3 perlguts - Introduction to the Perl API
7 This document attempts to describe how to use the Perl API, as well as
8 to provide some info on the basic workings of the Perl core. It is far
9 from complete and probably contains many errors. Please refer any
10 questions or comments to the author below.
16 Perl has three typedefs that handle Perl's three main data types:
22 Each typedef has specific routines that manipulate the various data types.
24 =head2 What is an "IV"?
26 Perl uses a special typedef IV which is a simple signed integer type that is
27 guaranteed to be large enough to hold a pointer (as well as an integer).
28 Additionally, there is the UV, which is simply an unsigned IV.
30 Perl also uses two special typedefs, I32 and I16, which will always be at
31 least 32-bits and 16-bits long, respectively. (Again, there are U32 and U16,
32 as well.) They will usually be exactly 32 and 16 bits long, but on Crays
33 they will both be 64 bits.
35 =head2 Working with SVs
37 An SV can be created and loaded with one command. There are five types of
38 values that can be loaded: an integer value (IV), an unsigned integer
39 value (UV), a double (NV), a string (PV), and another scalar (SV).
40 ("PV" stands for "Pointer Value". You might think that it is misnamed
41 because it is described as pointing only to strings. However, it is
42 possible to have it point to other things For example, it could point
43 to an array of UVs. But,
44 using it for non-strings requires care, as the underlying assumption of
45 much of the internals is that PVs are just for strings. Often, for
46 example, a trailing C<NUL> is tacked on automatically. The non-string use
47 is documented only in this paragraph.)
49 The seven routines are:
54 SV* newSVpv(const char*, STRLEN);
55 SV* newSVpvn(const char*, STRLEN);
56 SV* newSVpvf(const char*, ...);
59 C<STRLEN> is an integer type (Size_t, usually defined as size_t in
60 F<config.h>) guaranteed to be large enough to represent the size of
61 any string that perl can handle.
63 In the unlikely case of a SV requiring more complex initialization, you
64 can create an empty SV with newSV(len). If C<len> is 0 an empty SV of
65 type NULL is returned, else an SV of type PV is returned with len + 1 (for
66 the C<NUL>) bytes of storage allocated, accessible via SvPVX. In both cases
67 the SV has the undef value.
69 SV *sv = newSV(0); /* no storage allocated */
70 SV *sv = newSV(10); /* 10 (+1) bytes of uninitialised storage
73 To change the value of an I<already-existing> SV, there are eight routines:
75 void sv_setiv(SV*, IV);
76 void sv_setuv(SV*, UV);
77 void sv_setnv(SV*, double);
78 void sv_setpv(SV*, const char*);
79 void sv_setpvn(SV*, const char*, STRLEN)
80 void sv_setpvf(SV*, const char*, ...);
81 void sv_vsetpvfn(SV*, const char*, STRLEN, va_list *,
83 void sv_setsv(SV*, SV*);
85 Notice that you can choose to specify the length of the string to be
86 assigned by using C<sv_setpvn>, C<newSVpvn>, or C<newSVpv>, or you may
87 allow Perl to calculate the length by using C<sv_setpv> or by specifying
88 0 as the second argument to C<newSVpv>. Be warned, though, that Perl will
89 determine the string's length by using C<strlen>, which depends on the
90 string terminating with a C<NUL> character, and not otherwise containing
93 The arguments of C<sv_setpvf> are processed like C<sprintf>, and the
94 formatted output becomes the value.
96 C<sv_vsetpvfn> is an analogue of C<vsprintf>, but it allows you to specify
97 either a pointer to a variable argument list or the address and length of
98 an array of SVs. The last argument points to a boolean; on return, if that
99 boolean is true, then locale-specific information has been used to format
100 the string, and the string's contents are therefore untrustworthy (see
101 L<perlsec>). This pointer may be NULL if that information is not
102 important. Note that this function requires you to specify the length of
105 The C<sv_set*()> functions are not generic enough to operate on values
106 that have "magic". See L<Magic Virtual Tables> later in this document.
108 All SVs that contain strings should be terminated with a C<NUL> character.
109 If it is not C<NUL>-terminated there is a risk of
110 core dumps and corruptions from code which passes the string to C
111 functions or system calls which expect a C<NUL>-terminated string.
112 Perl's own functions typically add a trailing C<NUL> for this reason.
113 Nevertheless, you should be very careful when you pass a string stored
114 in an SV to a C function or system call.
116 To access the actual value that an SV points to, you can use the macros:
121 SvPV(SV*, STRLEN len)
124 which will automatically coerce the actual scalar type into an IV, UV, double,
127 In the C<SvPV> macro, the length of the string returned is placed into the
128 variable C<len> (this is a macro, so you do I<not> use C<&len>). If you do
129 not care what the length of the data is, use the C<SvPV_nolen> macro.
130 Historically the C<SvPV> macro with the global variable C<PL_na> has been
131 used in this case. But that can be quite inefficient because C<PL_na> must
132 be accessed in thread-local storage in threaded Perl. In any case, remember
133 that Perl allows arbitrary strings of data that may both contain NULs and
134 might not be terminated by a C<NUL>.
136 Also remember that C doesn't allow you to safely say C<foo(SvPV(s, len),
137 len);>. It might work with your
138 compiler, but it won't work for everyone.
139 Break this sort of statement up into separate assignments:
147 If you want to know if the scalar value is TRUE, you can use:
151 Although Perl will automatically grow strings for you, if you need to force
152 Perl to allocate more memory for your SV, you can use the macro
154 SvGROW(SV*, STRLEN newlen)
156 which will determine if more memory needs to be allocated. If so, it will
157 call the function C<sv_grow>. Note that C<SvGROW> can only increase, not
158 decrease, the allocated memory of an SV and that it does not automatically
159 add space for the trailing C<NUL> byte (perl's own string functions typically do
160 C<SvGROW(sv, len + 1)>).
162 If you want to write to an existing SV's buffer and set its value to a
163 string, use SvPV_force() or one of its variants to force the SV to be
164 a PV. This will remove any of various types of non-stringness from
165 the SV while preserving the content of the SV in the PV. This can be
166 used, for example, to append data from an API function to a buffer
167 without extra copying:
169 (void)SvPVbyte_force(sv, len);
170 s = SvGROW(sv, len + needlen + 1);
171 /* something that modifies up to needlen bytes at s+len, but
172 modifies newlen bytes
173 eg. newlen = read(fd, s + len, needlen);
174 ignoring errors for these examples
176 s[len + newlen] = '\0';
177 SvCUR_set(sv, len + newlen);
181 If you already have the data in memory or if you want to keep your
182 code simple, you can use one of the sv_cat*() variants, such as
183 sv_catpvn(). If you want to insert anywhere in the string you can use
184 sv_insert() or sv_insert_flags().
186 If you don't need the existing content of the SV, you can avoid some
189 sv_setpvn(sv, "", 0);
190 s = SvGROW(sv, needlen + 1);
191 /* something that modifies up to needlen bytes at s, but modifies
193 eg. newlen = read(fd, s. needlen);
196 SvCUR_set(sv, newlen);
197 SvPOK_only(sv); /* also clears SVf_UTF8 */
200 Again, if you already have the data in memory or want to avoid the
201 complexity of the above, you can use sv_setpvn().
203 If you have a buffer allocated with Newx() and want to set that as the
204 SV's value, you can use sv_usepvn_flags(). That has some requirements
205 if you want to avoid perl re-allocating the buffer to fit the trailing
208 Newx(buf, somesize+1, char);
209 /* ... fill in buf ... */
210 buf[somesize] = '\0';
211 sv_usepvn_flags(sv, buf, somesize, SV_SMAGIC | SV_HAS_TRAILING_NUL);
212 /* buf now belongs to perl, don't release it */
214 If you have an SV and want to know what kind of data Perl thinks is stored
215 in it, you can use the following macros to check the type of SV you have.
221 You can get and set the current length of the string stored in an SV with
222 the following macros:
225 SvCUR_set(SV*, I32 val)
227 You can also get a pointer to the end of the string stored in the SV
232 But note that these last three macros are valid only if C<SvPOK()> is true.
234 If you want to append something to the end of string stored in an C<SV*>,
235 you can use the following functions:
237 void sv_catpv(SV*, const char*);
238 void sv_catpvn(SV*, const char*, STRLEN);
239 void sv_catpvf(SV*, const char*, ...);
240 void sv_vcatpvfn(SV*, const char*, STRLEN, va_list *, SV **,
242 void sv_catsv(SV*, SV*);
244 The first function calculates the length of the string to be appended by
245 using C<strlen>. In the second, you specify the length of the string
246 yourself. The third function processes its arguments like C<sprintf> and
247 appends the formatted output. The fourth function works like C<vsprintf>.
248 You can specify the address and length of an array of SVs instead of the
249 va_list argument. The fifth function
250 extends the string stored in the first
251 SV with the string stored in the second SV. It also forces the second SV
252 to be interpreted as a string.
254 The C<sv_cat*()> functions are not generic enough to operate on values that
255 have "magic". See L<Magic Virtual Tables> later in this document.
257 If you know the name of a scalar variable, you can get a pointer to its SV
258 by using the following:
260 SV* get_sv("package::varname", 0);
262 This returns NULL if the variable does not exist.
264 If you want to know if this variable (or any other SV) is actually C<defined>,
269 The scalar C<undef> value is stored in an SV instance called C<PL_sv_undef>.
271 Its address can be used whenever an C<SV*> is needed. Make sure that
272 you don't try to compare a random sv with C<&PL_sv_undef>. For example
273 when interfacing Perl code, it'll work correctly for:
277 But won't work when called as:
282 So to repeat always use SvOK() to check whether an sv is defined.
284 Also you have to be careful when using C<&PL_sv_undef> as a value in
285 AVs or HVs (see L<AVs, HVs and undefined values>).
287 There are also the two values C<PL_sv_yes> and C<PL_sv_no>, which contain
288 boolean TRUE and FALSE values, respectively. Like C<PL_sv_undef>, their
289 addresses can be used whenever an C<SV*> is needed.
291 Do not be fooled into thinking that C<(SV *) 0> is the same as C<&PL_sv_undef>.
295 if (I-am-to-return-a-real-value) {
296 sv = sv_2mortal(newSViv(42));
300 This code tries to return a new SV (which contains the value 42) if it should
301 return a real value, or undef otherwise. Instead it has returned a NULL
302 pointer which, somewhere down the line, will cause a segmentation violation,
303 bus error, or just weird results. Change the zero to C<&PL_sv_undef> in the
304 first line and all will be well.
306 To free an SV that you've created, call C<SvREFCNT_dec(SV*)>. Normally this
307 call is not necessary (see L<Reference Counts and Mortality>).
311 Perl provides the function C<sv_chop> to efficiently remove characters
312 from the beginning of a string; you give it an SV and a pointer to
313 somewhere inside the PV, and it discards everything before the
314 pointer. The efficiency comes by means of a little hack: instead of
315 actually removing the characters, C<sv_chop> sets the flag C<OOK>
316 (offset OK) to signal to other functions that the offset hack is in
317 effect, and it moves the PV pointer (called C<SvPVX>) forward
318 by the number of bytes chopped off, and adjusts C<SvCUR> and C<SvLEN>
319 accordingly. (A portion of the space between the old and new PV
320 pointers is used to store the count of chopped bytes.)
322 Hence, at this point, the start of the buffer that we allocated lives
323 at C<SvPVX(sv) - SvIV(sv)> in memory and the PV pointer is pointing
324 into the middle of this allocated storage.
326 This is best demonstrated by example:
328 % ./perl -Ilib -MDevel::Peek -le '$a="12345"; $a=~s/.//; Dump($a)'
329 SV = PVIV(0x8128450) at 0x81340f0
331 FLAGS = (POK,OOK,pPOK)
333 PV = 0x8135781 ( "1" . ) "2345"\0
337 Here the number of bytes chopped off (1) is put into IV, and
338 C<Devel::Peek::Dump> helpfully reminds us that this is an offset. The
339 portion of the string between the "real" and the "fake" beginnings is
340 shown in parentheses, and the values of C<SvCUR> and C<SvLEN> reflect
341 the fake beginning, not the real one.
343 Something similar to the offset hack is performed on AVs to enable
344 efficient shifting and splicing off the beginning of the array; while
345 C<AvARRAY> points to the first element in the array that is visible from
346 Perl, C<AvALLOC> points to the real start of the C array. These are
347 usually the same, but a C<shift> operation can be carried out by
348 increasing C<AvARRAY> by one and decreasing C<AvFILL> and C<AvMAX>.
349 Again, the location of the real start of the C array only comes into
350 play when freeing the array. See C<av_shift> in F<av.c>.
352 =head2 What's Really Stored in an SV?
354 Recall that the usual method of determining the type of scalar you have is
355 to use C<Sv*OK> macros. Because a scalar can be both a number and a string,
356 usually these macros will always return TRUE and calling the C<Sv*V>
357 macros will do the appropriate conversion of string to integer/double or
358 integer/double to string.
360 If you I<really> need to know if you have an integer, double, or string
361 pointer in an SV, you can use the following three macros instead:
367 These will tell you if you truly have an integer, double, or string pointer
368 stored in your SV. The "p" stands for private.
370 There are various ways in which the private and public flags may differ.
371 For example, in perl 5.16 and earlier a tied SV may have a valid
372 underlying value in the IV slot (so SvIOKp is true), but the data
373 should be accessed via the FETCH routine rather than directly,
374 so SvIOK is false. (In perl 5.18 onwards, tied scalars use
375 the flags the same way as untied scalars.) Another is when
376 numeric conversion has occurred and precision has been lost: only the
377 private flag is set on 'lossy' values. So when an NV is converted to an
378 IV with loss, SvIOKp, SvNOKp and SvNOK will be set, while SvIOK wont be.
380 In general, though, it's best to use the C<Sv*V> macros.
382 =head2 Working with AVs
384 There are two ways to create and load an AV. The first method creates an
389 The second method both creates the AV and initially populates it with SVs:
391 AV* av_make(SSize_t num, SV **ptr);
393 The second argument points to an array containing C<num> C<SV*>'s. Once the
394 AV has been created, the SVs can be destroyed, if so desired.
396 Once the AV has been created, the following operations are possible on it:
398 void av_push(AV*, SV*);
401 void av_unshift(AV*, SSize_t num);
403 These should be familiar operations, with the exception of C<av_unshift>.
404 This routine adds C<num> elements at the front of the array with the C<undef>
405 value. You must then use C<av_store> (described below) to assign values
406 to these new elements.
408 Here are some other functions:
410 SSize_t av_top_index(AV*);
411 SV** av_fetch(AV*, SSize_t key, I32 lval);
412 SV** av_store(AV*, SSize_t key, SV* val);
414 The C<av_top_index> function returns the highest index value in an array (just
415 like $#array in Perl). If the array is empty, -1 is returned. The
416 C<av_fetch> function returns the value at index C<key>, but if C<lval>
417 is non-zero, then C<av_fetch> will store an undef value at that index.
418 The C<av_store> function stores the value C<val> at index C<key>, and does
419 not increment the reference count of C<val>. Thus the caller is responsible
420 for taking care of that, and if C<av_store> returns NULL, the caller will
421 have to decrement the reference count to avoid a memory leak. Note that
422 C<av_fetch> and C<av_store> both return C<SV**>'s, not C<SV*>'s as their
429 void av_extend(AV*, SSize_t key);
431 The C<av_clear> function deletes all the elements in the AV* array, but
432 does not actually delete the array itself. The C<av_undef> function will
433 delete all the elements in the array plus the array itself. The
434 C<av_extend> function extends the array so that it contains at least C<key+1>
435 elements. If C<key+1> is less than the currently allocated length of the array,
436 then nothing is done.
438 If you know the name of an array variable, you can get a pointer to its AV
439 by using the following:
441 AV* get_av("package::varname", 0);
443 This returns NULL if the variable does not exist.
445 See L<Understanding the Magic of Tied Hashes and Arrays> for more
446 information on how to use the array access functions on tied arrays.
448 =head2 Working with HVs
450 To create an HV, you use the following routine:
454 Once the HV has been created, the following operations are possible on it:
456 SV** hv_store(HV*, const char* key, U32 klen, SV* val, U32 hash);
457 SV** hv_fetch(HV*, const char* key, U32 klen, I32 lval);
459 The C<klen> parameter is the length of the key being passed in (Note that
460 you cannot pass 0 in as a value of C<klen> to tell Perl to measure the
461 length of the key). The C<val> argument contains the SV pointer to the
462 scalar being stored, and C<hash> is the precomputed hash value (zero if
463 you want C<hv_store> to calculate it for you). The C<lval> parameter
464 indicates whether this fetch is actually a part of a store operation, in
465 which case a new undefined value will be added to the HV with the supplied
466 key and C<hv_fetch> will return as if the value had already existed.
468 Remember that C<hv_store> and C<hv_fetch> return C<SV**>'s and not just
469 C<SV*>. To access the scalar value, you must first dereference the return
470 value. However, you should check to make sure that the return value is
471 not NULL before dereferencing it.
473 The first of these two functions checks if a hash table entry exists, and the
476 bool hv_exists(HV*, const char* key, U32 klen);
477 SV* hv_delete(HV*, const char* key, U32 klen, I32 flags);
479 If C<flags> does not include the C<G_DISCARD> flag then C<hv_delete> will
480 create and return a mortal copy of the deleted value.
482 And more miscellaneous functions:
487 Like their AV counterparts, C<hv_clear> deletes all the entries in the hash
488 table but does not actually delete the hash table. The C<hv_undef> deletes
489 both the entries and the hash table itself.
491 Perl keeps the actual data in a linked list of structures with a typedef of HE.
492 These contain the actual key and value pointers (plus extra administrative
493 overhead). The key is a string pointer; the value is an C<SV*>. However,
494 once you have an C<HE*>, to get the actual key and value, use the routines
497 I32 hv_iterinit(HV*);
498 /* Prepares starting point to traverse hash table */
499 HE* hv_iternext(HV*);
500 /* Get the next entry, and return a pointer to a
501 structure that has both the key and value */
502 char* hv_iterkey(HE* entry, I32* retlen);
503 /* Get the key from an HE structure and also return
504 the length of the key string */
505 SV* hv_iterval(HV*, HE* entry);
506 /* Return an SV pointer to the value of the HE
508 SV* hv_iternextsv(HV*, char** key, I32* retlen);
509 /* This convenience routine combines hv_iternext,
510 hv_iterkey, and hv_iterval. The key and retlen
511 arguments are return values for the key and its
512 length. The value is returned in the SV* argument */
514 If you know the name of a hash variable, you can get a pointer to its HV
515 by using the following:
517 HV* get_hv("package::varname", 0);
519 This returns NULL if the variable does not exist.
521 The hash algorithm is defined in the C<PERL_HASH> macro:
523 PERL_HASH(hash, key, klen)
525 The exact implementation of this macro varies by architecture and version
526 of perl, and the return value may change per invocation, so the value
527 is only valid for the duration of a single perl process.
529 See L<Understanding the Magic of Tied Hashes and Arrays> for more
530 information on how to use the hash access functions on tied hashes.
532 =head2 Hash API Extensions
534 Beginning with version 5.004, the following functions are also supported:
536 HE* hv_fetch_ent (HV* tb, SV* key, I32 lval, U32 hash);
537 HE* hv_store_ent (HV* tb, SV* key, SV* val, U32 hash);
539 bool hv_exists_ent (HV* tb, SV* key, U32 hash);
540 SV* hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
542 SV* hv_iterkeysv (HE* entry);
544 Note that these functions take C<SV*> keys, which simplifies writing
545 of extension code that deals with hash structures. These functions
546 also allow passing of C<SV*> keys to C<tie> functions without forcing
547 you to stringify the keys (unlike the previous set of functions).
549 They also return and accept whole hash entries (C<HE*>), making their
550 use more efficient (since the hash number for a particular string
551 doesn't have to be recomputed every time). See L<perlapi> for detailed
554 The following macros must always be used to access the contents of hash
555 entries. Note that the arguments to these macros must be simple
556 variables, since they may get evaluated more than once. See
557 L<perlapi> for detailed descriptions of these macros.
559 HePV(HE* he, STRLEN len)
563 HeSVKEY_force(HE* he)
564 HeSVKEY_set(HE* he, SV* sv)
566 These two lower level macros are defined, but must only be used when
567 dealing with keys that are not C<SV*>s:
572 Note that both C<hv_store> and C<hv_store_ent> do not increment the
573 reference count of the stored C<val>, which is the caller's responsibility.
574 If these functions return a NULL value, the caller will usually have to
575 decrement the reference count of C<val> to avoid a memory leak.
577 =head2 AVs, HVs and undefined values
579 Sometimes you have to store undefined values in AVs or HVs. Although
580 this may be a rare case, it can be tricky. That's because you're
581 used to using C<&PL_sv_undef> if you need an undefined SV.
583 For example, intuition tells you that this XS code:
586 av_store( av, 0, &PL_sv_undef );
588 is equivalent to this Perl code:
593 Unfortunately, this isn't true. In perl 5.18 and earlier, AVs use C<&PL_sv_undef> as a marker
594 for indicating that an array element has not yet been initialized.
595 Thus, C<exists $av[0]> would be true for the above Perl code, but
596 false for the array generated by the XS code. In perl 5.20, storing
597 &PL_sv_undef will create a read-only element, because the scalar
598 &PL_sv_undef itself is stored, not a copy.
600 Similar problems can occur when storing C<&PL_sv_undef> in HVs:
602 hv_store( hv, "key", 3, &PL_sv_undef, 0 );
604 This will indeed make the value C<undef>, but if you try to modify
605 the value of C<key>, you'll get the following error:
607 Modification of non-creatable hash value attempted
609 In perl 5.8.0, C<&PL_sv_undef> was also used to mark placeholders
610 in restricted hashes. This caused such hash entries not to appear
611 when iterating over the hash or when checking for the keys
612 with the C<hv_exists> function.
614 You can run into similar problems when you store C<&PL_sv_yes> or
615 C<&PL_sv_no> into AVs or HVs. Trying to modify such elements
616 will give you the following error:
618 Modification of a read-only value attempted
620 To make a long story short, you can use the special variables
621 C<&PL_sv_undef>, C<&PL_sv_yes> and C<&PL_sv_no> with AVs and
622 HVs, but you have to make sure you know what you're doing.
624 Generally, if you want to store an undefined value in an AV
625 or HV, you should not use C<&PL_sv_undef>, but rather create a
626 new undefined value using the C<newSV> function, for example:
628 av_store( av, 42, newSV(0) );
629 hv_store( hv, "foo", 3, newSV(0), 0 );
633 References are a special type of scalar that point to other data types
634 (including other references).
636 To create a reference, use either of the following functions:
638 SV* newRV_inc((SV*) thing);
639 SV* newRV_noinc((SV*) thing);
641 The C<thing> argument can be any of an C<SV*>, C<AV*>, or C<HV*>. The
642 functions are identical except that C<newRV_inc> increments the reference
643 count of the C<thing>, while C<newRV_noinc> does not. For historical
644 reasons, C<newRV> is a synonym for C<newRV_inc>.
646 Once you have a reference, you can use the following macro to dereference
651 then call the appropriate routines, casting the returned C<SV*> to either an
652 C<AV*> or C<HV*>, if required.
654 To determine if an SV is a reference, you can use the following macro:
658 To discover what type of value the reference refers to, use the following
659 macro and then check the return value.
663 The most useful types that will be returned are:
669 SVt_PVGV Glob (possibly a file handle)
671 See L<perlapi/svtype> for more details.
673 =head2 Blessed References and Class Objects
675 References are also used to support object-oriented programming. In perl's
676 OO lexicon, an object is simply a reference that has been blessed into a
677 package (or class). Once blessed, the programmer may now use the reference
678 to access the various methods in the class.
680 A reference can be blessed into a package with the following function:
682 SV* sv_bless(SV* sv, HV* stash);
684 The C<sv> argument must be a reference value. The C<stash> argument
685 specifies which class the reference will belong to. See
686 L<Stashes and Globs> for information on converting class names into stashes.
688 /* Still under construction */
690 The following function upgrades rv to reference if not already one.
691 Creates a new SV for rv to point to. If C<classname> is non-null, the SV
692 is blessed into the specified class. SV is returned.
694 SV* newSVrv(SV* rv, const char* classname);
696 The following three functions copy integer, unsigned integer or double
697 into an SV whose reference is C<rv>. SV is blessed if C<classname> is
700 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
701 SV* sv_setref_uv(SV* rv, const char* classname, UV uv);
702 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
704 The following function copies the pointer value (I<the address, not the
705 string!>) into an SV whose reference is rv. SV is blessed if C<classname>
708 SV* sv_setref_pv(SV* rv, const char* classname, void* pv);
710 The following function copies a string into an SV whose reference is C<rv>.
711 Set length to 0 to let Perl calculate the string length. SV is blessed if
712 C<classname> is non-null.
714 SV* sv_setref_pvn(SV* rv, const char* classname, char* pv,
717 The following function tests whether the SV is blessed into the specified
718 class. It does not check inheritance relationships.
720 int sv_isa(SV* sv, const char* name);
722 The following function tests whether the SV is a reference to a blessed object.
724 int sv_isobject(SV* sv);
726 The following function tests whether the SV is derived from the specified
727 class. SV can be either a reference to a blessed object or a string
728 containing a class name. This is the function implementing the
729 C<UNIVERSAL::isa> functionality.
731 bool sv_derived_from(SV* sv, const char* name);
733 To check if you've got an object derived from a specific class you have
736 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
738 =head2 Creating New Variables
740 To create a new Perl variable with an undef value which can be accessed from
741 your Perl script, use the following routines, depending on the variable type.
743 SV* get_sv("package::varname", GV_ADD);
744 AV* get_av("package::varname", GV_ADD);
745 HV* get_hv("package::varname", GV_ADD);
747 Notice the use of GV_ADD as the second parameter. The new variable can now
748 be set, using the routines appropriate to the data type.
750 There are additional macros whose values may be bitwise OR'ed with the
751 C<GV_ADD> argument to enable certain extra features. Those bits are:
757 Marks the variable as multiply defined, thus preventing the:
759 Name <varname> used only once: possible typo
767 Had to create <varname> unexpectedly
769 if the variable did not exist before the function was called.
773 If you do not specify a package name, the variable is created in the current
776 =head2 Reference Counts and Mortality
778 Perl uses a reference count-driven garbage collection mechanism. SVs,
779 AVs, or HVs (xV for short in the following) start their life with a
780 reference count of 1. If the reference count of an xV ever drops to 0,
781 then it will be destroyed and its memory made available for reuse.
783 This normally doesn't happen at the Perl level unless a variable is
784 undef'ed or the last variable holding a reference to it is changed or
785 overwritten. At the internal level, however, reference counts can be
786 manipulated with the following macros:
788 int SvREFCNT(SV* sv);
789 SV* SvREFCNT_inc(SV* sv);
790 void SvREFCNT_dec(SV* sv);
792 However, there is one other function which manipulates the reference
793 count of its argument. The C<newRV_inc> function, you will recall,
794 creates a reference to the specified argument. As a side effect,
795 it increments the argument's reference count. If this is not what
796 you want, use C<newRV_noinc> instead.
798 For example, imagine you want to return a reference from an XSUB function.
799 Inside the XSUB routine, you create an SV which initially has a reference
800 count of one. Then you call C<newRV_inc>, passing it the just-created SV.
801 This returns the reference as a new SV, but the reference count of the
802 SV you passed to C<newRV_inc> has been incremented to two. Now you
803 return the reference from the XSUB routine and forget about the SV.
804 But Perl hasn't! Whenever the returned reference is destroyed, the
805 reference count of the original SV is decreased to one and nothing happens.
806 The SV will hang around without any way to access it until Perl itself
807 terminates. This is a memory leak.
809 The correct procedure, then, is to use C<newRV_noinc> instead of
810 C<newRV_inc>. Then, if and when the last reference is destroyed,
811 the reference count of the SV will go to zero and it will be destroyed,
812 stopping any memory leak.
814 There are some convenience functions available that can help with the
815 destruction of xVs. These functions introduce the concept of "mortality".
816 An xV that is mortal has had its reference count marked to be decremented,
817 but not actually decremented, until "a short time later". Generally the
818 term "short time later" means a single Perl statement, such as a call to
819 an XSUB function. The actual determinant for when mortal xVs have their
820 reference count decremented depends on two macros, SAVETMPS and FREETMPS.
821 See L<perlcall> and L<perlxs> for more details on these macros.
823 "Mortalization" then is at its simplest a deferred C<SvREFCNT_dec>.
824 However, if you mortalize a variable twice, the reference count will
825 later be decremented twice.
827 "Mortal" SVs are mainly used for SVs that are placed on perl's stack.
828 For example an SV which is created just to pass a number to a called sub
829 is made mortal to have it cleaned up automatically when it's popped off
830 the stack. Similarly, results returned by XSUBs (which are pushed on the
831 stack) are often made mortal.
833 To create a mortal variable, use the functions:
837 SV* sv_mortalcopy(SV*)
839 The first call creates a mortal SV (with no value), the second converts an existing
840 SV to a mortal SV (and thus defers a call to C<SvREFCNT_dec>), and the
841 third creates a mortal copy of an existing SV.
842 Because C<sv_newmortal> gives the new SV no value, it must normally be given one
843 via C<sv_setpv>, C<sv_setiv>, etc. :
845 SV *tmp = sv_newmortal();
846 sv_setiv(tmp, an_integer);
848 As that is multiple C statements it is quite common so see this idiom instead:
850 SV *tmp = sv_2mortal(newSViv(an_integer));
853 You should be careful about creating mortal variables. Strange things
854 can happen if you make the same value mortal within multiple contexts,
855 or if you make a variable mortal multiple
856 times. Thinking of "Mortalization"
857 as deferred C<SvREFCNT_dec> should help to minimize such problems.
858 For example if you are passing an SV which you I<know> has a high enough REFCNT
859 to survive its use on the stack you need not do any mortalization.
860 If you are not sure then doing an C<SvREFCNT_inc> and C<sv_2mortal>, or
861 making a C<sv_mortalcopy> is safer.
863 The mortal routines are not just for SVs; AVs and HVs can be
864 made mortal by passing their address (type-casted to C<SV*>) to the
865 C<sv_2mortal> or C<sv_mortalcopy> routines.
867 =head2 Stashes and Globs
869 A B<stash> is a hash that contains all variables that are defined
870 within a package. Each key of the stash is a symbol
871 name (shared by all the different types of objects that have the same
872 name), and each value in the hash table is a GV (Glob Value). This GV
873 in turn contains references to the various objects of that name,
874 including (but not limited to) the following:
883 There is a single stash called C<PL_defstash> that holds the items that exist
884 in the C<main> package. To get at the items in other packages, append the
885 string "::" to the package name. The items in the C<Foo> package are in
886 the stash C<Foo::> in PL_defstash. The items in the C<Bar::Baz> package are
887 in the stash C<Baz::> in C<Bar::>'s stash.
889 To get the stash pointer for a particular package, use the function:
891 HV* gv_stashpv(const char* name, I32 flags)
892 HV* gv_stashsv(SV*, I32 flags)
894 The first function takes a literal string, the second uses the string stored
895 in the SV. Remember that a stash is just a hash table, so you get back an
896 C<HV*>. The C<flags> flag will create a new package if it is set to GV_ADD.
898 The name that C<gv_stash*v> wants is the name of the package whose symbol table
899 you want. The default package is called C<main>. If you have multiply nested
900 packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
903 Alternately, if you have an SV that is a blessed reference, you can find
904 out the stash pointer by using:
906 HV* SvSTASH(SvRV(SV*));
908 then use the following to get the package name itself:
910 char* HvNAME(HV* stash);
912 If you need to bless or re-bless an object you can use the following
915 SV* sv_bless(SV*, HV* stash)
917 where the first argument, an C<SV*>, must be a reference, and the second
918 argument is a stash. The returned C<SV*> can now be used in the same way
921 For more information on references and blessings, consult L<perlref>.
923 =head2 Double-Typed SVs
925 Scalar variables normally contain only one type of value, an integer,
926 double, pointer, or reference. Perl will automatically convert the
927 actual scalar data from the stored type into the requested type.
929 Some scalar variables contain more than one type of scalar data. For
930 example, the variable C<$!> contains either the numeric value of C<errno>
931 or its string equivalent from either C<strerror> or C<sys_errlist[]>.
933 To force multiple data values into an SV, you must do two things: use the
934 C<sv_set*v> routines to add the additional scalar type, then set a flag
935 so that Perl will believe it contains more than one type of data. The
936 four macros to set the flags are:
943 The particular macro you must use depends on which C<sv_set*v> routine
944 you called first. This is because every C<sv_set*v> routine turns on
945 only the bit for the particular type of data being set, and turns off
948 For example, to create a new Perl variable called "dberror" that contains
949 both the numeric and descriptive string error values, you could use the
953 extern char *dberror_list;
955 SV* sv = get_sv("dberror", GV_ADD);
956 sv_setiv(sv, (IV) dberror);
957 sv_setpv(sv, dberror_list[dberror]);
960 If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
961 macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
963 =head2 Read-Only Values
965 In Perl 5.16 and earlier, copy-on-write (see the next section) shared a
966 flag bit with read-only scalars. So the only way to test whether
967 C<sv_setsv>, etc., will raise a "Modification of a read-only value" error
968 in those versions is:
970 SvREADONLY(sv) && !SvIsCOW(sv)
972 Under Perl 5.18 and later, SvREADONLY only applies to read-only variables,
973 and, under 5.20, copy-on-write scalars can also be read-only, so the above
974 check is incorrect. You just want:
978 If you need to do this check often, define your own macro like this:
980 #if PERL_VERSION >= 18
981 # define SvTRULYREADONLY(sv) SvREADONLY(sv)
983 # define SvTRULYREADONLY(sv) (SvREADONLY(sv) && !SvIsCOW(sv))
988 Perl implements a copy-on-write (COW) mechanism for scalars, in which
989 string copies are not immediately made when requested, but are deferred
990 until made necessary by one or the other scalar changing. This is mostly
991 transparent, but one must take care not to modify string buffers that are
992 shared by multiple SVs.
994 You can test whether an SV is using copy-on-write with C<SvIsCOW(sv)>.
996 You can force an SV to make its own copy of its string buffer by calling C<sv_force_normal(sv)> or SvPV_force_nolen(sv).
998 If you want to make the SV drop its string buffer, use
999 C<sv_force_normal_flags(sv, SV_COW_DROP_PV)> or simply
1000 C<sv_setsv(sv, NULL)>.
1002 All of these functions will croak on read-only scalars (see the previous
1003 section for more on those).
1005 To test that your code is behaving correctly and not modifying COW buffers,
1006 on systems that support L<mmap(2)> (i.e., Unix) you can configure perl with
1007 C<-Accflags=-DPERL_DEBUG_READONLY_COW> and it will turn buffer violations
1008 into crashes. You will find it to be marvellously slow, so you may want to
1009 skip perl's own tests.
1011 =head2 Magic Variables
1013 [This section still under construction. Ignore everything here. Post no
1014 bills. Everything not permitted is forbidden.]
1016 Any SV may be magical, that is, it has special features that a normal
1017 SV does not have. These features are stored in the SV structure in a
1018 linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
1021 MAGIC* mg_moremagic;
1031 Note this is current as of patchlevel 0, and could change at any time.
1033 =head2 Assigning Magic
1035 Perl adds magic to an SV using the sv_magic function:
1037 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
1039 The C<sv> argument is a pointer to the SV that is to acquire a new magical
1042 If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
1043 convert C<sv> to type C<SVt_PVMG>.
1044 Perl then continues by adding new magic
1045 to the beginning of the linked list of magical features. Any prior entry
1046 of the same type of magic is deleted. Note that this can be overridden,
1047 and multiple instances of the same type of magic can be associated with an
1050 The C<name> and C<namlen> arguments are used to associate a string with
1051 the magic, typically the name of a variable. C<namlen> is stored in the
1052 C<mg_len> field and if C<name> is non-null then either a C<savepvn> copy of
1053 C<name> or C<name> itself is stored in the C<mg_ptr> field, depending on
1054 whether C<namlen> is greater than zero or equal to zero respectively. As a
1055 special case, if C<(name && namlen == HEf_SVKEY)> then C<name> is assumed
1056 to contain an C<SV*> and is stored as-is with its REFCNT incremented.
1058 The sv_magic function uses C<how> to determine which, if any, predefined
1059 "Magic Virtual Table" should be assigned to the C<mg_virtual> field.
1060 See the L<Magic Virtual Tables> section below. The C<how> argument is also
1061 stored in the C<mg_type> field. The value of
1062 C<how> should be chosen from the set of macros
1063 C<PERL_MAGIC_foo> found in F<perl.h>. Note that before
1064 these macros were added, Perl internals used to directly use character
1065 literals, so you may occasionally come across old code or documentation
1066 referring to 'U' magic rather than C<PERL_MAGIC_uvar> for example.
1068 The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
1069 structure. If it is not the same as the C<sv> argument, the reference
1070 count of the C<obj> object is incremented. If it is the same, or if
1071 the C<how> argument is C<PERL_MAGIC_arylen>, or if it is a NULL pointer,
1072 then C<obj> is merely stored, without the reference count being incremented.
1074 See also C<sv_magicext> in L<perlapi> for a more flexible way to add magic
1077 There is also a function to add magic to an C<HV>:
1079 void hv_magic(HV *hv, GV *gv, int how);
1081 This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
1083 To remove the magic from an SV, call the function sv_unmagic:
1085 int sv_unmagic(SV *sv, int type);
1087 The C<type> argument should be equal to the C<how> value when the C<SV>
1088 was initially made magical.
1090 However, note that C<sv_unmagic> removes all magic of a certain C<type> from the
1091 C<SV>. If you want to remove only certain
1092 magic of a C<type> based on the magic
1093 virtual table, use C<sv_unmagicext> instead:
1095 int sv_unmagicext(SV *sv, int type, MGVTBL *vtbl);
1097 =head2 Magic Virtual Tables
1099 The C<mg_virtual> field in the C<MAGIC> structure is a pointer to an
1100 C<MGVTBL>, which is a structure of function pointers and stands for
1101 "Magic Virtual Table" to handle the various operations that might be
1102 applied to that variable.
1104 The C<MGVTBL> has five (or sometimes eight) pointers to the following
1107 int (*svt_get)(SV* sv, MAGIC* mg);
1108 int (*svt_set)(SV* sv, MAGIC* mg);
1109 U32 (*svt_len)(SV* sv, MAGIC* mg);
1110 int (*svt_clear)(SV* sv, MAGIC* mg);
1111 int (*svt_free)(SV* sv, MAGIC* mg);
1113 int (*svt_copy)(SV *sv, MAGIC* mg, SV *nsv,
1114 const char *name, I32 namlen);
1115 int (*svt_dup)(MAGIC *mg, CLONE_PARAMS *param);
1116 int (*svt_local)(SV *nsv, MAGIC *mg);
1119 This MGVTBL structure is set at compile-time in F<perl.h> and there are
1120 currently 32 types. These different structures contain pointers to various
1121 routines that perform additional actions depending on which function is
1124 Function pointer Action taken
1125 ---------------- ------------
1126 svt_get Do something before the value of the SV is
1128 svt_set Do something after the SV is assigned a value.
1129 svt_len Report on the SV's length.
1130 svt_clear Clear something the SV represents.
1131 svt_free Free any extra storage associated with the SV.
1133 svt_copy copy tied variable magic to a tied element
1134 svt_dup duplicate a magic structure during thread cloning
1135 svt_local copy magic to local value during 'local'
1137 For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
1138 to an C<mg_type> of C<PERL_MAGIC_sv>) contains:
1140 { magic_get, magic_set, magic_len, 0, 0 }
1142 Thus, when an SV is determined to be magical and of type C<PERL_MAGIC_sv>,
1143 if a get operation is being performed, the routine C<magic_get> is
1144 called. All the various routines for the various magical types begin
1145 with C<magic_>. NOTE: the magic routines are not considered part of
1146 the Perl API, and may not be exported by the Perl library.
1148 The last three slots are a recent addition, and for source code
1149 compatibility they are only checked for if one of the three flags
1150 MGf_COPY, MGf_DUP or MGf_LOCAL is set in mg_flags.
1151 This means that most code can continue declaring
1152 a vtable as a 5-element value. These three are
1153 currently used exclusively by the threading code, and are highly subject
1156 The current kinds of Magic Virtual Tables are:
1159 This table is generated by regen/mg_vtable.pl. Any changes made here
1162 =for mg_vtable.pl begin
1165 (old-style char and macro) MGVTBL Type of magic
1166 -------------------------- ------ -------------
1167 \0 PERL_MAGIC_sv vtbl_sv Special scalar variable
1168 # PERL_MAGIC_arylen vtbl_arylen Array length ($#ary)
1169 % PERL_MAGIC_rhash (none) Extra data for restricted
1171 * PERL_MAGIC_debugvar vtbl_debugvar $DB::single, signal, trace
1173 . PERL_MAGIC_pos vtbl_pos pos() lvalue
1174 : PERL_MAGIC_symtab (none) Extra data for symbol
1176 < PERL_MAGIC_backref vtbl_backref For weak ref data
1177 @ PERL_MAGIC_arylen_p (none) To move arylen out of XPVAV
1178 B PERL_MAGIC_bm vtbl_regexp Boyer-Moore
1179 (fast string search)
1180 c PERL_MAGIC_overload_table vtbl_ovrld Holds overload table
1182 D PERL_MAGIC_regdata vtbl_regdata Regex match position data
1184 d PERL_MAGIC_regdatum vtbl_regdatum Regex match position data
1186 E PERL_MAGIC_env vtbl_env %ENV hash
1187 e PERL_MAGIC_envelem vtbl_envelem %ENV hash element
1188 f PERL_MAGIC_fm vtbl_regexp Formline
1190 g PERL_MAGIC_regex_global vtbl_mglob m//g target
1191 H PERL_MAGIC_hints vtbl_hints %^H hash
1192 h PERL_MAGIC_hintselem vtbl_hintselem %^H hash element
1193 I PERL_MAGIC_isa vtbl_isa @ISA array
1194 i PERL_MAGIC_isaelem vtbl_isaelem @ISA array element
1195 k PERL_MAGIC_nkeys vtbl_nkeys scalar(keys()) lvalue
1196 L PERL_MAGIC_dbfile (none) Debugger %_<filename
1197 l PERL_MAGIC_dbline vtbl_dbline Debugger %_<filename
1199 N PERL_MAGIC_shared (none) Shared between threads
1200 n PERL_MAGIC_shared_scalar (none) Shared between threads
1201 o PERL_MAGIC_collxfrm vtbl_collxfrm Locale transformation
1202 P PERL_MAGIC_tied vtbl_pack Tied array or hash
1203 p PERL_MAGIC_tiedelem vtbl_packelem Tied array or hash element
1204 q PERL_MAGIC_tiedscalar vtbl_packelem Tied scalar or handle
1205 r PERL_MAGIC_qr vtbl_regexp Precompiled qr// regex
1206 S PERL_MAGIC_sig (none) %SIG hash
1207 s PERL_MAGIC_sigelem vtbl_sigelem %SIG hash element
1208 t PERL_MAGIC_taint vtbl_taint Taintedness
1209 U PERL_MAGIC_uvar vtbl_uvar Available for use by
1211 u PERL_MAGIC_uvar_elem (none) Reserved for use by
1213 V PERL_MAGIC_vstring (none) SV was vstring literal
1214 v PERL_MAGIC_vec vtbl_vec vec() lvalue
1215 w PERL_MAGIC_utf8 vtbl_utf8 Cached UTF-8 information
1216 x PERL_MAGIC_substr vtbl_substr substr() lvalue
1217 y PERL_MAGIC_defelem vtbl_defelem Shadow "foreach" iterator
1218 variable / smart parameter
1220 \ PERL_MAGIC_lvref vtbl_lvref Lvalue reference
1222 ] PERL_MAGIC_checkcall vtbl_checkcall Inlining/mutation of call
1224 ~ PERL_MAGIC_ext (none) Available for use by
1227 =for mg_vtable.pl end
1229 When an uppercase and lowercase letter both exist in the table, then the
1230 uppercase letter is typically used to represent some kind of composite type
1231 (a list or a hash), and the lowercase letter is used to represent an element
1232 of that composite type. Some internals code makes use of this case
1233 relationship. However, 'v' and 'V' (vec and v-string) are in no way related.
1235 The C<PERL_MAGIC_ext> and C<PERL_MAGIC_uvar> magic types are defined
1236 specifically for use by extensions and will not be used by perl itself.
1237 Extensions can use C<PERL_MAGIC_ext> magic to 'attach' private information
1238 to variables (typically objects). This is especially useful because
1239 there is no way for normal perl code to corrupt this private information
1240 (unlike using extra elements of a hash object).
1242 Similarly, C<PERL_MAGIC_uvar> magic can be used much like tie() to call a
1243 C function any time a scalar's value is used or changed. The C<MAGIC>'s
1244 C<mg_ptr> field points to a C<ufuncs> structure:
1247 I32 (*uf_val)(pTHX_ IV, SV*);
1248 I32 (*uf_set)(pTHX_ IV, SV*);
1252 When the SV is read from or written to, the C<uf_val> or C<uf_set>
1253 function will be called with C<uf_index> as the first arg and a pointer to
1254 the SV as the second. A simple example of how to add C<PERL_MAGIC_uvar>
1255 magic is shown below. Note that the ufuncs structure is copied by
1256 sv_magic, so you can safely allocate it on the stack.
1264 uf.uf_val = &my_get_fn;
1265 uf.uf_set = &my_set_fn;
1267 sv_magic(sv, 0, PERL_MAGIC_uvar, (char*)&uf, sizeof(uf));
1269 Attaching C<PERL_MAGIC_uvar> to arrays is permissible but has no effect.
1271 For hashes there is a specialized hook that gives control over hash
1272 keys (but not values). This hook calls C<PERL_MAGIC_uvar> 'get' magic
1273 if the "set" function in the C<ufuncs> structure is NULL. The hook
1274 is activated whenever the hash is accessed with a key specified as
1275 an C<SV> through the functions C<hv_store_ent>, C<hv_fetch_ent>,
1276 C<hv_delete_ent>, and C<hv_exists_ent>. Accessing the key as a string
1277 through the functions without the C<..._ent> suffix circumvents the
1278 hook. See L<Hash::Util::FieldHash/GUTS> for a detailed description.
1280 Note that because multiple extensions may be using C<PERL_MAGIC_ext>
1281 or C<PERL_MAGIC_uvar> magic, it is important for extensions to take
1282 extra care to avoid conflict. Typically only using the magic on
1283 objects blessed into the same class as the extension is sufficient.
1284 For C<PERL_MAGIC_ext> magic, it is usually a good idea to define an
1285 C<MGVTBL>, even if all its fields will be C<0>, so that individual
1286 C<MAGIC> pointers can be identified as a particular kind of magic
1287 using their magic virtual table. C<mg_findext> provides an easy way
1290 STATIC MGVTBL my_vtbl = { 0, 0, 0, 0, 0, 0, 0, 0 };
1293 if ((mg = mg_findext(sv, PERL_MAGIC_ext, &my_vtbl))) {
1294 /* this is really ours, not another module's PERL_MAGIC_ext */
1295 my_priv_data_t *priv = (my_priv_data_t *)mg->mg_ptr;
1299 Also note that the C<sv_set*()> and C<sv_cat*()> functions described
1300 earlier do B<not> invoke 'set' magic on their targets. This must
1301 be done by the user either by calling the C<SvSETMAGIC()> macro after
1302 calling these functions, or by using one of the C<sv_set*_mg()> or
1303 C<sv_cat*_mg()> functions. Similarly, generic C code must call the
1304 C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
1305 obtained from external sources in functions that don't handle magic.
1306 See L<perlapi> for a description of these functions.
1307 For example, calls to the C<sv_cat*()> functions typically need to be
1308 followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
1309 since their implementation handles 'get' magic.
1311 =head2 Finding Magic
1313 MAGIC *mg_find(SV *sv, int type); /* Finds the magic pointer of that
1316 This routine returns a pointer to a C<MAGIC> structure stored in the SV.
1317 If the SV does not have that magical
1318 feature, C<NULL> is returned. If the
1319 SV has multiple instances of that magical feature, the first one will be
1320 returned. C<mg_findext> can be used
1321 to find a C<MAGIC> structure of an SV
1322 based on both its magic type and its magic virtual table:
1324 MAGIC *mg_findext(SV *sv, int type, MGVTBL *vtbl);
1326 Also, if the SV passed to C<mg_find> or C<mg_findext> is not of type
1327 SVt_PVMG, Perl may core dump.
1329 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
1331 This routine checks to see what types of magic C<sv> has. If the mg_type
1332 field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
1333 the mg_type field is changed to be the lowercase letter.
1335 =head2 Understanding the Magic of Tied Hashes and Arrays
1337 Tied hashes and arrays are magical beasts of the C<PERL_MAGIC_tied>
1340 WARNING: As of the 5.004 release, proper usage of the array and hash
1341 access functions requires understanding a few caveats. Some
1342 of these caveats are actually considered bugs in the API, to be fixed
1343 in later releases, and are bracketed with [MAYCHANGE] below. If
1344 you find yourself actually applying such information in this section, be
1345 aware that the behavior may change in the future, umm, without warning.
1347 The perl tie function associates a variable with an object that implements
1348 the various GET, SET, etc methods. To perform the equivalent of the perl
1349 tie function from an XSUB, you must mimic this behaviour. The code below
1350 carries out the necessary steps - firstly it creates a new hash, and then
1351 creates a second hash which it blesses into the class which will implement
1352 the tie methods. Lastly it ties the two hashes together, and returns a
1353 reference to the new tied hash. Note that the code below does NOT call the
1354 TIEHASH method in the MyTie class -
1355 see L<Calling Perl Routines from within C Programs> for details on how
1366 tie = newRV_noinc((SV*)newHV());
1367 stash = gv_stashpv("MyTie", GV_ADD);
1368 sv_bless(tie, stash);
1369 hv_magic(hash, (GV*)tie, PERL_MAGIC_tied);
1370 RETVAL = newRV_noinc(hash);
1374 The C<av_store> function, when given a tied array argument, merely
1375 copies the magic of the array onto the value to be "stored", using
1376 C<mg_copy>. It may also return NULL, indicating that the value did not
1377 actually need to be stored in the array. [MAYCHANGE] After a call to
1378 C<av_store> on a tied array, the caller will usually need to call
1379 C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1380 TIEARRAY object. If C<av_store> did return NULL, a call to
1381 C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1384 The previous paragraph is applicable verbatim to tied hash access using the
1385 C<hv_store> and C<hv_store_ent> functions as well.
1387 C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1388 C<hv_fetch_ent> actually return an undefined mortal value whose magic
1389 has been initialized using C<mg_copy>. Note the value so returned does not
1390 need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1391 need to call C<mg_get()> on the returned value in order to actually invoke
1392 the perl level "FETCH" method on the underlying TIE object. Similarly,
1393 you may also call C<mg_set()> on the return value after possibly assigning
1394 a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
1395 method on the TIE object. [/MAYCHANGE]
1398 In other words, the array or hash fetch/store functions don't really
1399 fetch and store actual values in the case of tied arrays and hashes. They
1400 merely call C<mg_copy> to attach magic to the values that were meant to be
1401 "stored" or "fetched". Later calls to C<mg_get> and C<mg_set> actually
1402 do the job of invoking the TIE methods on the underlying objects. Thus
1403 the magic mechanism currently implements a kind of lazy access to arrays
1406 Currently (as of perl version 5.004), use of the hash and array access
1407 functions requires the user to be aware of whether they are operating on
1408 "normal" hashes and arrays, or on their tied variants. The API may be
1409 changed to provide more transparent access to both tied and normal data
1410 types in future versions.
1413 You would do well to understand that the TIEARRAY and TIEHASH interfaces
1414 are mere sugar to invoke some perl method calls while using the uniform hash
1415 and array syntax. The use of this sugar imposes some overhead (typically
1416 about two to four extra opcodes per FETCH/STORE operation, in addition to
1417 the creation of all the mortal variables required to invoke the methods).
1418 This overhead will be comparatively small if the TIE methods are themselves
1419 substantial, but if they are only a few statements long, the overhead
1420 will not be insignificant.
1422 =head2 Localizing changes
1424 Perl has a very handy construction
1431 This construction is I<approximately> equivalent to
1440 The biggest difference is that the first construction would
1441 reinstate the initial value of $var, irrespective of how control exits
1442 the block: C<goto>, C<return>, C<die>/C<eval>, etc. It is a little bit
1443 more efficient as well.
1445 There is a way to achieve a similar task from C via Perl API: create a
1446 I<pseudo-block>, and arrange for some changes to be automatically
1447 undone at the end of it, either explicit, or via a non-local exit (via
1448 die()). A I<block>-like construct is created by a pair of
1449 C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1450 Such a construct may be created specially for some important localized
1451 task, or an existing one (like boundaries of enclosing Perl
1452 subroutine/block, or an existing pair for freeing TMPs) may be
1453 used. (In the second case the overhead of additional localization must
1454 be almost negligible.) Note that any XSUB is automatically enclosed in
1455 an C<ENTER>/C<LEAVE> pair.
1457 Inside such a I<pseudo-block> the following service is available:
1461 =item C<SAVEINT(int i)>
1463 =item C<SAVEIV(IV i)>
1465 =item C<SAVEI32(I32 i)>
1467 =item C<SAVELONG(long i)>
1469 These macros arrange things to restore the value of integer variable
1470 C<i> at the end of enclosing I<pseudo-block>.
1472 =item C<SAVESPTR(s)>
1474 =item C<SAVEPPTR(p)>
1476 These macros arrange things to restore the value of pointers C<s> and
1477 C<p>. C<s> must be a pointer of a type which survives conversion to
1478 C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1481 =item C<SAVEFREESV(SV *sv)>
1483 The refcount of C<sv> would be decremented at the end of
1484 I<pseudo-block>. This is similar to C<sv_2mortal> in that it is also a
1485 mechanism for doing a delayed C<SvREFCNT_dec>. However, while C<sv_2mortal>
1486 extends the lifetime of C<sv> until the beginning of the next statement,
1487 C<SAVEFREESV> extends it until the end of the enclosing scope. These
1488 lifetimes can be wildly different.
1490 Also compare C<SAVEMORTALIZESV>.
1492 =item C<SAVEMORTALIZESV(SV *sv)>
1494 Just like C<SAVEFREESV>, but mortalizes C<sv> at the end of the current
1495 scope instead of decrementing its reference count. This usually has the
1496 effect of keeping C<sv> alive until the statement that called the currently
1497 live scope has finished executing.
1499 =item C<SAVEFREEOP(OP *op)>
1501 The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1503 =item C<SAVEFREEPV(p)>
1505 The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1506 end of I<pseudo-block>.
1508 =item C<SAVECLEARSV(SV *sv)>
1510 Clears a slot in the current scratchpad which corresponds to C<sv> at
1511 the end of I<pseudo-block>.
1513 =item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1515 The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
1516 string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1517 short-lived storage, the corresponding string may be reallocated like
1520 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
1522 =item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
1524 At the end of I<pseudo-block> the function C<f> is called with the
1527 =item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1529 At the end of I<pseudo-block> the function C<f> is called with the
1530 implicit context argument (if any), and C<p>.
1532 =item C<SAVESTACK_POS()>
1534 The current offset on the Perl internal stack (cf. C<SP>) is restored
1535 at the end of I<pseudo-block>.
1539 The following API list contains functions, thus one needs to
1540 provide pointers to the modifiable data explicitly (either C pointers,
1541 or Perlish C<GV *>s). Where the above macros take C<int>, a similar
1542 function takes C<int *>.
1546 =item C<SV* save_scalar(GV *gv)>
1548 Equivalent to Perl code C<local $gv>.
1550 =item C<AV* save_ary(GV *gv)>
1552 =item C<HV* save_hash(GV *gv)>
1554 Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1556 =item C<void save_item(SV *item)>
1558 Duplicates the current value of C<SV>, on the exit from the current
1559 C<ENTER>/C<LEAVE> I<pseudo-block> will restore the value of C<SV>
1560 using the stored value. It doesn't handle magic. Use C<save_scalar> if
1563 =item C<void save_list(SV **sarg, I32 maxsarg)>
1565 A variant of C<save_item> which takes multiple arguments via an array
1566 C<sarg> of C<SV*> of length C<maxsarg>.
1568 =item C<SV* save_svref(SV **sptr)>
1570 Similar to C<save_scalar>, but will reinstate an C<SV *>.
1572 =item C<void save_aptr(AV **aptr)>
1574 =item C<void save_hptr(HV **hptr)>
1576 Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1580 The C<Alias> module implements localization of the basic types within the
1581 I<caller's scope>. People who are interested in how to localize things in
1582 the containing scope should take a look there too.
1586 =head2 XSUBs and the Argument Stack
1588 The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1589 An XSUB routine will have a stack that contains the arguments from the Perl
1590 program, and a way to map from the Perl data structures to a C equivalent.
1592 The stack arguments are accessible through the C<ST(n)> macro, which returns
1593 the C<n>'th stack argument. Argument 0 is the first argument passed in the
1594 Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1597 Most of the time, output from the C routine can be handled through use of
1598 the RETVAL and OUTPUT directives. However, there are some cases where the
1599 argument stack is not already long enough to handle all the return values.
1600 An example is the POSIX tzname() call, which takes no arguments, but returns
1601 two, the local time zone's standard and summer time abbreviations.
1603 To handle this situation, the PPCODE directive is used and the stack is
1604 extended using the macro:
1608 where C<SP> is the macro that represents the local copy of the stack pointer,
1609 and C<num> is the number of elements the stack should be extended by.
1611 Now that there is room on the stack, values can be pushed on it using C<PUSHs>
1612 macro. The pushed values will often need to be "mortal" (See
1613 L</Reference Counts and Mortality>):
1615 PUSHs(sv_2mortal(newSViv(an_integer)))
1616 PUSHs(sv_2mortal(newSVuv(an_unsigned_integer)))
1617 PUSHs(sv_2mortal(newSVnv(a_double)))
1618 PUSHs(sv_2mortal(newSVpv("Some String",0)))
1619 /* Although the last example is better written as the more
1621 PUSHs(newSVpvs_flags("Some String", SVs_TEMP))
1623 And now the Perl program calling C<tzname>, the two values will be assigned
1626 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1628 An alternate (and possibly simpler) method to pushing values on the stack is
1633 This macro automatically adjusts the stack for you, if needed. Thus, you
1634 do not need to call C<EXTEND> to extend the stack.
1636 Despite their suggestions in earlier versions of this document the macros
1637 C<(X)PUSH[iunp]> are I<not> suited to XSUBs which return multiple results.
1638 For that, either stick to the C<(X)PUSHs> macros shown above, or use the new
1639 C<m(X)PUSH[iunp]> macros instead; see L</Putting a C value on Perl stack>.
1641 For more information, consult L<perlxs> and L<perlxstut>.
1643 =head2 Autoloading with XSUBs
1645 If an AUTOLOAD routine is an XSUB, as with Perl subroutines, Perl puts the
1646 fully-qualified name of the autoloaded subroutine in the $AUTOLOAD variable
1647 of the XSUB's package.
1649 But it also puts the same information in certain fields of the XSUB itself:
1651 HV *stash = CvSTASH(cv);
1652 const char *subname = SvPVX(cv);
1653 STRLEN name_length = SvCUR(cv); /* in bytes */
1654 U32 is_utf8 = SvUTF8(cv);
1656 C<SvPVX(cv)> contains just the sub name itself, not including the package.
1657 For an AUTOLOAD routine in UNIVERSAL or one of its superclasses,
1658 C<CvSTASH(cv)> returns NULL during a method call on a nonexistent package.
1660 B<Note>: Setting $AUTOLOAD stopped working in 5.6.1, which did not support
1661 XS AUTOLOAD subs at all. Perl 5.8.0 introduced the use of fields in the
1662 XSUB itself. Perl 5.16.0 restored the setting of $AUTOLOAD. If you need
1663 to support 5.8-5.14, use the XSUB's fields.
1665 =head2 Calling Perl Routines from within C Programs
1667 There are four routines that can be used to call a Perl subroutine from
1668 within a C program. These four are:
1670 I32 call_sv(SV*, I32);
1671 I32 call_pv(const char*, I32);
1672 I32 call_method(const char*, I32);
1673 I32 call_argv(const char*, I32, char**);
1675 The routine most often used is C<call_sv>. The C<SV*> argument
1676 contains either the name of the Perl subroutine to be called, or a
1677 reference to the subroutine. The second argument consists of flags
1678 that control the context in which the subroutine is called, whether
1679 or not the subroutine is being passed arguments, how errors should be
1680 trapped, and how to treat return values.
1682 All four routines return the number of arguments that the subroutine returned
1685 These routines used to be called C<perl_call_sv>, etc., before Perl v5.6.0,
1686 but those names are now deprecated; macros of the same name are provided for
1689 When using any of these routines (except C<call_argv>), the programmer
1690 must manipulate the Perl stack. These include the following macros and
1705 For a detailed description of calling conventions from C to Perl,
1706 consult L<perlcall>.
1708 =head2 Putting a C value on Perl stack
1710 A lot of opcodes (this is an elementary operation in the internal perl
1711 stack machine) put an SV* on the stack. However, as an optimization
1712 the corresponding SV is (usually) not recreated each time. The opcodes
1713 reuse specially assigned SVs (I<target>s) which are (as a corollary)
1714 not constantly freed/created.
1716 Each of the targets is created only once (but see
1717 L<Scratchpads and recursion> below), and when an opcode needs to put
1718 an integer, a double, or a string on stack, it just sets the
1719 corresponding parts of its I<target> and puts the I<target> on stack.
1721 The macro to put this target on stack is C<PUSHTARG>, and it is
1722 directly used in some opcodes, as well as indirectly in zillions of
1723 others, which use it via C<(X)PUSH[iunp]>.
1725 Because the target is reused, you must be careful when pushing multiple
1726 values on the stack. The following code will not do what you think:
1731 This translates as "set C<TARG> to 10, push a pointer to C<TARG> onto
1732 the stack; set C<TARG> to 20, push a pointer to C<TARG> onto the stack".
1733 At the end of the operation, the stack does not contain the values 10
1734 and 20, but actually contains two pointers to C<TARG>, which we have set
1737 If you need to push multiple different values then you should either use
1738 the C<(X)PUSHs> macros, or else use the new C<m(X)PUSH[iunp]> macros,
1739 none of which make use of C<TARG>. The C<(X)PUSHs> macros simply push an
1740 SV* on the stack, which, as noted under L</XSUBs and the Argument Stack>,
1741 will often need to be "mortal". The new C<m(X)PUSH[iunp]> macros make
1742 this a little easier to achieve by creating a new mortal for you (via
1743 C<(X)PUSHmortal>), pushing that onto the stack (extending it if necessary
1744 in the case of the C<mXPUSH[iunp]> macros), and then setting its value.
1745 Thus, instead of writing this to "fix" the example above:
1747 XPUSHs(sv_2mortal(newSViv(10)))
1748 XPUSHs(sv_2mortal(newSViv(20)))
1750 you can simply write:
1755 On a related note, if you do use C<(X)PUSH[iunp]>, then you're going to
1756 need a C<dTARG> in your variable declarations so that the C<*PUSH*>
1757 macros can make use of the local variable C<TARG>. See also C<dTARGET>
1762 The question remains on when the SVs which are I<target>s for opcodes
1763 are created. The answer is that they are created when the current
1764 unit--a subroutine or a file (for opcodes for statements outside of
1765 subroutines)--is compiled. During this time a special anonymous Perl
1766 array is created, which is called a scratchpad for the current unit.
1768 A scratchpad keeps SVs which are lexicals for the current unit and are
1769 targets for opcodes. A previous version of this document
1770 stated that one can deduce that an SV lives on a scratchpad
1771 by looking on its flags: lexicals have C<SVs_PADMY> set, and
1772 I<target>s have C<SVs_PADTMP> set. But this has never been fully true.
1773 C<SVs_PADMY> could be set on a variable that no longer resides in any pad.
1774 While I<target>s do have C<SVs_PADTMP> set, it can also be set on variables
1775 that have never resided in a pad, but nonetheless act like I<target>s. As
1776 of perl 5.21.5, the C<SVs_PADMY> flag is no longer used and is defined as
1777 0. C<SvPADMY()> now returns true for anything without C<SVs_PADTMP>.
1779 The correspondence between OPs and I<target>s is not 1-to-1. Different
1780 OPs in the compile tree of the unit can use the same target, if this
1781 would not conflict with the expected life of the temporary.
1783 =head2 Scratchpads and recursion
1785 In fact it is not 100% true that a compiled unit contains a pointer to
1786 the scratchpad AV. In fact it contains a pointer to an AV of
1787 (initially) one element, and this element is the scratchpad AV. Why do
1788 we need an extra level of indirection?
1790 The answer is B<recursion>, and maybe B<threads>. Both
1791 these can create several execution pointers going into the same
1792 subroutine. For the subroutine-child not write over the temporaries
1793 for the subroutine-parent (lifespan of which covers the call to the
1794 child), the parent and the child should have different
1795 scratchpads. (I<And> the lexicals should be separate anyway!)
1797 So each subroutine is born with an array of scratchpads (of length 1).
1798 On each entry to the subroutine it is checked that the current
1799 depth of the recursion is not more than the length of this array, and
1800 if it is, new scratchpad is created and pushed into the array.
1802 The I<target>s on this scratchpad are C<undef>s, but they are already
1803 marked with correct flags.
1805 =head1 Memory Allocation
1809 All memory meant to be used with the Perl API functions should be manipulated
1810 using the macros described in this section. The macros provide the necessary
1811 transparency between differences in the actual malloc implementation that is
1814 It is suggested that you enable the version of malloc that is distributed
1815 with Perl. It keeps pools of various sizes of unallocated memory in
1816 order to satisfy allocation requests more quickly. However, on some
1817 platforms, it may cause spurious malloc or free errors.
1819 The following three macros are used to initially allocate memory :
1821 Newx(pointer, number, type);
1822 Newxc(pointer, number, type, cast);
1823 Newxz(pointer, number, type);
1825 The first argument C<pointer> should be the name of a variable that will
1826 point to the newly allocated memory.
1828 The second and third arguments C<number> and C<type> specify how many of
1829 the specified type of data structure should be allocated. The argument
1830 C<type> is passed to C<sizeof>. The final argument to C<Newxc>, C<cast>,
1831 should be used if the C<pointer> argument is different from the C<type>
1834 Unlike the C<Newx> and C<Newxc> macros, the C<Newxz> macro calls C<memzero>
1835 to zero out all the newly allocated memory.
1839 Renew(pointer, number, type);
1840 Renewc(pointer, number, type, cast);
1843 These three macros are used to change a memory buffer size or to free a
1844 piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1845 match those of C<New> and C<Newc> with the exception of not needing the
1846 "magic cookie" argument.
1850 Move(source, dest, number, type);
1851 Copy(source, dest, number, type);
1852 Zero(dest, number, type);
1854 These three macros are used to move, copy, or zero out previously allocated
1855 memory. The C<source> and C<dest> arguments point to the source and
1856 destination starting points. Perl will move, copy, or zero out C<number>
1857 instances of the size of the C<type> data structure (using the C<sizeof>
1862 The most recent development releases of Perl have been experimenting with
1863 removing Perl's dependency on the "normal" standard I/O suite and allowing
1864 other stdio implementations to be used. This involves creating a new
1865 abstraction layer that then calls whichever implementation of stdio Perl
1866 was compiled with. All XSUBs should now use the functions in the PerlIO
1867 abstraction layer and not make any assumptions about what kind of stdio
1870 For a complete description of the PerlIO abstraction, consult L<perlapio>.
1872 =head1 Compiled code
1876 Here we describe the internal form your code is converted to by
1877 Perl. Start with a simple example:
1881 This is converted to a tree similar to this one:
1889 (but slightly more complicated). This tree reflects the way Perl
1890 parsed your code, but has nothing to do with the execution order.
1891 There is an additional "thread" going through the nodes of the tree
1892 which shows the order of execution of the nodes. In our simplified
1893 example above it looks like:
1895 $b ---> $c ---> + ---> $a ---> assign-to
1897 But with the actual compile tree for C<$a = $b + $c> it is different:
1898 some nodes I<optimized away>. As a corollary, though the actual tree
1899 contains more nodes than our simplified example, the execution order
1900 is the same as in our example.
1902 =head2 Examining the tree
1904 If you have your perl compiled for debugging (usually done with
1905 C<-DDEBUGGING> on the C<Configure> command line), you may examine the
1906 compiled tree by specifying C<-Dx> on the Perl command line. The
1907 output takes several lines per node, and for C<$b+$c> it looks like
1912 FLAGS = (SCALAR,KIDS)
1914 TYPE = null ===> (4)
1916 FLAGS = (SCALAR,KIDS)
1918 3 TYPE = gvsv ===> 4
1924 TYPE = null ===> (5)
1926 FLAGS = (SCALAR,KIDS)
1928 4 TYPE = gvsv ===> 5
1934 This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
1935 not optimized away (one per number in the left column). The immediate
1936 children of the given node correspond to C<{}> pairs on the same level
1937 of indentation, thus this listing corresponds to the tree:
1945 The execution order is indicated by C<===E<gt>> marks, thus it is C<3
1946 4 5 6> (node C<6> is not included into above listing), i.e.,
1947 C<gvsv gvsv add whatever>.
1949 Each of these nodes represents an op, a fundamental operation inside the
1950 Perl core. The code which implements each operation can be found in the
1951 F<pp*.c> files; the function which implements the op with type C<gvsv>
1952 is C<pp_gvsv>, and so on. As the tree above shows, different ops have
1953 different numbers of children: C<add> is a binary operator, as one would
1954 expect, and so has two children. To accommodate the various different
1955 numbers of children, there are various types of op data structure, and
1956 they link together in different ways.
1958 The simplest type of op structure is C<OP>: this has no children. Unary
1959 operators, C<UNOP>s, have one child, and this is pointed to by the
1960 C<op_first> field. Binary operators (C<BINOP>s) have not only an
1961 C<op_first> field but also an C<op_last> field. The most complex type of
1962 op is a C<LISTOP>, which has any number of children. In this case, the
1963 first child is pointed to by C<op_first> and the last child by
1964 C<op_last>. The children in between can be found by iteratively
1965 following the C<op_sibling> pointer from the first child to the last 9but
1968 There are also some other op types: a C<PMOP> holds a regular expression,
1969 and has no children, and a C<LOOP> may or may not have children. If the
1970 C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
1971 complicate matters, if a C<UNOP> is actually a C<null> op after
1972 optimization (see L</Compile pass 2: context propagation>) it will still
1973 have children in accordance with its former type.
1975 Finally, there is a C<LOGOP>, or logic op. Like a C<LISTOP>, this has one
1976 or more children, but it doesn't have an C<op_last> field: so you have to
1977 follow C<op_first> and then the C<op_sibling> chain itself to find the
1978 last child. Instead it has an C<op_other> field, which is comparable to
1979 the C<op_next> field described below, and represents an alternate
1980 execution path. Operators like C<and>, C<or> and C<?> are C<LOGOP>s. Note
1981 that in general, C<op_other> may not point to any of the direct children
1984 Starting in version 5.21.2, perls built with the experimental
1985 define C<-DPERL_OP_PARENT> add an extra boolean flag for each op,
1986 C<op_lastsib>. When set, this indicates that this is the last op in an
1987 C<op_sibling> chain. This frees up the C<op_sibling> field on the last
1988 sibling to point back to the parent op. The macro C<OpSIBLING(o)> wraps
1989 this special behaviour, and always returns NULL on the last sibling.
1990 With this build the C<op_parent(o)> function can be used to find the
1993 Another way to examine the tree is to use a compiler back-end module, such
1996 =head2 Compile pass 1: check routines
1998 The tree is created by the compiler while I<yacc> code feeds it
1999 the constructions it recognizes. Since I<yacc> works bottom-up, so does
2000 the first pass of perl compilation.
2002 What makes this pass interesting for perl developers is that some
2003 optimization may be performed on this pass. This is optimization by
2004 so-called "check routines". The correspondence between node names
2005 and corresponding check routines is described in F<opcode.pl> (do not
2006 forget to run C<make regen_headers> if you modify this file).
2008 A check routine is called when the node is fully constructed except
2009 for the execution-order thread. Since at this time there are no
2010 back-links to the currently constructed node, one can do most any
2011 operation to the top-level node, including freeing it and/or creating
2012 new nodes above/below it.
2014 The check routine returns the node which should be inserted into the
2015 tree (if the top-level node was not modified, check routine returns
2018 By convention, check routines have names C<ck_*>. They are usually
2019 called from C<new*OP> subroutines (or C<convert>) (which in turn are
2020 called from F<perly.y>).
2022 =head2 Compile pass 1a: constant folding
2024 Immediately after the check routine is called the returned node is
2025 checked for being compile-time executable. If it is (the value is
2026 judged to be constant) it is immediately executed, and a I<constant>
2027 node with the "return value" of the corresponding subtree is
2028 substituted instead. The subtree is deleted.
2030 If constant folding was not performed, the execution-order thread is
2033 =head2 Compile pass 2: context propagation
2035 When a context for a part of compile tree is known, it is propagated
2036 down through the tree. At this time the context can have 5 values
2037 (instead of 2 for runtime context): void, boolean, scalar, list, and
2038 lvalue. In contrast with the pass 1 this pass is processed from top
2039 to bottom: a node's context determines the context for its children.
2041 Additional context-dependent optimizations are performed at this time.
2042 Since at this moment the compile tree contains back-references (via
2043 "thread" pointers), nodes cannot be free()d now. To allow
2044 optimized-away nodes at this stage, such nodes are null()ified instead
2045 of free()ing (i.e. their type is changed to OP_NULL).
2047 =head2 Compile pass 3: peephole optimization
2049 After the compile tree for a subroutine (or for an C<eval> or a file)
2050 is created, an additional pass over the code is performed. This pass
2051 is neither top-down or bottom-up, but in the execution order (with
2052 additional complications for conditionals). Optimizations performed
2053 at this stage are subject to the same restrictions as in the pass 2.
2055 Peephole optimizations are done by calling the function pointed to
2056 by the global variable C<PL_peepp>. By default, C<PL_peepp> just
2057 calls the function pointed to by the global variable C<PL_rpeepp>.
2058 By default, that performs some basic op fixups and optimisations along
2059 the execution-order op chain, and recursively calls C<PL_rpeepp> for
2060 each side chain of ops (resulting from conditionals). Extensions may
2061 provide additional optimisations or fixups, hooking into either the
2062 per-subroutine or recursive stage, like this:
2064 static peep_t prev_peepp;
2065 static void my_peep(pTHX_ OP *o)
2067 /* custom per-subroutine optimisation goes here */
2068 prev_peepp(aTHX_ o);
2069 /* custom per-subroutine optimisation may also go here */
2072 prev_peepp = PL_peepp;
2075 static peep_t prev_rpeepp;
2076 static void my_rpeep(pTHX_ OP *o)
2079 for(; o; o = o->op_next) {
2080 /* custom per-op optimisation goes here */
2082 prev_rpeepp(aTHX_ orig_o);
2085 prev_rpeepp = PL_rpeepp;
2086 PL_rpeepp = my_rpeep;
2088 =head2 Pluggable runops
2090 The compile tree is executed in a runops function. There are two runops
2091 functions, in F<run.c> and in F<dump.c>. C<Perl_runops_debug> is used
2092 with DEBUGGING and C<Perl_runops_standard> is used otherwise. For fine
2093 control over the execution of the compile tree it is possible to provide
2094 your own runops function.
2096 It's probably best to copy one of the existing runops functions and
2097 change it to suit your needs. Then, in the BOOT section of your XS
2100 PL_runops = my_runops;
2102 This function should be as efficient as possible to keep your programs
2103 running as fast as possible.
2105 =head2 Compile-time scope hooks
2107 As of perl 5.14 it is possible to hook into the compile-time lexical
2108 scope mechanism using C<Perl_blockhook_register>. This is used like
2111 STATIC void my_start_hook(pTHX_ int full);
2112 STATIC BHK my_hooks;
2115 BhkENTRY_set(&my_hooks, bhk_start, my_start_hook);
2116 Perl_blockhook_register(aTHX_ &my_hooks);
2118 This will arrange to have C<my_start_hook> called at the start of
2119 compiling every lexical scope. The available hooks are:
2123 =item C<void bhk_start(pTHX_ int full)>
2125 This is called just after starting a new lexical scope. Note that Perl
2130 creates two scopes: the first starts at the C<(> and has C<full == 1>,
2131 the second starts at the C<{> and has C<full == 0>. Both end at the
2132 C<}>, so calls to C<start> and C<pre/post_end> will match. Anything
2133 pushed onto the save stack by this hook will be popped just before the
2134 scope ends (between the C<pre_> and C<post_end> hooks, in fact).
2136 =item C<void bhk_pre_end(pTHX_ OP **o)>
2138 This is called at the end of a lexical scope, just before unwinding the
2139 stack. I<o> is the root of the optree representing the scope; it is a
2140 double pointer so you can replace the OP if you need to.
2142 =item C<void bhk_post_end(pTHX_ OP **o)>
2144 This is called at the end of a lexical scope, just after unwinding the
2145 stack. I<o> is as above. Note that it is possible for calls to C<pre_>
2146 and C<post_end> to nest, if there is something on the save stack that
2149 =item C<void bhk_eval(pTHX_ OP *const o)>
2151 This is called just before starting to compile an C<eval STRING>, C<do
2152 FILE>, C<require> or C<use>, after the eval has been set up. I<o> is the
2153 OP that requested the eval, and will normally be an C<OP_ENTEREVAL>,
2154 C<OP_DOFILE> or C<OP_REQUIRE>.
2158 Once you have your hook functions, you need a C<BHK> structure to put
2159 them in. It's best to allocate it statically, since there is no way to
2160 free it once it's registered. The function pointers should be inserted
2161 into this structure using the C<BhkENTRY_set> macro, which will also set
2162 flags indicating which entries are valid. If you do need to allocate
2163 your C<BHK> dynamically for some reason, be sure to zero it before you
2166 Once registered, there is no mechanism to switch these hooks off, so if
2167 that is necessary you will need to do this yourself. An entry in C<%^H>
2168 is probably the best way, so the effect is lexically scoped; however it
2169 is also possible to use the C<BhkDISABLE> and C<BhkENABLE> macros to
2170 temporarily switch entries on and off. You should also be aware that
2171 generally speaking at least one scope will have opened before your
2172 extension is loaded, so you will see some C<pre/post_end> pairs that
2173 didn't have a matching C<start>.
2175 =head1 Examining internal data structures with the C<dump> functions
2177 To aid debugging, the source file F<dump.c> contains a number of
2178 functions which produce formatted output of internal data structures.
2180 The most commonly used of these functions is C<Perl_sv_dump>; it's used
2181 for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
2182 C<sv_dump> to produce debugging output from Perl-space, so users of that
2183 module should already be familiar with its format.
2185 C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
2186 derivatives, and produces output similar to C<perl -Dx>; in fact,
2187 C<Perl_dump_eval> will dump the main root of the code being evaluated,
2188 exactly like C<-Dx>.
2190 Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
2191 op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
2192 subroutines in a package like so: (Thankfully, these are all xsubs, so
2193 there is no op tree)
2195 (gdb) print Perl_dump_packsubs(PL_defstash)
2197 SUB attributes::bootstrap = (xsub 0x811fedc 0)
2199 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
2201 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
2203 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
2205 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
2207 and C<Perl_dump_all>, which dumps all the subroutines in the stash and
2208 the op tree of the main root.
2210 =head1 How multiple interpreters and concurrency are supported
2212 =head2 Background and PERL_IMPLICIT_CONTEXT
2214 The Perl interpreter can be regarded as a closed box: it has an API
2215 for feeding it code or otherwise making it do things, but it also has
2216 functions for its own use. This smells a lot like an object, and
2217 there are ways for you to build Perl so that you can have multiple
2218 interpreters, with one interpreter represented either as a C structure,
2219 or inside a thread-specific structure. These structures contain all
2220 the context, the state of that interpreter.
2222 One macro controls the major Perl build flavor: MULTIPLICITY. The
2223 MULTIPLICITY build has a C structure that packages all the interpreter
2224 state. With multiplicity-enabled perls, PERL_IMPLICIT_CONTEXT is also
2225 normally defined, and enables the support for passing in a "hidden" first
2226 argument that represents all three data structures. MULTIPLICITY makes
2227 multi-threaded perls possible (with the ithreads threading model, related
2228 to the macro USE_ITHREADS.)
2230 Two other "encapsulation" macros are the PERL_GLOBAL_STRUCT and
2231 PERL_GLOBAL_STRUCT_PRIVATE (the latter turns on the former, and the
2232 former turns on MULTIPLICITY.) The PERL_GLOBAL_STRUCT causes all the
2233 internal variables of Perl to be wrapped inside a single global struct,
2234 struct perl_vars, accessible as (globals) &PL_Vars or PL_VarsPtr or
2235 the function Perl_GetVars(). The PERL_GLOBAL_STRUCT_PRIVATE goes
2236 one step further, there is still a single struct (allocated in main()
2237 either from heap or from stack) but there are no global data symbols
2238 pointing to it. In either case the global struct should be initialized
2239 as the very first thing in main() using Perl_init_global_struct() and
2240 correspondingly tear it down after perl_free() using Perl_free_global_struct(),
2241 please see F<miniperlmain.c> for usage details. You may also need
2242 to use C<dVAR> in your coding to "declare the global variables"
2243 when you are using them. dTHX does this for you automatically.
2245 To see whether you have non-const data you can use a BSD (or GNU)
2248 nm libperl.a | grep -v ' [TURtr] '
2250 If this displays any C<D> or C<d> symbols (or possibly C<C> or C<c>),
2251 you have non-const data. The symbols the C<grep> removed are as follows:
2252 C<Tt> are I<text>, or code, the C<Rr> are I<read-only> (const) data,
2253 and the C<U> is <undefined>, external symbols referred to.
2255 The test F<t/porting/libperl.t> does this kind of symbol sanity
2256 checking on C<libperl.a>.
2258 For backward compatibility reasons defining just PERL_GLOBAL_STRUCT
2259 doesn't actually hide all symbols inside a big global struct: some
2260 PerlIO_xxx vtables are left visible. The PERL_GLOBAL_STRUCT_PRIVATE
2261 then hides everything (see how the PERLIO_FUNCS_DECL is used).
2263 All this obviously requires a way for the Perl internal functions to be
2264 either subroutines taking some kind of structure as the first
2265 argument, or subroutines taking nothing as the first argument. To
2266 enable these two very different ways of building the interpreter,
2267 the Perl source (as it does in so many other situations) makes heavy
2268 use of macros and subroutine naming conventions.
2270 First problem: deciding which functions will be public API functions and
2271 which will be private. All functions whose names begin C<S_> are private
2272 (think "S" for "secret" or "static"). All other functions begin with
2273 "Perl_", but just because a function begins with "Perl_" does not mean it is
2274 part of the API. (See L</Internal
2275 Functions>.) The easiest way to be B<sure> a
2276 function is part of the API is to find its entry in L<perlapi>.
2277 If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
2278 think it should be (i.e., you need it for your extension), send mail via
2279 L<perlbug> explaining why you think it should be.
2281 Second problem: there must be a syntax so that the same subroutine
2282 declarations and calls can pass a structure as their first argument,
2283 or pass nothing. To solve this, the subroutines are named and
2284 declared in a particular way. Here's a typical start of a static
2285 function used within the Perl guts:
2288 S_incline(pTHX_ char *s)
2290 STATIC becomes "static" in C, and may be #define'd to nothing in some
2291 configurations in the future.
2293 A public function (i.e. part of the internal API, but not necessarily
2294 sanctioned for use in extensions) begins like this:
2297 Perl_sv_setiv(pTHX_ SV* dsv, IV num)
2299 C<pTHX_> is one of a number of macros (in F<perl.h>) that hide the
2300 details of the interpreter's context. THX stands for "thread", "this",
2301 or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
2302 The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
2303 or 'd' for B<d>eclaration, so we have C<pTHX>, C<aTHX> and C<dTHX>, and
2306 When Perl is built without options that set PERL_IMPLICIT_CONTEXT, there is no
2307 first argument containing the interpreter's context. The trailing underscore
2308 in the pTHX_ macro indicates that the macro expansion needs a comma
2309 after the context argument because other arguments follow it. If
2310 PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
2311 subroutine is not prototyped to take the extra argument. The form of the
2312 macro without the trailing underscore is used when there are no additional
2315 When a core function calls another, it must pass the context. This
2316 is normally hidden via macros. Consider C<sv_setiv>. It expands into
2317 something like this:
2319 #ifdef PERL_IMPLICIT_CONTEXT
2320 #define sv_setiv(a,b) Perl_sv_setiv(aTHX_ a, b)
2321 /* can't do this for vararg functions, see below */
2323 #define sv_setiv Perl_sv_setiv
2326 This works well, and means that XS authors can gleefully write:
2330 and still have it work under all the modes Perl could have been
2333 This doesn't work so cleanly for varargs functions, though, as macros
2334 imply that the number of arguments is known in advance. Instead we
2335 either need to spell them out fully, passing C<aTHX_> as the first
2336 argument (the Perl core tends to do this with functions like
2337 Perl_warner), or use a context-free version.
2339 The context-free version of Perl_warner is called
2340 Perl_warner_nocontext, and does not take the extra argument. Instead
2341 it does dTHX; to get the context from thread-local storage. We
2342 C<#define warner Perl_warner_nocontext> so that extensions get source
2343 compatibility at the expense of performance. (Passing an arg is
2344 cheaper than grabbing it from thread-local storage.)
2346 You can ignore [pad]THXx when browsing the Perl headers/sources.
2347 Those are strictly for use within the core. Extensions and embedders
2348 need only be aware of [pad]THX.
2350 =head2 So what happened to dTHR?
2352 C<dTHR> was introduced in perl 5.005 to support the older thread model.
2353 The older thread model now uses the C<THX> mechanism to pass context
2354 pointers around, so C<dTHR> is not useful any more. Perl 5.6.0 and
2355 later still have it for backward source compatibility, but it is defined
2358 =head2 How do I use all this in extensions?
2360 When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
2361 any functions in the Perl API will need to pass the initial context
2362 argument somehow. The kicker is that you will need to write it in
2363 such a way that the extension still compiles when Perl hasn't been
2364 built with PERL_IMPLICIT_CONTEXT enabled.
2366 There are three ways to do this. First, the easy but inefficient way,
2367 which is also the default, in order to maintain source compatibility
2368 with extensions: whenever F<XSUB.h> is #included, it redefines the aTHX
2369 and aTHX_ macros to call a function that will return the context.
2370 Thus, something like:
2374 in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
2377 Perl_sv_setiv(Perl_get_context(), sv, num);
2379 or to this otherwise:
2381 Perl_sv_setiv(sv, num);
2383 You don't have to do anything new in your extension to get this; since
2384 the Perl library provides Perl_get_context(), it will all just
2387 The second, more efficient way is to use the following template for
2390 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2395 STATIC void my_private_function(int arg1, int arg2);
2398 my_private_function(int arg1, int arg2)
2400 dTHX; /* fetch context */
2401 ... call many Perl API functions ...
2406 MODULE = Foo PACKAGE = Foo
2414 my_private_function(arg, 10);
2416 Note that the only two changes from the normal way of writing an
2417 extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
2418 including the Perl headers, followed by a C<dTHX;> declaration at
2419 the start of every function that will call the Perl API. (You'll
2420 know which functions need this, because the C compiler will complain
2421 that there's an undeclared identifier in those functions.) No changes
2422 are needed for the XSUBs themselves, because the XS() macro is
2423 correctly defined to pass in the implicit context if needed.
2425 The third, even more efficient way is to ape how it is done within
2429 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2434 /* pTHX_ only needed for functions that call Perl API */
2435 STATIC void my_private_function(pTHX_ int arg1, int arg2);
2438 my_private_function(pTHX_ int arg1, int arg2)
2440 /* dTHX; not needed here, because THX is an argument */
2441 ... call Perl API functions ...
2446 MODULE = Foo PACKAGE = Foo
2454 my_private_function(aTHX_ arg, 10);
2456 This implementation never has to fetch the context using a function
2457 call, since it is always passed as an extra argument. Depending on
2458 your needs for simplicity or efficiency, you may mix the previous
2459 two approaches freely.
2461 Never add a comma after C<pTHX> yourself--always use the form of the
2462 macro with the underscore for functions that take explicit arguments,
2463 or the form without the argument for functions with no explicit arguments.
2465 If one is compiling Perl with the C<-DPERL_GLOBAL_STRUCT> the C<dVAR>
2466 definition is needed if the Perl global variables (see F<perlvars.h>
2467 or F<globvar.sym>) are accessed in the function and C<dTHX> is not
2468 used (the C<dTHX> includes the C<dVAR> if necessary). One notices
2469 the need for C<dVAR> only with the said compile-time define, because
2470 otherwise the Perl global variables are visible as-is.
2472 =head2 Should I do anything special if I call perl from multiple threads?
2474 If you create interpreters in one thread and then proceed to call them in
2475 another, you need to make sure perl's own Thread Local Storage (TLS) slot is
2476 initialized correctly in each of those threads.
2478 The C<perl_alloc> and C<perl_clone> API functions will automatically set
2479 the TLS slot to the interpreter they created, so that there is no need to do
2480 anything special if the interpreter is always accessed in the same thread that
2481 created it, and that thread did not create or call any other interpreters
2482 afterwards. If that is not the case, you have to set the TLS slot of the
2483 thread before calling any functions in the Perl API on that particular
2484 interpreter. This is done by calling the C<PERL_SET_CONTEXT> macro in that
2485 thread as the first thing you do:
2487 /* do this before doing anything else with some_perl */
2488 PERL_SET_CONTEXT(some_perl);
2490 ... other Perl API calls on some_perl go here ...
2492 =head2 Future Plans and PERL_IMPLICIT_SYS
2494 Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
2495 that the interpreter knows about itself and pass it around, so too are
2496 there plans to allow the interpreter to bundle up everything it knows
2497 about the environment it's running on. This is enabled with the
2498 PERL_IMPLICIT_SYS macro. Currently it only works with USE_ITHREADS on
2501 This allows the ability to provide an extra pointer (called the "host"
2502 environment) for all the system calls. This makes it possible for
2503 all the system stuff to maintain their own state, broken down into
2504 seven C structures. These are thin wrappers around the usual system
2505 calls (see F<win32/perllib.c>) for the default perl executable, but for a
2506 more ambitious host (like the one that would do fork() emulation) all
2507 the extra work needed to pretend that different interpreters are
2508 actually different "processes", would be done here.
2510 The Perl engine/interpreter and the host are orthogonal entities.
2511 There could be one or more interpreters in a process, and one or
2512 more "hosts", with free association between them.
2514 =head1 Internal Functions
2516 All of Perl's internal functions which will be exposed to the outside
2517 world are prefixed by C<Perl_> so that they will not conflict with XS
2518 functions or functions used in a program in which Perl is embedded.
2519 Similarly, all global variables begin with C<PL_>. (By convention,
2520 static functions start with C<S_>.)
2522 Inside the Perl core (C<PERL_CORE> defined), you can get at the functions
2523 either with or without the C<Perl_> prefix, thanks to a bunch of defines
2524 that live in F<embed.h>. Note that extension code should I<not> set
2525 C<PERL_CORE>; this exposes the full perl internals, and is likely to cause
2526 breakage of the XS in each new perl release.
2528 The file F<embed.h> is generated automatically from
2529 F<embed.pl> and F<embed.fnc>. F<embed.pl> also creates the prototyping
2530 header files for the internal functions, generates the documentation
2531 and a lot of other bits and pieces. It's important that when you add
2532 a new function to the core or change an existing one, you change the
2533 data in the table in F<embed.fnc> as well. Here's a sample entry from
2536 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
2538 The second column is the return type, the third column the name. Columns
2539 after that are the arguments. The first column is a set of flags:
2545 This function is a part of the public
2546 API. All such functions should also
2547 have 'd', very few do not.
2551 This function has a C<Perl_> prefix; i.e. it is defined as
2556 This function has documentation using the C<apidoc> feature which we'll
2557 look at in a second. Some functions have 'd' but not 'A'; docs are good.
2561 Other available flags are:
2567 This is a static function and is defined as C<STATIC S_whatever>, and
2568 usually called within the sources as C<whatever(...)>.
2572 This does not need an interpreter context, so the definition has no
2573 C<pTHX>, and it follows that callers don't use C<aTHX>. (See
2574 L</Background and PERL_IMPLICIT_CONTEXT>.)
2578 This function never returns; C<croak>, C<exit> and friends.
2582 This function takes a variable number of arguments, C<printf> style.
2583 The argument list should end with C<...>, like this:
2585 Afprd |void |croak |const char* pat|...
2589 This function is part of the experimental development API, and may change
2590 or disappear without notice.
2594 This function should not have a compatibility macro to define, say,
2595 C<Perl_parse> to C<parse>. It must be called as C<Perl_parse>.
2599 This function isn't exported out of the Perl core.
2603 This is implemented as a macro.
2607 This function is explicitly exported.
2611 This function is visible to extensions included in the Perl core.
2615 Binary backward compatibility; this function is a macro but also has
2616 a C<Perl_> implementation (which is exported).
2620 See the comments at the top of C<embed.fnc> for others.
2624 If you edit F<embed.pl> or F<embed.fnc>, you will need to run
2625 C<make regen_headers> to force a rebuild of F<embed.h> and other
2626 auto-generated files.
2628 =head2 Formatted Printing of IVs, UVs, and NVs
2630 If you are printing IVs, UVs, or NVS instead of the stdio(3) style
2631 formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
2632 following macros for portability
2637 UVxf UV in hexadecimal
2642 These will take care of 64-bit integers and long doubles.
2645 printf("IV is %"IVdf"\n", iv);
2647 The IVdf will expand to whatever is the correct format for the IVs.
2649 Note that there are different "long doubles": Perl will use
2650 whatever the compiler has.
2652 If you are printing addresses of pointers, use UVxf combined
2653 with PTR2UV(), do not use %lx or %p.
2655 =head2 Pointer-To-Integer and Integer-To-Pointer
2657 Because pointer size does not necessarily equal integer size,
2658 use the follow macros to do it right.
2663 INT2PTR(pointertotype, integer)
2668 SV *sv = INT2PTR(SV*, iv);
2675 =head2 Exception Handling
2677 There are a couple of macros to do very basic exception handling in XS
2678 modules. You have to define C<NO_XSLOCKS> before including F<XSUB.h> to
2679 be able to use these macros:
2684 You can use these macros if you call code that may croak, but you need
2685 to do some cleanup before giving control back to Perl. For example:
2687 dXCPT; /* set up necessary variables */
2690 code_that_may_croak();
2695 /* do cleanup here */
2699 Note that you always have to rethrow an exception that has been
2700 caught. Using these macros, it is not possible to just catch the
2701 exception and ignore it. If you have to ignore the exception, you
2702 have to use the C<call_*> function.
2704 The advantage of using the above macros is that you don't have
2705 to setup an extra function for C<call_*>, and that using these
2706 macros is faster than using C<call_*>.
2708 =head2 Source Documentation
2710 There's an effort going on to document the internal functions and
2711 automatically produce reference manuals from them - L<perlapi> is one
2712 such manual which details all the functions which are available to XS
2713 writers. L<perlintern> is the autogenerated manual for the functions
2714 which are not part of the API and are supposedly for internal use only.
2716 Source documentation is created by putting POD comments into the C
2720 =for apidoc sv_setiv
2722 Copies an integer into the given SV. Does not handle 'set' magic. See
2728 Please try and supply some documentation if you add functions to the
2731 =head2 Backwards compatibility
2733 The Perl API changes over time. New functions are
2734 added or the interfaces of existing functions are
2735 changed. The C<Devel::PPPort> module tries to
2736 provide compatibility code for some of these changes, so XS writers don't
2737 have to code it themselves when supporting multiple versions of Perl.
2739 C<Devel::PPPort> generates a C header file F<ppport.h> that can also
2740 be run as a Perl script. To generate F<ppport.h>, run:
2742 perl -MDevel::PPPort -eDevel::PPPort::WriteFile
2744 Besides checking existing XS code, the script can also be used to retrieve
2745 compatibility information for various API calls using the C<--api-info>
2746 command line switch. For example:
2748 % perl ppport.h --api-info=sv_magicext
2750 For details, see C<perldoc ppport.h>.
2752 =head1 Unicode Support
2754 Perl 5.6.0 introduced Unicode support. It's important for porters and XS
2755 writers to understand this support and make sure that the code they
2756 write does not corrupt Unicode data.
2758 =head2 What B<is> Unicode, anyway?
2760 In the olden, less enlightened times, we all used to use ASCII. Most of
2761 us did, anyway. The big problem with ASCII is that it's American. Well,
2762 no, that's not actually the problem; the problem is that it's not
2763 particularly useful for people who don't use the Roman alphabet. What
2764 used to happen was that particular languages would stick their own
2765 alphabet in the upper range of the sequence, between 128 and 255. Of
2766 course, we then ended up with plenty of variants that weren't quite
2767 ASCII, and the whole point of it being a standard was lost.
2769 Worse still, if you've got a language like Chinese or
2770 Japanese that has hundreds or thousands of characters, then you really
2771 can't fit them into a mere 256, so they had to forget about ASCII
2772 altogether, and build their own systems using pairs of numbers to refer
2775 To fix this, some people formed Unicode, Inc. and
2776 produced a new character set containing all the characters you can
2777 possibly think of and more. There are several ways of representing these
2778 characters, and the one Perl uses is called UTF-8. UTF-8 uses
2779 a variable number of bytes to represent a character. You can learn more
2780 about Unicode and Perl's Unicode model in L<perlunicode>.
2782 =head2 How can I recognise a UTF-8 string?
2784 You can't. This is because UTF-8 data is stored in bytes just like
2785 non-UTF-8 data. The Unicode character 200, (C<0xC8> for you hex types)
2786 capital E with a grave accent, is represented by the two bytes
2787 C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
2788 has that byte sequence as well. So you can't tell just by looking - this
2789 is what makes Unicode input an interesting problem.
2791 In general, you either have to know what you're dealing with, or you
2792 have to guess. The API function C<is_utf8_string> can help; it'll tell
2793 you if a string contains only valid UTF-8 characters. However, it can't
2794 do the work for you. On a character-by-character basis,
2796 will tell you whether the current character in a string is valid UTF-8.
2798 =head2 How does UTF-8 represent Unicode characters?
2800 As mentioned above, UTF-8 uses a variable number of bytes to store a
2801 character. Characters with values 0...127 are stored in one
2802 byte, just like good ol' ASCII. Character 128 is stored as
2803 C<v194.128>; this continues up to character 191, which is
2804 C<v194.191>. Now we've run out of bits (191 is binary
2805 C<10111111>) so we move on; 192 is C<v195.128>. And
2806 so it goes on, moving to three bytes at character 2048.
2808 Assuming you know you're dealing with a UTF-8 string, you can find out
2809 how long the first character in it is with the C<UTF8SKIP> macro:
2811 char *utf = "\305\233\340\240\201";
2814 len = UTF8SKIP(utf); /* len is 2 here */
2816 len = UTF8SKIP(utf); /* len is 3 here */
2818 Another way to skip over characters in a UTF-8 string is to use
2819 C<utf8_hop>, which takes a string and a number of characters to skip
2820 over. You're on your own about bounds checking, though, so don't use it
2823 All bytes in a multi-byte UTF-8 character will have the high bit set,
2824 so you can test if you need to do something special with this
2825 character like this (the UTF8_IS_INVARIANT() is a macro that tests
2826 whether the byte is encoded as a single byte even in UTF-8):
2829 U8 *utf_end; /* 1 beyond buffer pointed to by utf */
2830 UV uv; /* Note: a UV, not a U8, not a char */
2831 STRLEN len; /* length of character in bytes */
2833 if (!UTF8_IS_INVARIANT(*utf))
2834 /* Must treat this as UTF-8 */
2835 uv = utf8_to_uvchr_buf(utf, utf_end, &len);
2837 /* OK to treat this character as a byte */
2840 You can also see in that example that we use C<utf8_to_uvchr_buf> to get the
2841 value of the character; the inverse function C<uvchr_to_utf8> is available
2842 for putting a UV into UTF-8:
2844 if (!UTF8_IS_INVARIANT(uv))
2845 /* Must treat this as UTF8 */
2846 utf8 = uvchr_to_utf8(utf8, uv);
2848 /* OK to treat this character as a byte */
2851 You B<must> convert characters to UVs using the above functions if
2852 you're ever in a situation where you have to match UTF-8 and non-UTF-8
2853 characters. You may not skip over UTF-8 characters in this case. If you
2854 do this, you'll lose the ability to match hi-bit non-UTF-8 characters;
2855 for instance, if your UTF-8 string contains C<v196.172>, and you skip
2856 that character, you can never match a C<chr(200)> in a non-UTF-8 string.
2859 =head2 How does Perl store UTF-8 strings?
2861 Currently, Perl deals with Unicode strings and non-Unicode strings
2862 slightly differently. A flag in the SV, C<SVf_UTF8>, indicates that the
2863 string is internally encoded as UTF-8. Without it, the byte value is the
2864 codepoint number and vice versa (in other words, the string is encoded
2865 as iso-8859-1, but C<use feature 'unicode_strings'> is needed to get iso-8859-1
2866 semantics). This flag is only meaningful if the SV is C<SvPOK>
2867 or immediately after stringification via C<SvPV> or a similar
2868 macro. You can check and manipulate this flag with the
2875 This flag has an important effect on Perl's treatment of the string: if
2876 Unicode data is not properly distinguished, regular expressions,
2877 C<length>, C<substr> and other string handling operations will have
2878 undesirable results.
2880 The problem comes when you have, for instance, a string that isn't
2881 flagged as UTF-8, and contains a byte sequence that could be UTF-8 -
2882 especially when combining non-UTF-8 and UTF-8 strings.
2884 Never forget that the C<SVf_UTF8> flag is separate to the PV value; you
2885 need be sure you don't accidentally knock it off while you're
2886 manipulating SVs. More specifically, you cannot expect to do this:
2895 nsv = newSVpvn(p, len);
2897 The C<char*> string does not tell you the whole story, and you can't
2898 copy or reconstruct an SV just by copying the string value. Check if the
2899 old SV has the UTF8 flag set (I<after> the C<SvPV> call), and act
2904 nsv = newSVpvn(p, len);
2908 In fact, your C<frobnicate> function should be made aware of whether or
2909 not it's dealing with UTF-8 data, so that it can handle the string
2912 Since just passing an SV to an XS function and copying the data of
2913 the SV is not enough to copy the UTF8 flags, even less right is just
2914 passing a C<char *> to an XS function.
2916 =head2 How do I convert a string to UTF-8?
2918 If you're mixing UTF-8 and non-UTF-8 strings, it is necessary to upgrade
2919 one of the strings to UTF-8. If you've got an SV, the easiest way to do
2922 sv_utf8_upgrade(sv);
2924 However, you must not do this, for example:
2927 sv_utf8_upgrade(left);
2929 If you do this in a binary operator, you will actually change one of the
2930 strings that came into the operator, and, while it shouldn't be noticeable
2931 by the end user, it can cause problems in deficient code.
2933 Instead, C<bytes_to_utf8> will give you a UTF-8-encoded B<copy> of its
2934 string argument. This is useful for having the data available for
2935 comparisons and so on, without harming the original SV. There's also
2936 C<utf8_to_bytes> to go the other way, but naturally, this will fail if
2937 the string contains any characters above 255 that can't be represented
2940 =head2 Is there anything else I need to know?
2942 Not really. Just remember these things:
2948 There's no way to tell if a string is UTF-8 or not. You can tell if an SV
2949 is UTF-8 by looking at its C<SvUTF8> flag after stringifying it
2950 with C<SvPV> or a similar macro. Don't forget to set the flag if
2951 something should be UTF-8. Treat the flag as part of the PV, even though
2952 it's not - if you pass on the PV to somewhere, pass on the flag too.
2956 If a string is UTF-8, B<always> use C<utf8_to_uvchr_buf> to get at the value,
2957 unless C<UTF8_IS_INVARIANT(*s)> in which case you can use C<*s>.
2961 When writing a character C<uv> to a UTF-8 string, B<always> use
2962 C<uvchr_to_utf8>, unless C<UTF8_IS_INVARIANT(uv))> in which case
2963 you can use C<*s = uv>.
2967 Mixing UTF-8 and non-UTF-8 strings is
2968 tricky. Use C<bytes_to_utf8> to get
2969 a new string which is UTF-8 encoded, and then combine them.
2973 =head1 Custom Operators
2975 Custom operator support is an experimental feature that allows you to
2976 define your own ops. This is primarily to allow the building of
2977 interpreters for other languages in the Perl core, but it also allows
2978 optimizations through the creation of "macro-ops" (ops which perform the
2979 functions of multiple ops which are usually executed together, such as
2980 C<gvsv, gvsv, add>.)
2982 This feature is implemented as a new op type, C<OP_CUSTOM>. The Perl
2983 core does not "know" anything special about this op type, and so it will
2984 not be involved in any optimizations. This also means that you can
2985 define your custom ops to be any op structure - unary, binary, list and
2988 It's important to know what custom operators won't do for you. They
2989 won't let you add new syntax to Perl, directly. They won't even let you
2990 add new keywords, directly. In fact, they won't change the way Perl
2991 compiles a program at all. You have to do those changes yourself, after
2992 Perl has compiled the program. You do this either by manipulating the op
2993 tree using a C<CHECK> block and the C<B::Generate> module, or by adding
2994 a custom peephole optimizer with the C<optimize> module.
2996 When you do this, you replace ordinary Perl ops with custom ops by
2997 creating ops with the type C<OP_CUSTOM> and the C<op_ppaddr> of your own
2998 PP function. This should be defined in XS code, and should look like
2999 the PP ops in C<pp_*.c>. You are responsible for ensuring that your op
3000 takes the appropriate number of values from the stack, and you are
3001 responsible for adding stack marks if necessary.
3003 You should also "register" your op with the Perl interpreter so that it
3004 can produce sensible error and warning messages. Since it is possible to
3005 have multiple custom ops within the one "logical" op type C<OP_CUSTOM>,
3006 Perl uses the value of C<< o->op_ppaddr >> to determine which custom op
3007 it is dealing with. You should create an C<XOP> structure for each
3008 ppaddr you use, set the properties of the custom op with
3009 C<XopENTRY_set>, and register the structure against the ppaddr using
3010 C<Perl_custom_op_register>. A trivial example might look like:
3013 static OP *my_pp(pTHX);
3016 XopENTRY_set(&my_xop, xop_name, "myxop");
3017 XopENTRY_set(&my_xop, xop_desc, "Useless custom op");
3018 Perl_custom_op_register(aTHX_ my_pp, &my_xop);
3020 The available fields in the structure are:
3026 A short name for your op. This will be included in some error messages,
3027 and will also be returned as C<< $op->name >> by the L<B|B> module, so
3028 it will appear in the output of module like L<B::Concise|B::Concise>.
3032 A short description of the function of the op.
3036 Which of the various C<*OP> structures this op uses. This should be one of
3037 the C<OA_*> constants from F<op.h>, namely
3057 =item OA_PVOP_OR_SVOP
3059 This should be interpreted as 'C<PVOP>' only. The C<_OR_SVOP> is because
3060 the only core C<PVOP>, C<OP_TRANS>, can sometimes be a C<SVOP> instead.
3068 The other C<OA_*> constants should not be used.
3072 This member is of type C<Perl_cpeep_t>, which expands to C<void
3073 (*Perl_cpeep_t)(aTHX_ OP *o, OP *oldop)>. If it is set, this function
3074 will be called from C<Perl_rpeep> when ops of this type are encountered
3075 by the peephole optimizer. I<o> is the OP that needs optimizing;
3076 I<oldop> is the previous OP optimized, whose C<op_next> points to I<o>.
3080 C<B::Generate> directly supports the creation of custom ops by name.
3084 Until May 1997, this document was maintained by Jeff Okamoto
3085 E<lt>okamoto@corp.hp.comE<gt>. It is now maintained as part of Perl
3086 itself by the Perl 5 Porters E<lt>perl5-porters@perl.orgE<gt>.
3088 With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
3089 Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
3090 Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
3091 Stephen McCamant, and Gurusamy Sarathy.
3095 L<perlapi>, L<perlintern>, L<perlxs>, L<perlembed>