This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
DLL descriptions on OS/2
[perl5.git] / pod / perlguts.pod
CommitLineData
a0d0e21e
LW
1=head1 NAME
2
954c1994 3perlguts - Introduction to the Perl API
a0d0e21e
LW
4
5=head1 DESCRIPTION
6
b3b6085d
PP
7This document attempts to describe how to use the Perl API, as well as
8containing some info on the basic workings of the Perl core. It is far
9from complete and probably contains many errors. Please refer any
10questions or comments to the author below.
a0d0e21e 11
0a753a76
PP
12=head1 Variables
13
5f05dabc 14=head2 Datatypes
a0d0e21e
LW
15
16Perl has three typedefs that handle Perl's three main data types:
17
18 SV Scalar Value
19 AV Array Value
20 HV Hash Value
21
d1b91892 22Each typedef has specific routines that manipulate the various data types.
a0d0e21e
LW
23
24=head2 What is an "IV"?
25
954c1994 26Perl uses a special typedef IV which is a simple signed integer type that is
5f05dabc 27guaranteed to be large enough to hold a pointer (as well as an integer).
954c1994 28Additionally, there is the UV, which is simply an unsigned IV.
a0d0e21e 29
d1b91892 30Perl also uses two special typedefs, I32 and I16, which will always be at
954c1994
GS
31least 32-bits and 16-bits long, respectively. (Again, there are U32 and U16,
32as well.)
a0d0e21e 33
54310121 34=head2 Working with SVs
a0d0e21e
LW
35
36An SV can be created and loaded with one command. There are four types of
a7dfe00a
MS
37values that can be loaded: an integer value (IV), a double (NV),
38a string (PV), and another scalar (SV).
a0d0e21e 39
9da1e3b5 40The six routines are:
a0d0e21e
LW
41
42 SV* newSViv(IV);
43 SV* newSVnv(double);
08105a92
GS
44 SV* newSVpv(const char*, int);
45 SV* newSVpvn(const char*, int);
46fc3d4c 46 SV* newSVpvf(const char*, ...);
a0d0e21e
LW
47 SV* newSVsv(SV*);
48
deb3007b 49To change the value of an *already-existing* SV, there are seven routines:
a0d0e21e
LW
50
51 void sv_setiv(SV*, IV);
deb3007b 52 void sv_setuv(SV*, UV);
a0d0e21e 53 void sv_setnv(SV*, double);
08105a92
GS
54 void sv_setpv(SV*, const char*);
55 void sv_setpvn(SV*, const char*, int)
46fc3d4c 56 void sv_setpvf(SV*, const char*, ...);
9abd00ed 57 void sv_setpvfn(SV*, const char*, STRLEN, va_list *, SV **, I32, bool);
a0d0e21e
LW
58 void sv_setsv(SV*, SV*);
59
60Notice that you can choose to specify the length of the string to be
9da1e3b5
MUN
61assigned by using C<sv_setpvn>, C<newSVpvn>, or C<newSVpv>, or you may
62allow Perl to calculate the length by using C<sv_setpv> or by specifying
630 as the second argument to C<newSVpv>. Be warned, though, that Perl will
64determine the string's length by using C<strlen>, which depends on the
9abd00ed
GS
65string terminating with a NUL character.
66
67The arguments of C<sv_setpvf> are processed like C<sprintf>, and the
68formatted output becomes the value.
69
70C<sv_setpvfn> is an analogue of C<vsprintf>, but it allows you to specify
71either a pointer to a variable argument list or the address and length of
72an array of SVs. The last argument points to a boolean; on return, if that
73boolean is true, then locale-specific information has been used to format
c2611fb3 74the string, and the string's contents are therefore untrustworthy (see
9abd00ed
GS
75L<perlsec>). This pointer may be NULL if that information is not
76important. Note that this function requires you to specify the length of
77the format.
78
7c906a97
AD
79STRLEN is an integer type (Size_t, usually defined as size_t in
80config.h) guaranteed to be large enough to represent the size of
81any string that perl can handle.
82
9da1e3b5
MUN
83The C<sv_set*()> functions are not generic enough to operate on values
84that have "magic". See L<Magic Virtual Tables> later in this document.
a0d0e21e 85
a3cb178b
GS
86All SVs that contain strings should be terminated with a NUL character.
87If it is not NUL-terminated there is a risk of
5f05dabc
PP
88core dumps and corruptions from code which passes the string to C
89functions or system calls which expect a NUL-terminated string.
90Perl's own functions typically add a trailing NUL for this reason.
91Nevertheless, you should be very careful when you pass a string stored
92in an SV to a C function or system call.
93
a0d0e21e
LW
94To access the actual value that an SV points to, you can use the macros:
95
96 SvIV(SV*)
954c1994 97 SvUV(SV*)
a0d0e21e
LW
98 SvNV(SV*)
99 SvPV(SV*, STRLEN len)
1fa8b10d 100 SvPV_nolen(SV*)
a0d0e21e 101
954c1994 102which will automatically coerce the actual scalar type into an IV, UV, double,
a0d0e21e
LW
103or string.
104
105In the C<SvPV> macro, the length of the string returned is placed into the
1fa8b10d
JD
106variable C<len> (this is a macro, so you do I<not> use C<&len>). If you do
107not care what the length of the data is, use the C<SvPV_nolen> macro.
108Historically the C<SvPV> macro with the global variable C<PL_na> has been
109used in this case. But that can be quite inefficient because C<PL_na> must
110be accessed in thread-local storage in threaded Perl. In any case, remember
111that Perl allows arbitrary strings of data that may both contain NULs and
112might not be terminated by a NUL.
a0d0e21e 113
ce2f5d8f
KA
114Also remember that C doesn't allow you to safely say C<foo(SvPV(s, len),
115len);>. It might work with your compiler, but it won't work for everyone.
116Break this sort of statement up into separate assignments:
117
b2f5ed49 118 SV *s;
ce2f5d8f
KA
119 STRLEN len;
120 char * ptr;
b2f5ed49 121 ptr = SvPV(s, len);
ce2f5d8f
KA
122 foo(ptr, len);
123
07fa94a1 124If you want to know if the scalar value is TRUE, you can use:
a0d0e21e
LW
125
126 SvTRUE(SV*)
127
128Although Perl will automatically grow strings for you, if you need to force
129Perl to allocate more memory for your SV, you can use the macro
130
131 SvGROW(SV*, STRLEN newlen)
132
133which will determine if more memory needs to be allocated. If so, it will
134call the function C<sv_grow>. Note that C<SvGROW> can only increase, not
5f05dabc
PP
135decrease, the allocated memory of an SV and that it does not automatically
136add a byte for the a trailing NUL (perl's own string functions typically do
8ebc5c01 137C<SvGROW(sv, len + 1)>).
a0d0e21e
LW
138
139If you have an SV and want to know what kind of data Perl thinks is stored
140in it, you can use the following macros to check the type of SV you have.
141
142 SvIOK(SV*)
143 SvNOK(SV*)
144 SvPOK(SV*)
145
146You can get and set the current length of the string stored in an SV with
147the following macros:
148
149 SvCUR(SV*)
150 SvCUR_set(SV*, I32 val)
151
cb1a09d0
AD
152You can also get a pointer to the end of the string stored in the SV
153with the macro:
154
155 SvEND(SV*)
156
157But note that these last three macros are valid only if C<SvPOK()> is true.
a0d0e21e 158
d1b91892
AD
159If you want to append something to the end of string stored in an C<SV*>,
160you can use the following functions:
161
08105a92 162 void sv_catpv(SV*, const char*);
e65f3abd 163 void sv_catpvn(SV*, const char*, STRLEN);
46fc3d4c 164 void sv_catpvf(SV*, const char*, ...);
9abd00ed 165 void sv_catpvfn(SV*, const char*, STRLEN, va_list *, SV **, I32, bool);
d1b91892
AD
166 void sv_catsv(SV*, SV*);
167
168The first function calculates the length of the string to be appended by
169using C<strlen>. In the second, you specify the length of the string
46fc3d4c 170yourself. The third function processes its arguments like C<sprintf> and
9abd00ed
GS
171appends the formatted output. The fourth function works like C<vsprintf>.
172You can specify the address and length of an array of SVs instead of the
173va_list argument. The fifth function extends the string stored in the first
174SV with the string stored in the second SV. It also forces the second SV
175to be interpreted as a string.
176
177The C<sv_cat*()> functions are not generic enough to operate on values that
178have "magic". See L<Magic Virtual Tables> later in this document.
d1b91892 179
a0d0e21e
LW
180If you know the name of a scalar variable, you can get a pointer to its SV
181by using the following:
182
4929bf7b 183 SV* get_sv("package::varname", FALSE);
a0d0e21e
LW
184
185This returns NULL if the variable does not exist.
186
d1b91892 187If you want to know if this variable (or any other SV) is actually C<defined>,
a0d0e21e
LW
188you can call:
189
190 SvOK(SV*)
191
9cde0e7f 192The scalar C<undef> value is stored in an SV instance called C<PL_sv_undef>. Its
a0d0e21e
LW
193address can be used whenever an C<SV*> is needed.
194
9cde0e7f
GS
195There are also the two values C<PL_sv_yes> and C<PL_sv_no>, which contain Boolean
196TRUE and FALSE values, respectively. Like C<PL_sv_undef>, their addresses can
a0d0e21e
LW
197be used whenever an C<SV*> is needed.
198
9cde0e7f 199Do not be fooled into thinking that C<(SV *) 0> is the same as C<&PL_sv_undef>.
a0d0e21e
LW
200Take this code:
201
202 SV* sv = (SV*) 0;
203 if (I-am-to-return-a-real-value) {
204 sv = sv_2mortal(newSViv(42));
205 }
206 sv_setsv(ST(0), sv);
207
208This code tries to return a new SV (which contains the value 42) if it should
04343c6d 209return a real value, or undef otherwise. Instead it has returned a NULL
a0d0e21e 210pointer which, somewhere down the line, will cause a segmentation violation,
9cde0e7f 211bus error, or just weird results. Change the zero to C<&PL_sv_undef> in the first
5f05dabc 212line and all will be well.
a0d0e21e
LW
213
214To free an SV that you've created, call C<SvREFCNT_dec(SV*)>. Normally this
3fe9a6f1 215call is not necessary (see L<Reference Counts and Mortality>).
a0d0e21e 216
94dde4fb
SC
217=head2 Offsets
218
219Perl provides the function C<sv_chop> to efficiently remove characters
220from the beginning of a string; you give it an SV and a pointer to
221somewhere inside the the PV, and it discards everything before the
222pointer. The efficiency comes by means of a little hack: instead of
223actually removing the characters, C<sv_chop> sets the flag C<OOK>
224(offset OK) to signal to other functions that the offset hack is in
225effect, and it puts the number of bytes chopped off into the IV field
226of the SV. It then moves the PV pointer (called C<SvPVX>) forward that
227many bytes, and adjusts C<SvCUR> and C<SvLEN>.
228
229Hence, at this point, the start of the buffer that we allocated lives
230at C<SvPVX(sv) - SvIV(sv)> in memory and the PV pointer is pointing
231into the middle of this allocated storage.
232
233This is best demonstrated by example:
234
235 % ./perl -Ilib -MDevel::Peek -le '$a="12345"; $a=~s/.//; Dump($a)'
236 SV = PVIV(0x8128450) at 0x81340f0
237 REFCNT = 1
238 FLAGS = (POK,OOK,pPOK)
239 IV = 1 (OFFSET)
240 PV = 0x8135781 ( "1" . ) "2345"\0
241 CUR = 4
242 LEN = 5
243
244Here the number of bytes chopped off (1) is put into IV, and
245C<Devel::Peek::Dump> helpfully reminds us that this is an offset. The
246portion of the string between the "real" and the "fake" beginnings is
247shown in parentheses, and the values of C<SvCUR> and C<SvLEN> reflect
248the fake beginning, not the real one.
249
319cef53
SC
250Something similar to the offset hack is perfomed on AVs to enable
251efficient shifting and splicing off the beginning of the array; while
252C<AvARRAY> points to the first element in the array that is visible from
253Perl, C<AvALLOC> points to the real start of the C array. These are
254usually the same, but a C<shift> operation can be carried out by
255increasing C<AvARRAY> by one and decreasing C<AvFILL> and C<AvLEN>.
256Again, the location of the real start of the C array only comes into
257play when freeing the array. See C<av_shift> in F<av.c>.
258
d1b91892 259=head2 What's Really Stored in an SV?
a0d0e21e
LW
260
261Recall that the usual method of determining the type of scalar you have is
5f05dabc 262to use C<Sv*OK> macros. Because a scalar can be both a number and a string,
d1b91892 263usually these macros will always return TRUE and calling the C<Sv*V>
a0d0e21e
LW
264macros will do the appropriate conversion of string to integer/double or
265integer/double to string.
266
267If you I<really> need to know if you have an integer, double, or string
268pointer in an SV, you can use the following three macros instead:
269
270 SvIOKp(SV*)
271 SvNOKp(SV*)
272 SvPOKp(SV*)
273
274These will tell you if you truly have an integer, double, or string pointer
d1b91892 275stored in your SV. The "p" stands for private.
a0d0e21e 276
07fa94a1 277In general, though, it's best to use the C<Sv*V> macros.
a0d0e21e 278
54310121 279=head2 Working with AVs
a0d0e21e 280
07fa94a1
JO
281There are two ways to create and load an AV. The first method creates an
282empty AV:
a0d0e21e
LW
283
284 AV* newAV();
285
54310121 286The second method both creates the AV and initially populates it with SVs:
a0d0e21e
LW
287
288 AV* av_make(I32 num, SV **ptr);
289
5f05dabc 290The second argument points to an array containing C<num> C<SV*>'s. Once the
54310121 291AV has been created, the SVs can be destroyed, if so desired.
a0d0e21e 292
54310121 293Once the AV has been created, the following operations are possible on AVs:
a0d0e21e
LW
294
295 void av_push(AV*, SV*);
296 SV* av_pop(AV*);
297 SV* av_shift(AV*);
298 void av_unshift(AV*, I32 num);
299
300These should be familiar operations, with the exception of C<av_unshift>.
301This routine adds C<num> elements at the front of the array with the C<undef>
302value. You must then use C<av_store> (described below) to assign values
303to these new elements.
304
305Here are some other functions:
306
5f05dabc 307 I32 av_len(AV*);
a0d0e21e 308 SV** av_fetch(AV*, I32 key, I32 lval);
a0d0e21e 309 SV** av_store(AV*, I32 key, SV* val);
a0d0e21e 310
5f05dabc
PP
311The C<av_len> function returns the highest index value in array (just
312like $#array in Perl). If the array is empty, -1 is returned. The
313C<av_fetch> function returns the value at index C<key>, but if C<lval>
314is non-zero, then C<av_fetch> will store an undef value at that index.
04343c6d
GS
315The C<av_store> function stores the value C<val> at index C<key>, and does
316not increment the reference count of C<val>. Thus the caller is responsible
317for taking care of that, and if C<av_store> returns NULL, the caller will
318have to decrement the reference count to avoid a memory leak. Note that
319C<av_fetch> and C<av_store> both return C<SV**>'s, not C<SV*>'s as their
320return value.
d1b91892 321
a0d0e21e 322 void av_clear(AV*);
a0d0e21e 323 void av_undef(AV*);
cb1a09d0 324 void av_extend(AV*, I32 key);
5f05dabc
PP
325
326The C<av_clear> function deletes all the elements in the AV* array, but
327does not actually delete the array itself. The C<av_undef> function will
328delete all the elements in the array plus the array itself. The
adc882cf
GS
329C<av_extend> function extends the array so that it contains at least C<key+1>
330elements. If C<key+1> is less than the currently allocated length of the array,
331then nothing is done.
a0d0e21e
LW
332
333If you know the name of an array variable, you can get a pointer to its AV
334by using the following:
335
4929bf7b 336 AV* get_av("package::varname", FALSE);
a0d0e21e
LW
337
338This returns NULL if the variable does not exist.
339
04343c6d
GS
340See L<Understanding the Magic of Tied Hashes and Arrays> for more
341information on how to use the array access functions on tied arrays.
342
54310121 343=head2 Working with HVs
a0d0e21e
LW
344
345To create an HV, you use the following routine:
346
347 HV* newHV();
348
54310121 349Once the HV has been created, the following operations are possible on HVs:
a0d0e21e 350
08105a92
GS
351 SV** hv_store(HV*, const char* key, U32 klen, SV* val, U32 hash);
352 SV** hv_fetch(HV*, const char* key, U32 klen, I32 lval);
a0d0e21e 353
5f05dabc
PP
354The C<klen> parameter is the length of the key being passed in (Note that
355you cannot pass 0 in as a value of C<klen> to tell Perl to measure the
356length of the key). The C<val> argument contains the SV pointer to the
54310121 357scalar being stored, and C<hash> is the precomputed hash value (zero if
5f05dabc
PP
358you want C<hv_store> to calculate it for you). The C<lval> parameter
359indicates whether this fetch is actually a part of a store operation, in
360which case a new undefined value will be added to the HV with the supplied
361key and C<hv_fetch> will return as if the value had already existed.
a0d0e21e 362
5f05dabc
PP
363Remember that C<hv_store> and C<hv_fetch> return C<SV**>'s and not just
364C<SV*>. To access the scalar value, you must first dereference the return
365value. However, you should check to make sure that the return value is
366not NULL before dereferencing it.
a0d0e21e
LW
367
368These two functions check if a hash table entry exists, and deletes it.
369
08105a92
GS
370 bool hv_exists(HV*, const char* key, U32 klen);
371 SV* hv_delete(HV*, const char* key, U32 klen, I32 flags);
a0d0e21e 372
5f05dabc
PP
373If C<flags> does not include the C<G_DISCARD> flag then C<hv_delete> will
374create and return a mortal copy of the deleted value.
375
a0d0e21e
LW
376And more miscellaneous functions:
377
378 void hv_clear(HV*);
a0d0e21e 379 void hv_undef(HV*);
5f05dabc
PP
380
381Like their AV counterparts, C<hv_clear> deletes all the entries in the hash
382table but does not actually delete the hash table. The C<hv_undef> deletes
383both the entries and the hash table itself.
a0d0e21e 384
d1b91892
AD
385Perl keeps the actual data in linked list of structures with a typedef of HE.
386These contain the actual key and value pointers (plus extra administrative
387overhead). The key is a string pointer; the value is an C<SV*>. However,
388once you have an C<HE*>, to get the actual key and value, use the routines
389specified below.
390
a0d0e21e
LW
391 I32 hv_iterinit(HV*);
392 /* Prepares starting point to traverse hash table */
393 HE* hv_iternext(HV*);
394 /* Get the next entry, and return a pointer to a
395 structure that has both the key and value */
396 char* hv_iterkey(HE* entry, I32* retlen);
397 /* Get the key from an HE structure and also return
398 the length of the key string */
cb1a09d0 399 SV* hv_iterval(HV*, HE* entry);
a0d0e21e
LW
400 /* Return a SV pointer to the value of the HE
401 structure */
cb1a09d0 402 SV* hv_iternextsv(HV*, char** key, I32* retlen);
d1b91892
AD
403 /* This convenience routine combines hv_iternext,
404 hv_iterkey, and hv_iterval. The key and retlen
405 arguments are return values for the key and its
406 length. The value is returned in the SV* argument */
a0d0e21e
LW
407
408If you know the name of a hash variable, you can get a pointer to its HV
409by using the following:
410
4929bf7b 411 HV* get_hv("package::varname", FALSE);
a0d0e21e
LW
412
413This returns NULL if the variable does not exist.
414
8ebc5c01 415The hash algorithm is defined in the C<PERL_HASH(hash, key, klen)> macro:
a0d0e21e 416
a0d0e21e 417 hash = 0;
ab192400
GS
418 while (klen--)
419 hash = (hash * 33) + *key++;
87275199 420 hash = hash + (hash >> 5); /* after 5.6 */
ab192400 421
87275199 422The last step was added in version 5.6 to improve distribution of
ab192400 423lower bits in the resulting hash value.
a0d0e21e 424
04343c6d
GS
425See L<Understanding the Magic of Tied Hashes and Arrays> for more
426information on how to use the hash access functions on tied hashes.
427
1e422769
PP
428=head2 Hash API Extensions
429
430Beginning with version 5.004, the following functions are also supported:
431
432 HE* hv_fetch_ent (HV* tb, SV* key, I32 lval, U32 hash);
433 HE* hv_store_ent (HV* tb, SV* key, SV* val, U32 hash);
c47ff5f1 434
1e422769
PP
435 bool hv_exists_ent (HV* tb, SV* key, U32 hash);
436 SV* hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
c47ff5f1 437
1e422769
PP
438 SV* hv_iterkeysv (HE* entry);
439
440Note that these functions take C<SV*> keys, which simplifies writing
441of extension code that deals with hash structures. These functions
442also allow passing of C<SV*> keys to C<tie> functions without forcing
443you to stringify the keys (unlike the previous set of functions).
444
445They also return and accept whole hash entries (C<HE*>), making their
446use more efficient (since the hash number for a particular string
4a4eefd0
GS
447doesn't have to be recomputed every time). See L<perlapi> for detailed
448descriptions.
1e422769
PP
449
450The following macros must always be used to access the contents of hash
451entries. Note that the arguments to these macros must be simple
452variables, since they may get evaluated more than once. See
4a4eefd0 453L<perlapi> for detailed descriptions of these macros.
1e422769
PP
454
455 HePV(HE* he, STRLEN len)
456 HeVAL(HE* he)
457 HeHASH(HE* he)
458 HeSVKEY(HE* he)
459 HeSVKEY_force(HE* he)
460 HeSVKEY_set(HE* he, SV* sv)
461
462These two lower level macros are defined, but must only be used when
463dealing with keys that are not C<SV*>s:
464
465 HeKEY(HE* he)
466 HeKLEN(HE* he)
467
04343c6d
GS
468Note that both C<hv_store> and C<hv_store_ent> do not increment the
469reference count of the stored C<val>, which is the caller's responsibility.
470If these functions return a NULL value, the caller will usually have to
471decrement the reference count of C<val> to avoid a memory leak.
1e422769 472
a0d0e21e
LW
473=head2 References
474
d1b91892
AD
475References are a special type of scalar that point to other data types
476(including references).
a0d0e21e 477
07fa94a1 478To create a reference, use either of the following functions:
a0d0e21e 479
5f05dabc
PP
480 SV* newRV_inc((SV*) thing);
481 SV* newRV_noinc((SV*) thing);
a0d0e21e 482
5f05dabc 483The C<thing> argument can be any of an C<SV*>, C<AV*>, or C<HV*>. The
07fa94a1
JO
484functions are identical except that C<newRV_inc> increments the reference
485count of the C<thing>, while C<newRV_noinc> does not. For historical
486reasons, C<newRV> is a synonym for C<newRV_inc>.
487
488Once you have a reference, you can use the following macro to dereference
489the reference:
a0d0e21e
LW
490
491 SvRV(SV*)
492
493then call the appropriate routines, casting the returned C<SV*> to either an
d1b91892 494C<AV*> or C<HV*>, if required.
a0d0e21e 495
d1b91892 496To determine if an SV is a reference, you can use the following macro:
a0d0e21e
LW
497
498 SvROK(SV*)
499
07fa94a1
JO
500To discover what type of value the reference refers to, use the following
501macro and then check the return value.
d1b91892
AD
502
503 SvTYPE(SvRV(SV*))
504
505The most useful types that will be returned are:
506
507 SVt_IV Scalar
508 SVt_NV Scalar
509 SVt_PV Scalar
5f05dabc 510 SVt_RV Scalar
d1b91892
AD
511 SVt_PVAV Array
512 SVt_PVHV Hash
513 SVt_PVCV Code
5f05dabc
PP
514 SVt_PVGV Glob (possible a file handle)
515 SVt_PVMG Blessed or Magical Scalar
516
517 See the sv.h header file for more details.
d1b91892 518
cb1a09d0
AD
519=head2 Blessed References and Class Objects
520
521References are also used to support object-oriented programming. In the
522OO lexicon, an object is simply a reference that has been blessed into a
523package (or class). Once blessed, the programmer may now use the reference
524to access the various methods in the class.
525
526A reference can be blessed into a package with the following function:
527
528 SV* sv_bless(SV* sv, HV* stash);
529
530The C<sv> argument must be a reference. The C<stash> argument specifies
3fe9a6f1 531which class the reference will belong to. See
2ae324a7 532L<Stashes and Globs> for information on converting class names into stashes.
cb1a09d0
AD
533
534/* Still under construction */
535
536Upgrades rv to reference if not already one. Creates new SV for rv to
8ebc5c01
PP
537point to. If C<classname> is non-null, the SV is blessed into the specified
538class. SV is returned.
cb1a09d0 539
08105a92 540 SV* newSVrv(SV* rv, const char* classname);
cb1a09d0 541
e1c57cef 542Copies integer, unsigned integer or double into an SV whose reference is C<rv>. SV is blessed
8ebc5c01 543if C<classname> is non-null.
cb1a09d0 544
08105a92 545 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
e1c57cef 546 SV* sv_setref_uv(SV* rv, const char* classname, UV uv);
08105a92 547 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
cb1a09d0 548
5f05dabc 549Copies the pointer value (I<the address, not the string!>) into an SV whose
8ebc5c01 550reference is rv. SV is blessed if C<classname> is non-null.
cb1a09d0 551
08105a92 552 SV* sv_setref_pv(SV* rv, const char* classname, PV iv);
cb1a09d0 553
8ebc5c01
PP
554Copies string into an SV whose reference is C<rv>. Set length to 0 to let
555Perl calculate the string length. SV is blessed if C<classname> is non-null.
cb1a09d0 556
e65f3abd 557 SV* sv_setref_pvn(SV* rv, const char* classname, PV iv, STRLEN length);
cb1a09d0 558
9abd00ed
GS
559Tests whether the SV is blessed into the specified class. It does not
560check inheritance relationships.
561
08105a92 562 int sv_isa(SV* sv, const char* name);
9abd00ed
GS
563
564Tests whether the SV is a reference to a blessed object.
565
566 int sv_isobject(SV* sv);
567
568Tests whether the SV is derived from the specified class. SV can be either
569a reference to a blessed object or a string containing a class name. This
570is the function implementing the C<UNIVERSAL::isa> functionality.
571
08105a92 572 bool sv_derived_from(SV* sv, const char* name);
9abd00ed
GS
573
574To check if you've got an object derived from a specific class you have
575to write:
576
577 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
cb1a09d0 578
5f05dabc 579=head2 Creating New Variables
cb1a09d0 580
5f05dabc
PP
581To create a new Perl variable with an undef value which can be accessed from
582your Perl script, use the following routines, depending on the variable type.
cb1a09d0 583
4929bf7b
GS
584 SV* get_sv("package::varname", TRUE);
585 AV* get_av("package::varname", TRUE);
586 HV* get_hv("package::varname", TRUE);
cb1a09d0
AD
587
588Notice the use of TRUE as the second parameter. The new variable can now
589be set, using the routines appropriate to the data type.
590
5f05dabc
PP
591There are additional macros whose values may be bitwise OR'ed with the
592C<TRUE> argument to enable certain extra features. Those bits are:
cb1a09d0 593
5f05dabc 594 GV_ADDMULTI Marks the variable as multiply defined, thus preventing the
54310121 595 "Name <varname> used only once: possible typo" warning.
07fa94a1
JO
596 GV_ADDWARN Issues the warning "Had to create <varname> unexpectedly" if
597 the variable did not exist before the function was called.
cb1a09d0 598
07fa94a1
JO
599If you do not specify a package name, the variable is created in the current
600package.
cb1a09d0 601
5f05dabc 602=head2 Reference Counts and Mortality
a0d0e21e 603
54310121
PP
604Perl uses an reference count-driven garbage collection mechanism. SVs,
605AVs, or HVs (xV for short in the following) start their life with a
55497cff 606reference count of 1. If the reference count of an xV ever drops to 0,
07fa94a1 607then it will be destroyed and its memory made available for reuse.
55497cff
PP
608
609This normally doesn't happen at the Perl level unless a variable is
5f05dabc
PP
610undef'ed or the last variable holding a reference to it is changed or
611overwritten. At the internal level, however, reference counts can be
55497cff
PP
612manipulated with the following macros:
613
614 int SvREFCNT(SV* sv);
5f05dabc 615 SV* SvREFCNT_inc(SV* sv);
55497cff
PP
616 void SvREFCNT_dec(SV* sv);
617
618However, there is one other function which manipulates the reference
07fa94a1
JO
619count of its argument. The C<newRV_inc> function, you will recall,
620creates a reference to the specified argument. As a side effect,
621it increments the argument's reference count. If this is not what
622you want, use C<newRV_noinc> instead.
623
624For example, imagine you want to return a reference from an XSUB function.
625Inside the XSUB routine, you create an SV which initially has a reference
626count of one. Then you call C<newRV_inc>, passing it the just-created SV.
5f05dabc
PP
627This returns the reference as a new SV, but the reference count of the
628SV you passed to C<newRV_inc> has been incremented to two. Now you
07fa94a1
JO
629return the reference from the XSUB routine and forget about the SV.
630But Perl hasn't! Whenever the returned reference is destroyed, the
631reference count of the original SV is decreased to one and nothing happens.
632The SV will hang around without any way to access it until Perl itself
633terminates. This is a memory leak.
5f05dabc
PP
634
635The correct procedure, then, is to use C<newRV_noinc> instead of
faed5253
JO
636C<newRV_inc>. Then, if and when the last reference is destroyed,
637the reference count of the SV will go to zero and it will be destroyed,
07fa94a1 638stopping any memory leak.
55497cff 639
5f05dabc 640There are some convenience functions available that can help with the
54310121 641destruction of xVs. These functions introduce the concept of "mortality".
07fa94a1
JO
642An xV that is mortal has had its reference count marked to be decremented,
643but not actually decremented, until "a short time later". Generally the
644term "short time later" means a single Perl statement, such as a call to
54310121 645an XSUB function. The actual determinant for when mortal xVs have their
07fa94a1
JO
646reference count decremented depends on two macros, SAVETMPS and FREETMPS.
647See L<perlcall> and L<perlxs> for more details on these macros.
55497cff
PP
648
649"Mortalization" then is at its simplest a deferred C<SvREFCNT_dec>.
650However, if you mortalize a variable twice, the reference count will
651later be decremented twice.
652
653You should be careful about creating mortal variables. Strange things
654can happen if you make the same value mortal within multiple contexts,
5f05dabc 655or if you make a variable mortal multiple times.
a0d0e21e
LW
656
657To create a mortal variable, use the functions:
658
659 SV* sv_newmortal()
660 SV* sv_2mortal(SV*)
661 SV* sv_mortalcopy(SV*)
662
5f05dabc
PP
663The first call creates a mortal SV, the second converts an existing
664SV to a mortal SV (and thus defers a call to C<SvREFCNT_dec>), and the
665third creates a mortal copy of an existing SV.
a0d0e21e 666
54310121 667The mortal routines are not just for SVs -- AVs and HVs can be
faed5253 668made mortal by passing their address (type-casted to C<SV*>) to the
07fa94a1 669C<sv_2mortal> or C<sv_mortalcopy> routines.
a0d0e21e 670
5f05dabc 671=head2 Stashes and Globs
a0d0e21e 672
aa689395
PP
673A "stash" is a hash that contains all of the different objects that
674are contained within a package. Each key of the stash is a symbol
675name (shared by all the different types of objects that have the same
676name), and each value in the hash table is a GV (Glob Value). This GV
677in turn contains references to the various objects of that name,
678including (but not limited to) the following:
cb1a09d0 679
a0d0e21e
LW
680 Scalar Value
681 Array Value
682 Hash Value
a3cb178b 683 I/O Handle
a0d0e21e
LW
684 Format
685 Subroutine
686
9cde0e7f 687There is a single stash called "PL_defstash" that holds the items that exist
5f05dabc
PP
688in the "main" package. To get at the items in other packages, append the
689string "::" to the package name. The items in the "Foo" package are in
9cde0e7f 690the stash "Foo::" in PL_defstash. The items in the "Bar::Baz" package are
5f05dabc 691in the stash "Baz::" in "Bar::"'s stash.
a0d0e21e 692
d1b91892 693To get the stash pointer for a particular package, use the function:
a0d0e21e 694
08105a92 695 HV* gv_stashpv(const char* name, I32 create)
a0d0e21e
LW
696 HV* gv_stashsv(SV*, I32 create)
697
698The first function takes a literal string, the second uses the string stored
d1b91892 699in the SV. Remember that a stash is just a hash table, so you get back an
cb1a09d0 700C<HV*>. The C<create> flag will create a new package if it is set.
a0d0e21e
LW
701
702The name that C<gv_stash*v> wants is the name of the package whose symbol table
703you want. The default package is called C<main>. If you have multiply nested
d1b91892
AD
704packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
705language itself.
a0d0e21e
LW
706
707Alternately, if you have an SV that is a blessed reference, you can find
708out the stash pointer by using:
709
710 HV* SvSTASH(SvRV(SV*));
711
712then use the following to get the package name itself:
713
714 char* HvNAME(HV* stash);
715
5f05dabc
PP
716If you need to bless or re-bless an object you can use the following
717function:
a0d0e21e
LW
718
719 SV* sv_bless(SV*, HV* stash)
720
721where the first argument, an C<SV*>, must be a reference, and the second
722argument is a stash. The returned C<SV*> can now be used in the same way
723as any other SV.
724
d1b91892
AD
725For more information on references and blessings, consult L<perlref>.
726
54310121 727=head2 Double-Typed SVs
0a753a76
PP
728
729Scalar variables normally contain only one type of value, an integer,
730double, pointer, or reference. Perl will automatically convert the
731actual scalar data from the stored type into the requested type.
732
733Some scalar variables contain more than one type of scalar data. For
734example, the variable C<$!> contains either the numeric value of C<errno>
735or its string equivalent from either C<strerror> or C<sys_errlist[]>.
736
737To force multiple data values into an SV, you must do two things: use the
738C<sv_set*v> routines to add the additional scalar type, then set a flag
739so that Perl will believe it contains more than one type of data. The
740four macros to set the flags are:
741
742 SvIOK_on
743 SvNOK_on
744 SvPOK_on
745 SvROK_on
746
747The particular macro you must use depends on which C<sv_set*v> routine
748you called first. This is because every C<sv_set*v> routine turns on
749only the bit for the particular type of data being set, and turns off
750all the rest.
751
752For example, to create a new Perl variable called "dberror" that contains
753both the numeric and descriptive string error values, you could use the
754following code:
755
756 extern int dberror;
757 extern char *dberror_list;
758
4929bf7b 759 SV* sv = get_sv("dberror", TRUE);
0a753a76
PP
760 sv_setiv(sv, (IV) dberror);
761 sv_setpv(sv, dberror_list[dberror]);
762 SvIOK_on(sv);
763
764If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
765macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
766
767=head2 Magic Variables
a0d0e21e 768
d1b91892
AD
769[This section still under construction. Ignore everything here. Post no
770bills. Everything not permitted is forbidden.]
771
d1b91892
AD
772Any SV may be magical, that is, it has special features that a normal
773SV does not have. These features are stored in the SV structure in a
5f05dabc 774linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
d1b91892
AD
775
776 struct magic {
777 MAGIC* mg_moremagic;
778 MGVTBL* mg_virtual;
779 U16 mg_private;
780 char mg_type;
781 U8 mg_flags;
782 SV* mg_obj;
783 char* mg_ptr;
784 I32 mg_len;
785 };
786
787Note this is current as of patchlevel 0, and could change at any time.
788
789=head2 Assigning Magic
790
791Perl adds magic to an SV using the sv_magic function:
792
08105a92 793 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
d1b91892
AD
794
795The C<sv> argument is a pointer to the SV that is to acquire a new magical
796feature.
797
798If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
799set the C<SVt_PVMG> flag for the C<sv>. Perl then continues by adding
800it to the beginning of the linked list of magical features. Any prior
801entry of the same type of magic is deleted. Note that this can be
5fb8527f 802overridden, and multiple instances of the same type of magic can be
d1b91892
AD
803associated with an SV.
804
54310121
PP
805The C<name> and C<namlen> arguments are used to associate a string with
806the magic, typically the name of a variable. C<namlen> is stored in the
807C<mg_len> field and if C<name> is non-null and C<namlen> >= 0 a malloc'd
d1b91892
AD
808copy of the name is stored in C<mg_ptr> field.
809
810The sv_magic function uses C<how> to determine which, if any, predefined
811"Magic Virtual Table" should be assigned to the C<mg_virtual> field.
cb1a09d0
AD
812See the "Magic Virtual Table" section below. The C<how> argument is also
813stored in the C<mg_type> field.
d1b91892
AD
814
815The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
816structure. If it is not the same as the C<sv> argument, the reference
817count of the C<obj> object is incremented. If it is the same, or if
04343c6d 818the C<how> argument is "#", or if it is a NULL pointer, then C<obj> is
d1b91892
AD
819merely stored, without the reference count being incremented.
820
cb1a09d0
AD
821There is also a function to add magic to an C<HV>:
822
823 void hv_magic(HV *hv, GV *gv, int how);
824
825This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
826
827To remove the magic from an SV, call the function sv_unmagic:
828
829 void sv_unmagic(SV *sv, int type);
830
831The C<type> argument should be equal to the C<how> value when the C<SV>
832was initially made magical.
833
d1b91892
AD
834=head2 Magic Virtual Tables
835
836The C<mg_virtual> field in the C<MAGIC> structure is a pointer to a
837C<MGVTBL>, which is a structure of function pointers and stands for
838"Magic Virtual Table" to handle the various operations that might be
839applied to that variable.
840
841The C<MGVTBL> has five pointers to the following routine types:
842
843 int (*svt_get)(SV* sv, MAGIC* mg);
844 int (*svt_set)(SV* sv, MAGIC* mg);
845 U32 (*svt_len)(SV* sv, MAGIC* mg);
846 int (*svt_clear)(SV* sv, MAGIC* mg);
847 int (*svt_free)(SV* sv, MAGIC* mg);
848
849This MGVTBL structure is set at compile-time in C<perl.h> and there are
850currently 19 types (or 21 with overloading turned on). These different
851structures contain pointers to various routines that perform additional
852actions depending on which function is being called.
853
854 Function pointer Action taken
855 ---------------- ------------
856 svt_get Do something after the value of the SV is retrieved.
857 svt_set Do something after the SV is assigned a value.
858 svt_len Report on the SV's length.
859 svt_clear Clear something the SV represents.
860 svt_free Free any extra storage associated with the SV.
861
862For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
863to an C<mg_type> of '\0') contains:
864
865 { magic_get, magic_set, magic_len, 0, 0 }
866
867Thus, when an SV is determined to be magical and of type '\0', if a get
868operation is being performed, the routine C<magic_get> is called. All
869the various routines for the various magical types begin with C<magic_>.
954c1994
GS
870NOTE: the magic routines are not considered part of the Perl API, and may
871not be exported by the Perl library.
d1b91892
AD
872
873The current kinds of Magic Virtual Tables are:
874
bdbeb323 875 mg_type MGVTBL Type of magic
5f05dabc 876 ------- ------ ----------------------------
bdbeb323
SM
877 \0 vtbl_sv Special scalar variable
878 A vtbl_amagic %OVERLOAD hash
879 a vtbl_amagicelem %OVERLOAD hash element
880 c (none) Holds overload table (AMT) on stash
881 B vtbl_bm Boyer-Moore (fast string search)
c2e66d9e
GS
882 D vtbl_regdata Regex match position data (@+ and @- vars)
883 d vtbl_regdatum Regex match position data element
d1b91892
AD
884 E vtbl_env %ENV hash
885 e vtbl_envelem %ENV hash element
bdbeb323
SM
886 f vtbl_fm Formline ('compiled' format)
887 g vtbl_mglob m//g target / study()ed string
d1b91892
AD
888 I vtbl_isa @ISA array
889 i vtbl_isaelem @ISA array element
bdbeb323
SM
890 k vtbl_nkeys scalar(keys()) lvalue
891 L (none) Debugger %_<filename
892 l vtbl_dbline Debugger %_<filename element
44a8e56a 893 o vtbl_collxfrm Locale transformation
bdbeb323
SM
894 P vtbl_pack Tied array or hash
895 p vtbl_packelem Tied array or hash element
896 q vtbl_packelem Tied scalar or handle
897 S vtbl_sig %SIG hash
898 s vtbl_sigelem %SIG hash element
d1b91892 899 t vtbl_taint Taintedness
bdbeb323
SM
900 U vtbl_uvar Available for use by extensions
901 v vtbl_vec vec() lvalue
902 x vtbl_substr substr() lvalue
903 y vtbl_defelem Shadow "foreach" iterator variable /
904 smart parameter vivification
905 * vtbl_glob GV (typeglob)
906 # vtbl_arylen Array length ($#ary)
907 . vtbl_pos pos() lvalue
908 ~ (none) Available for use by extensions
d1b91892 909
68dc0745
PP
910When an uppercase and lowercase letter both exist in the table, then the
911uppercase letter is used to represent some kind of composite type (a list
912or a hash), and the lowercase letter is used to represent an element of
d1b91892
AD
913that composite type.
914
bdbeb323
SM
915The '~' and 'U' magic types are defined specifically for use by
916extensions and will not be used by perl itself. Extensions can use
917'~' magic to 'attach' private information to variables (typically
918objects). This is especially useful because there is no way for
919normal perl code to corrupt this private information (unlike using
920extra elements of a hash object).
921
922Similarly, 'U' magic can be used much like tie() to call a C function
923any time a scalar's value is used or changed. The C<MAGIC>'s
924C<mg_ptr> field points to a C<ufuncs> structure:
925
926 struct ufuncs {
927 I32 (*uf_val)(IV, SV*);
928 I32 (*uf_set)(IV, SV*);
929 IV uf_index;
930 };
931
932When the SV is read from or written to, the C<uf_val> or C<uf_set>
933function will be called with C<uf_index> as the first arg and a
1526ead6
AB
934pointer to the SV as the second. A simple example of how to add 'U'
935magic is shown below. Note that the ufuncs structure is copied by
936sv_magic, so you can safely allocate it on the stack.
937
938 void
939 Umagic(sv)
940 SV *sv;
941 PREINIT:
942 struct ufuncs uf;
943 CODE:
944 uf.uf_val = &my_get_fn;
945 uf.uf_set = &my_set_fn;
946 uf.uf_index = 0;
947 sv_magic(sv, 0, 'U', (char*)&uf, sizeof(uf));
5f05dabc 948
bdbeb323
SM
949Note that because multiple extensions may be using '~' or 'U' magic,
950it is important for extensions to take extra care to avoid conflict.
951Typically only using the magic on objects blessed into the same class
952as the extension is sufficient. For '~' magic, it may also be
953appropriate to add an I32 'signature' at the top of the private data
954area and check that.
5f05dabc 955
ef50df4b
GS
956Also note that the C<sv_set*()> and C<sv_cat*()> functions described
957earlier do B<not> invoke 'set' magic on their targets. This must
958be done by the user either by calling the C<SvSETMAGIC()> macro after
959calling these functions, or by using one of the C<sv_set*_mg()> or
960C<sv_cat*_mg()> functions. Similarly, generic C code must call the
961C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
962obtained from external sources in functions that don't handle magic.
4a4eefd0 963See L<perlapi> for a description of these functions.
189b2af5
GS
964For example, calls to the C<sv_cat*()> functions typically need to be
965followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
966since their implementation handles 'get' magic.
967
d1b91892
AD
968=head2 Finding Magic
969
970 MAGIC* mg_find(SV*, int type); /* Finds the magic pointer of that type */
971
972This routine returns a pointer to the C<MAGIC> structure stored in the SV.
973If the SV does not have that magical feature, C<NULL> is returned. Also,
54310121 974if the SV is not of type SVt_PVMG, Perl may core dump.
d1b91892 975
08105a92 976 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
d1b91892
AD
977
978This routine checks to see what types of magic C<sv> has. If the mg_type
68dc0745
PP
979field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
980the mg_type field is changed to be the lowercase letter.
a0d0e21e 981
04343c6d
GS
982=head2 Understanding the Magic of Tied Hashes and Arrays
983
984Tied hashes and arrays are magical beasts of the 'P' magic type.
9edb2b46
GS
985
986WARNING: As of the 5.004 release, proper usage of the array and hash
987access functions requires understanding a few caveats. Some
988of these caveats are actually considered bugs in the API, to be fixed
989in later releases, and are bracketed with [MAYCHANGE] below. If
990you find yourself actually applying such information in this section, be
991aware that the behavior may change in the future, umm, without warning.
04343c6d 992
1526ead6
AB
993The perl tie function associates a variable with an object that implements
994the various GET, SET etc methods. To perform the equivalent of the perl
995tie function from an XSUB, you must mimic this behaviour. The code below
996carries out the necessary steps - firstly it creates a new hash, and then
997creates a second hash which it blesses into the class which will implement
998the tie methods. Lastly it ties the two hashes together, and returns a
999reference to the new tied hash. Note that the code below does NOT call the
1000TIEHASH method in the MyTie class -
1001see L<Calling Perl Routines from within C Programs> for details on how
1002to do this.
1003
1004 SV*
1005 mytie()
1006 PREINIT:
1007 HV *hash;
1008 HV *stash;
1009 SV *tie;
1010 CODE:
1011 hash = newHV();
1012 tie = newRV_noinc((SV*)newHV());
1013 stash = gv_stashpv("MyTie", TRUE);
1014 sv_bless(tie, stash);
1015 hv_magic(hash, tie, 'P');
1016 RETVAL = newRV_noinc(hash);
1017 OUTPUT:
1018 RETVAL
1019
04343c6d
GS
1020The C<av_store> function, when given a tied array argument, merely
1021copies the magic of the array onto the value to be "stored", using
1022C<mg_copy>. It may also return NULL, indicating that the value did not
9edb2b46
GS
1023actually need to be stored in the array. [MAYCHANGE] After a call to
1024C<av_store> on a tied array, the caller will usually need to call
1025C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1026TIEARRAY object. If C<av_store> did return NULL, a call to
1027C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1028leak. [/MAYCHANGE]
04343c6d
GS
1029
1030The previous paragraph is applicable verbatim to tied hash access using the
1031C<hv_store> and C<hv_store_ent> functions as well.
1032
1033C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1034C<hv_fetch_ent> actually return an undefined mortal value whose magic
1035has been initialized using C<mg_copy>. Note the value so returned does not
9edb2b46
GS
1036need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1037need to call C<mg_get()> on the returned value in order to actually invoke
1038the perl level "FETCH" method on the underlying TIE object. Similarly,
04343c6d
GS
1039you may also call C<mg_set()> on the return value after possibly assigning
1040a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
9edb2b46 1041method on the TIE object. [/MAYCHANGE]
04343c6d 1042
9edb2b46 1043[MAYCHANGE]
04343c6d
GS
1044In other words, the array or hash fetch/store functions don't really
1045fetch and store actual values in the case of tied arrays and hashes. They
1046merely call C<mg_copy> to attach magic to the values that were meant to be
1047"stored" or "fetched". Later calls to C<mg_get> and C<mg_set> actually
1048do the job of invoking the TIE methods on the underlying objects. Thus
9edb2b46 1049the magic mechanism currently implements a kind of lazy access to arrays
04343c6d
GS
1050and hashes.
1051
1052Currently (as of perl version 5.004), use of the hash and array access
1053functions requires the user to be aware of whether they are operating on
9edb2b46
GS
1054"normal" hashes and arrays, or on their tied variants. The API may be
1055changed to provide more transparent access to both tied and normal data
1056types in future versions.
1057[/MAYCHANGE]
04343c6d
GS
1058
1059You would do well to understand that the TIEARRAY and TIEHASH interfaces
1060are mere sugar to invoke some perl method calls while using the uniform hash
1061and array syntax. The use of this sugar imposes some overhead (typically
1062about two to four extra opcodes per FETCH/STORE operation, in addition to
1063the creation of all the mortal variables required to invoke the methods).
1064This overhead will be comparatively small if the TIE methods are themselves
1065substantial, but if they are only a few statements long, the overhead
1066will not be insignificant.
1067
d1c897a1
IZ
1068=head2 Localizing changes
1069
1070Perl has a very handy construction
1071
1072 {
1073 local $var = 2;
1074 ...
1075 }
1076
1077This construction is I<approximately> equivalent to
1078
1079 {
1080 my $oldvar = $var;
1081 $var = 2;
1082 ...
1083 $var = $oldvar;
1084 }
1085
1086The biggest difference is that the first construction would
1087reinstate the initial value of $var, irrespective of how control exits
1088the block: C<goto>, C<return>, C<die>/C<eval> etc. It is a little bit
1089more efficient as well.
1090
1091There is a way to achieve a similar task from C via Perl API: create a
1092I<pseudo-block>, and arrange for some changes to be automatically
1093undone at the end of it, either explicit, or via a non-local exit (via
1094die()). A I<block>-like construct is created by a pair of
b687b08b
TC
1095C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1096Such a construct may be created specially for some important localized
1097task, or an existing one (like boundaries of enclosing Perl
1098subroutine/block, or an existing pair for freeing TMPs) may be
1099used. (In the second case the overhead of additional localization must
1100be almost negligible.) Note that any XSUB is automatically enclosed in
1101an C<ENTER>/C<LEAVE> pair.
d1c897a1
IZ
1102
1103Inside such a I<pseudo-block> the following service is available:
1104
13a2d996 1105=over 4
d1c897a1
IZ
1106
1107=item C<SAVEINT(int i)>
1108
1109=item C<SAVEIV(IV i)>
1110
1111=item C<SAVEI32(I32 i)>
1112
1113=item C<SAVELONG(long i)>
1114
1115These macros arrange things to restore the value of integer variable
1116C<i> at the end of enclosing I<pseudo-block>.
1117
1118=item C<SAVESPTR(s)>
1119
1120=item C<SAVEPPTR(p)>
1121
1122These macros arrange things to restore the value of pointers C<s> and
1123C<p>. C<s> must be a pointer of a type which survives conversion to
1124C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1125and back.
1126
1127=item C<SAVEFREESV(SV *sv)>
1128
1129The refcount of C<sv> would be decremented at the end of
26d9b02f
JH
1130I<pseudo-block>. This is similar to C<sv_2mortal> in that it is also a
1131mechanism for doing a delayed C<SvREFCNT_dec>. However, while C<sv_2mortal>
1132extends the lifetime of C<sv> until the beginning of the next statement,
1133C<SAVEFREESV> extends it until the end of the enclosing scope. These
1134lifetimes can be wildly different.
1135
1136Also compare C<SAVEMORTALIZESV>.
1137
1138=item C<SAVEMORTALIZESV(SV *sv)>
1139
1140Just like C<SAVEFREESV>, but mortalizes C<sv> at the end of the current
1141scope instead of decrementing its reference count. This usually has the
1142effect of keeping C<sv> alive until the statement that called the currently
1143live scope has finished executing.
d1c897a1
IZ
1144
1145=item C<SAVEFREEOP(OP *op)>
1146
1147The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1148
1149=item C<SAVEFREEPV(p)>
1150
1151The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1152end of I<pseudo-block>.
1153
1154=item C<SAVECLEARSV(SV *sv)>
1155
1156Clears a slot in the current scratchpad which corresponds to C<sv> at
1157the end of I<pseudo-block>.
1158
1159=item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1160
1161The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
1162string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1163short-lived storage, the corresponding string may be reallocated like
1164this:
1165
9cde0e7f 1166 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
d1c897a1 1167
c76ac1ee 1168=item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
d1c897a1
IZ
1169
1170At the end of I<pseudo-block> the function C<f> is called with the
c76ac1ee
GS
1171only argument C<p>.
1172
1173=item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1174
1175At the end of I<pseudo-block> the function C<f> is called with the
1176implicit context argument (if any), and C<p>.
d1c897a1
IZ
1177
1178=item C<SAVESTACK_POS()>
1179
1180The current offset on the Perl internal stack (cf. C<SP>) is restored
1181at the end of I<pseudo-block>.
1182
1183=back
1184
1185The following API list contains functions, thus one needs to
1186provide pointers to the modifiable data explicitly (either C pointers,
1187or Perlish C<GV *>s). Where the above macros take C<int>, a similar
1188function takes C<int *>.
1189
13a2d996 1190=over 4
d1c897a1
IZ
1191
1192=item C<SV* save_scalar(GV *gv)>
1193
1194Equivalent to Perl code C<local $gv>.
1195
1196=item C<AV* save_ary(GV *gv)>
1197
1198=item C<HV* save_hash(GV *gv)>
1199
1200Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1201
1202=item C<void save_item(SV *item)>
1203
1204Duplicates the current value of C<SV>, on the exit from the current
1205C<ENTER>/C<LEAVE> I<pseudo-block> will restore the value of C<SV>
1206using the stored value.
1207
1208=item C<void save_list(SV **sarg, I32 maxsarg)>
1209
1210A variant of C<save_item> which takes multiple arguments via an array
1211C<sarg> of C<SV*> of length C<maxsarg>.
1212
1213=item C<SV* save_svref(SV **sptr)>
1214
1215Similar to C<save_scalar>, but will reinstate a C<SV *>.
1216
1217=item C<void save_aptr(AV **aptr)>
1218
1219=item C<void save_hptr(HV **hptr)>
1220
1221Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1222
1223=back
1224
1225The C<Alias> module implements localization of the basic types within the
1226I<caller's scope>. People who are interested in how to localize things in
1227the containing scope should take a look there too.
1228
0a753a76 1229=head1 Subroutines
a0d0e21e 1230
68dc0745 1231=head2 XSUBs and the Argument Stack
5f05dabc
PP
1232
1233The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1234An XSUB routine will have a stack that contains the arguments from the Perl
1235program, and a way to map from the Perl data structures to a C equivalent.
1236
1237The stack arguments are accessible through the C<ST(n)> macro, which returns
1238the C<n>'th stack argument. Argument 0 is the first argument passed in the
1239Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1240an C<SV*> is used.
1241
1242Most of the time, output from the C routine can be handled through use of
1243the RETVAL and OUTPUT directives. However, there are some cases where the
1244argument stack is not already long enough to handle all the return values.
1245An example is the POSIX tzname() call, which takes no arguments, but returns
1246two, the local time zone's standard and summer time abbreviations.
1247
1248To handle this situation, the PPCODE directive is used and the stack is
1249extended using the macro:
1250
924508f0 1251 EXTEND(SP, num);
5f05dabc 1252
924508f0
GS
1253where C<SP> is the macro that represents the local copy of the stack pointer,
1254and C<num> is the number of elements the stack should be extended by.
5f05dabc
PP
1255
1256Now that there is room on the stack, values can be pushed on it using the
54310121 1257macros to push IVs, doubles, strings, and SV pointers respectively:
5f05dabc
PP
1258
1259 PUSHi(IV)
1260 PUSHn(double)
1261 PUSHp(char*, I32)
1262 PUSHs(SV*)
1263
1264And now the Perl program calling C<tzname>, the two values will be assigned
1265as in:
1266
1267 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1268
1269An alternate (and possibly simpler) method to pushing values on the stack is
1270to use the macros:
1271
1272 XPUSHi(IV)
1273 XPUSHn(double)
1274 XPUSHp(char*, I32)
1275 XPUSHs(SV*)
1276
1277These macros automatically adjust the stack for you, if needed. Thus, you
1278do not need to call C<EXTEND> to extend the stack.
1bd1c0d5 1279However, see L</Putting a C value on Perl stack>
5f05dabc
PP
1280
1281For more information, consult L<perlxs> and L<perlxstut>.
1282
1283=head2 Calling Perl Routines from within C Programs
a0d0e21e
LW
1284
1285There are four routines that can be used to call a Perl subroutine from
1286within a C program. These four are:
1287
954c1994
GS
1288 I32 call_sv(SV*, I32);
1289 I32 call_pv(const char*, I32);
1290 I32 call_method(const char*, I32);
1291 I32 call_argv(const char*, I32, register char**);
a0d0e21e 1292
954c1994 1293The routine most often used is C<call_sv>. The C<SV*> argument
d1b91892
AD
1294contains either the name of the Perl subroutine to be called, or a
1295reference to the subroutine. The second argument consists of flags
1296that control the context in which the subroutine is called, whether
1297or not the subroutine is being passed arguments, how errors should be
1298trapped, and how to treat return values.
a0d0e21e
LW
1299
1300All four routines return the number of arguments that the subroutine returned
1301on the Perl stack.
1302
954c1994
GS
1303These routines used to be called C<perl_call_sv> etc., before Perl v5.6.0,
1304but those names are now deprecated; macros of the same name are provided for
1305compatibility.
1306
1307When using any of these routines (except C<call_argv>), the programmer
d1b91892
AD
1308must manipulate the Perl stack. These include the following macros and
1309functions:
a0d0e21e
LW
1310
1311 dSP
924508f0 1312 SP
a0d0e21e
LW
1313 PUSHMARK()
1314 PUTBACK
1315 SPAGAIN
1316 ENTER
1317 SAVETMPS
1318 FREETMPS
1319 LEAVE
1320 XPUSH*()
cb1a09d0 1321 POP*()
a0d0e21e 1322
5f05dabc
PP
1323For a detailed description of calling conventions from C to Perl,
1324consult L<perlcall>.
a0d0e21e 1325
5f05dabc 1326=head2 Memory Allocation
a0d0e21e 1327
86058a2d
GS
1328All memory meant to be used with the Perl API functions should be manipulated
1329using the macros described in this section. The macros provide the necessary
1330transparency between differences in the actual malloc implementation that is
1331used within perl.
1332
1333It is suggested that you enable the version of malloc that is distributed
5f05dabc 1334with Perl. It keeps pools of various sizes of unallocated memory in
07fa94a1
JO
1335order to satisfy allocation requests more quickly. However, on some
1336platforms, it may cause spurious malloc or free errors.
d1b91892
AD
1337
1338 New(x, pointer, number, type);
1339 Newc(x, pointer, number, type, cast);
1340 Newz(x, pointer, number, type);
1341
07fa94a1 1342These three macros are used to initially allocate memory.
5f05dabc
PP
1343
1344The first argument C<x> was a "magic cookie" that was used to keep track
1345of who called the macro, to help when debugging memory problems. However,
07fa94a1
JO
1346the current code makes no use of this feature (most Perl developers now
1347use run-time memory checkers), so this argument can be any number.
5f05dabc
PP
1348
1349The second argument C<pointer> should be the name of a variable that will
1350point to the newly allocated memory.
d1b91892 1351
d1b91892
AD
1352The third and fourth arguments C<number> and C<type> specify how many of
1353the specified type of data structure should be allocated. The argument
1354C<type> is passed to C<sizeof>. The final argument to C<Newc>, C<cast>,
1355should be used if the C<pointer> argument is different from the C<type>
1356argument.
1357
1358Unlike the C<New> and C<Newc> macros, the C<Newz> macro calls C<memzero>
1359to zero out all the newly allocated memory.
1360
1361 Renew(pointer, number, type);
1362 Renewc(pointer, number, type, cast);
1363 Safefree(pointer)
1364
1365These three macros are used to change a memory buffer size or to free a
1366piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1367match those of C<New> and C<Newc> with the exception of not needing the
1368"magic cookie" argument.
1369
1370 Move(source, dest, number, type);
1371 Copy(source, dest, number, type);
1372 Zero(dest, number, type);
1373
1374These three macros are used to move, copy, or zero out previously allocated
1375memory. The C<source> and C<dest> arguments point to the source and
1376destination starting points. Perl will move, copy, or zero out C<number>
1377instances of the size of the C<type> data structure (using the C<sizeof>
1378function).
a0d0e21e 1379
5f05dabc 1380=head2 PerlIO
ce3d39e2 1381
5f05dabc
PP
1382The most recent development releases of Perl has been experimenting with
1383removing Perl's dependency on the "normal" standard I/O suite and allowing
1384other stdio implementations to be used. This involves creating a new
1385abstraction layer that then calls whichever implementation of stdio Perl
68dc0745 1386was compiled with. All XSUBs should now use the functions in the PerlIO
5f05dabc
PP
1387abstraction layer and not make any assumptions about what kind of stdio
1388is being used.
1389
1390For a complete description of the PerlIO abstraction, consult L<perlapio>.
1391
8ebc5c01 1392=head2 Putting a C value on Perl stack
ce3d39e2
IZ
1393
1394A lot of opcodes (this is an elementary operation in the internal perl
1395stack machine) put an SV* on the stack. However, as an optimization
1396the corresponding SV is (usually) not recreated each time. The opcodes
1397reuse specially assigned SVs (I<target>s) which are (as a corollary)
1398not constantly freed/created.
1399
0a753a76 1400Each of the targets is created only once (but see
ce3d39e2
IZ
1401L<Scratchpads and recursion> below), and when an opcode needs to put
1402an integer, a double, or a string on stack, it just sets the
1403corresponding parts of its I<target> and puts the I<target> on stack.
1404
1405The macro to put this target on stack is C<PUSHTARG>, and it is
1406directly used in some opcodes, as well as indirectly in zillions of
1407others, which use it via C<(X)PUSH[pni]>.
1408
1bd1c0d5
SC
1409Because the target is reused, you must be careful when pushing multiple
1410values on the stack. The following code will not do what you think:
1411
1412 XPUSHi(10);
1413 XPUSHi(20);
1414
1415This translates as "set C<TARG> to 10, push a pointer to C<TARG> onto
1416the stack; set C<TARG> to 20, push a pointer to C<TARG> onto the stack".
1417At the end of the operation, the stack does not contain the values 10
1418and 20, but actually contains two pointers to C<TARG>, which we have set
1419to 20. If you need to push multiple different values, use C<XPUSHs>,
1420which bypasses C<TARG>.
1421
1422On a related note, if you do use C<(X)PUSH[npi]>, then you're going to
1423need a C<dTARG> in your variable declarations so that the C<*PUSH*>
1424macros can make use of the local variable C<TARG>.
1425
8ebc5c01 1426=head2 Scratchpads
ce3d39e2 1427
54310121 1428The question remains on when the SVs which are I<target>s for opcodes
5f05dabc
PP
1429are created. The answer is that they are created when the current unit --
1430a subroutine or a file (for opcodes for statements outside of
1431subroutines) -- is compiled. During this time a special anonymous Perl
ce3d39e2
IZ
1432array is created, which is called a scratchpad for the current
1433unit.
1434
54310121 1435A scratchpad keeps SVs which are lexicals for the current unit and are
ce3d39e2
IZ
1436targets for opcodes. One can deduce that an SV lives on a scratchpad
1437by looking on its flags: lexicals have C<SVs_PADMY> set, and
1438I<target>s have C<SVs_PADTMP> set.
1439
54310121
PP
1440The correspondence between OPs and I<target>s is not 1-to-1. Different
1441OPs in the compile tree of the unit can use the same target, if this
ce3d39e2
IZ
1442would not conflict with the expected life of the temporary.
1443
2ae324a7 1444=head2 Scratchpads and recursion
ce3d39e2
IZ
1445
1446In fact it is not 100% true that a compiled unit contains a pointer to
1447the scratchpad AV. In fact it contains a pointer to an AV of
1448(initially) one element, and this element is the scratchpad AV. Why do
1449we need an extra level of indirection?
1450
1451The answer is B<recursion>, and maybe (sometime soon) B<threads>. Both
1452these can create several execution pointers going into the same
1453subroutine. For the subroutine-child not write over the temporaries
1454for the subroutine-parent (lifespan of which covers the call to the
1455child), the parent and the child should have different
1456scratchpads. (I<And> the lexicals should be separate anyway!)
1457
5f05dabc
PP
1458So each subroutine is born with an array of scratchpads (of length 1).
1459On each entry to the subroutine it is checked that the current
ce3d39e2
IZ
1460depth of the recursion is not more than the length of this array, and
1461if it is, new scratchpad is created and pushed into the array.
1462
1463The I<target>s on this scratchpad are C<undef>s, but they are already
1464marked with correct flags.
1465
0a753a76
PP
1466=head1 Compiled code
1467
1468=head2 Code tree
1469
1470Here we describe the internal form your code is converted to by
1471Perl. Start with a simple example:
1472
1473 $a = $b + $c;
1474
1475This is converted to a tree similar to this one:
1476
1477 assign-to
1478 / \
1479 + $a
1480 / \
1481 $b $c
1482
7b8d334a 1483(but slightly more complicated). This tree reflects the way Perl
0a753a76
PP
1484parsed your code, but has nothing to do with the execution order.
1485There is an additional "thread" going through the nodes of the tree
1486which shows the order of execution of the nodes. In our simplified
1487example above it looks like:
1488
1489 $b ---> $c ---> + ---> $a ---> assign-to
1490
1491But with the actual compile tree for C<$a = $b + $c> it is different:
1492some nodes I<optimized away>. As a corollary, though the actual tree
1493contains more nodes than our simplified example, the execution order
1494is the same as in our example.
1495
1496=head2 Examining the tree
1497
1498If you have your perl compiled for debugging (usually done with C<-D
1499optimize=-g> on C<Configure> command line), you may examine the
1500compiled tree by specifying C<-Dx> on the Perl command line. The
1501output takes several lines per node, and for C<$b+$c> it looks like
1502this:
1503
1504 5 TYPE = add ===> 6
1505 TARG = 1
1506 FLAGS = (SCALAR,KIDS)
1507 {
1508 TYPE = null ===> (4)
1509 (was rv2sv)
1510 FLAGS = (SCALAR,KIDS)
1511 {
1512 3 TYPE = gvsv ===> 4
1513 FLAGS = (SCALAR)
1514 GV = main::b
1515 }
1516 }
1517 {
1518 TYPE = null ===> (5)
1519 (was rv2sv)
1520 FLAGS = (SCALAR,KIDS)
1521 {
1522 4 TYPE = gvsv ===> 5
1523 FLAGS = (SCALAR)
1524 GV = main::c
1525 }
1526 }
1527
1528This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
1529not optimized away (one per number in the left column). The immediate
1530children of the given node correspond to C<{}> pairs on the same level
1531of indentation, thus this listing corresponds to the tree:
1532
1533 add
1534 / \
1535 null null
1536 | |
1537 gvsv gvsv
1538
1539The execution order is indicated by C<===E<gt>> marks, thus it is C<3
15404 5 6> (node C<6> is not included into above listing), i.e.,
1541C<gvsv gvsv add whatever>.
1542
9afa14e3
SC
1543Each of these nodes represents an op, a fundamental operation inside the
1544Perl core. The code which implements each operation can be found in the
1545F<pp*.c> files; the function which implements the op with type C<gvsv>
1546is C<pp_gvsv>, and so on. As the tree above shows, different ops have
1547different numbers of children: C<add> is a binary operator, as one would
1548expect, and so has two children. To accommodate the various different
1549numbers of children, there are various types of op data structure, and
1550they link together in different ways.
1551
1552The simplest type of op structure is C<OP>: this has no children. Unary
1553operators, C<UNOP>s, have one child, and this is pointed to by the
1554C<op_first> field. Binary operators (C<BINOP>s) have not only an
1555C<op_first> field but also an C<op_last> field. The most complex type of
1556op is a C<LISTOP>, which has any number of children. In this case, the
1557first child is pointed to by C<op_first> and the last child by
1558C<op_last>. The children in between can be found by iteratively
1559following the C<op_sibling> pointer from the first child to the last.
1560
1561There are also two other op types: a C<PMOP> holds a regular expression,
1562and has no children, and a C<LOOP> may or may not have children. If the
1563C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
1564complicate matters, if a C<UNOP> is actually a C<null> op after
1565optimization (see L</Compile pass 2: context propagation>) it will still
1566have children in accordance with its former type.
1567
0a753a76
PP
1568=head2 Compile pass 1: check routines
1569
8870b5c7
GS
1570The tree is created by the compiler while I<yacc> code feeds it
1571the constructions it recognizes. Since I<yacc> works bottom-up, so does
0a753a76
PP
1572the first pass of perl compilation.
1573
1574What makes this pass interesting for perl developers is that some
1575optimization may be performed on this pass. This is optimization by
8870b5c7 1576so-called "check routines". The correspondence between node names
0a753a76
PP
1577and corresponding check routines is described in F<opcode.pl> (do not
1578forget to run C<make regen_headers> if you modify this file).
1579
1580A check routine is called when the node is fully constructed except
7b8d334a 1581for the execution-order thread. Since at this time there are no
0a753a76
PP
1582back-links to the currently constructed node, one can do most any
1583operation to the top-level node, including freeing it and/or creating
1584new nodes above/below it.
1585
1586The check routine returns the node which should be inserted into the
1587tree (if the top-level node was not modified, check routine returns
1588its argument).
1589
1590By convention, check routines have names C<ck_*>. They are usually
1591called from C<new*OP> subroutines (or C<convert>) (which in turn are
1592called from F<perly.y>).
1593
1594=head2 Compile pass 1a: constant folding
1595
1596Immediately after the check routine is called the returned node is
1597checked for being compile-time executable. If it is (the value is
1598judged to be constant) it is immediately executed, and a I<constant>
1599node with the "return value" of the corresponding subtree is
1600substituted instead. The subtree is deleted.
1601
1602If constant folding was not performed, the execution-order thread is
1603created.
1604
1605=head2 Compile pass 2: context propagation
1606
1607When a context for a part of compile tree is known, it is propagated
a3cb178b 1608down through the tree. At this time the context can have 5 values
0a753a76
PP
1609(instead of 2 for runtime context): void, boolean, scalar, list, and
1610lvalue. In contrast with the pass 1 this pass is processed from top
1611to bottom: a node's context determines the context for its children.
1612
1613Additional context-dependent optimizations are performed at this time.
1614Since at this moment the compile tree contains back-references (via
1615"thread" pointers), nodes cannot be free()d now. To allow
1616optimized-away nodes at this stage, such nodes are null()ified instead
1617of free()ing (i.e. their type is changed to OP_NULL).
1618
1619=head2 Compile pass 3: peephole optimization
1620
1621After the compile tree for a subroutine (or for an C<eval> or a file)
1622is created, an additional pass over the code is performed. This pass
1623is neither top-down or bottom-up, but in the execution order (with
7b8d334a 1624additional complications for conditionals). These optimizations are
0a753a76
PP
1625done in the subroutine peep(). Optimizations performed at this stage
1626are subject to the same restrictions as in the pass 2.
1627
9afa14e3
SC
1628=head1 Examining internal data structures with the C<dump> functions
1629
1630To aid debugging, the source file F<dump.c> contains a number of
1631functions which produce formatted output of internal data structures.
1632
1633The most commonly used of these functions is C<Perl_sv_dump>; it's used
1634for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
1635C<sv_dump> to produce debugging output from Perl-space, so users of that
1636module should already be familiar with its format.
1637
1638C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
1639derivatives, and produces output similiar to C<perl -Dx>; in fact,
1640C<Perl_dump_eval> will dump the main root of the code being evaluated,
1641exactly like C<-Dx>.
1642
1643Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
1644op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
1645subroutines in a package like so: (Thankfully, these are all xsubs, so
1646there is no op tree)
1647
1648 (gdb) print Perl_dump_packsubs(PL_defstash)
1649
1650 SUB attributes::bootstrap = (xsub 0x811fedc 0)
1651
1652 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
1653
1654 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
1655
1656 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
1657
1658 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
1659
1660and C<Perl_dump_all>, which dumps all the subroutines in the stash and
1661the op tree of the main root.
1662
954c1994 1663=head1 How multiple interpreters and concurrency are supported
ee072b34 1664
ee072b34
GS
1665=head2 Background and PERL_IMPLICIT_CONTEXT
1666
1667The Perl interpreter can be regarded as a closed box: it has an API
1668for feeding it code or otherwise making it do things, but it also has
1669functions for its own use. This smells a lot like an object, and
1670there are ways for you to build Perl so that you can have multiple
1671interpreters, with one interpreter represented either as a C++ object,
1672a C structure, or inside a thread. The thread, the C structure, or
1673the C++ object will contain all the context, the state of that
1674interpreter.
1675
54aff467
GS
1676Three macros control the major Perl build flavors: MULTIPLICITY,
1677USE_THREADS and PERL_OBJECT. The MULTIPLICITY build has a C structure
1678that packages all the interpreter state, there is a similar thread-specific
a7486cbb
JH
1679data structure under USE_THREADS, and the (now deprecated) PERL_OBJECT
1680build has a C++ class to maintain interpreter state. In all three cases,
54aff467
GS
1681PERL_IMPLICIT_CONTEXT is also normally defined, and enables the
1682support for passing in a "hidden" first argument that represents all three
651a3225 1683data structures.
54aff467
GS
1684
1685All this obviously requires a way for the Perl internal functions to be
ee072b34
GS
1686C++ methods, subroutines taking some kind of structure as the first
1687argument, or subroutines taking nothing as the first argument. To
1688enable these three very different ways of building the interpreter,
1689the Perl source (as it does in so many other situations) makes heavy
1690use of macros and subroutine naming conventions.
1691
54aff467 1692First problem: deciding which functions will be public API functions and
954c1994
GS
1693which will be private. All functions whose names begin C<S_> are private
1694(think "S" for "secret" or "static"). All other functions begin with
1695"Perl_", but just because a function begins with "Perl_" does not mean it is
a422fd2d
SC
1696part of the API. (See L</Internal Functions>.) The easiest way to be B<sure> a
1697function is part of the API is to find its entry in L<perlapi>.
1698If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
1699think it should be (i.e., you need it for your extension), send mail via
1700L<perlbug> explaining why you think it should be.
ee072b34
GS
1701
1702Second problem: there must be a syntax so that the same subroutine
1703declarations and calls can pass a structure as their first argument,
1704or pass nothing. To solve this, the subroutines are named and
1705declared in a particular way. Here's a typical start of a static
1706function used within the Perl guts:
1707
1708 STATIC void
1709 S_incline(pTHX_ char *s)
1710
1711STATIC becomes "static" in C, and is #define'd to nothing in C++.
1712
651a3225
GS
1713A public function (i.e. part of the internal API, but not necessarily
1714sanctioned for use in extensions) begins like this:
ee072b34
GS
1715
1716 void
1717 Perl_sv_setsv(pTHX_ SV* dsv, SV* ssv)
1718
1719C<pTHX_> is one of a number of macros (in perl.h) that hide the
1720details of the interpreter's context. THX stands for "thread", "this",
1721or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
1722The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
a7486cbb
JH
1723or 'd' for B<d>eclaration, so we have C<pTHX>, C<aTHX> and C<dTHX>, and
1724their variants.
ee072b34 1725
a7486cbb
JH
1726When Perl is built without options that set PERL_IMPLICIT_CONTEXT, there is no
1727first argument containing the interpreter's context. The trailing underscore
ee072b34
GS
1728in the pTHX_ macro indicates that the macro expansion needs a comma
1729after the context argument because other arguments follow it. If
1730PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
54aff467
GS
1731subroutine is not prototyped to take the extra argument. The form of the
1732macro without the trailing underscore is used when there are no additional
ee072b34
GS
1733explicit arguments.
1734
54aff467 1735When a core function calls another, it must pass the context. This
a7486cbb 1736is normally hidden via macros. Consider C<sv_setsv>. It expands into
ee072b34
GS
1737something like this:
1738
1739 ifdef PERL_IMPLICIT_CONTEXT
c52f9dcd 1740 define sv_setsv(a,b) Perl_sv_setsv(aTHX_ a, b)
ee072b34
GS
1741 /* can't do this for vararg functions, see below */
1742 else
c52f9dcd 1743 define sv_setsv Perl_sv_setsv
ee072b34
GS
1744 endif
1745
1746This works well, and means that XS authors can gleefully write:
1747
1748 sv_setsv(foo, bar);
1749
1750and still have it work under all the modes Perl could have been
1751compiled with.
1752
1753Under PERL_OBJECT in the core, that will translate to either:
1754
1755 CPerlObj::Perl_sv_setsv(foo,bar); # in CPerlObj functions,
1756 # C++ takes care of 'this'
1757 or
1758
1759 pPerl->Perl_sv_setsv(foo,bar); # in truly static functions,
1760 # see objXSUB.h
1761
1762Under PERL_OBJECT in extensions (aka PERL_CAPI), or under
a7486cbb
JH
1763MULTIPLICITY/USE_THREADS with PERL_IMPLICIT_CONTEXT in both core
1764and extensions, it will become:
ee072b34
GS
1765
1766 Perl_sv_setsv(aTHX_ foo, bar); # the canonical Perl "API"
1767 # for all build flavors
1768
1769This doesn't work so cleanly for varargs functions, though, as macros
1770imply that the number of arguments is known in advance. Instead we
1771either need to spell them out fully, passing C<aTHX_> as the first
1772argument (the Perl core tends to do this with functions like
1773Perl_warner), or use a context-free version.
1774
1775The context-free version of Perl_warner is called
1776Perl_warner_nocontext, and does not take the extra argument. Instead
1777it does dTHX; to get the context from thread-local storage. We
1778C<#define warner Perl_warner_nocontext> so that extensions get source
1779compatibility at the expense of performance. (Passing an arg is
1780cheaper than grabbing it from thread-local storage.)
1781
1782You can ignore [pad]THX[xo] when browsing the Perl headers/sources.
1783Those are strictly for use within the core. Extensions and embedders
1784need only be aware of [pad]THX.
1785
a7486cbb
JH
1786=head2 So what happened to dTHR?
1787
1788C<dTHR> was introduced in perl 5.005 to support the older thread model.
1789The older thread model now uses the C<THX> mechanism to pass context
1790pointers around, so C<dTHR> is not useful any more. Perl 5.6.0 and
1791later still have it for backward source compatibility, but it is defined
1792to be a no-op.
1793
ee072b34
GS
1794=head2 How do I use all this in extensions?
1795
1796When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
1797any functions in the Perl API will need to pass the initial context
1798argument somehow. The kicker is that you will need to write it in
1799such a way that the extension still compiles when Perl hasn't been
1800built with PERL_IMPLICIT_CONTEXT enabled.
1801
1802There are three ways to do this. First, the easy but inefficient way,
1803which is also the default, in order to maintain source compatibility
1804with extensions: whenever XSUB.h is #included, it redefines the aTHX
1805and aTHX_ macros to call a function that will return the context.
1806Thus, something like:
1807
1808 sv_setsv(asv, bsv);
1809
4375e838 1810in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
54aff467 1811in effect:
ee072b34 1812
2fa86c13 1813 Perl_sv_setsv(Perl_get_context(), asv, bsv);
ee072b34 1814
54aff467 1815or to this otherwise:
ee072b34
GS
1816
1817 Perl_sv_setsv(asv, bsv);
1818
1819You have to do nothing new in your extension to get this; since
2fa86c13 1820the Perl library provides Perl_get_context(), it will all just
ee072b34
GS
1821work.
1822
1823The second, more efficient way is to use the following template for
1824your Foo.xs:
1825
c52f9dcd
JH
1826 #define PERL_NO_GET_CONTEXT /* we want efficiency */
1827 #include "EXTERN.h"
1828 #include "perl.h"
1829 #include "XSUB.h"
ee072b34
GS
1830
1831 static my_private_function(int arg1, int arg2);
1832
c52f9dcd
JH
1833 static SV *
1834 my_private_function(int arg1, int arg2)
1835 {
1836 dTHX; /* fetch context */
1837 ... call many Perl API functions ...
1838 }
ee072b34
GS
1839
1840 [... etc ...]
1841
c52f9dcd 1842 MODULE = Foo PACKAGE = Foo
ee072b34 1843
c52f9dcd 1844 /* typical XSUB */
ee072b34 1845
c52f9dcd
JH
1846 void
1847 my_xsub(arg)
1848 int arg
1849 CODE:
1850 my_private_function(arg, 10);
ee072b34
GS
1851
1852Note that the only two changes from the normal way of writing an
1853extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
1854including the Perl headers, followed by a C<dTHX;> declaration at
1855the start of every function that will call the Perl API. (You'll
1856know which functions need this, because the C compiler will complain
1857that there's an undeclared identifier in those functions.) No changes
1858are needed for the XSUBs themselves, because the XS() macro is
1859correctly defined to pass in the implicit context if needed.
1860
1861The third, even more efficient way is to ape how it is done within
1862the Perl guts:
1863
1864
c52f9dcd
JH
1865 #define PERL_NO_GET_CONTEXT /* we want efficiency */
1866 #include "EXTERN.h"
1867 #include "perl.h"
1868 #include "XSUB.h"
ee072b34
GS
1869
1870 /* pTHX_ only needed for functions that call Perl API */
1871 static my_private_function(pTHX_ int arg1, int arg2);
1872
c52f9dcd
JH
1873 static SV *
1874 my_private_function(pTHX_ int arg1, int arg2)
1875 {
1876 /* dTHX; not needed here, because THX is an argument */
1877 ... call Perl API functions ...
1878 }
ee072b34
GS
1879
1880 [... etc ...]
1881
c52f9dcd 1882 MODULE = Foo PACKAGE = Foo
ee072b34 1883
c52f9dcd 1884 /* typical XSUB */
ee072b34 1885
c52f9dcd
JH
1886 void
1887 my_xsub(arg)
1888 int arg
1889 CODE:
1890 my_private_function(aTHX_ arg, 10);
ee072b34
GS
1891
1892This implementation never has to fetch the context using a function
1893call, since it is always passed as an extra argument. Depending on
1894your needs for simplicity or efficiency, you may mix the previous
1895two approaches freely.
1896
651a3225
GS
1897Never add a comma after C<pTHX> yourself--always use the form of the
1898macro with the underscore for functions that take explicit arguments,
1899or the form without the argument for functions with no explicit arguments.
ee072b34 1900
a7486cbb
JH
1901=head2 Should I do anything special if I call perl from multiple threads?
1902
1903If you create interpreters in one thread and then proceed to call them in
1904another, you need to make sure perl's own Thread Local Storage (TLS) slot is
1905initialized correctly in each of those threads.
1906
1907The C<perl_alloc> and C<perl_clone> API functions will automatically set
1908the TLS slot to the interpreter they created, so that there is no need to do
1909anything special if the interpreter is always accessed in the same thread that
1910created it, and that thread did not create or call any other interpreters
1911afterwards. If that is not the case, you have to set the TLS slot of the
1912thread before calling any functions in the Perl API on that particular
1913interpreter. This is done by calling the C<PERL_SET_CONTEXT> macro in that
1914thread as the first thing you do:
1915
1916 /* do this before doing anything else with some_perl */
1917 PERL_SET_CONTEXT(some_perl);
1918
1919 ... other Perl API calls on some_perl go here ...
1920
ee072b34
GS
1921=head2 Future Plans and PERL_IMPLICIT_SYS
1922
1923Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
1924that the interpreter knows about itself and pass it around, so too are
1925there plans to allow the interpreter to bundle up everything it knows
1926about the environment it's running on. This is enabled with the
a7486cbb
JH
1927PERL_IMPLICIT_SYS macro. Currently it only works with PERL_OBJECT
1928and USE_THREADS on Windows (see inside iperlsys.h).
ee072b34
GS
1929
1930This allows the ability to provide an extra pointer (called the "host"
1931environment) for all the system calls. This makes it possible for
1932all the system stuff to maintain their own state, broken down into
1933seven C structures. These are thin wrappers around the usual system
1934calls (see win32/perllib.c) for the default perl executable, but for a
1935more ambitious host (like the one that would do fork() emulation) all
1936the extra work needed to pretend that different interpreters are
1937actually different "processes", would be done here.
1938
1939The Perl engine/interpreter and the host are orthogonal entities.
1940There could be one or more interpreters in a process, and one or
1941more "hosts", with free association between them.
1942
a422fd2d
SC
1943=head1 Internal Functions
1944
1945All of Perl's internal functions which will be exposed to the outside
1946world are be prefixed by C<Perl_> so that they will not conflict with XS
1947functions or functions used in a program in which Perl is embedded.
1948Similarly, all global variables begin with C<PL_>. (By convention,
1949static functions start with C<S_>)
1950
1951Inside the Perl core, you can get at the functions either with or
1952without the C<Perl_> prefix, thanks to a bunch of defines that live in
1953F<embed.h>. This header file is generated automatically from
1954F<embed.pl>. F<embed.pl> also creates the prototyping header files for
1955the internal functions, generates the documentation and a lot of other
1956bits and pieces. It's important that when you add a new function to the
1957core or change an existing one, you change the data in the table at the
1958end of F<embed.pl> as well. Here's a sample entry from that table:
1959
1960 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
1961
1962The second column is the return type, the third column the name. Columns
1963after that are the arguments. The first column is a set of flags:
1964
1965=over 3
1966
1967=item A
1968
1969This function is a part of the public API.
1970
1971=item p
1972
1973This function has a C<Perl_> prefix; ie, it is defined as C<Perl_av_fetch>
1974
1975=item d
1976
1977This function has documentation using the C<apidoc> feature which we'll
1978look at in a second.
1979
1980=back
1981
1982Other available flags are:
1983
1984=over 3
1985
1986=item s
1987
a7486cbb
JH
1988This is a static function and is defined as C<S_whatever>, and usually
1989called within the sources as C<whatever(...)>.
a422fd2d
SC
1990
1991=item n
1992
1993This does not use C<aTHX_> and C<pTHX> to pass interpreter context. (See
1994L<perlguts/Background and PERL_IMPLICIT_CONTEXT>.)
1995
1996=item r
1997
1998This function never returns; C<croak>, C<exit> and friends.
1999
2000=item f
2001
2002This function takes a variable number of arguments, C<printf> style.
2003The argument list should end with C<...>, like this:
2004
2005 Afprd |void |croak |const char* pat|...
2006
a7486cbb 2007=item M
a422fd2d
SC
2008
2009This function is part of the experimental development API, and may change
2010or disappear without notice.
2011
2012=item o
2013
2014This function should not have a compatibility macro to define, say,
2015C<Perl_parse> to C<parse>. It must be called as C<Perl_parse>.
2016
2017=item j
2018
2019This function is not a member of C<CPerlObj>. If you don't know
2020what this means, don't use it.
2021
2022=item x
2023
2024This function isn't exported out of the Perl core.
2025
2026=back
2027
2028If you edit F<embed.pl>, you will need to run C<make regen_headers> to
2029force a rebuild of F<embed.h> and other auto-generated files.
2030
6b4667fc 2031=head2 Formatted Printing of IVs, UVs, and NVs
9dd9db0b 2032
6b4667fc
A
2033If you are printing IVs, UVs, or NVS instead of the stdio(3) style
2034formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
2035following macros for portability
9dd9db0b 2036
c52f9dcd
JH
2037 IVdf IV in decimal
2038 UVuf UV in decimal
2039 UVof UV in octal
2040 UVxf UV in hexadecimal
2041 NVef NV %e-like
2042 NVff NV %f-like
2043 NVgf NV %g-like
9dd9db0b 2044
6b4667fc
A
2045These will take care of 64-bit integers and long doubles.
2046For example:
2047
c52f9dcd 2048 printf("IV is %"IVdf"\n", iv);
6b4667fc
A
2049
2050The IVdf will expand to whatever is the correct format for the IVs.
9dd9db0b 2051
8908e76d
JH
2052If you are printing addresses of pointers, use UVxf combined
2053with PTR2UV(), do not use %lx or %p.
2054
2055=head2 Pointer-To-Integer and Integer-To-Pointer
2056
2057Because pointer size does not necessarily equal integer size,
2058use the follow macros to do it right.
2059
c52f9dcd
JH
2060 PTR2UV(pointer)
2061 PTR2IV(pointer)
2062 PTR2NV(pointer)
2063 INT2PTR(pointertotype, integer)
8908e76d
JH
2064
2065For example:
2066
c52f9dcd
JH
2067 IV iv = ...;
2068 SV *sv = INT2PTR(SV*, iv);
8908e76d
JH
2069
2070and
2071
c52f9dcd
JH
2072 AV *av = ...;
2073 UV uv = PTR2UV(av);
8908e76d 2074
a422fd2d
SC
2075=head2 Source Documentation
2076
2077There's an effort going on to document the internal functions and
2078automatically produce reference manuals from them - L<perlapi> is one
2079such manual which details all the functions which are available to XS
2080writers. L<perlintern> is the autogenerated manual for the functions
2081which are not part of the API and are supposedly for internal use only.
2082
2083Source documentation is created by putting POD comments into the C
2084source, like this:
2085
2086 /*
2087 =for apidoc sv_setiv
2088
2089 Copies an integer into the given SV. Does not handle 'set' magic. See
2090 C<sv_setiv_mg>.
2091
2092 =cut
2093 */
2094
2095Please try and supply some documentation if you add functions to the
2096Perl core.
2097
2098=head1 Unicode Support
2099
2100Perl 5.6.0 introduced Unicode support. It's important for porters and XS
2101writers to understand this support and make sure that the code they
2102write does not corrupt Unicode data.
2103
2104=head2 What B<is> Unicode, anyway?
2105
2106In the olden, less enlightened times, we all used to use ASCII. Most of
2107us did, anyway. The big problem with ASCII is that it's American. Well,
2108no, that's not actually the problem; the problem is that it's not
2109particularly useful for people who don't use the Roman alphabet. What
2110used to happen was that particular languages would stick their own
2111alphabet in the upper range of the sequence, between 128 and 255. Of
2112course, we then ended up with plenty of variants that weren't quite
2113ASCII, and the whole point of it being a standard was lost.
2114
2115Worse still, if you've got a language like Chinese or
2116Japanese that has hundreds or thousands of characters, then you really
2117can't fit them into a mere 256, so they had to forget about ASCII
2118altogether, and build their own systems using pairs of numbers to refer
2119to one character.
2120
2121To fix this, some people formed Unicode, Inc. and
2122produced a new character set containing all the characters you can
2123possibly think of and more. There are several ways of representing these
2124characters, and the one Perl uses is called UTF8. UTF8 uses
2125a variable number of bytes to represent a character, instead of just
b3b6085d 2126one. You can learn more about Unicode at http://www.unicode.org/
a422fd2d
SC
2127
2128=head2 How can I recognise a UTF8 string?
2129
2130You can't. This is because UTF8 data is stored in bytes just like
2131non-UTF8 data. The Unicode character 200, (C<0xC8> for you hex types)
2132capital E with a grave accent, is represented by the two bytes
2133C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
2134has that byte sequence as well. So you can't tell just by looking - this
2135is what makes Unicode input an interesting problem.
2136
2137The API function C<is_utf8_string> can help; it'll tell you if a string
2138contains only valid UTF8 characters. However, it can't do the work for
2139you. On a character-by-character basis, C<is_utf8_char> will tell you
2140whether the current character in a string is valid UTF8.
2141
2142=head2 How does UTF8 represent Unicode characters?
2143
2144As mentioned above, UTF8 uses a variable number of bytes to store a
2145character. Characters with values 1...128 are stored in one byte, just
2146like good ol' ASCII. Character 129 is stored as C<v194.129>; this
a31a806a 2147continues up to character 191, which is C<v194.191>. Now we've run out of
a422fd2d
SC
2148bits (191 is binary C<10111111>) so we move on; 192 is C<v195.128>. And
2149so it goes on, moving to three bytes at character 2048.
2150
2151Assuming you know you're dealing with a UTF8 string, you can find out
2152how long the first character in it is with the C<UTF8SKIP> macro:
2153
2154 char *utf = "\305\233\340\240\201";
2155 I32 len;
2156
2157 len = UTF8SKIP(utf); /* len is 2 here */
2158 utf += len;
2159 len = UTF8SKIP(utf); /* len is 3 here */
2160
2161Another way to skip over characters in a UTF8 string is to use
2162C<utf8_hop>, which takes a string and a number of characters to skip
2163over. You're on your own about bounds checking, though, so don't use it
2164lightly.
2165
2166All bytes in a multi-byte UTF8 character will have the high bit set, so
2167you can test if you need to do something special with this character
2168like this:
2169
2170 UV uv;
2171
2172 if (utf & 0x80)
2173 /* Must treat this as UTF8 */
2174 uv = utf8_to_uv(utf);
2175 else
2176 /* OK to treat this character as a byte */
2177 uv = *utf;
2178
2179You can also see in that example that we use C<utf8_to_uv> to get the
2180value of the character; the inverse function C<uv_to_utf8> is available
2181for putting a UV into UTF8:
2182
2183 if (uv > 0x80)
2184 /* Must treat this as UTF8 */
2185 utf8 = uv_to_utf8(utf8, uv);
2186 else
2187 /* OK to treat this character as a byte */
2188 *utf8++ = uv;
2189
2190You B<must> convert characters to UVs using the above functions if
2191you're ever in a situation where you have to match UTF8 and non-UTF8
2192characters. You may not skip over UTF8 characters in this case. If you
2193do this, you'll lose the ability to match hi-bit non-UTF8 characters;
2194for instance, if your UTF8 string contains C<v196.172>, and you skip
2195that character, you can never match a C<chr(200)> in a non-UTF8 string.
2196So don't do that!
2197
2198=head2 How does Perl store UTF8 strings?
2199
2200Currently, Perl deals with Unicode strings and non-Unicode strings
2201slightly differently. If a string has been identified as being UTF-8
2202encoded, Perl will set a flag in the SV, C<SVf_UTF8>. You can check and
2203manipulate this flag with the following macros:
2204
2205 SvUTF8(sv)
2206 SvUTF8_on(sv)
2207 SvUTF8_off(sv)
2208
2209This flag has an important effect on Perl's treatment of the string: if
2210Unicode data is not properly distinguished, regular expressions,
2211C<length>, C<substr> and other string handling operations will have
2212undesirable results.
2213
2214The problem comes when you have, for instance, a string that isn't
2215flagged is UTF8, and contains a byte sequence that could be UTF8 -
2216especially when combining non-UTF8 and UTF8 strings.
2217
2218Never forget that the C<SVf_UTF8> flag is separate to the PV value; you
2219need be sure you don't accidentally knock it off while you're
2220manipulating SVs. More specifically, you cannot expect to do this:
2221
2222 SV *sv;
2223 SV *nsv;
2224 STRLEN len;
2225 char *p;
2226
2227 p = SvPV(sv, len);
2228 frobnicate(p);
2229 nsv = newSVpvn(p, len);
2230
2231The C<char*> string does not tell you the whole story, and you can't
2232copy or reconstruct an SV just by copying the string value. Check if the
2233old SV has the UTF8 flag set, and act accordingly:
2234
2235 p = SvPV(sv, len);
2236 frobnicate(p);
2237 nsv = newSVpvn(p, len);
2238 if (SvUTF8(sv))
2239 SvUTF8_on(nsv);
2240
2241In fact, your C<frobnicate> function should be made aware of whether or
2242not it's dealing with UTF8 data, so that it can handle the string
2243appropriately.
2244
2245=head2 How do I convert a string to UTF8?
2246
2247If you're mixing UTF8 and non-UTF8 strings, you might find it necessary
2248to upgrade one of the strings to UTF8. If you've got an SV, the easiest
2249way to do this is:
2250
2251 sv_utf8_upgrade(sv);
2252
2253However, you must not do this, for example:
2254
2255 if (!SvUTF8(left))
2256 sv_utf8_upgrade(left);
2257
2258If you do this in a binary operator, you will actually change one of the
b1866b2d 2259strings that came into the operator, and, while it shouldn't be noticeable
a422fd2d
SC
2260by the end user, it can cause problems.
2261
2262Instead, C<bytes_to_utf8> will give you a UTF8-encoded B<copy> of its
2263string argument. This is useful for having the data available for
b1866b2d 2264comparisons and so on, without harming the original SV. There's also
a422fd2d
SC
2265C<utf8_to_bytes> to go the other way, but naturally, this will fail if
2266the string contains any characters above 255 that can't be represented
2267in a single byte.
2268
2269=head2 Is there anything else I need to know?
2270
2271Not really. Just remember these things:
2272
2273=over 3
2274
2275=item *
2276
2277There's no way to tell if a string is UTF8 or not. You can tell if an SV
2278is UTF8 by looking at is C<SvUTF8> flag. Don't forget to set the flag if
2279something should be UTF8. Treat the flag as part of the PV, even though
2280it's not - if you pass on the PV to somewhere, pass on the flag too.
2281
2282=item *
2283
2284If a string is UTF8, B<always> use C<utf8_to_uv> to get at the value,
2285unless C<!(*s & 0x80)> in which case you can use C<*s>.
2286
2287=item *
2288
2289When writing to a UTF8 string, B<always> use C<uv_to_utf8>, unless
2290C<uv < 0x80> in which case you can use C<*s = uv>.
2291
2292=item *
2293
2294Mixing UTF8 and non-UTF8 strings is tricky. Use C<bytes_to_utf8> to get
2295a new string which is UTF8 encoded. There are tricks you can use to
2296delay deciding whether you need to use a UTF8 string until you get to a
2297high character - C<HALF_UPGRADE> is one of those.
2298
2299=back
2300
954c1994 2301=head1 AUTHORS
e89caa19 2302
954c1994
GS
2303Until May 1997, this document was maintained by Jeff Okamoto
2304<okamoto@corp.hp.com>. It is now maintained as part of Perl itself
2305by the Perl 5 Porters <perl5-porters@perl.org>.
cb1a09d0 2306
954c1994
GS
2307With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
2308Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
2309Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
2310Stephen McCamant, and Gurusamy Sarathy.
cb1a09d0 2311
954c1994 2312API Listing originally by Dean Roehrich <roehrich@cray.com>.
cb1a09d0 2313
954c1994
GS
2314Modifications to autogenerate the API listing (L<perlapi>) by Benjamin
2315Stuhl.
cb1a09d0 2316
954c1994 2317=head1 SEE ALSO
cb1a09d0 2318
954c1994 2319perlapi(1), perlintern(1), perlxs(1), perlembed(1)