant */
11905 if (match) {
11906 if (SvMAGICAL(sv))
11907 break;
11908 if (obase->op_type == OP_HELEM) {
11909 HE* he = hv_fetch_ent((HV*)sv, cSVOPx_sv(kid), 0, 0);
11910 if (!he || HeVAL(he) != uninit_sv)
11911 break;
11912 }
11913 else {
00b6aa41 11914 SV * const * const svp = av_fetch((AV*)sv, SvIV(cSVOPx_sv(kid)), FALSE);
bd81e77b
NC
11915 if (!svp || *svp != uninit_sv)
11916 break;
11917 }
11918 }
11919 if (obase->op_type == OP_HELEM)
11920 return varname(gv, '%', o->op_targ,
11921 cSVOPx_sv(kid), 0, FUV_SUBSCRIPT_HASH);
11922 else
a0714e2c 11923 return varname(gv, '@', o->op_targ, NULL,
bd81e77b 11924 SvIV(cSVOPx_sv(kid)), FUV_SUBSCRIPT_ARRAY);
bd81e77b
NC
11925 }
11926 else {
11927 /* index is an expression;
11928 * attempt to find a match within the aggregate */
11929 if (obase->op_type == OP_HELEM) {
d4c19fe8 11930 SV * const keysv = find_hash_subscript((HV*)sv, uninit_sv);
bd81e77b
NC
11931 if (keysv)
11932 return varname(gv, '%', o->op_targ,
11933 keysv, 0, FUV_SUBSCRIPT_HASH);
11934 }
11935 else {
d4c19fe8 11936 const I32 index = find_array_subscript((AV*)sv, uninit_sv);
bd81e77b
NC
11937 if (index >= 0)
11938 return varname(gv, '@', o->op_targ,
a0714e2c 11939 NULL, index, FUV_SUBSCRIPT_ARRAY);
bd81e77b
NC
11940 }
11941 if (match)
11942 break;
11943 return varname(gv,
11944 (o->op_type == OP_PADAV || o->op_type == OP_RV2AV)
11945 ? '@' : '%',
a0714e2c 11946 o->op_targ, NULL, 0, FUV_SUBSCRIPT_WITHIN);
f284b03f 11947 }
bd81e77b 11948 break;
dc507217 11949
bd81e77b
NC
11950 case OP_AASSIGN:
11951 /* only examine RHS */
11952 return find_uninit_var(cBINOPx(obase)->op_first, uninit_sv, match);
6d26897e 11953
bd81e77b
NC
11954 case OP_OPEN:
11955 o = cUNOPx(obase)->op_first;
11956 if (o->op_type == OP_PUSHMARK)
11957 o = o->op_sibling;
1d7c1841 11958
bd81e77b
NC
11959 if (!o->op_sibling) {
11960 /* one-arg version of open is highly magical */
a0ae6670 11961
bd81e77b
NC
11962 if (o->op_type == OP_GV) { /* open FOO; */
11963 gv = cGVOPx_gv(o);
11964 if (match && GvSV(gv) != uninit_sv)
11965 break;
11966 return varname(gv, '$', 0,
a0714e2c 11967 NULL, 0, FUV_SUBSCRIPT_NONE);
bd81e77b
NC
11968 }
11969 /* other possibilities not handled are:
11970 * open $x; or open my $x; should return '${*$x}'
11971 * open expr; should return '$'.expr ideally
11972 */
11973 break;
11974 }
11975 goto do_op;
ccfc67b7 11976
bd81e77b
NC
11977 /* ops where $_ may be an implicit arg */
11978 case OP_TRANS:
11979 case OP_SUBST:
11980 case OP_MATCH:
11981 if ( !(obase->op_flags & OPf_STACKED)) {
11982 if (uninit_sv == ((obase->op_private & OPpTARGET_MY)
11983 ? PAD_SVl(obase->op_targ)
11984 : DEFSV))
11985 {
11986 sv = sv_newmortal();
11987 sv_setpvn(sv, "$_", 2);
11988 return sv;
11989 }
11990 }
11991 goto do_op;
9f4817db 11992
bd81e77b
NC
11993 case OP_PRTF:
11994 case OP_PRINT:
11995 /* skip filehandle as it can't produce 'undef' warning */
11996 o = cUNOPx(obase)->op_first;
11997 if ((obase->op_flags & OPf_STACKED) && o->op_type == OP_PUSHMARK)
11998 o = o->op_sibling->op_sibling;
11999 goto do_op2;
9f4817db 12000
9f4817db 12001
bd81e77b
NC
12002 case OP_RV2SV:
12003 case OP_CUSTOM:
12004 case OP_ENTERSUB:
12005 match = 1; /* XS or custom code could trigger random warnings */
12006 goto do_op;
9f4817db 12007
bd81e77b
NC
12008 case OP_SCHOMP:
12009 case OP_CHOMP:
12010 if (SvROK(PL_rs) && uninit_sv == SvRV(PL_rs))
396482e1 12011 return sv_2mortal(newSVpvs("${$/}"));
5f66b61c 12012 /*FALLTHROUGH*/
5d170f3a 12013
bd81e77b
NC
12014 default:
12015 do_op:
12016 if (!(obase->op_flags & OPf_KIDS))
12017 break;
12018 o = cUNOPx(obase)->op_first;
12019
12020 do_op2:
12021 if (!o)
12022 break;
f9893866 12023
bd81e77b
NC
12024 /* if all except one arg are constant, or have no side-effects,
12025 * or are optimized away, then it's unambiguous */
5f66b61c 12026 o2 = NULL;
bd81e77b 12027 for (kid=o; kid; kid = kid->op_sibling) {
e15d5972
AL
12028 if (kid) {
12029 const OPCODE type = kid->op_type;
12030 if ( (type == OP_CONST && SvOK(cSVOPx_sv(kid)))
12031 || (type == OP_NULL && ! (kid->op_flags & OPf_KIDS))
12032 || (type == OP_PUSHMARK)
bd81e77b 12033 )
bd81e77b 12034 continue;
e15d5972 12035 }
bd81e77b 12036 if (o2) { /* more than one found */
5f66b61c 12037 o2 = NULL;
bd81e77b
NC
12038 break;
12039 }
12040 o2 = kid;
12041 }
12042 if (o2)
12043 return find_uninit_var(o2, uninit_sv, match);
7a5fa8a2 12044
bd81e77b
NC
12045 /* scan all args */
12046 while (o) {
12047 sv = find_uninit_var(o, uninit_sv, 1);
12048 if (sv)
12049 return sv;
12050 o = o->op_sibling;
d0063567 12051 }
bd81e77b 12052 break;
f9893866 12053 }
a0714e2c 12054 return NULL;
9f4817db
JH
12055}
12056
220e2d4e 12057
bd81e77b
NC
12058/*
12059=for apidoc report_uninit
68795e93 12060
bd81e77b 12061Print appropriate "Use of uninitialized variable" warning
220e2d4e 12062
bd81e77b
NC
12063=cut
12064*/
220e2d4e 12065
bd81e77b
NC
12066void
12067Perl_report_uninit(pTHX_ SV* uninit_sv)
220e2d4e 12068{
97aff369 12069 dVAR;
bd81e77b 12070 if (PL_op) {
a0714e2c 12071 SV* varname = NULL;
bd81e77b
NC
12072 if (uninit_sv) {
12073 varname = find_uninit_var(PL_op, uninit_sv,0);
12074 if (varname)
12075 sv_insert(varname, 0, 0, " ", 1);
12076 }
12077 Perl_warner(aTHX_ packWARN(WARN_UNINITIALIZED), PL_warn_uninit,
12078 varname ? SvPV_nolen_const(varname) : "",
12079 " in ", OP_DESC(PL_op));
220e2d4e 12080 }
a73e8557 12081 else
bd81e77b
NC
12082 Perl_warner(aTHX_ packWARN(WARN_UNINITIALIZED), PL_warn_uninit,
12083 "", "", "");
220e2d4e 12084}
f9893866 12085
241d1a3b
NC
12086/*
12087 * Local variables:
12088 * c-indentation-style: bsd
12089 * c-basic-offset: 4
12090 * indent-tabs-mode: t
12091 * End:
12092 *
37442d52
RGS
12093 * ex: set ts=8 sts=4 sw=4 noet:
12094 */