This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Use test.pl functions in t/op/groups.t
[perl5.git] / inline.h
CommitLineData
25468daa
FC
1/* inline.h
2 *
3 * Copyright (C) 2012 by Larry Wall and others
4 *
5 * You may distribute under the terms of either the GNU General Public
6 * License or the Artistic License, as specified in the README file.
7 *
8 * This file is a home for static inline functions that cannot go in other
9 * headers files, because they depend on proto.h (included after most other
10 * headers) or struct definitions.
11 *
12 * Each section names the header file that the functions "belong" to.
13 */
27669aa4 14
be3a7a5d
KW
15/* ------------------------------- av.h ------------------------------- */
16
c70927a6 17PERL_STATIC_INLINE SSize_t
be3a7a5d
KW
18S_av_top_index(pTHX_ AV *av)
19{
20 PERL_ARGS_ASSERT_AV_TOP_INDEX;
21 assert(SvTYPE(av) == SVt_PVAV);
22
23 return AvFILL(av);
24}
25
1afe1db1
FC
26/* ------------------------------- cv.h ------------------------------- */
27
ae77754a
FC
28PERL_STATIC_INLINE GV *
29S_CvGV(pTHX_ CV *sv)
30{
31 return CvNAMED(sv)
32 ? Perl_cvgv_from_hek(aTHX_ sv)
33 : ((XPVCV*)MUTABLE_PTR(SvANY(sv)))->xcv_gv_u.xcv_gv;
34}
35
1afe1db1
FC
36PERL_STATIC_INLINE I32 *
37S_CvDEPTHp(const CV * const sv)
38{
39 assert(SvTYPE(sv) == SVt_PVCV || SvTYPE(sv) == SVt_PVFM);
8de47657 40 return &((XPVCV*)SvANY(sv))->xcv_depth;
1afe1db1
FC
41}
42
d16269d8
PM
43/*
44 CvPROTO returns the prototype as stored, which is not necessarily what
45 the interpreter should be using. Specifically, the interpreter assumes
46 that spaces have been stripped, which has been the case if the prototype
47 was added by toke.c, but is generally not the case if it was added elsewhere.
48 Since we can't enforce the spacelessness at assignment time, this routine
49 provides a temporary copy at parse time with spaces removed.
50 I<orig> is the start of the original buffer, I<len> is the length of the
51 prototype and will be updated when this returns.
52 */
53
5b67adb8 54#ifdef PERL_CORE
d16269d8
PM
55PERL_STATIC_INLINE char *
56S_strip_spaces(pTHX_ const char * orig, STRLEN * const len)
57{
58 SV * tmpsv;
59 char * tmps;
60 tmpsv = newSVpvn_flags(orig, *len, SVs_TEMP);
61 tmps = SvPVX(tmpsv);
62 while ((*len)--) {
63 if (!isSPACE(*orig))
64 *tmps++ = *orig;
65 orig++;
66 }
67 *tmps = '\0';
68 *len = tmps - SvPVX(tmpsv);
69 return SvPVX(tmpsv);
70}
5b67adb8 71#endif
d16269d8 72
25fdce4a
FC
73/* ------------------------------- mg.h ------------------------------- */
74
75#if defined(PERL_CORE) || defined(PERL_EXT)
76/* assumes get-magic and stringification have already occurred */
77PERL_STATIC_INLINE STRLEN
78S_MgBYTEPOS(pTHX_ MAGIC *mg, SV *sv, const char *s, STRLEN len)
79{
80 assert(mg->mg_type == PERL_MAGIC_regex_global);
81 assert(mg->mg_len != -1);
82 if (mg->mg_flags & MGf_BYTES || !DO_UTF8(sv))
83 return (STRLEN)mg->mg_len;
84 else {
85 const STRLEN pos = (STRLEN)mg->mg_len;
86 /* Without this check, we may read past the end of the buffer: */
87 if (pos > sv_or_pv_len_utf8(sv, s, len)) return len+1;
88 return sv_or_pv_pos_u2b(sv, s, pos, NULL);
89 }
90}
91#endif
92
03414f05
FC
93/* ------------------------------- pad.h ------------------------------ */
94
95#if defined(PERL_IN_PAD_C) || defined(PERL_IN_OP_C)
96PERL_STATIC_INLINE bool
97PadnameIN_SCOPE(const PADNAME * const pn, const U32 seq)
98{
99 /* is seq within the range _LOW to _HIGH ?
100 * This is complicated by the fact that PL_cop_seqmax
101 * may have wrapped around at some point */
102 if (COP_SEQ_RANGE_LOW(pn) == PERL_PADSEQ_INTRO)
103 return FALSE; /* not yet introduced */
104
105 if (COP_SEQ_RANGE_HIGH(pn) == PERL_PADSEQ_INTRO) {
106 /* in compiling scope */
107 if (
108 (seq > COP_SEQ_RANGE_LOW(pn))
109 ? (seq - COP_SEQ_RANGE_LOW(pn) < (U32_MAX >> 1))
110 : (COP_SEQ_RANGE_LOW(pn) - seq > (U32_MAX >> 1))
111 )
112 return TRUE;
113 }
114 else if (
115 (COP_SEQ_RANGE_LOW(pn) > COP_SEQ_RANGE_HIGH(pn))
116 ?
117 ( seq > COP_SEQ_RANGE_LOW(pn)
118 || seq <= COP_SEQ_RANGE_HIGH(pn))
119
120 : ( seq > COP_SEQ_RANGE_LOW(pn)
121 && seq <= COP_SEQ_RANGE_HIGH(pn))
122 )
123 return TRUE;
124 return FALSE;
125}
126#endif
127
33a4312b
FC
128/* ------------------------------- pp.h ------------------------------- */
129
130PERL_STATIC_INLINE I32
131S_TOPMARK(pTHX)
132{
133 DEBUG_s(DEBUG_v(PerlIO_printf(Perl_debug_log,
134 "MARK top %p %"IVdf"\n",
135 PL_markstack_ptr,
136 (IV)*PL_markstack_ptr)));
137 return *PL_markstack_ptr;
138}
139
140PERL_STATIC_INLINE I32
141S_POPMARK(pTHX)
142{
143 DEBUG_s(DEBUG_v(PerlIO_printf(Perl_debug_log,
144 "MARK pop %p %"IVdf"\n",
145 (PL_markstack_ptr-1),
146 (IV)*(PL_markstack_ptr-1))));
147 assert((PL_markstack_ptr > PL_markstack) || !"MARK underflow");
148 return *PL_markstack_ptr--;
149}
150
8d919b0a
FC
151/* ----------------------------- regexp.h ----------------------------- */
152
153PERL_STATIC_INLINE struct regexp *
154S_ReANY(const REGEXP * const re)
155{
156 assert(isREGEXP(re));
157 return re->sv_u.svu_rx;
158}
159
27669aa4
FC
160/* ------------------------------- sv.h ------------------------------- */
161
162PERL_STATIC_INLINE SV *
163S_SvREFCNT_inc(SV *sv)
164{
2439e033 165 if (LIKELY(sv != NULL))
27669aa4
FC
166 SvREFCNT(sv)++;
167 return sv;
168}
169PERL_STATIC_INLINE SV *
170S_SvREFCNT_inc_NN(SV *sv)
171{
172 SvREFCNT(sv)++;
173 return sv;
174}
175PERL_STATIC_INLINE void
176S_SvREFCNT_inc_void(SV *sv)
177{
2439e033 178 if (LIKELY(sv != NULL))
27669aa4
FC
179 SvREFCNT(sv)++;
180}
75e16a44
FC
181PERL_STATIC_INLINE void
182S_SvREFCNT_dec(pTHX_ SV *sv)
183{
2439e033 184 if (LIKELY(sv != NULL)) {
75a9bf96 185 U32 rc = SvREFCNT(sv);
79e2a32a 186 if (LIKELY(rc > 1))
75a9bf96
DM
187 SvREFCNT(sv) = rc - 1;
188 else
189 Perl_sv_free2(aTHX_ sv, rc);
75e16a44
FC
190 }
191}
541377b1
FC
192
193PERL_STATIC_INLINE void
4a9a56a7
DM
194S_SvREFCNT_dec_NN(pTHX_ SV *sv)
195{
196 U32 rc = SvREFCNT(sv);
79e2a32a 197 if (LIKELY(rc > 1))
4a9a56a7
DM
198 SvREFCNT(sv) = rc - 1;
199 else
200 Perl_sv_free2(aTHX_ sv, rc);
201}
202
203PERL_STATIC_INLINE void
541377b1
FC
204SvAMAGIC_on(SV *sv)
205{
206 assert(SvROK(sv));
207 if (SvOBJECT(SvRV(sv))) HvAMAGIC_on(SvSTASH(SvRV(sv)));
208}
209PERL_STATIC_INLINE void
210SvAMAGIC_off(SV *sv)
211{
212 if (SvROK(sv) && SvOBJECT(SvRV(sv)))
213 HvAMAGIC_off(SvSTASH(SvRV(sv)));
214}
215
216PERL_STATIC_INLINE U32
541377b1
FC
217S_SvPADSTALE_on(SV *sv)
218{
c0683843 219 assert(!(SvFLAGS(sv) & SVs_PADTMP));
541377b1
FC
220 return SvFLAGS(sv) |= SVs_PADSTALE;
221}
222PERL_STATIC_INLINE U32
223S_SvPADSTALE_off(SV *sv)
224{
c0683843 225 assert(!(SvFLAGS(sv) & SVs_PADTMP));
541377b1
FC
226 return SvFLAGS(sv) &= ~SVs_PADSTALE;
227}
25fdce4a 228#if defined(PERL_CORE) || defined (PERL_EXT)
4ddea69a 229PERL_STATIC_INLINE STRLEN
6964422a 230S_sv_or_pv_pos_u2b(pTHX_ SV *sv, const char *pv, STRLEN pos, STRLEN *lenp)
4ddea69a 231{
25fdce4a 232 PERL_ARGS_ASSERT_SV_OR_PV_POS_U2B;
4ddea69a
FC
233 if (SvGAMAGIC(sv)) {
234 U8 *hopped = utf8_hop((U8 *)pv, pos);
235 if (lenp) *lenp = (STRLEN)(utf8_hop(hopped, *lenp) - hopped);
236 return (STRLEN)(hopped - (U8 *)pv);
237 }
238 return sv_pos_u2b_flags(sv,pos,lenp,SV_CONST_RETURN);
239}
240#endif
f019c49e 241
d1decf2b
TC
242/* ------------------------------- handy.h ------------------------------- */
243
244/* saves machine code for a common noreturn idiom typically used in Newx*() */
c1d6452f 245#ifdef GCC_DIAG_PRAGMA
6ab56f1e 246GCC_DIAG_IGNORE(-Wunused-function) /* Intentionally left semicolonless. */
c1d6452f 247#endif
d1decf2b
TC
248static void
249S_croak_memory_wrap(void)
250{
251 Perl_croak_nocontext("%s",PL_memory_wrap);
252}
c1d6452f 253#ifdef GCC_DIAG_PRAGMA
6ab56f1e 254GCC_DIAG_RESTORE /* Intentionally left semicolonless. */
c1d6452f 255#endif
d1decf2b 256
a8a2ceaa
KW
257/* ------------------------------- utf8.h ------------------------------- */
258
2fe720e2
KW
259/*
260=head1 Unicode Support
261*/
262
55d09dc8
KW
263PERL_STATIC_INLINE void
264S_append_utf8_from_native_byte(const U8 byte, U8** dest)
265{
266 /* Takes an input 'byte' (Latin1 or EBCDIC) and appends it to the UTF-8
267 * encoded string at '*dest', updating '*dest' to include it */
268
55d09dc8
KW
269 PERL_ARGS_ASSERT_APPEND_UTF8_FROM_NATIVE_BYTE;
270
6f2d5cbc 271 if (NATIVE_BYTE_IS_INVARIANT(byte))
a09ec51a 272 *((*dest)++) = byte;
55d09dc8 273 else {
a09ec51a
KW
274 *((*dest)++) = UTF8_EIGHT_BIT_HI(byte);
275 *((*dest)++) = UTF8_EIGHT_BIT_LO(byte);
55d09dc8
KW
276 }
277}
278
e123187a 279/*
2fe720e2 280=for apidoc valid_utf8_to_uvchr
2717076a 281Like C<L</utf8_to_uvchr_buf>>, but should only be called when it is known that
2fe720e2
KW
282the next character in the input UTF-8 string C<s> is well-formed (I<e.g.>,
283it passes C<L</isUTF8_CHAR>>. Surrogates, non-character code points, and
284non-Unicode code points are allowed.
285
286=cut
287
288 */
289
290PERL_STATIC_INLINE UV
291Perl_valid_utf8_to_uvchr(const U8 *s, STRLEN *retlen)
292{
c41b2540 293 const UV expectlen = UTF8SKIP(s);
2fe720e2
KW
294 const U8* send = s + expectlen;
295 UV uv = *s;
296
297 PERL_ARGS_ASSERT_VALID_UTF8_TO_UVCHR;
298
299 if (retlen) {
300 *retlen = expectlen;
301 }
302
303 /* An invariant is trivially returned */
304 if (expectlen == 1) {
305 return uv;
306 }
307
308 /* Remove the leading bits that indicate the number of bytes, leaving just
309 * the bits that are part of the value */
310 uv = NATIVE_UTF8_TO_I8(uv) & UTF_START_MASK(expectlen);
311
312 /* Now, loop through the remaining bytes, accumulating each into the
313 * working total as we go. (I khw tried unrolling the loop for up to 4
314 * bytes, but there was no performance improvement) */
315 for (++s; s < send; s++) {
316 uv = UTF8_ACCUMULATE(uv, *s);
317 }
318
319 return UNI_TO_NATIVE(uv);
320
321}
322
1e599354
KW
323/*
324=for apidoc is_utf8_invariant_string
325
82c5d941 326Returns TRUE if the first C<len> bytes of the string C<s> are the same
1e599354 327regardless of the UTF-8 encoding of the string (or UTF-EBCDIC encoding on
82c5d941
KW
328EBCDIC machines); otherwise it returns FALSE. That is, it returns TRUE if they
329are UTF-8 invariant. On ASCII-ish machines, all the ASCII characters and only
330the ASCII characters fit this definition. On EBCDIC machines, the ASCII-range
331characters are invariant, but so also are the C1 controls.
1e599354
KW
332
333If C<len> is 0, it will be calculated using C<strlen(s)>, (which means if you
334use this option, that C<s> can't have embedded C<NUL> characters and has to
335have a terminating C<NUL> byte).
336
9f2abfde
KW
337See also
338C<L</is_utf8_string>>,
339C<L</is_utf8_string_flags>>,
340C<L</is_utf8_string_loc>>,
341C<L</is_utf8_string_loc_flags>>,
342C<L</is_utf8_string_loclen>>,
343C<L</is_utf8_string_loclen_flags>>,
8bc127bf
KW
344C<L</is_utf8_fixed_width_buf_flags>>,
345C<L</is_utf8_fixed_width_buf_loc_flags>>,
346C<L</is_utf8_fixed_width_buf_loclen_flags>>,
9f2abfde
KW
347C<L</is_strict_utf8_string>>,
348C<L</is_strict_utf8_string_loc>>,
349C<L</is_strict_utf8_string_loclen>>,
350C<L</is_c9strict_utf8_string>>,
351C<L</is_c9strict_utf8_string_loc>>,
352and
353C<L</is_c9strict_utf8_string_loclen>>.
1e599354
KW
354
355=cut
356*/
357
358PERL_STATIC_INLINE bool
359S_is_utf8_invariant_string(const U8* const s, const STRLEN len)
360{
361 const U8* const send = s + (len ? len : strlen((const char *)s));
362 const U8* x = s;
363
364 PERL_ARGS_ASSERT_IS_UTF8_INVARIANT_STRING;
365
366 for (; x < send; ++x) {
367 if (!UTF8_IS_INVARIANT(*x))
368 return FALSE;
369 }
370
371 return TRUE;
372}
373
7c93d8f0 374/*
5ff889fb
KW
375=for apidoc is_utf8_string
376
82c5d941
KW
377Returns TRUE if the first C<len> bytes of string C<s> form a valid
378Perl-extended-UTF-8 string; returns FALSE otherwise. If C<len> is 0, it will
379be calculated using C<strlen(s)> (which means if you use this option, that C<s>
380can't have embedded C<NUL> characters and has to have a terminating C<NUL>
381byte). Note that all characters being ASCII constitute 'a valid UTF-8 string'.
382
2717076a
KW
383This function considers Perl's extended UTF-8 to be valid. That means that
384code points above Unicode, surrogates, and non-character code points are
9f2abfde
KW
385considered valid by this function. Use C<L</is_strict_utf8_string>>,
386C<L</is_c9strict_utf8_string>>, or C<L</is_utf8_string_flags>> to restrict what
387code points are considered valid.
5ff889fb 388
9f2abfde
KW
389See also
390C<L</is_utf8_invariant_string>>,
391C<L</is_utf8_string_loc>>,
392C<L</is_utf8_string_loclen>>,
8bc127bf
KW
393C<L</is_utf8_fixed_width_buf_flags>>,
394C<L</is_utf8_fixed_width_buf_loc_flags>>,
395C<L</is_utf8_fixed_width_buf_loclen_flags>>,
5ff889fb
KW
396
397=cut
398*/
399
56e4cf64 400PERL_STATIC_INLINE bool
c41b2540 401Perl_is_utf8_string(const U8 *s, const STRLEN len)
5ff889fb 402{
35936d22
KW
403 /* This is now marked pure in embed.fnc, because isUTF8_CHAR now is pure.
404 * Be aware of possible changes to that */
405
5ff889fb
KW
406 const U8* const send = s + (len ? len : strlen((const char *)s));
407 const U8* x = s;
408
409 PERL_ARGS_ASSERT_IS_UTF8_STRING;
410
411 while (x < send) {
c41b2540
KW
412 const STRLEN cur_len = isUTF8_CHAR(x, send);
413 if (UNLIKELY(! cur_len)) {
5ff889fb
KW
414 return FALSE;
415 }
c41b2540 416 x += cur_len;
5ff889fb
KW
417 }
418
419 return TRUE;
420}
421
422/*
9f2abfde
KW
423=for apidoc is_strict_utf8_string
424
425Returns TRUE if the first C<len> bytes of string C<s> form a valid
426UTF-8-encoded string that is fully interchangeable by any application using
427Unicode rules; otherwise it returns FALSE. If C<len> is 0, it will be
428calculated using C<strlen(s)> (which means if you use this option, that C<s>
429can't have embedded C<NUL> characters and has to have a terminating C<NUL>
430byte). Note that all characters being ASCII constitute 'a valid UTF-8 string'.
431
432This function returns FALSE for strings containing any
433code points above the Unicode max of 0x10FFFF, surrogate code points, or
434non-character code points.
435
436See also
437C<L</is_utf8_invariant_string>>,
438C<L</is_utf8_string>>,
439C<L</is_utf8_string_flags>>,
440C<L</is_utf8_string_loc>>,
441C<L</is_utf8_string_loc_flags>>,
442C<L</is_utf8_string_loclen>>,
443C<L</is_utf8_string_loclen_flags>>,
8bc127bf
KW
444C<L</is_utf8_fixed_width_buf_flags>>,
445C<L</is_utf8_fixed_width_buf_loc_flags>>,
446C<L</is_utf8_fixed_width_buf_loclen_flags>>,
9f2abfde
KW
447C<L</is_strict_utf8_string_loc>>,
448C<L</is_strict_utf8_string_loclen>>,
449C<L</is_c9strict_utf8_string>>,
450C<L</is_c9strict_utf8_string_loc>>,
451and
452C<L</is_c9strict_utf8_string_loclen>>.
453
454=cut
455*/
456
457PERL_STATIC_INLINE bool
458S_is_strict_utf8_string(const U8 *s, const STRLEN len)
459{
460 const U8* const send = s + (len ? len : strlen((const char *)s));
461 const U8* x = s;
462
463 PERL_ARGS_ASSERT_IS_STRICT_UTF8_STRING;
464
465 while (x < send) {
466 const STRLEN cur_len = isSTRICT_UTF8_CHAR(x, send);
467 if (UNLIKELY(! cur_len)) {
468 return FALSE;
469 }
470 x += cur_len;
471 }
472
473 return TRUE;
474}
475
476/*
477=for apidoc is_c9strict_utf8_string
478
479Returns TRUE if the first C<len> bytes of string C<s> form a valid
480UTF-8-encoded string that conforms to
481L<Unicode Corrigendum #9|http://www.unicode.org/versions/corrigendum9.html>;
482otherwise it returns FALSE. If C<len> is 0, it will be calculated using
483C<strlen(s)> (which means if you use this option, that C<s> can't have embedded
484C<NUL> characters and has to have a terminating C<NUL> byte). Note that all
485characters being ASCII constitute 'a valid UTF-8 string'.
486
487This function returns FALSE for strings containing any code points above the
488Unicode max of 0x10FFFF or surrogate code points, but accepts non-character
489code points per
490L<Corrigendum #9|http://www.unicode.org/versions/corrigendum9.html>.
491
492See also
493C<L</is_utf8_invariant_string>>,
494C<L</is_utf8_string>>,
495C<L</is_utf8_string_flags>>,
496C<L</is_utf8_string_loc>>,
497C<L</is_utf8_string_loc_flags>>,
498C<L</is_utf8_string_loclen>>,
499C<L</is_utf8_string_loclen_flags>>,
8bc127bf
KW
500C<L</is_utf8_fixed_width_buf_flags>>,
501C<L</is_utf8_fixed_width_buf_loc_flags>>,
502C<L</is_utf8_fixed_width_buf_loclen_flags>>,
9f2abfde
KW
503C<L</is_strict_utf8_string>>,
504C<L</is_strict_utf8_string_loc>>,
505C<L</is_strict_utf8_string_loclen>>,
506C<L</is_c9strict_utf8_string_loc>>,
507and
508C<L</is_c9strict_utf8_string_loclen>>.
509
510=cut
511*/
512
513PERL_STATIC_INLINE bool
514S_is_c9strict_utf8_string(const U8 *s, const STRLEN len)
515{
516 const U8* const send = s + (len ? len : strlen((const char *)s));
517 const U8* x = s;
518
519 PERL_ARGS_ASSERT_IS_C9STRICT_UTF8_STRING;
520
521 while (x < send) {
522 const STRLEN cur_len = isC9_STRICT_UTF8_CHAR(x, send);
523 if (UNLIKELY(! cur_len)) {
524 return FALSE;
525 }
526 x += cur_len;
527 }
528
529 return TRUE;
530}
531
532/* The above 3 functions could have been moved into the more general one just
533 * below, and made #defines that call it with the right 'flags'. They are
534 * currently kept separate to increase their chances of getting inlined */
535
536/*
537=for apidoc is_utf8_string_flags
538
539Returns TRUE if the first C<len> bytes of string C<s> form a valid
540UTF-8 string, subject to the restrictions imposed by C<flags>;
541returns FALSE otherwise. If C<len> is 0, it will be calculated
542using C<strlen(s)> (which means if you use this option, that C<s> can't have
543embedded C<NUL> characters and has to have a terminating C<NUL> byte). Note
544that all characters being ASCII constitute 'a valid UTF-8 string'.
545
546If C<flags> is 0, this gives the same results as C<L</is_utf8_string>>; if
547C<flags> is C<UTF8_DISALLOW_ILLEGAL_INTERCHANGE>, this gives the same results
548as C<L</is_strict_utf8_string>>; and if C<flags> is
549C<UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE>, this gives the same results as
550C<L</is_c9strict_utf8_string>>. Otherwise C<flags> may be any
551combination of the C<UTF8_DISALLOW_I<foo>> flags understood by
552C<L</utf8n_to_uvchr>>, with the same meanings.
553
554See also
555C<L</is_utf8_invariant_string>>,
556C<L</is_utf8_string>>,
557C<L</is_utf8_string_loc>>,
558C<L</is_utf8_string_loc_flags>>,
559C<L</is_utf8_string_loclen>>,
560C<L</is_utf8_string_loclen_flags>>,
8bc127bf
KW
561C<L</is_utf8_fixed_width_buf_flags>>,
562C<L</is_utf8_fixed_width_buf_loc_flags>>,
563C<L</is_utf8_fixed_width_buf_loclen_flags>>,
9f2abfde
KW
564C<L</is_strict_utf8_string>>,
565C<L</is_strict_utf8_string_loc>>,
566C<L</is_strict_utf8_string_loclen>>,
567C<L</is_c9strict_utf8_string>>,
568C<L</is_c9strict_utf8_string_loc>>,
569and
570C<L</is_c9strict_utf8_string_loclen>>.
571
572=cut
573*/
574
575PERL_STATIC_INLINE bool
576S_is_utf8_string_flags(const U8 *s, const STRLEN len, const U32 flags)
577{
578 const U8* const send = s + (len ? len : strlen((const char *)s));
579 const U8* x = s;
580
581 PERL_ARGS_ASSERT_IS_UTF8_STRING_FLAGS;
582 assert(0 == (flags & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE
583 |UTF8_DISALLOW_ABOVE_31_BIT)));
584
585 if (flags == 0) {
586 return is_utf8_string(s, len);
587 }
588
589 if ((flags & ~UTF8_DISALLOW_ABOVE_31_BIT)
590 == UTF8_DISALLOW_ILLEGAL_INTERCHANGE)
591 {
592 return is_strict_utf8_string(s, len);
593 }
594
595 if ((flags & ~UTF8_DISALLOW_ABOVE_31_BIT)
596 == UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE)
597 {
598 return is_c9strict_utf8_string(s, len);
599 }
600
601 while (x < send) {
602 STRLEN cur_len = isUTF8_CHAR_flags(x, send, flags);
603 if (UNLIKELY(! cur_len)) {
604 return FALSE;
605 }
606 x += cur_len;
607 }
608
609 return TRUE;
610}
611
612/*
5ff889fb
KW
613
614=for apidoc is_utf8_string_loc
615
2717076a 616Like C<L</is_utf8_string>> but stores the location of the failure (in the
5ff889fb 617case of "utf8ness failure") or the location C<s>+C<len> (in the case of
82c5d941 618"utf8ness success") in the C<ep> pointer.
5ff889fb 619
2717076a 620See also C<L</is_utf8_string_loclen>>.
5ff889fb 621
3964c812
KW
622=cut
623*/
624
625#define is_utf8_string_loc(s, len, ep) is_utf8_string_loclen(s, len, ep, 0)
626
627/*
628
5ff889fb
KW
629=for apidoc is_utf8_string_loclen
630
2717076a 631Like C<L</is_utf8_string>> but stores the location of the failure (in the
5ff889fb 632case of "utf8ness failure") or the location C<s>+C<len> (in the case of
9f2abfde 633"utf8ness success") in the C<ep> pointer, and the number of UTF-8
82c5d941 634encoded characters in the C<el> pointer.
5ff889fb 635
2717076a 636See also C<L</is_utf8_string_loc>>.
5ff889fb
KW
637
638=cut
639*/
640
56e4cf64 641PERL_STATIC_INLINE bool
c41b2540 642Perl_is_utf8_string_loclen(const U8 *s, const STRLEN len, const U8 **ep, STRLEN *el)
5ff889fb
KW
643{
644 const U8* const send = s + (len ? len : strlen((const char *)s));
645 const U8* x = s;
646 STRLEN outlen = 0;
647
648 PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN;
649
650 while (x < send) {
c41b2540
KW
651 const STRLEN cur_len = isUTF8_CHAR(x, send);
652 if (UNLIKELY(! cur_len)) {
5ff889fb
KW
653 break;
654 }
c41b2540 655 x += cur_len;
5ff889fb
KW
656 outlen++;
657 }
658
659 if (el)
660 *el = outlen;
661
662 if (ep) {
663 *ep = x;
664 }
665
666 return (x == send);
667}
668
669/*
9f2abfde
KW
670
671=for apidoc is_strict_utf8_string_loc
672
673Like C<L</is_strict_utf8_string>> but stores the location of the failure (in the
674case of "utf8ness failure") or the location C<s>+C<len> (in the case of
675"utf8ness success") in the C<ep> pointer.
676
677See also C<L</is_strict_utf8_string_loclen>>.
678
679=cut
680*/
681
682#define is_strict_utf8_string_loc(s, len, ep) \
683 is_strict_utf8_string_loclen(s, len, ep, 0)
684
685/*
686
687=for apidoc is_strict_utf8_string_loclen
688
689Like C<L</is_strict_utf8_string>> but stores the location of the failure (in the
690case of "utf8ness failure") or the location C<s>+C<len> (in the case of
691"utf8ness success") in the C<ep> pointer, and the number of UTF-8
692encoded characters in the C<el> pointer.
693
694See also C<L</is_strict_utf8_string_loc>>.
695
696=cut
697*/
698
699PERL_STATIC_INLINE bool
700S_is_strict_utf8_string_loclen(const U8 *s, const STRLEN len, const U8 **ep, STRLEN *el)
701{
702 const U8* const send = s + (len ? len : strlen((const char *)s));
703 const U8* x = s;
704 STRLEN outlen = 0;
705
706 PERL_ARGS_ASSERT_IS_STRICT_UTF8_STRING_LOCLEN;
707
708 while (x < send) {
709 const STRLEN cur_len = isSTRICT_UTF8_CHAR(x, send);
710 if (UNLIKELY(! cur_len)) {
711 break;
712 }
713 x += cur_len;
714 outlen++;
715 }
716
717 if (el)
718 *el = outlen;
719
720 if (ep) {
721 *ep = x;
722 }
723
724 return (x == send);
725}
726
727/*
728
729=for apidoc is_c9strict_utf8_string_loc
730
731Like C<L</is_c9strict_utf8_string>> but stores the location of the failure (in
732the case of "utf8ness failure") or the location C<s>+C<len> (in the case of
733"utf8ness success") in the C<ep> pointer.
734
735See also C<L</is_c9strict_utf8_string_loclen>>.
736
737=cut
738*/
739
740#define is_c9strict_utf8_string_loc(s, len, ep) \
741 is_c9strict_utf8_string_loclen(s, len, ep, 0)
742
743/*
744
745=for apidoc is_c9strict_utf8_string_loclen
746
747Like C<L</is_c9strict_utf8_string>> but stores the location of the failure (in
748the case of "utf8ness failure") or the location C<s>+C<len> (in the case of
749"utf8ness success") in the C<ep> pointer, and the number of UTF-8 encoded
750characters in the C<el> pointer.
751
752See also C<L</is_c9strict_utf8_string_loc>>.
753
754=cut
755*/
756
757PERL_STATIC_INLINE bool
758S_is_c9strict_utf8_string_loclen(const U8 *s, const STRLEN len, const U8 **ep, STRLEN *el)
759{
760 const U8* const send = s + (len ? len : strlen((const char *)s));
761 const U8* x = s;
762 STRLEN outlen = 0;
763
764 PERL_ARGS_ASSERT_IS_C9STRICT_UTF8_STRING_LOCLEN;
765
766 while (x < send) {
767 const STRLEN cur_len = isC9_STRICT_UTF8_CHAR(x, send);
768 if (UNLIKELY(! cur_len)) {
769 break;
770 }
771 x += cur_len;
772 outlen++;
773 }
774
775 if (el)
776 *el = outlen;
777
778 if (ep) {
779 *ep = x;
780 }
781
782 return (x == send);
783}
784
785/*
786
787=for apidoc is_utf8_string_loc_flags
788
789Like C<L</is_utf8_string_flags>> but stores the location of the failure (in the
790case of "utf8ness failure") or the location C<s>+C<len> (in the case of
791"utf8ness success") in the C<ep> pointer.
792
793See also C<L</is_utf8_string_loclen_flags>>.
794
795=cut
796*/
797
798#define is_utf8_string_loc_flags(s, len, ep, flags) \
799 is_utf8_string_loclen_flags(s, len, ep, 0, flags)
800
801
802/* The above 3 actual functions could have been moved into the more general one
803 * just below, and made #defines that call it with the right 'flags'. They are
804 * currently kept separate to increase their chances of getting inlined */
805
806/*
807
808=for apidoc is_utf8_string_loclen_flags
809
810Like C<L</is_utf8_string_flags>> but stores the location of the failure (in the
811case of "utf8ness failure") or the location C<s>+C<len> (in the case of
812"utf8ness success") in the C<ep> pointer, and the number of UTF-8
813encoded characters in the C<el> pointer.
814
815See also C<L</is_utf8_string_loc_flags>>.
816
817=cut
818*/
819
820PERL_STATIC_INLINE bool
821S_is_utf8_string_loclen_flags(const U8 *s, const STRLEN len, const U8 **ep, STRLEN *el, const U32 flags)
822{
823 const U8* const send = s + (len ? len : strlen((const char *)s));
824 const U8* x = s;
825 STRLEN outlen = 0;
826
827 PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN_FLAGS;
828 assert(0 == (flags & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE
829 |UTF8_DISALLOW_ABOVE_31_BIT)));
830
831 if (flags == 0) {
832 return is_utf8_string_loclen(s, len, ep, el);
833 }
834
835 if ((flags & ~UTF8_DISALLOW_ABOVE_31_BIT)
836 == UTF8_DISALLOW_ILLEGAL_INTERCHANGE)
837 {
838 return is_strict_utf8_string_loclen(s, len, ep, el);
839 }
840
841 if ((flags & ~UTF8_DISALLOW_ABOVE_31_BIT)
842 == UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE)
843 {
844 return is_c9strict_utf8_string_loclen(s, len, ep, el);
845 }
846
847 while (x < send) {
848 const STRLEN cur_len = isUTF8_CHAR_flags(x, send, flags);
849 if (UNLIKELY(! cur_len)) {
850 break;
851 }
852 x += cur_len;
853 outlen++;
854 }
855
856 if (el)
857 *el = outlen;
858
859 if (ep) {
860 *ep = x;
861 }
862
863 return (x == send);
864}
865
866/*
7c93d8f0
KW
867=for apidoc utf8_distance
868
869Returns the number of UTF-8 characters between the UTF-8 pointers C<a>
870and C<b>.
871
872WARNING: use only if you *know* that the pointers point inside the
873same UTF-8 buffer.
874
875=cut
876*/
877
878PERL_STATIC_INLINE IV
879Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b)
880{
881 PERL_ARGS_ASSERT_UTF8_DISTANCE;
882
883 return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a);
884}
885
886/*
887=for apidoc utf8_hop
888
889Return the UTF-8 pointer C<s> displaced by C<off> characters, either
890forward or backward.
891
892WARNING: do not use the following unless you *know* C<off> is within
893the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
894on the first byte of character or just after the last byte of a character.
895
896=cut
897*/
898
899PERL_STATIC_INLINE U8 *
900Perl_utf8_hop(const U8 *s, SSize_t off)
901{
902 PERL_ARGS_ASSERT_UTF8_HOP;
903
904 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
905 * the bitops (especially ~) can create illegal UTF-8.
906 * In other words: in Perl UTF-8 is not just for Unicode. */
907
908 if (off >= 0) {
909 while (off--)
910 s += UTF8SKIP(s);
911 }
912 else {
913 while (off++) {
914 s--;
915 while (UTF8_IS_CONTINUATION(*s))
916 s--;
917 }
918 }
919 return (U8 *)s;
920}
921
4dab108f 922/*
65df57a8
TC
923=for apidoc utf8_hop_forward
924
925Return the UTF-8 pointer C<s> displaced by up to C<off> characters,
926forward.
927
928C<off> must be non-negative.
929
930C<s> must be before or equal to C<end>.
931
932When moving forward it will not move beyond C<end>.
933
934Will not exceed this limit even if the string is not valid "UTF-8".
935
936=cut
937*/
938
939PERL_STATIC_INLINE U8 *
940Perl_utf8_hop_forward(const U8 *s, SSize_t off, const U8 *end)
941{
942 PERL_ARGS_ASSERT_UTF8_HOP_FORWARD;
943
944 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
945 * the bitops (especially ~) can create illegal UTF-8.
946 * In other words: in Perl UTF-8 is not just for Unicode. */
947
948 assert(s <= end);
949 assert(off >= 0);
950
951 while (off--) {
952 STRLEN skip = UTF8SKIP(s);
953 if ((STRLEN)(end - s) <= skip)
954 return (U8 *)end;
955 s += skip;
956 }
957
958 return (U8 *)s;
959}
960
961/*
962=for apidoc utf8_hop_back
963
964Return the UTF-8 pointer C<s> displaced by up to C<off> characters,
965backward.
966
967C<off> must be non-positive.
968
969C<s> must be after or equal to C<start>.
970
971When moving backward it will not move before C<start>.
972
973Will not exceed this limit even if the string is not valid "UTF-8".
974
975=cut
976*/
977
978PERL_STATIC_INLINE U8 *
979Perl_utf8_hop_back(const U8 *s, SSize_t off, const U8 *start)
980{
981 PERL_ARGS_ASSERT_UTF8_HOP_BACK;
982
983 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
984 * the bitops (especially ~) can create illegal UTF-8.
985 * In other words: in Perl UTF-8 is not just for Unicode. */
986
987 assert(start <= s);
988 assert(off <= 0);
989
990 while (off++ && s > start) {
991 s--;
992 while (UTF8_IS_CONTINUATION(*s) && s > start)
993 s--;
994 }
995
996 return (U8 *)s;
997}
998
999/*
1000=for apidoc utf8_hop_safe
1001
1002Return the UTF-8 pointer C<s> displaced by up to C<off> characters,
1003either forward or backward.
1004
1005When moving backward it will not move before C<start>.
1006
1007When moving forward it will not move beyond C<end>.
1008
1009Will not exceed those limits even if the string is not valid "UTF-8".
1010
1011=cut
1012*/
1013
1014PERL_STATIC_INLINE U8 *
1015Perl_utf8_hop_safe(const U8 *s, SSize_t off, const U8 *start, const U8 *end)
1016{
1017 PERL_ARGS_ASSERT_UTF8_HOP_SAFE;
1018
1019 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
1020 * the bitops (especially ~) can create illegal UTF-8.
1021 * In other words: in Perl UTF-8 is not just for Unicode. */
1022
1023 assert(start <= s && s <= end);
1024
1025 if (off >= 0) {
1026 return utf8_hop_forward(s, off, end);
1027 }
1028 else {
1029 return utf8_hop_back(s, off, start);
1030 }
1031}
1032
1033/*
4dab108f
KW
1034
1035=for apidoc is_utf8_valid_partial_char
1036
6cbb9248
KW
1037Returns 0 if the sequence of bytes starting at C<s> and looking no further than
1038S<C<e - 1>> is the UTF-8 encoding, as extended by Perl, for one or more code
1039points. Otherwise, it returns 1 if there exists at least one non-empty
1040sequence of bytes that when appended to sequence C<s>, starting at position
1041C<e> causes the entire sequence to be the well-formed UTF-8 of some code point;
1042otherwise returns 0.
1043
1044In other words this returns TRUE if C<s> points to a partial UTF-8-encoded code
1045point.
1046
1047This is useful when a fixed-length buffer is being tested for being well-formed
1048UTF-8, but the final few bytes in it don't comprise a full character; that is,
1049it is split somewhere in the middle of the final code point's UTF-8
1050representation. (Presumably when the buffer is refreshed with the next chunk
1051of data, the new first bytes will complete the partial code point.) This
1052function is used to verify that the final bytes in the current buffer are in
1053fact the legal beginning of some code point, so that if they aren't, the
1054failure can be signalled without having to wait for the next read.
4dab108f
KW
1055
1056=cut
1057*/
2717076a
KW
1058#define is_utf8_valid_partial_char(s, e) \
1059 is_utf8_valid_partial_char_flags(s, e, 0)
f1c999a7
KW
1060
1061/*
1062
1063=for apidoc is_utf8_valid_partial_char_flags
1064
1065Like C<L</is_utf8_valid_partial_char>>, it returns a boolean giving whether
1066or not the input is a valid UTF-8 encoded partial character, but it takes an
1067extra parameter, C<flags>, which can further restrict which code points are
1068considered valid.
1069
1070If C<flags> is 0, this behaves identically to
1071C<L</is_utf8_valid_partial_char>>. Otherwise C<flags> can be any combination
1072of the C<UTF8_DISALLOW_I<foo>> flags accepted by C<L</utf8n_to_uvchr>>. If
1073there is any sequence of bytes that can complete the input partial character in
1074such a way that a non-prohibited character is formed, the function returns
2717076a
KW
1075TRUE; otherwise FALSE. Non character code points cannot be determined based on
1076partial character input. But many of the other possible excluded types can be
f1c999a7
KW
1077determined from just the first one or two bytes.
1078
1079=cut
1080 */
1081
56e4cf64 1082PERL_STATIC_INLINE bool
f1c999a7 1083S_is_utf8_valid_partial_char_flags(const U8 * const s, const U8 * const e, const U32 flags)
4dab108f 1084{
f1c999a7 1085 PERL_ARGS_ASSERT_IS_UTF8_VALID_PARTIAL_CHAR_FLAGS;
4dab108f 1086
f1c999a7
KW
1087 assert(0 == (flags & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE
1088 |UTF8_DISALLOW_ABOVE_31_BIT)));
4dab108f 1089
8875bd48 1090 if (s >= e || s + UTF8SKIP(s) <= e) {
4dab108f
KW
1091 return FALSE;
1092 }
1093
f1c999a7 1094 return cBOOL(_is_utf8_char_helper(s, e, flags));
4dab108f
KW
1095}
1096
8bc127bf
KW
1097/*
1098
1099=for apidoc is_utf8_fixed_width_buf_flags
1100
1101Returns TRUE if the fixed-width buffer starting at C<s> with length C<len>
1102is entirely valid UTF-8, subject to the restrictions given by C<flags>;
1103otherwise it returns FALSE.
1104
1105If C<flags> is 0, any well-formed UTF-8, as extended by Perl, is accepted
1106without restriction. If the final few bytes of the buffer do not form a
1107complete code point, this will return TRUE anyway, provided that
1108C<L</is_utf8_valid_partial_char_flags>> returns TRUE for them.
1109
1110If C<flags> in non-zero, it can be any combination of the
1111C<UTF8_DISALLOW_I<foo>> flags accepted by C<L</utf8n_to_uvchr>>, and with the
1112same meanings.
1113
1114This function differs from C<L</is_utf8_string_flags>> only in that the latter
1115returns FALSE if the final few bytes of the string don't form a complete code
1116point.
1117
1118=cut
1119 */
1120#define is_utf8_fixed_width_buf_flags(s, len, flags) \
1121 is_utf8_fixed_width_buf_loclen_flags(s, len, 0, 0, flags)
1122
1123/*
1124
1125=for apidoc is_utf8_fixed_width_buf_loc_flags
1126
1127Like C<L</is_utf8_fixed_width_buf_flags>> but stores the location of the
1128failure in the C<ep> pointer. If the function returns TRUE, C<*ep> will point
1129to the beginning of any partial character at the end of the buffer; if there is
1130no partial character C<*ep> will contain C<s>+C<len>.
1131
1132See also C<L</is_utf8_fixed_width_buf_loclen_flags>>.
1133
1134=cut
1135*/
1136
1137#define is_utf8_fixed_width_buf_loc_flags(s, len, loc, flags) \
1138 is_utf8_fixed_width_buf_loclen_flags(s, len, loc, 0, flags)
1139
1140/*
1141
1142=for apidoc is_utf8_fixed_width_buf_loclen_flags
1143
1144Like C<L</is_utf8_fixed_width_buf_loc_flags>> but stores the number of
1145complete, valid characters found in the C<el> pointer.
1146
1147=cut
1148*/
1149
1150PERL_STATIC_INLINE bool
1151S_is_utf8_fixed_width_buf_loclen_flags(const U8 * const s,
1152 const STRLEN len,
1153 const U8 **ep,
1154 STRLEN *el,
1155 const U32 flags)
1156{
1157 const U8 * maybe_partial;
1158
1159 PERL_ARGS_ASSERT_IS_UTF8_FIXED_WIDTH_BUF_LOCLEN_FLAGS;
1160
1161 if (! ep) {
1162 ep = &maybe_partial;
1163 }
1164
1165 /* If it's entirely valid, return that; otherwise see if the only error is
1166 * that the final few bytes are for a partial character */
1167 return is_utf8_string_loclen_flags(s, len, ep, el, flags)
1168 || is_utf8_valid_partial_char_flags(*ep, s + len, flags);
1169}
1170
c8028aa6
TC
1171/* ------------------------------- perl.h ----------------------------- */
1172
1173/*
dcccc8ff
KW
1174=head1 Miscellaneous Functions
1175
41188aa0 1176=for apidoc AiR|bool|is_safe_syscall|const char *pv|STRLEN len|const char *what|const char *op_name
c8028aa6 1177
6602b933 1178Test that the given C<pv> doesn't contain any internal C<NUL> characters.
796b6530 1179If it does, set C<errno> to C<ENOENT>, optionally warn, and return FALSE.
c8028aa6
TC
1180
1181Return TRUE if the name is safe.
1182
796b6530 1183Used by the C<IS_SAFE_SYSCALL()> macro.
c8028aa6
TC
1184
1185=cut
1186*/
1187
1188PERL_STATIC_INLINE bool
41188aa0 1189S_is_safe_syscall(pTHX_ const char *pv, STRLEN len, const char *what, const char *op_name) {
c8028aa6
TC
1190 /* While the Windows CE API provides only UCS-16 (or UTF-16) APIs
1191 * perl itself uses xce*() functions which accept 8-bit strings.
1192 */
1193
1194 PERL_ARGS_ASSERT_IS_SAFE_SYSCALL;
1195
6c4650b3 1196 if (len > 1) {
c8028aa6 1197 char *null_at;
41188aa0 1198 if (UNLIKELY((null_at = (char *)memchr(pv, 0, len-1)) != NULL)) {
c8028aa6 1199 SETERRNO(ENOENT, LIB_INVARG);
1d505182 1200 Perl_ck_warner(aTHX_ packWARN(WARN_SYSCALLS),
c8028aa6 1201 "Invalid \\0 character in %s for %s: %s\\0%s",
41188aa0 1202 what, op_name, pv, null_at+1);
c8028aa6
TC
1203 return FALSE;
1204 }
1205 }
1206
1207 return TRUE;
1208}
1209
1210/*
7cb3f959
TC
1211
1212Return true if the supplied filename has a newline character
fa6c7d00 1213immediately before the first (hopefully only) NUL.
7cb3f959
TC
1214
1215My original look at this incorrectly used the len from SvPV(), but
1216that's incorrect, since we allow for a NUL in pv[len-1].
1217
1218So instead, strlen() and work from there.
1219
1220This allow for the user reading a filename, forgetting to chomp it,
1221then calling:
1222
1223 open my $foo, "$file\0";
1224
1225*/
1226
1227#ifdef PERL_CORE
1228
1229PERL_STATIC_INLINE bool
1230S_should_warn_nl(const char *pv) {
1231 STRLEN len;
1232
1233 PERL_ARGS_ASSERT_SHOULD_WARN_NL;
1234
1235 len = strlen(pv);
1236
1237 return len > 0 && pv[len-1] == '\n';
1238}
1239
1240#endif
1241
81d52ecd
JH
1242/* ------------------ pp.c, regcomp.c, toke.c, universal.c ------------ */
1243
1244#define MAX_CHARSET_NAME_LENGTH 2
1245
1246PERL_STATIC_INLINE const char *
1247get_regex_charset_name(const U32 flags, STRLEN* const lenp)
1248{
1249 /* Returns a string that corresponds to the name of the regex character set
1250 * given by 'flags', and *lenp is set the length of that string, which
1251 * cannot exceed MAX_CHARSET_NAME_LENGTH characters */
1252
1253 *lenp = 1;
1254 switch (get_regex_charset(flags)) {
1255 case REGEX_DEPENDS_CHARSET: return DEPENDS_PAT_MODS;
1256 case REGEX_LOCALE_CHARSET: return LOCALE_PAT_MODS;
1257 case REGEX_UNICODE_CHARSET: return UNICODE_PAT_MODS;
1258 case REGEX_ASCII_RESTRICTED_CHARSET: return ASCII_RESTRICT_PAT_MODS;
1259 case REGEX_ASCII_MORE_RESTRICTED_CHARSET:
1260 *lenp = 2;
1261 return ASCII_MORE_RESTRICT_PAT_MODS;
1262 }
1263 /* The NOT_REACHED; hides an assert() which has a rather complex
1264 * definition in perl.h. */
1265 NOT_REACHED; /* NOTREACHED */
1266 return "?"; /* Unknown */
1267}
1268
7cb3f959 1269/*
ed382232
TC
1270
1271Return false if any get magic is on the SV other than taint magic.
1272
1273*/
1274
1275PERL_STATIC_INLINE bool
1276S_sv_only_taint_gmagic(SV *sv) {
1277 MAGIC *mg = SvMAGIC(sv);
1278
1279 PERL_ARGS_ASSERT_SV_ONLY_TAINT_GMAGIC;
1280
1281 while (mg) {
1282 if (mg->mg_type != PERL_MAGIC_taint
1283 && !(mg->mg_flags & MGf_GSKIP)
1284 && mg->mg_virtual->svt_get) {
1285 return FALSE;
1286 }
1287 mg = mg->mg_moremagic;
1288 }
1289
1290 return TRUE;
1291}
1292
ed8ff0f3
DM
1293/* ------------------ cop.h ------------------------------------------- */
1294
1295
1296/* Enter a block. Push a new base context and return its address. */
1297
1298PERL_STATIC_INLINE PERL_CONTEXT *
1299S_cx_pushblock(pTHX_ U8 type, U8 gimme, SV** sp, I32 saveix)
1300{
1301 PERL_CONTEXT * cx;
1302
1303 PERL_ARGS_ASSERT_CX_PUSHBLOCK;
1304
1305 CXINC;
1306 cx = CX_CUR();
1307 cx->cx_type = type;
1308 cx->blk_gimme = gimme;
1309 cx->blk_oldsaveix = saveix;
4caf7d8c 1310 cx->blk_oldsp = (I32)(sp - PL_stack_base);
ed8ff0f3 1311 cx->blk_oldcop = PL_curcop;
4caf7d8c 1312 cx->blk_oldmarksp = (I32)(PL_markstack_ptr - PL_markstack);
ed8ff0f3
DM
1313 cx->blk_oldscopesp = PL_scopestack_ix;
1314 cx->blk_oldpm = PL_curpm;
ce8bb8d8 1315 cx->blk_old_tmpsfloor = PL_tmps_floor;
ed8ff0f3
DM
1316
1317 PL_tmps_floor = PL_tmps_ix;
1318 CX_DEBUG(cx, "PUSH");
1319 return cx;
1320}
1321
1322
1323/* Exit a block (RETURN and LAST). */
1324
1325PERL_STATIC_INLINE void
1326S_cx_popblock(pTHX_ PERL_CONTEXT *cx)
1327{
1328 PERL_ARGS_ASSERT_CX_POPBLOCK;
1329
1330 CX_DEBUG(cx, "POP");
1331 /* these 3 are common to cx_popblock and cx_topblock */
1332 PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
1333 PL_scopestack_ix = cx->blk_oldscopesp;
1334 PL_curpm = cx->blk_oldpm;
1335
1336 /* LEAVE_SCOPE() should have made this true. /(?{})/ cheats
1337 * and leaves a CX entry lying around for repeated use, so
1338 * skip for multicall */ \
1339 assert( (CxTYPE(cx) == CXt_SUB && CxMULTICALL(cx))
1340 || PL_savestack_ix == cx->blk_oldsaveix);
1341 PL_curcop = cx->blk_oldcop;
ce8bb8d8 1342 PL_tmps_floor = cx->blk_old_tmpsfloor;
ed8ff0f3
DM
1343}
1344
1345/* Continue a block elsewhere (e.g. NEXT, REDO, GOTO).
1346 * Whereas cx_popblock() restores the state to the point just before
1347 * cx_pushblock() was called, cx_topblock() restores it to the point just
1348 * *after* cx_pushblock() was called. */
1349
1350PERL_STATIC_INLINE void
1351S_cx_topblock(pTHX_ PERL_CONTEXT *cx)
1352{
1353 PERL_ARGS_ASSERT_CX_TOPBLOCK;
1354
1355 CX_DEBUG(cx, "TOP");
1356 /* these 3 are common to cx_popblock and cx_topblock */
1357 PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
1358 PL_scopestack_ix = cx->blk_oldscopesp;
1359 PL_curpm = cx->blk_oldpm;
1360
1361 PL_stack_sp = PL_stack_base + cx->blk_oldsp;
1362}
1363
1364
a73d8813
DM
1365PERL_STATIC_INLINE void
1366S_cx_pushsub(pTHX_ PERL_CONTEXT *cx, CV *cv, OP *retop, bool hasargs)
1367{
1368 U8 phlags = CX_PUSHSUB_GET_LVALUE_MASK(Perl_was_lvalue_sub);
1369
1370 PERL_ARGS_ASSERT_CX_PUSHSUB;
1371
3f6bd23a 1372 PERL_DTRACE_PROBE_ENTRY(cv);
a73d8813
DM
1373 cx->blk_sub.cv = cv;
1374 cx->blk_sub.olddepth = CvDEPTH(cv);
1375 cx->blk_sub.prevcomppad = PL_comppad;
1376 cx->cx_type |= (hasargs) ? CXp_HASARGS : 0;
1377 cx->blk_sub.retop = retop;
1378 SvREFCNT_inc_simple_void_NN(cv);
1379 cx->blk_u16 = PL_op->op_private & (phlags|OPpDEREF);
1380}
1381
1382
1383/* subsets of cx_popsub() */
1384
1385PERL_STATIC_INLINE void
1386S_cx_popsub_common(pTHX_ PERL_CONTEXT *cx)
1387{
1388 CV *cv;
1389
1390 PERL_ARGS_ASSERT_CX_POPSUB_COMMON;
1391 assert(CxTYPE(cx) == CXt_SUB);
1392
1393 PL_comppad = cx->blk_sub.prevcomppad;
1394 PL_curpad = LIKELY(PL_comppad) ? AvARRAY(PL_comppad) : NULL;
1395 cv = cx->blk_sub.cv;
1396 CvDEPTH(cv) = cx->blk_sub.olddepth;
1397 cx->blk_sub.cv = NULL;
1398 SvREFCNT_dec(cv);
1399}
1400
1401
1402/* handle the @_ part of leaving a sub */
1403
1404PERL_STATIC_INLINE void
1405S_cx_popsub_args(pTHX_ PERL_CONTEXT *cx)
1406{
1407 AV *av;
1408
1409 PERL_ARGS_ASSERT_CX_POPSUB_ARGS;
1410 assert(CxTYPE(cx) == CXt_SUB);
1411 assert(AvARRAY(MUTABLE_AV(
1412 PadlistARRAY(CvPADLIST(cx->blk_sub.cv))[
1413 CvDEPTH(cx->blk_sub.cv)])) == PL_curpad);
1414
1415 CX_POP_SAVEARRAY(cx);
1416 av = MUTABLE_AV(PAD_SVl(0));
1417 if (UNLIKELY(AvREAL(av)))
1418 /* abandon @_ if it got reified */
1419 clear_defarray(av, 0);
1420 else {
1421 CLEAR_ARGARRAY(av);
1422 }
1423}
1424
1425
1426PERL_STATIC_INLINE void
1427S_cx_popsub(pTHX_ PERL_CONTEXT *cx)
1428{
1429 PERL_ARGS_ASSERT_CX_POPSUB;
1430 assert(CxTYPE(cx) == CXt_SUB);
1431
3f6bd23a 1432 PERL_DTRACE_PROBE_RETURN(cx->blk_sub.cv);
a73d8813
DM
1433
1434 if (CxHASARGS(cx))
1435 cx_popsub_args(cx);
1436 cx_popsub_common(cx);
1437}
1438
1439
6a7d52cc
DM
1440PERL_STATIC_INLINE void
1441S_cx_pushformat(pTHX_ PERL_CONTEXT *cx, CV *cv, OP *retop, GV *gv)
1442{
1443 PERL_ARGS_ASSERT_CX_PUSHFORMAT;
1444
1445 cx->blk_format.cv = cv;
1446 cx->blk_format.retop = retop;
1447 cx->blk_format.gv = gv;
1448 cx->blk_format.dfoutgv = PL_defoutgv;
1449 cx->blk_format.prevcomppad = PL_comppad;
1450 cx->blk_u16 = 0;
1451
1452 SvREFCNT_inc_simple_void_NN(cv);
1453 CvDEPTH(cv)++;
1454 SvREFCNT_inc_void(cx->blk_format.dfoutgv);
1455}
1456
1457
1458PERL_STATIC_INLINE void
1459S_cx_popformat(pTHX_ PERL_CONTEXT *cx)
1460{
1461 CV *cv;
1462 GV *dfout;
1463
1464 PERL_ARGS_ASSERT_CX_POPFORMAT;
1465 assert(CxTYPE(cx) == CXt_FORMAT);
1466
1467 dfout = cx->blk_format.dfoutgv;
1468 setdefout(dfout);
1469 cx->blk_format.dfoutgv = NULL;
1470 SvREFCNT_dec_NN(dfout);
1471
1472 PL_comppad = cx->blk_format.prevcomppad;
1473 PL_curpad = LIKELY(PL_comppad) ? AvARRAY(PL_comppad) : NULL;
1474 cv = cx->blk_format.cv;
1475 cx->blk_format.cv = NULL;
1476 --CvDEPTH(cv);
1477 SvREFCNT_dec_NN(cv);
1478}
1479
1480
13febba5
DM
1481PERL_STATIC_INLINE void
1482S_cx_pusheval(pTHX_ PERL_CONTEXT *cx, OP *retop, SV *namesv)
1483{
1484 PERL_ARGS_ASSERT_CX_PUSHEVAL;
1485
1486 cx->blk_eval.retop = retop;
1487 cx->blk_eval.old_namesv = namesv;
1488 cx->blk_eval.old_eval_root = PL_eval_root;
1489 cx->blk_eval.cur_text = PL_parser ? PL_parser->linestr : NULL;
1490 cx->blk_eval.cv = NULL; /* later set by doeval_compile() */
1491 cx->blk_eval.cur_top_env = PL_top_env;
1492
4c57ced5 1493 assert(!(PL_in_eval & ~ 0x3F));
13febba5 1494 assert(!(PL_op->op_type & ~0x1FF));
4c57ced5 1495 cx->blk_u16 = (PL_in_eval & 0x3F) | ((U16)PL_op->op_type << 7);
13febba5
DM
1496}
1497
1498
1499PERL_STATIC_INLINE void
1500S_cx_popeval(pTHX_ PERL_CONTEXT *cx)
1501{
1502 SV *sv;
1503
1504 PERL_ARGS_ASSERT_CX_POPEVAL;
1505 assert(CxTYPE(cx) == CXt_EVAL);
1506
1507 PL_in_eval = CxOLD_IN_EVAL(cx);
4c57ced5 1508 assert(!(PL_in_eval & 0xc0));
13febba5
DM
1509 PL_eval_root = cx->blk_eval.old_eval_root;
1510 sv = cx->blk_eval.cur_text;
4c57ced5 1511 if (sv && CxEVAL_TXT_REFCNTED(cx)) {
13febba5
DM
1512 cx->blk_eval.cur_text = NULL;
1513 SvREFCNT_dec_NN(sv);
1514 }
1515
1516 sv = cx->blk_eval.old_namesv;
2a1e0dfe
DM
1517 if (sv) {
1518 cx->blk_eval.old_namesv = NULL;
1519 SvREFCNT_dec_NN(sv);
1520 }
13febba5 1521}
6a7d52cc 1522
a73d8813 1523
d1b6bf72
DM
1524/* push a plain loop, i.e.
1525 * { block }
1526 * while (cond) { block }
1527 * for (init;cond;continue) { block }
1528 * This loop can be last/redo'ed etc.
1529 */
1530
1531PERL_STATIC_INLINE void
1532S_cx_pushloop_plain(pTHX_ PERL_CONTEXT *cx)
1533{
1534 PERL_ARGS_ASSERT_CX_PUSHLOOP_PLAIN;
1535 cx->blk_loop.my_op = cLOOP;
1536}
1537
1538
1539/* push a true for loop, i.e.
1540 * for var (list) { block }
1541 */
1542
1543PERL_STATIC_INLINE void
1544S_cx_pushloop_for(pTHX_ PERL_CONTEXT *cx, void *itervarp, SV* itersave)
1545{
1546 PERL_ARGS_ASSERT_CX_PUSHLOOP_FOR;
1547
1548 /* this one line is common with cx_pushloop_plain */
1549 cx->blk_loop.my_op = cLOOP;
1550
1551 cx->blk_loop.itervar_u.svp = (SV**)itervarp;
1552 cx->blk_loop.itersave = itersave;
1553#ifdef USE_ITHREADS
1554 cx->blk_loop.oldcomppad = PL_comppad;
1555#endif
1556}
1557
1558
1559/* pop all loop types, including plain */
1560
1561PERL_STATIC_INLINE void
1562S_cx_poploop(pTHX_ PERL_CONTEXT *cx)
1563{
1564 PERL_ARGS_ASSERT_CX_POPLOOP;
1565
1566 assert(CxTYPE_is_LOOP(cx));
1567 if ( CxTYPE(cx) == CXt_LOOP_ARY
1568 || CxTYPE(cx) == CXt_LOOP_LAZYSV)
1569 {
1570 /* Free ary or cur. This assumes that state_u.ary.ary
1571 * aligns with state_u.lazysv.cur. See cx_dup() */
1572 SV *sv = cx->blk_loop.state_u.lazysv.cur;
1573 cx->blk_loop.state_u.lazysv.cur = NULL;
1574 SvREFCNT_dec_NN(sv);
1575 if (CxTYPE(cx) == CXt_LOOP_LAZYSV) {
1576 sv = cx->blk_loop.state_u.lazysv.end;
1577 cx->blk_loop.state_u.lazysv.end = NULL;
1578 SvREFCNT_dec_NN(sv);
1579 }
1580 }
1581 if (cx->cx_type & (CXp_FOR_PAD|CXp_FOR_GV)) {
1582 SV *cursv;
1583 SV **svp = (cx)->blk_loop.itervar_u.svp;
1584 if ((cx->cx_type & CXp_FOR_GV))
1585 svp = &GvSV((GV*)svp);
1586 cursv = *svp;
1587 *svp = cx->blk_loop.itersave;
1588 cx->blk_loop.itersave = NULL;
1589 SvREFCNT_dec(cursv);
1590 }
1591}
1592
2a7b7c61
DM
1593
1594PERL_STATIC_INLINE void
1595S_cx_pushwhen(pTHX_ PERL_CONTEXT *cx)
1596{
1597 PERL_ARGS_ASSERT_CX_PUSHWHEN;
1598
1599 cx->blk_givwhen.leave_op = cLOGOP->op_other;
1600}
1601
1602
1603PERL_STATIC_INLINE void
1604S_cx_popwhen(pTHX_ PERL_CONTEXT *cx)
1605{
1606 PERL_ARGS_ASSERT_CX_POPWHEN;
1607 assert(CxTYPE(cx) == CXt_WHEN);
1608
1609 PERL_UNUSED_ARG(cx);
59a14f30 1610 PERL_UNUSED_CONTEXT;
2a7b7c61
DM
1611 /* currently NOOP */
1612}
1613
1614
1615PERL_STATIC_INLINE void
1616S_cx_pushgiven(pTHX_ PERL_CONTEXT *cx, SV *orig_defsv)
1617{
1618 PERL_ARGS_ASSERT_CX_PUSHGIVEN;
1619
1620 cx->blk_givwhen.leave_op = cLOGOP->op_other;
1621 cx->blk_givwhen.defsv_save = orig_defsv;
1622}
1623
1624
1625PERL_STATIC_INLINE void
1626S_cx_popgiven(pTHX_ PERL_CONTEXT *cx)
1627{
1628 SV *sv;
1629
1630 PERL_ARGS_ASSERT_CX_POPGIVEN;
1631 assert(CxTYPE(cx) == CXt_GIVEN);
1632
1633 sv = GvSV(PL_defgv);
1634 GvSV(PL_defgv) = cx->blk_givwhen.defsv_save;
1635 cx->blk_givwhen.defsv_save = NULL;
1636 SvREFCNT_dec(sv);
1637}
1638
ed382232 1639/*
c8028aa6
TC
1640 * ex: set ts=8 sts=4 sw=4 et:
1641 */