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 (C<Size_t>, usually defined as C<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 *,
82 SV **, Size_t, bool *);
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
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. Normally copy-on-write will prevent
327 the substitution from operator from using this hack, but if you can craft a
328 string for which copy-on-write is not possible, you can see it in play. In
329 the current implementation, the final byte of a string buffer is used as a
330 copy-on-write reference count. If the buffer is not big enough, then
331 copy-on-write is skipped. First have a look at an empty string:
333 % ./perl -Ilib -MDevel::Peek -le '$a=""; $a .= ""; Dump $a'
334 SV = PV(0x7ffb7c008a70) at 0x7ffb7c030390
337 PV = 0x7ffb7bc05b50 ""\0
341 Notice here the LEN is 10. (It may differ on your platform.) Extend the
342 length of the string to one less than 10, and do a substitution:
344 % ./perl -Ilib -MDevel::Peek -le '$a=""; $a.="123456789"; $a=~s/.//; \
346 SV = PV(0x7ffa04008a70) at 0x7ffa04030390
348 FLAGS = (POK,OOK,pPOK)
350 PV = 0x7ffa03c05b61 ( "\1" . ) "23456789"\0
354 Here the number of bytes chopped off (1) is shown next as the OFFSET. The
355 portion of the string between the "real" and the "fake" beginnings is
356 shown in parentheses, and the values of C<SvCUR> and C<SvLEN> reflect
357 the fake beginning, not the real one. (The first character of the string
358 buffer happens to have changed to "\1" here, not "1", because the current
359 implementation stores the offset count in the string buffer. This is
362 Something similar to the offset hack is performed on AVs to enable
363 efficient shifting and splicing off the beginning of the array; while
364 C<AvARRAY> points to the first element in the array that is visible from
365 Perl, C<AvALLOC> points to the real start of the C array. These are
366 usually the same, but a C<shift> operation can be carried out by
367 increasing C<AvARRAY> by one and decreasing C<AvFILL> and C<AvMAX>.
368 Again, the location of the real start of the C array only comes into
369 play when freeing the array. See C<av_shift> in F<av.c>.
371 =head2 What's Really Stored in an SV?
373 Recall that the usual method of determining the type of scalar you have is
374 to use C<Sv*OK> macros. Because a scalar can be both a number and a string,
375 usually these macros will always return TRUE and calling the C<Sv*V>
376 macros will do the appropriate conversion of string to integer/double or
377 integer/double to string.
379 If you I<really> need to know if you have an integer, double, or string
380 pointer in an SV, you can use the following three macros instead:
386 These will tell you if you truly have an integer, double, or string pointer
387 stored in your SV. The "p" stands for private.
389 There are various ways in which the private and public flags may differ.
390 For example, in perl 5.16 and earlier a tied SV may have a valid
391 underlying value in the IV slot (so SvIOKp is true), but the data
392 should be accessed via the FETCH routine rather than directly,
393 so SvIOK is false. (In perl 5.18 onwards, tied scalars use
394 the flags the same way as untied scalars.) Another is when
395 numeric conversion has occurred and precision has been lost: only the
396 private flag is set on 'lossy' values. So when an NV is converted to an
397 IV with loss, SvIOKp, SvNOKp and SvNOK will be set, while SvIOK wont be.
399 In general, though, it's best to use the C<Sv*V> macros.
401 =head2 Working with AVs
403 There are two ways to create and load an AV. The first method creates an
408 The second method both creates the AV and initially populates it with SVs:
410 AV* av_make(SSize_t num, SV **ptr);
412 The second argument points to an array containing C<num> C<SV*>'s. Once the
413 AV has been created, the SVs can be destroyed, if so desired.
415 Once the AV has been created, the following operations are possible on it:
417 void av_push(AV*, SV*);
420 void av_unshift(AV*, SSize_t num);
422 These should be familiar operations, with the exception of C<av_unshift>.
423 This routine adds C<num> elements at the front of the array with the C<undef>
424 value. You must then use C<av_store> (described below) to assign values
425 to these new elements.
427 Here are some other functions:
429 SSize_t av_top_index(AV*);
430 SV** av_fetch(AV*, SSize_t key, I32 lval);
431 SV** av_store(AV*, SSize_t key, SV* val);
433 The C<av_top_index> function returns the highest index value in an array (just
434 like $#array in Perl). If the array is empty, -1 is returned. The
435 C<av_fetch> function returns the value at index C<key>, but if C<lval>
436 is non-zero, then C<av_fetch> will store an undef value at that index.
437 The C<av_store> function stores the value C<val> at index C<key>, and does
438 not increment the reference count of C<val>. Thus the caller is responsible
439 for taking care of that, and if C<av_store> returns NULL, the caller will
440 have to decrement the reference count to avoid a memory leak. Note that
441 C<av_fetch> and C<av_store> both return C<SV**>'s, not C<SV*>'s as their
448 void av_extend(AV*, SSize_t key);
450 The C<av_clear> function deletes all the elements in the AV* array, but
451 does not actually delete the array itself. The C<av_undef> function will
452 delete all the elements in the array plus the array itself. The
453 C<av_extend> function extends the array so that it contains at least C<key+1>
454 elements. If C<key+1> is less than the currently allocated length of the array,
455 then nothing is done.
457 If you know the name of an array variable, you can get a pointer to its AV
458 by using the following:
460 AV* get_av("package::varname", 0);
462 This returns NULL if the variable does not exist.
464 See L</Understanding the Magic of Tied Hashes and Arrays> for more
465 information on how to use the array access functions on tied arrays.
467 =head2 Working with HVs
469 To create an HV, you use the following routine:
473 Once the HV has been created, the following operations are possible on it:
475 SV** hv_store(HV*, const char* key, U32 klen, SV* val, U32 hash);
476 SV** hv_fetch(HV*, const char* key, U32 klen, I32 lval);
478 The C<klen> parameter is the length of the key being passed in (Note that
479 you cannot pass 0 in as a value of C<klen> to tell Perl to measure the
480 length of the key). The C<val> argument contains the SV pointer to the
481 scalar being stored, and C<hash> is the precomputed hash value (zero if
482 you want C<hv_store> to calculate it for you). The C<lval> parameter
483 indicates whether this fetch is actually a part of a store operation, in
484 which case a new undefined value will be added to the HV with the supplied
485 key and C<hv_fetch> will return as if the value had already existed.
487 Remember that C<hv_store> and C<hv_fetch> return C<SV**>'s and not just
488 C<SV*>. To access the scalar value, you must first dereference the return
489 value. However, you should check to make sure that the return value is
490 not NULL before dereferencing it.
492 The first of these two functions checks if a hash table entry exists, and the
495 bool hv_exists(HV*, const char* key, U32 klen);
496 SV* hv_delete(HV*, const char* key, U32 klen, I32 flags);
498 If C<flags> does not include the C<G_DISCARD> flag then C<hv_delete> will
499 create and return a mortal copy of the deleted value.
501 And more miscellaneous functions:
506 Like their AV counterparts, C<hv_clear> deletes all the entries in the hash
507 table but does not actually delete the hash table. The C<hv_undef> deletes
508 both the entries and the hash table itself.
510 Perl keeps the actual data in a linked list of structures with a typedef of HE.
511 These contain the actual key and value pointers (plus extra administrative
512 overhead). The key is a string pointer; the value is an C<SV*>. However,
513 once you have an C<HE*>, to get the actual key and value, use the routines
516 I32 hv_iterinit(HV*);
517 /* Prepares starting point to traverse hash table */
518 HE* hv_iternext(HV*);
519 /* Get the next entry, and return a pointer to a
520 structure that has both the key and value */
521 char* hv_iterkey(HE* entry, I32* retlen);
522 /* Get the key from an HE structure and also return
523 the length of the key string */
524 SV* hv_iterval(HV*, HE* entry);
525 /* Return an SV pointer to the value of the HE
527 SV* hv_iternextsv(HV*, char** key, I32* retlen);
528 /* This convenience routine combines hv_iternext,
529 hv_iterkey, and hv_iterval. The key and retlen
530 arguments are return values for the key and its
531 length. The value is returned in the SV* argument */
533 If you know the name of a hash variable, you can get a pointer to its HV
534 by using the following:
536 HV* get_hv("package::varname", 0);
538 This returns NULL if the variable does not exist.
540 The hash algorithm is defined in the C<PERL_HASH> macro:
542 PERL_HASH(hash, key, klen)
544 The exact implementation of this macro varies by architecture and version
545 of perl, and the return value may change per invocation, so the value
546 is only valid for the duration of a single perl process.
548 See L</Understanding the Magic of Tied Hashes and Arrays> for more
549 information on how to use the hash access functions on tied hashes.
551 =head2 Hash API Extensions
553 Beginning with version 5.004, the following functions are also supported:
555 HE* hv_fetch_ent (HV* tb, SV* key, I32 lval, U32 hash);
556 HE* hv_store_ent (HV* tb, SV* key, SV* val, U32 hash);
558 bool hv_exists_ent (HV* tb, SV* key, U32 hash);
559 SV* hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
561 SV* hv_iterkeysv (HE* entry);
563 Note that these functions take C<SV*> keys, which simplifies writing
564 of extension code that deals with hash structures. These functions
565 also allow passing of C<SV*> keys to C<tie> functions without forcing
566 you to stringify the keys (unlike the previous set of functions).
568 They also return and accept whole hash entries (C<HE*>), making their
569 use more efficient (since the hash number for a particular string
570 doesn't have to be recomputed every time). See L<perlapi> for detailed
573 The following macros must always be used to access the contents of hash
574 entries. Note that the arguments to these macros must be simple
575 variables, since they may get evaluated more than once. See
576 L<perlapi> for detailed descriptions of these macros.
578 HePV(HE* he, STRLEN len)
582 HeSVKEY_force(HE* he)
583 HeSVKEY_set(HE* he, SV* sv)
585 These two lower level macros are defined, but must only be used when
586 dealing with keys that are not C<SV*>s:
591 Note that both C<hv_store> and C<hv_store_ent> do not increment the
592 reference count of the stored C<val>, which is the caller's responsibility.
593 If these functions return a NULL value, the caller will usually have to
594 decrement the reference count of C<val> to avoid a memory leak.
596 =head2 AVs, HVs and undefined values
598 Sometimes you have to store undefined values in AVs or HVs. Although
599 this may be a rare case, it can be tricky. That's because you're
600 used to using C<&PL_sv_undef> if you need an undefined SV.
602 For example, intuition tells you that this XS code:
605 av_store( av, 0, &PL_sv_undef );
607 is equivalent to this Perl code:
612 Unfortunately, this isn't true. In perl 5.18 and earlier, AVs use C<&PL_sv_undef> as a marker
613 for indicating that an array element has not yet been initialized.
614 Thus, C<exists $av[0]> would be true for the above Perl code, but
615 false for the array generated by the XS code. In perl 5.20, storing
616 &PL_sv_undef will create a read-only element, because the scalar
617 &PL_sv_undef itself is stored, not a copy.
619 Similar problems can occur when storing C<&PL_sv_undef> in HVs:
621 hv_store( hv, "key", 3, &PL_sv_undef, 0 );
623 This will indeed make the value C<undef>, but if you try to modify
624 the value of C<key>, you'll get the following error:
626 Modification of non-creatable hash value attempted
628 In perl 5.8.0, C<&PL_sv_undef> was also used to mark placeholders
629 in restricted hashes. This caused such hash entries not to appear
630 when iterating over the hash or when checking for the keys
631 with the C<hv_exists> function.
633 You can run into similar problems when you store C<&PL_sv_yes> or
634 C<&PL_sv_no> into AVs or HVs. Trying to modify such elements
635 will give you the following error:
637 Modification of a read-only value attempted
639 To make a long story short, you can use the special variables
640 C<&PL_sv_undef>, C<&PL_sv_yes> and C<&PL_sv_no> with AVs and
641 HVs, but you have to make sure you know what you're doing.
643 Generally, if you want to store an undefined value in an AV
644 or HV, you should not use C<&PL_sv_undef>, but rather create a
645 new undefined value using the C<newSV> function, for example:
647 av_store( av, 42, newSV(0) );
648 hv_store( hv, "foo", 3, newSV(0), 0 );
652 References are a special type of scalar that point to other data types
653 (including other references).
655 To create a reference, use either of the following functions:
657 SV* newRV_inc((SV*) thing);
658 SV* newRV_noinc((SV*) thing);
660 The C<thing> argument can be any of an C<SV*>, C<AV*>, or C<HV*>. The
661 functions are identical except that C<newRV_inc> increments the reference
662 count of the C<thing>, while C<newRV_noinc> does not. For historical
663 reasons, C<newRV> is a synonym for C<newRV_inc>.
665 Once you have a reference, you can use the following macro to dereference
670 then call the appropriate routines, casting the returned C<SV*> to either an
671 C<AV*> or C<HV*>, if required.
673 To determine if an SV is a reference, you can use the following macro:
677 To discover what type of value the reference refers to, use the following
678 macro and then check the return value.
682 The most useful types that will be returned are:
687 SVt_PVGV Glob (possibly a file handle)
689 Any numerical value returned which is less than SVt_PVAV will be a scalar
692 See L<perlapi/svtype> for more details.
694 =head2 Blessed References and Class Objects
696 References are also used to support object-oriented programming. In perl's
697 OO lexicon, an object is simply a reference that has been blessed into a
698 package (or class). Once blessed, the programmer may now use the reference
699 to access the various methods in the class.
701 A reference can be blessed into a package with the following function:
703 SV* sv_bless(SV* sv, HV* stash);
705 The C<sv> argument must be a reference value. The C<stash> argument
706 specifies which class the reference will belong to. See
707 L</Stashes and Globs> for information on converting class names into stashes.
709 /* Still under construction */
711 The following function upgrades rv to reference if not already one.
712 Creates a new SV for rv to point to. If C<classname> is non-null, the SV
713 is blessed into the specified class. SV is returned.
715 SV* newSVrv(SV* rv, const char* classname);
717 The following three functions copy integer, unsigned integer or double
718 into an SV whose reference is C<rv>. SV is blessed if C<classname> is
721 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
722 SV* sv_setref_uv(SV* rv, const char* classname, UV uv);
723 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
725 The following function copies the pointer value (I<the address, not the
726 string!>) into an SV whose reference is rv. SV is blessed if C<classname>
729 SV* sv_setref_pv(SV* rv, const char* classname, void* pv);
731 The following function copies a string into an SV whose reference is C<rv>.
732 Set length to 0 to let Perl calculate the string length. SV is blessed if
733 C<classname> is non-null.
735 SV* sv_setref_pvn(SV* rv, const char* classname, char* pv,
738 The following function tests whether the SV is blessed into the specified
739 class. It does not check inheritance relationships.
741 int sv_isa(SV* sv, const char* name);
743 The following function tests whether the SV is a reference to a blessed object.
745 int sv_isobject(SV* sv);
747 The following function tests whether the SV is derived from the specified
748 class. SV can be either a reference to a blessed object or a string
749 containing a class name. This is the function implementing the
750 C<UNIVERSAL::isa> functionality.
752 bool sv_derived_from(SV* sv, const char* name);
754 To check if you've got an object derived from a specific class you have
757 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
759 =head2 Creating New Variables
761 To create a new Perl variable with an undef value which can be accessed from
762 your Perl script, use the following routines, depending on the variable type.
764 SV* get_sv("package::varname", GV_ADD);
765 AV* get_av("package::varname", GV_ADD);
766 HV* get_hv("package::varname", GV_ADD);
768 Notice the use of GV_ADD as the second parameter. The new variable can now
769 be set, using the routines appropriate to the data type.
771 There are additional macros whose values may be bitwise OR'ed with the
772 C<GV_ADD> argument to enable certain extra features. Those bits are:
778 Marks the variable as multiply defined, thus preventing the:
780 Name <varname> used only once: possible typo
788 Had to create <varname> unexpectedly
790 if the variable did not exist before the function was called.
794 If you do not specify a package name, the variable is created in the current
797 =head2 Reference Counts and Mortality
799 Perl uses a reference count-driven garbage collection mechanism. SVs,
800 AVs, or HVs (xV for short in the following) start their life with a
801 reference count of 1. If the reference count of an xV ever drops to 0,
802 then it will be destroyed and its memory made available for reuse.
803 At the most basic internal level, reference counts can be manipulated
804 with the following macros:
806 int SvREFCNT(SV* sv);
807 SV* SvREFCNT_inc(SV* sv);
808 void SvREFCNT_dec(SV* sv);
810 (There are also suffixed versions of the increment and decrement macros,
811 for situations where the full generality of these basic macros can be
812 exchanged for some performance.)
814 However, the way a programmer should think about references is not so
815 much in terms of the bare reference count, but in terms of I<ownership>
816 of references. A reference to an xV can be owned by any of a variety
817 of entities: another xV, the Perl interpreter, an XS data structure,
818 a piece of running code, or a dynamic scope. An xV generally does not
819 know what entities own the references to it; it only knows how many
820 references there are, which is the reference count.
822 To correctly maintain reference counts, it is essential to keep track
823 of what references the XS code is manipulating. The programmer should
824 always know where a reference has come from and who owns it, and be
825 aware of any creation or destruction of references, and any transfers
826 of ownership. Because ownership isn't represented explicitly in the xV
827 data structures, only the reference count need be actually maintained
828 by the code, and that means that this understanding of ownership is not
829 actually evident in the code. For example, transferring ownership of a
830 reference from one owner to another doesn't change the reference count
831 at all, so may be achieved with no actual code. (The transferring code
832 doesn't touch the referenced object, but does need to ensure that the
833 former owner knows that it no longer owns the reference, and that the
834 new owner knows that it now does.)
836 An xV that is visible at the Perl level should not become unreferenced
837 and thus be destroyed. Normally, an object will only become unreferenced
838 when it is no longer visible, often by the same means that makes it
839 invisible. For example, a Perl reference value (RV) owns a reference to
840 its referent, so if the RV is overwritten that reference gets destroyed,
841 and the no-longer-reachable referent may be destroyed as a result.
843 Many functions have some kind of reference manipulation as
844 part of their purpose. Sometimes this is documented in terms
845 of ownership of references, and sometimes it is (less helpfully)
846 documented in terms of changes to reference counts. For example, the
847 L<newRV_inc()|perlapi/newRV_inc> function is documented to create a new RV
848 (with reference count 1) and increment the reference count of the referent
849 that was supplied by the caller. This is best understood as creating
850 a new reference to the referent, which is owned by the created RV,
851 and returning to the caller ownership of the sole reference to the RV.
852 The L<newRV_noinc()|perlapi/newRV_noinc> function instead does not
853 increment the reference count of the referent, but the RV nevertheless
854 ends up owning a reference to the referent. It is therefore implied
855 that the caller of C<newRV_noinc()> is relinquishing a reference to the
856 referent, making this conceptually a more complicated operation even
857 though it does less to the data structures.
859 For example, imagine you want to return a reference from an XSUB
860 function. Inside the XSUB routine, you create an SV which initially
861 has just a single reference, owned by the XSUB routine. This reference
862 needs to be disposed of before the routine is complete, otherwise it
863 will leak, preventing the SV from ever being destroyed. So to create
864 an RV referencing the SV, it is most convenient to pass the SV to
865 C<newRV_noinc()>, which consumes that reference. Now the XSUB routine
866 no longer owns a reference to the SV, but does own a reference to the RV,
867 which in turn owns a reference to the SV. The ownership of the reference
868 to the RV is then transferred by the process of returning the RV from
871 There are some convenience functions available that can help with the
872 destruction of xVs. These functions introduce the concept of "mortality".
873 Much documentation speaks of an xV itself being mortal, but this is
874 misleading. It is really I<a reference to> an xV that is mortal, and it
875 is possible for there to be more than one mortal reference to a single xV.
876 For a reference to be mortal means that it is owned by the temps stack,
877 one of perl's many internal stacks, which will destroy that reference
878 "a short time later". Usually the "short time later" is the end of
879 the current Perl statement. However, it gets more complicated around
880 dynamic scopes: there can be multiple sets of mortal references hanging
881 around at the same time, with different death dates. Internally, the
882 actual determinant for when mortal xV references are destroyed depends
883 on two macros, SAVETMPS and FREETMPS. See L<perlcall> and L<perlxs>
884 for more details on these macros.
886 Mortal references are mainly used for xVs that are placed on perl's
887 main stack. The stack is problematic for reference tracking, because it
888 contains a lot of xV references, but doesn't own those references: they
889 are not counted. Currently, there are many bugs resulting from xVs being
890 destroyed while referenced by the stack, because the stack's uncounted
891 references aren't enough to keep the xVs alive. So when putting an
892 (uncounted) reference on the stack, it is vitally important to ensure that
893 there will be a counted reference to the same xV that will last at least
894 as long as the uncounted reference. But it's also important that that
895 counted reference be cleaned up at an appropriate time, and not unduly
896 prolong the xV's life. For there to be a mortal reference is often the
897 best way to satisfy this requirement, especially if the xV was created
898 especially to be put on the stack and would otherwise be unreferenced.
900 To create a mortal reference, use the functions:
903 SV* sv_mortalcopy(SV*)
906 C<sv_newmortal()> creates an SV (with the undefined value) whose sole
907 reference is mortal. C<sv_mortalcopy()> creates an xV whose value is a
908 copy of a supplied xV and whose sole reference is mortal. C<sv_2mortal()>
909 mortalises an existing xV reference: it transfers ownership of a reference
910 from the caller to the temps stack. Because C<sv_newmortal> gives the new
911 SV no value, it must normally be given one via C<sv_setpv>, C<sv_setiv>,
914 SV *tmp = sv_newmortal();
915 sv_setiv(tmp, an_integer);
917 As that is multiple C statements it is quite common so see this idiom instead:
919 SV *tmp = sv_2mortal(newSViv(an_integer));
921 The mortal routines are not just for SVs; AVs and HVs can be
922 made mortal by passing their address (type-casted to C<SV*>) to the
923 C<sv_2mortal> or C<sv_mortalcopy> routines.
925 =head2 Stashes and Globs
927 A B<stash> is a hash that contains all variables that are defined
928 within a package. Each key of the stash is a symbol
929 name (shared by all the different types of objects that have the same
930 name), and each value in the hash table is a GV (Glob Value). This GV
931 in turn contains references to the various objects of that name,
932 including (but not limited to) the following:
941 There is a single stash called C<PL_defstash> that holds the items that exist
942 in the C<main> package. To get at the items in other packages, append the
943 string "::" to the package name. The items in the C<Foo> package are in
944 the stash C<Foo::> in PL_defstash. The items in the C<Bar::Baz> package are
945 in the stash C<Baz::> in C<Bar::>'s stash.
947 To get the stash pointer for a particular package, use the function:
949 HV* gv_stashpv(const char* name, I32 flags)
950 HV* gv_stashsv(SV*, I32 flags)
952 The first function takes a literal string, the second uses the string stored
953 in the SV. Remember that a stash is just a hash table, so you get back an
954 C<HV*>. The C<flags> flag will create a new package if it is set to GV_ADD.
956 The name that C<gv_stash*v> wants is the name of the package whose symbol table
957 you want. The default package is called C<main>. If you have multiply nested
958 packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
961 Alternately, if you have an SV that is a blessed reference, you can find
962 out the stash pointer by using:
964 HV* SvSTASH(SvRV(SV*));
966 then use the following to get the package name itself:
968 char* HvNAME(HV* stash);
970 If you need to bless or re-bless an object you can use the following
973 SV* sv_bless(SV*, HV* stash)
975 where the first argument, an C<SV*>, must be a reference, and the second
976 argument is a stash. The returned C<SV*> can now be used in the same way
979 For more information on references and blessings, consult L<perlref>.
981 =head2 Double-Typed SVs
983 Scalar variables normally contain only one type of value, an integer,
984 double, pointer, or reference. Perl will automatically convert the
985 actual scalar data from the stored type into the requested type.
987 Some scalar variables contain more than one type of scalar data. For
988 example, the variable C<$!> contains either the numeric value of C<errno>
989 or its string equivalent from either C<strerror> or C<sys_errlist[]>.
991 To force multiple data values into an SV, you must do two things: use the
992 C<sv_set*v> routines to add the additional scalar type, then set a flag
993 so that Perl will believe it contains more than one type of data. The
994 four macros to set the flags are:
1001 The particular macro you must use depends on which C<sv_set*v> routine
1002 you called first. This is because every C<sv_set*v> routine turns on
1003 only the bit for the particular type of data being set, and turns off
1006 For example, to create a new Perl variable called "dberror" that contains
1007 both the numeric and descriptive string error values, you could use the
1011 extern char *dberror_list;
1013 SV* sv = get_sv("dberror", GV_ADD);
1014 sv_setiv(sv, (IV) dberror);
1015 sv_setpv(sv, dberror_list[dberror]);
1018 If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
1019 macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
1021 =head2 Read-Only Values
1023 In Perl 5.16 and earlier, copy-on-write (see the next section) shared a
1024 flag bit with read-only scalars. So the only way to test whether
1025 C<sv_setsv>, etc., will raise a "Modification of a read-only value" error
1026 in those versions is:
1028 SvREADONLY(sv) && !SvIsCOW(sv)
1030 Under Perl 5.18 and later, SvREADONLY only applies to read-only variables,
1031 and, under 5.20, copy-on-write scalars can also be read-only, so the above
1032 check is incorrect. You just want:
1036 If you need to do this check often, define your own macro like this:
1038 #if PERL_VERSION >= 18
1039 # define SvTRULYREADONLY(sv) SvREADONLY(sv)
1041 # define SvTRULYREADONLY(sv) (SvREADONLY(sv) && !SvIsCOW(sv))
1044 =head2 Copy on Write
1046 Perl implements a copy-on-write (COW) mechanism for scalars, in which
1047 string copies are not immediately made when requested, but are deferred
1048 until made necessary by one or the other scalar changing. This is mostly
1049 transparent, but one must take care not to modify string buffers that are
1050 shared by multiple SVs.
1052 You can test whether an SV is using copy-on-write with C<SvIsCOW(sv)>.
1054 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).
1056 If you want to make the SV drop its string buffer, use
1057 C<sv_force_normal_flags(sv, SV_COW_DROP_PV)> or simply
1058 C<sv_setsv(sv, NULL)>.
1060 All of these functions will croak on read-only scalars (see the previous
1061 section for more on those).
1063 To test that your code is behaving correctly and not modifying COW buffers,
1064 on systems that support L<mmap(2)> (i.e., Unix) you can configure perl with
1065 C<-Accflags=-DPERL_DEBUG_READONLY_COW> and it will turn buffer violations
1066 into crashes. You will find it to be marvellously slow, so you may want to
1067 skip perl's own tests.
1069 =head2 Magic Variables
1071 [This section still under construction. Ignore everything here. Post no
1072 bills. Everything not permitted is forbidden.]
1074 Any SV may be magical, that is, it has special features that a normal
1075 SV does not have. These features are stored in the SV structure in a
1076 linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
1079 MAGIC* mg_moremagic;
1089 Note this is current as of patchlevel 0, and could change at any time.
1091 =head2 Assigning Magic
1093 Perl adds magic to an SV using the sv_magic function:
1095 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
1097 The C<sv> argument is a pointer to the SV that is to acquire a new magical
1100 If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
1101 convert C<sv> to type C<SVt_PVMG>.
1102 Perl then continues by adding new magic
1103 to the beginning of the linked list of magical features. Any prior entry
1104 of the same type of magic is deleted. Note that this can be overridden,
1105 and multiple instances of the same type of magic can be associated with an
1108 The C<name> and C<namlen> arguments are used to associate a string with
1109 the magic, typically the name of a variable. C<namlen> is stored in the
1110 C<mg_len> field and if C<name> is non-null then either a C<savepvn> copy of
1111 C<name> or C<name> itself is stored in the C<mg_ptr> field, depending on
1112 whether C<namlen> is greater than zero or equal to zero respectively. As a
1113 special case, if C<(name && namlen == HEf_SVKEY)> then C<name> is assumed
1114 to contain an C<SV*> and is stored as-is with its REFCNT incremented.
1116 The sv_magic function uses C<how> to determine which, if any, predefined
1117 "Magic Virtual Table" should be assigned to the C<mg_virtual> field.
1118 See the L</Magic Virtual Tables> section below. The C<how> argument is also
1119 stored in the C<mg_type> field. The value of
1120 C<how> should be chosen from the set of macros
1121 C<PERL_MAGIC_foo> found in F<perl.h>. Note that before
1122 these macros were added, Perl internals used to directly use character
1123 literals, so you may occasionally come across old code or documentation
1124 referring to 'U' magic rather than C<PERL_MAGIC_uvar> for example.
1126 The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
1127 structure. If it is not the same as the C<sv> argument, the reference
1128 count of the C<obj> object is incremented. If it is the same, or if
1129 the C<how> argument is C<PERL_MAGIC_arylen>, C<PERL_MAGIC_regdatum>,
1130 C<PERL_MAGIC_regdata>, or if it is a NULL pointer, then C<obj> is merely
1131 stored, without the reference count being incremented.
1133 See also C<sv_magicext> in L<perlapi> for a more flexible way to add magic
1136 There is also a function to add magic to an C<HV>:
1138 void hv_magic(HV *hv, GV *gv, int how);
1140 This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
1142 To remove the magic from an SV, call the function sv_unmagic:
1144 int sv_unmagic(SV *sv, int type);
1146 The C<type> argument should be equal to the C<how> value when the C<SV>
1147 was initially made magical.
1149 However, note that C<sv_unmagic> removes all magic of a certain C<type> from the
1150 C<SV>. If you want to remove only certain
1151 magic of a C<type> based on the magic
1152 virtual table, use C<sv_unmagicext> instead:
1154 int sv_unmagicext(SV *sv, int type, MGVTBL *vtbl);
1156 =head2 Magic Virtual Tables
1158 The C<mg_virtual> field in the C<MAGIC> structure is a pointer to an
1159 C<MGVTBL>, which is a structure of function pointers and stands for
1160 "Magic Virtual Table" to handle the various operations that might be
1161 applied to that variable.
1163 The C<MGVTBL> has five (or sometimes eight) pointers to the following
1166 int (*svt_get) (pTHX_ SV* sv, MAGIC* mg);
1167 int (*svt_set) (pTHX_ SV* sv, MAGIC* mg);
1168 U32 (*svt_len) (pTHX_ SV* sv, MAGIC* mg);
1169 int (*svt_clear)(pTHX_ SV* sv, MAGIC* mg);
1170 int (*svt_free) (pTHX_ SV* sv, MAGIC* mg);
1172 int (*svt_copy) (pTHX_ SV *sv, MAGIC* mg, SV *nsv,
1173 const char *name, I32 namlen);
1174 int (*svt_dup) (pTHX_ MAGIC *mg, CLONE_PARAMS *param);
1175 int (*svt_local)(pTHX_ SV *nsv, MAGIC *mg);
1178 This MGVTBL structure is set at compile-time in F<perl.h> and there are
1179 currently 32 types. These different structures contain pointers to various
1180 routines that perform additional actions depending on which function is
1183 Function pointer Action taken
1184 ---------------- ------------
1185 svt_get Do something before the value of the SV is
1187 svt_set Do something after the SV is assigned a value.
1188 svt_len Report on the SV's length.
1189 svt_clear Clear something the SV represents.
1190 svt_free Free any extra storage associated with the SV.
1192 svt_copy copy tied variable magic to a tied element
1193 svt_dup duplicate a magic structure during thread cloning
1194 svt_local copy magic to local value during 'local'
1196 For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
1197 to an C<mg_type> of C<PERL_MAGIC_sv>) contains:
1199 { magic_get, magic_set, magic_len, 0, 0 }
1201 Thus, when an SV is determined to be magical and of type C<PERL_MAGIC_sv>,
1202 if a get operation is being performed, the routine C<magic_get> is
1203 called. All the various routines for the various magical types begin
1204 with C<magic_>. NOTE: the magic routines are not considered part of
1205 the Perl API, and may not be exported by the Perl library.
1207 The last three slots are a recent addition, and for source code
1208 compatibility they are only checked for if one of the three flags
1209 MGf_COPY, MGf_DUP or MGf_LOCAL is set in mg_flags.
1210 This means that most code can continue declaring
1211 a vtable as a 5-element value. These three are
1212 currently used exclusively by the threading code, and are highly subject
1215 The current kinds of Magic Virtual Tables are:
1218 This table is generated by regen/mg_vtable.pl. Any changes made here
1221 =for mg_vtable.pl begin
1224 (old-style char and macro) MGVTBL Type of magic
1225 -------------------------- ------ -------------
1226 \0 PERL_MAGIC_sv vtbl_sv Special scalar variable
1227 # PERL_MAGIC_arylen vtbl_arylen Array length ($#ary)
1228 % PERL_MAGIC_rhash (none) Extra data for restricted
1230 * PERL_MAGIC_debugvar vtbl_debugvar $DB::single, signal, trace
1232 . PERL_MAGIC_pos vtbl_pos pos() lvalue
1233 : PERL_MAGIC_symtab (none) Extra data for symbol
1235 < PERL_MAGIC_backref vtbl_backref For weak ref data
1236 @ PERL_MAGIC_arylen_p (none) To move arylen out of XPVAV
1237 B PERL_MAGIC_bm vtbl_regexp Boyer-Moore
1238 (fast string search)
1239 c PERL_MAGIC_overload_table vtbl_ovrld Holds overload table
1241 D PERL_MAGIC_regdata vtbl_regdata Regex match position data
1243 d PERL_MAGIC_regdatum vtbl_regdatum Regex match position data
1245 E PERL_MAGIC_env vtbl_env %ENV hash
1246 e PERL_MAGIC_envelem vtbl_envelem %ENV hash element
1247 f PERL_MAGIC_fm vtbl_regexp Formline
1249 g PERL_MAGIC_regex_global vtbl_mglob m//g target
1250 H PERL_MAGIC_hints vtbl_hints %^H hash
1251 h PERL_MAGIC_hintselem vtbl_hintselem %^H hash element
1252 I PERL_MAGIC_isa vtbl_isa @ISA array
1253 i PERL_MAGIC_isaelem vtbl_isaelem @ISA array element
1254 k PERL_MAGIC_nkeys vtbl_nkeys scalar(keys()) lvalue
1255 L PERL_MAGIC_dbfile (none) Debugger %_<filename
1256 l PERL_MAGIC_dbline vtbl_dbline Debugger %_<filename
1258 N PERL_MAGIC_shared (none) Shared between threads
1259 n PERL_MAGIC_shared_scalar (none) Shared between threads
1260 o PERL_MAGIC_collxfrm vtbl_collxfrm Locale transformation
1261 P PERL_MAGIC_tied vtbl_pack Tied array or hash
1262 p PERL_MAGIC_tiedelem vtbl_packelem Tied array or hash element
1263 q PERL_MAGIC_tiedscalar vtbl_packelem Tied scalar or handle
1264 r PERL_MAGIC_qr vtbl_regexp Precompiled qr// regex
1265 S PERL_MAGIC_sig (none) %SIG hash
1266 s PERL_MAGIC_sigelem vtbl_sigelem %SIG hash element
1267 t PERL_MAGIC_taint vtbl_taint Taintedness
1268 U PERL_MAGIC_uvar vtbl_uvar Available for use by
1270 u PERL_MAGIC_uvar_elem (none) Reserved for use by
1272 V PERL_MAGIC_vstring (none) SV was vstring literal
1273 v PERL_MAGIC_vec vtbl_vec vec() lvalue
1274 w PERL_MAGIC_utf8 vtbl_utf8 Cached UTF-8 information
1275 x PERL_MAGIC_substr vtbl_substr substr() lvalue
1276 Y PERL_MAGIC_nonelem vtbl_nonelem Array element that does not
1278 y PERL_MAGIC_defelem vtbl_defelem Shadow "foreach" iterator
1279 variable / smart parameter
1281 \ PERL_MAGIC_lvref vtbl_lvref Lvalue reference
1283 ] PERL_MAGIC_checkcall vtbl_checkcall Inlining/mutation of call
1285 ~ PERL_MAGIC_ext (none) Available for use by
1288 =for mg_vtable.pl end
1290 When an uppercase and lowercase letter both exist in the table, then the
1291 uppercase letter is typically used to represent some kind of composite type
1292 (a list or a hash), and the lowercase letter is used to represent an element
1293 of that composite type. Some internals code makes use of this case
1294 relationship. However, 'v' and 'V' (vec and v-string) are in no way related.
1296 The C<PERL_MAGIC_ext> and C<PERL_MAGIC_uvar> magic types are defined
1297 specifically for use by extensions and will not be used by perl itself.
1298 Extensions can use C<PERL_MAGIC_ext> magic to 'attach' private information
1299 to variables (typically objects). This is especially useful because
1300 there is no way for normal perl code to corrupt this private information
1301 (unlike using extra elements of a hash object).
1303 Similarly, C<PERL_MAGIC_uvar> magic can be used much like tie() to call a
1304 C function any time a scalar's value is used or changed. The C<MAGIC>'s
1305 C<mg_ptr> field points to a C<ufuncs> structure:
1308 I32 (*uf_val)(pTHX_ IV, SV*);
1309 I32 (*uf_set)(pTHX_ IV, SV*);
1313 When the SV is read from or written to, the C<uf_val> or C<uf_set>
1314 function will be called with C<uf_index> as the first arg and a pointer to
1315 the SV as the second. A simple example of how to add C<PERL_MAGIC_uvar>
1316 magic is shown below. Note that the ufuncs structure is copied by
1317 sv_magic, so you can safely allocate it on the stack.
1325 uf.uf_val = &my_get_fn;
1326 uf.uf_set = &my_set_fn;
1328 sv_magic(sv, 0, PERL_MAGIC_uvar, (char*)&uf, sizeof(uf));
1330 Attaching C<PERL_MAGIC_uvar> to arrays is permissible but has no effect.
1332 For hashes there is a specialized hook that gives control over hash
1333 keys (but not values). This hook calls C<PERL_MAGIC_uvar> 'get' magic
1334 if the "set" function in the C<ufuncs> structure is NULL. The hook
1335 is activated whenever the hash is accessed with a key specified as
1336 an C<SV> through the functions C<hv_store_ent>, C<hv_fetch_ent>,
1337 C<hv_delete_ent>, and C<hv_exists_ent>. Accessing the key as a string
1338 through the functions without the C<..._ent> suffix circumvents the
1339 hook. See L<Hash::Util::FieldHash/GUTS> for a detailed description.
1341 Note that because multiple extensions may be using C<PERL_MAGIC_ext>
1342 or C<PERL_MAGIC_uvar> magic, it is important for extensions to take
1343 extra care to avoid conflict. Typically only using the magic on
1344 objects blessed into the same class as the extension is sufficient.
1345 For C<PERL_MAGIC_ext> magic, it is usually a good idea to define an
1346 C<MGVTBL>, even if all its fields will be C<0>, so that individual
1347 C<MAGIC> pointers can be identified as a particular kind of magic
1348 using their magic virtual table. C<mg_findext> provides an easy way
1351 STATIC MGVTBL my_vtbl = { 0, 0, 0, 0, 0, 0, 0, 0 };
1354 if ((mg = mg_findext(sv, PERL_MAGIC_ext, &my_vtbl))) {
1355 /* this is really ours, not another module's PERL_MAGIC_ext */
1356 my_priv_data_t *priv = (my_priv_data_t *)mg->mg_ptr;
1360 Also note that the C<sv_set*()> and C<sv_cat*()> functions described
1361 earlier do B<not> invoke 'set' magic on their targets. This must
1362 be done by the user either by calling the C<SvSETMAGIC()> macro after
1363 calling these functions, or by using one of the C<sv_set*_mg()> or
1364 C<sv_cat*_mg()> functions. Similarly, generic C code must call the
1365 C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
1366 obtained from external sources in functions that don't handle magic.
1367 See L<perlapi> for a description of these functions.
1368 For example, calls to the C<sv_cat*()> functions typically need to be
1369 followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
1370 since their implementation handles 'get' magic.
1372 =head2 Finding Magic
1374 MAGIC *mg_find(SV *sv, int type); /* Finds the magic pointer of that
1377 This routine returns a pointer to a C<MAGIC> structure stored in the SV.
1378 If the SV does not have that magical
1379 feature, C<NULL> is returned. If the
1380 SV has multiple instances of that magical feature, the first one will be
1381 returned. C<mg_findext> can be used
1382 to find a C<MAGIC> structure of an SV
1383 based on both its magic type and its magic virtual table:
1385 MAGIC *mg_findext(SV *sv, int type, MGVTBL *vtbl);
1387 Also, if the SV passed to C<mg_find> or C<mg_findext> is not of type
1388 SVt_PVMG, Perl may core dump.
1390 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
1392 This routine checks to see what types of magic C<sv> has. If the mg_type
1393 field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
1394 the mg_type field is changed to be the lowercase letter.
1396 =head2 Understanding the Magic of Tied Hashes and Arrays
1398 Tied hashes and arrays are magical beasts of the C<PERL_MAGIC_tied>
1401 WARNING: As of the 5.004 release, proper usage of the array and hash
1402 access functions requires understanding a few caveats. Some
1403 of these caveats are actually considered bugs in the API, to be fixed
1404 in later releases, and are bracketed with [MAYCHANGE] below. If
1405 you find yourself actually applying such information in this section, be
1406 aware that the behavior may change in the future, umm, without warning.
1408 The perl tie function associates a variable with an object that implements
1409 the various GET, SET, etc methods. To perform the equivalent of the perl
1410 tie function from an XSUB, you must mimic this behaviour. The code below
1411 carries out the necessary steps -- firstly it creates a new hash, and then
1412 creates a second hash which it blesses into the class which will implement
1413 the tie methods. Lastly it ties the two hashes together, and returns a
1414 reference to the new tied hash. Note that the code below does NOT call the
1415 TIEHASH method in the MyTie class -
1416 see L</Calling Perl Routines from within C Programs> for details on how
1427 tie = newRV_noinc((SV*)newHV());
1428 stash = gv_stashpv("MyTie", GV_ADD);
1429 sv_bless(tie, stash);
1430 hv_magic(hash, (GV*)tie, PERL_MAGIC_tied);
1431 RETVAL = newRV_noinc(hash);
1435 The C<av_store> function, when given a tied array argument, merely
1436 copies the magic of the array onto the value to be "stored", using
1437 C<mg_copy>. It may also return NULL, indicating that the value did not
1438 actually need to be stored in the array. [MAYCHANGE] After a call to
1439 C<av_store> on a tied array, the caller will usually need to call
1440 C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1441 TIEARRAY object. If C<av_store> did return NULL, a call to
1442 C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1445 The previous paragraph is applicable verbatim to tied hash access using the
1446 C<hv_store> and C<hv_store_ent> functions as well.
1448 C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1449 C<hv_fetch_ent> actually return an undefined mortal value whose magic
1450 has been initialized using C<mg_copy>. Note the value so returned does not
1451 need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1452 need to call C<mg_get()> on the returned value in order to actually invoke
1453 the perl level "FETCH" method on the underlying TIE object. Similarly,
1454 you may also call C<mg_set()> on the return value after possibly assigning
1455 a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
1456 method on the TIE object. [/MAYCHANGE]
1459 In other words, the array or hash fetch/store functions don't really
1460 fetch and store actual values in the case of tied arrays and hashes. They
1461 merely call C<mg_copy> to attach magic to the values that were meant to be
1462 "stored" or "fetched". Later calls to C<mg_get> and C<mg_set> actually
1463 do the job of invoking the TIE methods on the underlying objects. Thus
1464 the magic mechanism currently implements a kind of lazy access to arrays
1467 Currently (as of perl version 5.004), use of the hash and array access
1468 functions requires the user to be aware of whether they are operating on
1469 "normal" hashes and arrays, or on their tied variants. The API may be
1470 changed to provide more transparent access to both tied and normal data
1471 types in future versions.
1474 You would do well to understand that the TIEARRAY and TIEHASH interfaces
1475 are mere sugar to invoke some perl method calls while using the uniform hash
1476 and array syntax. The use of this sugar imposes some overhead (typically
1477 about two to four extra opcodes per FETCH/STORE operation, in addition to
1478 the creation of all the mortal variables required to invoke the methods).
1479 This overhead will be comparatively small if the TIE methods are themselves
1480 substantial, but if they are only a few statements long, the overhead
1481 will not be insignificant.
1483 =head2 Localizing changes
1485 Perl has a very handy construction
1492 This construction is I<approximately> equivalent to
1501 The biggest difference is that the first construction would
1502 reinstate the initial value of $var, irrespective of how control exits
1503 the block: C<goto>, C<return>, C<die>/C<eval>, etc. It is a little bit
1504 more efficient as well.
1506 There is a way to achieve a similar task from C via Perl API: create a
1507 I<pseudo-block>, and arrange for some changes to be automatically
1508 undone at the end of it, either explicit, or via a non-local exit (via
1509 die()). A I<block>-like construct is created by a pair of
1510 C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1511 Such a construct may be created specially for some important localized
1512 task, or an existing one (like boundaries of enclosing Perl
1513 subroutine/block, or an existing pair for freeing TMPs) may be
1514 used. (In the second case the overhead of additional localization must
1515 be almost negligible.) Note that any XSUB is automatically enclosed in
1516 an C<ENTER>/C<LEAVE> pair.
1518 Inside such a I<pseudo-block> the following service is available:
1522 =item C<SAVEINT(int i)>
1524 =item C<SAVEIV(IV i)>
1526 =item C<SAVEI32(I32 i)>
1528 =item C<SAVELONG(long i)>
1530 These macros arrange things to restore the value of integer variable
1531 C<i> at the end of enclosing I<pseudo-block>.
1533 =item C<SAVESPTR(s)>
1535 =item C<SAVEPPTR(p)>
1537 These macros arrange things to restore the value of pointers C<s> and
1538 C<p>. C<s> must be a pointer of a type which survives conversion to
1539 C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1542 =item C<SAVEFREESV(SV *sv)>
1544 The refcount of C<sv> will be decremented at the end of
1545 I<pseudo-block>. This is similar to C<sv_2mortal> in that it is also a
1546 mechanism for doing a delayed C<SvREFCNT_dec>. However, while C<sv_2mortal>
1547 extends the lifetime of C<sv> until the beginning of the next statement,
1548 C<SAVEFREESV> extends it until the end of the enclosing scope. These
1549 lifetimes can be wildly different.
1551 Also compare C<SAVEMORTALIZESV>.
1553 =item C<SAVEMORTALIZESV(SV *sv)>
1555 Just like C<SAVEFREESV>, but mortalizes C<sv> at the end of the current
1556 scope instead of decrementing its reference count. This usually has the
1557 effect of keeping C<sv> alive until the statement that called the currently
1558 live scope has finished executing.
1560 =item C<SAVEFREEOP(OP *op)>
1562 The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1564 =item C<SAVEFREEPV(p)>
1566 The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1567 end of I<pseudo-block>.
1569 =item C<SAVECLEARSV(SV *sv)>
1571 Clears a slot in the current scratchpad which corresponds to C<sv> at
1572 the end of I<pseudo-block>.
1574 =item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1576 The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
1577 string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1578 short-lived storage, the corresponding string may be reallocated like
1581 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
1583 =item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
1585 At the end of I<pseudo-block> the function C<f> is called with the
1588 =item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1590 At the end of I<pseudo-block> the function C<f> is called with the
1591 implicit context argument (if any), and C<p>.
1593 =item C<SAVESTACK_POS()>
1595 The current offset on the Perl internal stack (cf. C<SP>) is restored
1596 at the end of I<pseudo-block>.
1600 The following API list contains functions, thus one needs to
1601 provide pointers to the modifiable data explicitly (either C pointers,
1602 or Perlish C<GV *>s). Where the above macros take C<int>, a similar
1603 function takes C<int *>.
1607 =item C<SV* save_scalar(GV *gv)>
1609 Equivalent to Perl code C<local $gv>.
1611 =item C<AV* save_ary(GV *gv)>
1613 =item C<HV* save_hash(GV *gv)>
1615 Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1617 =item C<void save_item(SV *item)>
1619 Duplicates the current value of C<SV>, on the exit from the current
1620 C<ENTER>/C<LEAVE> I<pseudo-block> will restore the value of C<SV>
1621 using the stored value. It doesn't handle magic. Use C<save_scalar> if
1624 =item C<void save_list(SV **sarg, I32 maxsarg)>
1626 A variant of C<save_item> which takes multiple arguments via an array
1627 C<sarg> of C<SV*> of length C<maxsarg>.
1629 =item C<SV* save_svref(SV **sptr)>
1631 Similar to C<save_scalar>, but will reinstate an C<SV *>.
1633 =item C<void save_aptr(AV **aptr)>
1635 =item C<void save_hptr(HV **hptr)>
1637 Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1641 The C<Alias> module implements localization of the basic types within the
1642 I<caller's scope>. People who are interested in how to localize things in
1643 the containing scope should take a look there too.
1647 =head2 XSUBs and the Argument Stack
1649 The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1650 An XSUB routine will have a stack that contains the arguments from the Perl
1651 program, and a way to map from the Perl data structures to a C equivalent.
1653 The stack arguments are accessible through the C<ST(n)> macro, which returns
1654 the C<n>'th stack argument. Argument 0 is the first argument passed in the
1655 Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1658 Most of the time, output from the C routine can be handled through use of
1659 the RETVAL and OUTPUT directives. However, there are some cases where the
1660 argument stack is not already long enough to handle all the return values.
1661 An example is the POSIX tzname() call, which takes no arguments, but returns
1662 two, the local time zone's standard and summer time abbreviations.
1664 To handle this situation, the PPCODE directive is used and the stack is
1665 extended using the macro:
1669 where C<SP> is the macro that represents the local copy of the stack pointer,
1670 and C<num> is the number of elements the stack should be extended by.
1672 Now that there is room on the stack, values can be pushed on it using C<PUSHs>
1673 macro. The pushed values will often need to be "mortal" (See
1674 L</Reference Counts and Mortality>):
1676 PUSHs(sv_2mortal(newSViv(an_integer)))
1677 PUSHs(sv_2mortal(newSVuv(an_unsigned_integer)))
1678 PUSHs(sv_2mortal(newSVnv(a_double)))
1679 PUSHs(sv_2mortal(newSVpv("Some String",0)))
1680 /* Although the last example is better written as the more
1682 PUSHs(newSVpvs_flags("Some String", SVs_TEMP))
1684 And now the Perl program calling C<tzname>, the two values will be assigned
1687 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1689 An alternate (and possibly simpler) method to pushing values on the stack is
1694 This macro automatically adjusts the stack for you, if needed. Thus, you
1695 do not need to call C<EXTEND> to extend the stack.
1697 Despite their suggestions in earlier versions of this document the macros
1698 C<(X)PUSH[iunp]> are I<not> suited to XSUBs which return multiple results.
1699 For that, either stick to the C<(X)PUSHs> macros shown above, or use the new
1700 C<m(X)PUSH[iunp]> macros instead; see L</Putting a C value on Perl stack>.
1702 For more information, consult L<perlxs> and L<perlxstut>.
1704 =head2 Autoloading with XSUBs
1706 If an AUTOLOAD routine is an XSUB, as with Perl subroutines, Perl puts the
1707 fully-qualified name of the autoloaded subroutine in the $AUTOLOAD variable
1708 of the XSUB's package.
1710 But it also puts the same information in certain fields of the XSUB itself:
1712 HV *stash = CvSTASH(cv);
1713 const char *subname = SvPVX(cv);
1714 STRLEN name_length = SvCUR(cv); /* in bytes */
1715 U32 is_utf8 = SvUTF8(cv);
1717 C<SvPVX(cv)> contains just the sub name itself, not including the package.
1718 For an AUTOLOAD routine in UNIVERSAL or one of its superclasses,
1719 C<CvSTASH(cv)> returns NULL during a method call on a nonexistent package.
1721 B<Note>: Setting $AUTOLOAD stopped working in 5.6.1, which did not support
1722 XS AUTOLOAD subs at all. Perl 5.8.0 introduced the use of fields in the
1723 XSUB itself. Perl 5.16.0 restored the setting of $AUTOLOAD. If you need
1724 to support 5.8-5.14, use the XSUB's fields.
1726 =head2 Calling Perl Routines from within C Programs
1728 There are four routines that can be used to call a Perl subroutine from
1729 within a C program. These four are:
1731 I32 call_sv(SV*, I32);
1732 I32 call_pv(const char*, I32);
1733 I32 call_method(const char*, I32);
1734 I32 call_argv(const char*, I32, char**);
1736 The routine most often used is C<call_sv>. The C<SV*> argument
1737 contains either the name of the Perl subroutine to be called, or a
1738 reference to the subroutine. The second argument consists of flags
1739 that control the context in which the subroutine is called, whether
1740 or not the subroutine is being passed arguments, how errors should be
1741 trapped, and how to treat return values.
1743 All four routines return the number of arguments that the subroutine returned
1746 These routines used to be called C<perl_call_sv>, etc., before Perl v5.6.0,
1747 but those names are now deprecated; macros of the same name are provided for
1750 When using any of these routines (except C<call_argv>), the programmer
1751 must manipulate the Perl stack. These include the following macros and
1766 For a detailed description of calling conventions from C to Perl,
1767 consult L<perlcall>.
1769 =head2 Putting a C value on Perl stack
1771 A lot of opcodes (this is an elementary operation in the internal perl
1772 stack machine) put an SV* on the stack. However, as an optimization
1773 the corresponding SV is (usually) not recreated each time. The opcodes
1774 reuse specially assigned SVs (I<target>s) which are (as a corollary)
1775 not constantly freed/created.
1777 Each of the targets is created only once (but see
1778 L</Scratchpads and recursion> below), and when an opcode needs to put
1779 an integer, a double, or a string on stack, it just sets the
1780 corresponding parts of its I<target> and puts the I<target> on stack.
1782 The macro to put this target on stack is C<PUSHTARG>, and it is
1783 directly used in some opcodes, as well as indirectly in zillions of
1784 others, which use it via C<(X)PUSH[iunp]>.
1786 Because the target is reused, you must be careful when pushing multiple
1787 values on the stack. The following code will not do what you think:
1792 This translates as "set C<TARG> to 10, push a pointer to C<TARG> onto
1793 the stack; set C<TARG> to 20, push a pointer to C<TARG> onto the stack".
1794 At the end of the operation, the stack does not contain the values 10
1795 and 20, but actually contains two pointers to C<TARG>, which we have set
1798 If you need to push multiple different values then you should either use
1799 the C<(X)PUSHs> macros, or else use the new C<m(X)PUSH[iunp]> macros,
1800 none of which make use of C<TARG>. The C<(X)PUSHs> macros simply push an
1801 SV* on the stack, which, as noted under L</XSUBs and the Argument Stack>,
1802 will often need to be "mortal". The new C<m(X)PUSH[iunp]> macros make
1803 this a little easier to achieve by creating a new mortal for you (via
1804 C<(X)PUSHmortal>), pushing that onto the stack (extending it if necessary
1805 in the case of the C<mXPUSH[iunp]> macros), and then setting its value.
1806 Thus, instead of writing this to "fix" the example above:
1808 XPUSHs(sv_2mortal(newSViv(10)))
1809 XPUSHs(sv_2mortal(newSViv(20)))
1811 you can simply write:
1816 On a related note, if you do use C<(X)PUSH[iunp]>, then you're going to
1817 need a C<dTARG> in your variable declarations so that the C<*PUSH*>
1818 macros can make use of the local variable C<TARG>. See also C<dTARGET>
1823 The question remains on when the SVs which are I<target>s for opcodes
1824 are created. The answer is that they are created when the current
1825 unit--a subroutine or a file (for opcodes for statements outside of
1826 subroutines)--is compiled. During this time a special anonymous Perl
1827 array is created, which is called a scratchpad for the current unit.
1829 A scratchpad keeps SVs which are lexicals for the current unit and are
1830 targets for opcodes. A previous version of this document
1831 stated that one can deduce that an SV lives on a scratchpad
1832 by looking on its flags: lexicals have C<SVs_PADMY> set, and
1833 I<target>s have C<SVs_PADTMP> set. But this has never been fully true.
1834 C<SVs_PADMY> could be set on a variable that no longer resides in any pad.
1835 While I<target>s do have C<SVs_PADTMP> set, it can also be set on variables
1836 that have never resided in a pad, but nonetheless act like I<target>s. As
1837 of perl 5.21.5, the C<SVs_PADMY> flag is no longer used and is defined as
1838 0. C<SvPADMY()> now returns true for anything without C<SVs_PADTMP>.
1840 The correspondence between OPs and I<target>s is not 1-to-1. Different
1841 OPs in the compile tree of the unit can use the same target, if this
1842 would not conflict with the expected life of the temporary.
1844 =head2 Scratchpads and recursion
1846 In fact it is not 100% true that a compiled unit contains a pointer to
1847 the scratchpad AV. In fact it contains a pointer to an AV of
1848 (initially) one element, and this element is the scratchpad AV. Why do
1849 we need an extra level of indirection?
1851 The answer is B<recursion>, and maybe B<threads>. Both
1852 these can create several execution pointers going into the same
1853 subroutine. For the subroutine-child not write over the temporaries
1854 for the subroutine-parent (lifespan of which covers the call to the
1855 child), the parent and the child should have different
1856 scratchpads. (I<And> the lexicals should be separate anyway!)
1858 So each subroutine is born with an array of scratchpads (of length 1).
1859 On each entry to the subroutine it is checked that the current
1860 depth of the recursion is not more than the length of this array, and
1861 if it is, new scratchpad is created and pushed into the array.
1863 The I<target>s on this scratchpad are C<undef>s, but they are already
1864 marked with correct flags.
1866 =head1 Memory Allocation
1870 All memory meant to be used with the Perl API functions should be manipulated
1871 using the macros described in this section. The macros provide the necessary
1872 transparency between differences in the actual malloc implementation that is
1875 It is suggested that you enable the version of malloc that is distributed
1876 with Perl. It keeps pools of various sizes of unallocated memory in
1877 order to satisfy allocation requests more quickly. However, on some
1878 platforms, it may cause spurious malloc or free errors.
1880 The following three macros are used to initially allocate memory :
1882 Newx(pointer, number, type);
1883 Newxc(pointer, number, type, cast);
1884 Newxz(pointer, number, type);
1886 The first argument C<pointer> should be the name of a variable that will
1887 point to the newly allocated memory.
1889 The second and third arguments C<number> and C<type> specify how many of
1890 the specified type of data structure should be allocated. The argument
1891 C<type> is passed to C<sizeof>. The final argument to C<Newxc>, C<cast>,
1892 should be used if the C<pointer> argument is different from the C<type>
1895 Unlike the C<Newx> and C<Newxc> macros, the C<Newxz> macro calls C<memzero>
1896 to zero out all the newly allocated memory.
1900 Renew(pointer, number, type);
1901 Renewc(pointer, number, type, cast);
1904 These three macros are used to change a memory buffer size or to free a
1905 piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1906 match those of C<New> and C<Newc> with the exception of not needing the
1907 "magic cookie" argument.
1911 Move(source, dest, number, type);
1912 Copy(source, dest, number, type);
1913 Zero(dest, number, type);
1915 These three macros are used to move, copy, or zero out previously allocated
1916 memory. The C<source> and C<dest> arguments point to the source and
1917 destination starting points. Perl will move, copy, or zero out C<number>
1918 instances of the size of the C<type> data structure (using the C<sizeof>
1923 The most recent development releases of Perl have been experimenting with
1924 removing Perl's dependency on the "normal" standard I/O suite and allowing
1925 other stdio implementations to be used. This involves creating a new
1926 abstraction layer that then calls whichever implementation of stdio Perl
1927 was compiled with. All XSUBs should now use the functions in the PerlIO
1928 abstraction layer and not make any assumptions about what kind of stdio
1931 For a complete description of the PerlIO abstraction, consult L<perlapio>.
1933 =head1 Compiled code
1937 Here we describe the internal form your code is converted to by
1938 Perl. Start with a simple example:
1942 This is converted to a tree similar to this one:
1950 (but slightly more complicated). This tree reflects the way Perl
1951 parsed your code, but has nothing to do with the execution order.
1952 There is an additional "thread" going through the nodes of the tree
1953 which shows the order of execution of the nodes. In our simplified
1954 example above it looks like:
1956 $b ---> $c ---> + ---> $a ---> assign-to
1958 But with the actual compile tree for C<$a = $b + $c> it is different:
1959 some nodes I<optimized away>. As a corollary, though the actual tree
1960 contains more nodes than our simplified example, the execution order
1961 is the same as in our example.
1963 =head2 Examining the tree
1965 If you have your perl compiled for debugging (usually done with
1966 C<-DDEBUGGING> on the C<Configure> command line), you may examine the
1967 compiled tree by specifying C<-Dx> on the Perl command line. The
1968 output takes several lines per node, and for C<$b+$c> it looks like
1973 FLAGS = (SCALAR,KIDS)
1975 TYPE = null ===> (4)
1977 FLAGS = (SCALAR,KIDS)
1979 3 TYPE = gvsv ===> 4
1985 TYPE = null ===> (5)
1987 FLAGS = (SCALAR,KIDS)
1989 4 TYPE = gvsv ===> 5
1995 This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
1996 not optimized away (one per number in the left column). The immediate
1997 children of the given node correspond to C<{}> pairs on the same level
1998 of indentation, thus this listing corresponds to the tree:
2006 The execution order is indicated by C<===E<gt>> marks, thus it is C<3
2007 4 5 6> (node C<6> is not included into above listing), i.e.,
2008 C<gvsv gvsv add whatever>.
2010 Each of these nodes represents an op, a fundamental operation inside the
2011 Perl core. The code which implements each operation can be found in the
2012 F<pp*.c> files; the function which implements the op with type C<gvsv>
2013 is C<pp_gvsv>, and so on. As the tree above shows, different ops have
2014 different numbers of children: C<add> is a binary operator, as one would
2015 expect, and so has two children. To accommodate the various different
2016 numbers of children, there are various types of op data structure, and
2017 they link together in different ways.
2019 The simplest type of op structure is C<OP>: this has no children. Unary
2020 operators, C<UNOP>s, have one child, and this is pointed to by the
2021 C<op_first> field. Binary operators (C<BINOP>s) have not only an
2022 C<op_first> field but also an C<op_last> field. The most complex type of
2023 op is a C<LISTOP>, which has any number of children. In this case, the
2024 first child is pointed to by C<op_first> and the last child by
2025 C<op_last>. The children in between can be found by iteratively
2026 following the C<OpSIBLING> pointer from the first child to the last (but
2029 There are also some other op types: a C<PMOP> holds a regular expression,
2030 and has no children, and a C<LOOP> may or may not have children. If the
2031 C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
2032 complicate matters, if a C<UNOP> is actually a C<null> op after
2033 optimization (see L</Compile pass 2: context propagation>) it will still
2034 have children in accordance with its former type.
2036 Finally, there is a C<LOGOP>, or logic op. Like a C<LISTOP>, this has one
2037 or more children, but it doesn't have an C<op_last> field: so you have to
2038 follow C<op_first> and then the C<OpSIBLING> chain itself to find the
2039 last child. Instead it has an C<op_other> field, which is comparable to
2040 the C<op_next> field described below, and represents an alternate
2041 execution path. Operators like C<and>, C<or> and C<?> are C<LOGOP>s. Note
2042 that in general, C<op_other> may not point to any of the direct children
2045 Starting in version 5.21.2, perls built with the experimental
2046 define C<-DPERL_OP_PARENT> add an extra boolean flag for each op,
2047 C<op_moresib>. When not set, this indicates that this is the last op in an
2048 C<OpSIBLING> chain. This frees up the C<op_sibling> field on the last
2049 sibling to point back to the parent op. Under this build, that field is
2050 also renamed C<op_sibparent> to reflect its joint role. The macro
2051 C<OpSIBLING(o)> wraps this special behaviour, and always returns NULL on
2052 the last sibling. With this build the C<op_parent(o)> function can be
2053 used to find the parent of any op. Thus for forward compatibility, you
2054 should always use the C<OpSIBLING(o)> macro rather than accessing
2055 C<op_sibling> directly.
2057 Another way to examine the tree is to use a compiler back-end module, such
2060 =head2 Compile pass 1: check routines
2062 The tree is created by the compiler while I<yacc> code feeds it
2063 the constructions it recognizes. Since I<yacc> works bottom-up, so does
2064 the first pass of perl compilation.
2066 What makes this pass interesting for perl developers is that some
2067 optimization may be performed on this pass. This is optimization by
2068 so-called "check routines". The correspondence between node names
2069 and corresponding check routines is described in F<opcode.pl> (do not
2070 forget to run C<make regen_headers> if you modify this file).
2072 A check routine is called when the node is fully constructed except
2073 for the execution-order thread. Since at this time there are no
2074 back-links to the currently constructed node, one can do most any
2075 operation to the top-level node, including freeing it and/or creating
2076 new nodes above/below it.
2078 The check routine returns the node which should be inserted into the
2079 tree (if the top-level node was not modified, check routine returns
2082 By convention, check routines have names C<ck_*>. They are usually
2083 called from C<new*OP> subroutines (or C<convert>) (which in turn are
2084 called from F<perly.y>).
2086 =head2 Compile pass 1a: constant folding
2088 Immediately after the check routine is called the returned node is
2089 checked for being compile-time executable. If it is (the value is
2090 judged to be constant) it is immediately executed, and a I<constant>
2091 node with the "return value" of the corresponding subtree is
2092 substituted instead. The subtree is deleted.
2094 If constant folding was not performed, the execution-order thread is
2097 =head2 Compile pass 2: context propagation
2099 When a context for a part of compile tree is known, it is propagated
2100 down through the tree. At this time the context can have 5 values
2101 (instead of 2 for runtime context): void, boolean, scalar, list, and
2102 lvalue. In contrast with the pass 1 this pass is processed from top
2103 to bottom: a node's context determines the context for its children.
2105 Additional context-dependent optimizations are performed at this time.
2106 Since at this moment the compile tree contains back-references (via
2107 "thread" pointers), nodes cannot be free()d now. To allow
2108 optimized-away nodes at this stage, such nodes are null()ified instead
2109 of free()ing (i.e. their type is changed to OP_NULL).
2111 =head2 Compile pass 3: peephole optimization
2113 After the compile tree for a subroutine (or for an C<eval> or a file)
2114 is created, an additional pass over the code is performed. This pass
2115 is neither top-down or bottom-up, but in the execution order (with
2116 additional complications for conditionals). Optimizations performed
2117 at this stage are subject to the same restrictions as in the pass 2.
2119 Peephole optimizations are done by calling the function pointed to
2120 by the global variable C<PL_peepp>. By default, C<PL_peepp> just
2121 calls the function pointed to by the global variable C<PL_rpeepp>.
2122 By default, that performs some basic op fixups and optimisations along
2123 the execution-order op chain, and recursively calls C<PL_rpeepp> for
2124 each side chain of ops (resulting from conditionals). Extensions may
2125 provide additional optimisations or fixups, hooking into either the
2126 per-subroutine or recursive stage, like this:
2128 static peep_t prev_peepp;
2129 static void my_peep(pTHX_ OP *o)
2131 /* custom per-subroutine optimisation goes here */
2132 prev_peepp(aTHX_ o);
2133 /* custom per-subroutine optimisation may also go here */
2136 prev_peepp = PL_peepp;
2139 static peep_t prev_rpeepp;
2140 static void my_rpeep(pTHX_ OP *o)
2143 for(; o; o = o->op_next) {
2144 /* custom per-op optimisation goes here */
2146 prev_rpeepp(aTHX_ orig_o);
2149 prev_rpeepp = PL_rpeepp;
2150 PL_rpeepp = my_rpeep;
2152 =head2 Pluggable runops
2154 The compile tree is executed in a runops function. There are two runops
2155 functions, in F<run.c> and in F<dump.c>. C<Perl_runops_debug> is used
2156 with DEBUGGING and C<Perl_runops_standard> is used otherwise. For fine
2157 control over the execution of the compile tree it is possible to provide
2158 your own runops function.
2160 It's probably best to copy one of the existing runops functions and
2161 change it to suit your needs. Then, in the BOOT section of your XS
2164 PL_runops = my_runops;
2166 This function should be as efficient as possible to keep your programs
2167 running as fast as possible.
2169 =head2 Compile-time scope hooks
2171 As of perl 5.14 it is possible to hook into the compile-time lexical
2172 scope mechanism using C<Perl_blockhook_register>. This is used like
2175 STATIC void my_start_hook(pTHX_ int full);
2176 STATIC BHK my_hooks;
2179 BhkENTRY_set(&my_hooks, bhk_start, my_start_hook);
2180 Perl_blockhook_register(aTHX_ &my_hooks);
2182 This will arrange to have C<my_start_hook> called at the start of
2183 compiling every lexical scope. The available hooks are:
2187 =item C<void bhk_start(pTHX_ int full)>
2189 This is called just after starting a new lexical scope. Note that Perl
2194 creates two scopes: the first starts at the C<(> and has C<full == 1>,
2195 the second starts at the C<{> and has C<full == 0>. Both end at the
2196 C<}>, so calls to C<start> and C<pre>/C<post_end> will match. Anything
2197 pushed onto the save stack by this hook will be popped just before the
2198 scope ends (between the C<pre_> and C<post_end> hooks, in fact).
2200 =item C<void bhk_pre_end(pTHX_ OP **o)>
2202 This is called at the end of a lexical scope, just before unwinding the
2203 stack. I<o> is the root of the optree representing the scope; it is a
2204 double pointer so you can replace the OP if you need to.
2206 =item C<void bhk_post_end(pTHX_ OP **o)>
2208 This is called at the end of a lexical scope, just after unwinding the
2209 stack. I<o> is as above. Note that it is possible for calls to C<pre_>
2210 and C<post_end> to nest, if there is something on the save stack that
2213 =item C<void bhk_eval(pTHX_ OP *const o)>
2215 This is called just before starting to compile an C<eval STRING>, C<do
2216 FILE>, C<require> or C<use>, after the eval has been set up. I<o> is the
2217 OP that requested the eval, and will normally be an C<OP_ENTEREVAL>,
2218 C<OP_DOFILE> or C<OP_REQUIRE>.
2222 Once you have your hook functions, you need a C<BHK> structure to put
2223 them in. It's best to allocate it statically, since there is no way to
2224 free it once it's registered. The function pointers should be inserted
2225 into this structure using the C<BhkENTRY_set> macro, which will also set
2226 flags indicating which entries are valid. If you do need to allocate
2227 your C<BHK> dynamically for some reason, be sure to zero it before you
2230 Once registered, there is no mechanism to switch these hooks off, so if
2231 that is necessary you will need to do this yourself. An entry in C<%^H>
2232 is probably the best way, so the effect is lexically scoped; however it
2233 is also possible to use the C<BhkDISABLE> and C<BhkENABLE> macros to
2234 temporarily switch entries on and off. You should also be aware that
2235 generally speaking at least one scope will have opened before your
2236 extension is loaded, so you will see some C<pre>/C<post_end> pairs that
2237 didn't have a matching C<start>.
2239 =head1 Examining internal data structures with the C<dump> functions
2241 To aid debugging, the source file F<dump.c> contains a number of
2242 functions which produce formatted output of internal data structures.
2244 The most commonly used of these functions is C<Perl_sv_dump>; it's used
2245 for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
2246 C<sv_dump> to produce debugging output from Perl-space, so users of that
2247 module should already be familiar with its format.
2249 C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
2250 derivatives, and produces output similar to C<perl -Dx>; in fact,
2251 C<Perl_dump_eval> will dump the main root of the code being evaluated,
2252 exactly like C<-Dx>.
2254 Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
2255 op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
2256 subroutines in a package like so: (Thankfully, these are all xsubs, so
2257 there is no op tree)
2259 (gdb) print Perl_dump_packsubs(PL_defstash)
2261 SUB attributes::bootstrap = (xsub 0x811fedc 0)
2263 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
2265 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
2267 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
2269 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
2271 and C<Perl_dump_all>, which dumps all the subroutines in the stash and
2272 the op tree of the main root.
2274 =head1 How multiple interpreters and concurrency are supported
2276 =head2 Background and PERL_IMPLICIT_CONTEXT
2278 The Perl interpreter can be regarded as a closed box: it has an API
2279 for feeding it code or otherwise making it do things, but it also has
2280 functions for its own use. This smells a lot like an object, and
2281 there are ways for you to build Perl so that you can have multiple
2282 interpreters, with one interpreter represented either as a C structure,
2283 or inside a thread-specific structure. These structures contain all
2284 the context, the state of that interpreter.
2286 One macro controls the major Perl build flavor: MULTIPLICITY. The
2287 MULTIPLICITY build has a C structure that packages all the interpreter
2288 state. With multiplicity-enabled perls, PERL_IMPLICIT_CONTEXT is also
2289 normally defined, and enables the support for passing in a "hidden" first
2290 argument that represents all three data structures. MULTIPLICITY makes
2291 multi-threaded perls possible (with the ithreads threading model, related
2292 to the macro USE_ITHREADS.)
2294 Two other "encapsulation" macros are the PERL_GLOBAL_STRUCT and
2295 PERL_GLOBAL_STRUCT_PRIVATE (the latter turns on the former, and the
2296 former turns on MULTIPLICITY.) The PERL_GLOBAL_STRUCT causes all the
2297 internal variables of Perl to be wrapped inside a single global struct,
2298 struct perl_vars, accessible as (globals) &PL_Vars or PL_VarsPtr or
2299 the function Perl_GetVars(). The PERL_GLOBAL_STRUCT_PRIVATE goes
2300 one step further, there is still a single struct (allocated in main()
2301 either from heap or from stack) but there are no global data symbols
2302 pointing to it. In either case the global struct should be initialized
2303 as the very first thing in main() using Perl_init_global_struct() and
2304 correspondingly tear it down after perl_free() using Perl_free_global_struct(),
2305 please see F<miniperlmain.c> for usage details. You may also need
2306 to use C<dVAR> in your coding to "declare the global variables"
2307 when you are using them. dTHX does this for you automatically.
2309 To see whether you have non-const data you can use a BSD (or GNU)
2312 nm libperl.a | grep -v ' [TURtr] '
2314 If this displays any C<D> or C<d> symbols (or possibly C<C> or C<c>),
2315 you have non-const data. The symbols the C<grep> removed are as follows:
2316 C<Tt> are I<text>, or code, the C<Rr> are I<read-only> (const) data,
2317 and the C<U> is <undefined>, external symbols referred to.
2319 The test F<t/porting/libperl.t> does this kind of symbol sanity
2320 checking on C<libperl.a>.
2322 For backward compatibility reasons defining just PERL_GLOBAL_STRUCT
2323 doesn't actually hide all symbols inside a big global struct: some
2324 PerlIO_xxx vtables are left visible. The PERL_GLOBAL_STRUCT_PRIVATE
2325 then hides everything (see how the PERLIO_FUNCS_DECL is used).
2327 All this obviously requires a way for the Perl internal functions to be
2328 either subroutines taking some kind of structure as the first
2329 argument, or subroutines taking nothing as the first argument. To
2330 enable these two very different ways of building the interpreter,
2331 the Perl source (as it does in so many other situations) makes heavy
2332 use of macros and subroutine naming conventions.
2334 First problem: deciding which functions will be public API functions and
2335 which will be private. All functions whose names begin C<S_> are private
2336 (think "S" for "secret" or "static"). All other functions begin with
2337 "Perl_", but just because a function begins with "Perl_" does not mean it is
2338 part of the API. (See L</Internal
2339 Functions>.) The easiest way to be B<sure> a
2340 function is part of the API is to find its entry in L<perlapi>.
2341 If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
2342 think it should be (i.e., you need it for your extension), send mail via
2343 L<perlbug> explaining why you think it should be.
2345 Second problem: there must be a syntax so that the same subroutine
2346 declarations and calls can pass a structure as their first argument,
2347 or pass nothing. To solve this, the subroutines are named and
2348 declared in a particular way. Here's a typical start of a static
2349 function used within the Perl guts:
2352 S_incline(pTHX_ char *s)
2354 STATIC becomes "static" in C, and may be #define'd to nothing in some
2355 configurations in the future.
2357 A public function (i.e. part of the internal API, but not necessarily
2358 sanctioned for use in extensions) begins like this:
2361 Perl_sv_setiv(pTHX_ SV* dsv, IV num)
2363 C<pTHX_> is one of a number of macros (in F<perl.h>) that hide the
2364 details of the interpreter's context. THX stands for "thread", "this",
2365 or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
2366 The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
2367 or 'd' for B<d>eclaration, so we have C<pTHX>, C<aTHX> and C<dTHX>, and
2370 When Perl is built without options that set PERL_IMPLICIT_CONTEXT, there is no
2371 first argument containing the interpreter's context. The trailing underscore
2372 in the pTHX_ macro indicates that the macro expansion needs a comma
2373 after the context argument because other arguments follow it. If
2374 PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
2375 subroutine is not prototyped to take the extra argument. The form of the
2376 macro without the trailing underscore is used when there are no additional
2379 When a core function calls another, it must pass the context. This
2380 is normally hidden via macros. Consider C<sv_setiv>. It expands into
2381 something like this:
2383 #ifdef PERL_IMPLICIT_CONTEXT
2384 #define sv_setiv(a,b) Perl_sv_setiv(aTHX_ a, b)
2385 /* can't do this for vararg functions, see below */
2387 #define sv_setiv Perl_sv_setiv
2390 This works well, and means that XS authors can gleefully write:
2394 and still have it work under all the modes Perl could have been
2397 This doesn't work so cleanly for varargs functions, though, as macros
2398 imply that the number of arguments is known in advance. Instead we
2399 either need to spell them out fully, passing C<aTHX_> as the first
2400 argument (the Perl core tends to do this with functions like
2401 Perl_warner), or use a context-free version.
2403 The context-free version of Perl_warner is called
2404 Perl_warner_nocontext, and does not take the extra argument. Instead
2405 it does dTHX; to get the context from thread-local storage. We
2406 C<#define warner Perl_warner_nocontext> so that extensions get source
2407 compatibility at the expense of performance. (Passing an arg is
2408 cheaper than grabbing it from thread-local storage.)
2410 You can ignore [pad]THXx when browsing the Perl headers/sources.
2411 Those are strictly for use within the core. Extensions and embedders
2412 need only be aware of [pad]THX.
2414 =head2 So what happened to dTHR?
2416 C<dTHR> was introduced in perl 5.005 to support the older thread model.
2417 The older thread model now uses the C<THX> mechanism to pass context
2418 pointers around, so C<dTHR> is not useful any more. Perl 5.6.0 and
2419 later still have it for backward source compatibility, but it is defined
2422 =head2 How do I use all this in extensions?
2424 When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
2425 any functions in the Perl API will need to pass the initial context
2426 argument somehow. The kicker is that you will need to write it in
2427 such a way that the extension still compiles when Perl hasn't been
2428 built with PERL_IMPLICIT_CONTEXT enabled.
2430 There are three ways to do this. First, the easy but inefficient way,
2431 which is also the default, in order to maintain source compatibility
2432 with extensions: whenever F<XSUB.h> is #included, it redefines the aTHX
2433 and aTHX_ macros to call a function that will return the context.
2434 Thus, something like:
2438 in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
2441 Perl_sv_setiv(Perl_get_context(), sv, num);
2443 or to this otherwise:
2445 Perl_sv_setiv(sv, num);
2447 You don't have to do anything new in your extension to get this; since
2448 the Perl library provides Perl_get_context(), it will all just
2451 The second, more efficient way is to use the following template for
2454 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2459 STATIC void my_private_function(int arg1, int arg2);
2462 my_private_function(int arg1, int arg2)
2464 dTHX; /* fetch context */
2465 ... call many Perl API functions ...
2470 MODULE = Foo PACKAGE = Foo
2478 my_private_function(arg, 10);
2480 Note that the only two changes from the normal way of writing an
2481 extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
2482 including the Perl headers, followed by a C<dTHX;> declaration at
2483 the start of every function that will call the Perl API. (You'll
2484 know which functions need this, because the C compiler will complain
2485 that there's an undeclared identifier in those functions.) No changes
2486 are needed for the XSUBs themselves, because the XS() macro is
2487 correctly defined to pass in the implicit context if needed.
2489 The third, even more efficient way is to ape how it is done within
2493 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2498 /* pTHX_ only needed for functions that call Perl API */
2499 STATIC void my_private_function(pTHX_ int arg1, int arg2);
2502 my_private_function(pTHX_ int arg1, int arg2)
2504 /* dTHX; not needed here, because THX is an argument */
2505 ... call Perl API functions ...
2510 MODULE = Foo PACKAGE = Foo
2518 my_private_function(aTHX_ arg, 10);
2520 This implementation never has to fetch the context using a function
2521 call, since it is always passed as an extra argument. Depending on
2522 your needs for simplicity or efficiency, you may mix the previous
2523 two approaches freely.
2525 Never add a comma after C<pTHX> yourself--always use the form of the
2526 macro with the underscore for functions that take explicit arguments,
2527 or the form without the argument for functions with no explicit arguments.
2529 If one is compiling Perl with the C<-DPERL_GLOBAL_STRUCT> the C<dVAR>
2530 definition is needed if the Perl global variables (see F<perlvars.h>
2531 or F<globvar.sym>) are accessed in the function and C<dTHX> is not
2532 used (the C<dTHX> includes the C<dVAR> if necessary). One notices
2533 the need for C<dVAR> only with the said compile-time define, because
2534 otherwise the Perl global variables are visible as-is.
2536 =head2 Should I do anything special if I call perl from multiple threads?
2538 If you create interpreters in one thread and then proceed to call them in
2539 another, you need to make sure perl's own Thread Local Storage (TLS) slot is
2540 initialized correctly in each of those threads.
2542 The C<perl_alloc> and C<perl_clone> API functions will automatically set
2543 the TLS slot to the interpreter they created, so that there is no need to do
2544 anything special if the interpreter is always accessed in the same thread that
2545 created it, and that thread did not create or call any other interpreters
2546 afterwards. If that is not the case, you have to set the TLS slot of the
2547 thread before calling any functions in the Perl API on that particular
2548 interpreter. This is done by calling the C<PERL_SET_CONTEXT> macro in that
2549 thread as the first thing you do:
2551 /* do this before doing anything else with some_perl */
2552 PERL_SET_CONTEXT(some_perl);
2554 ... other Perl API calls on some_perl go here ...
2556 =head2 Future Plans and PERL_IMPLICIT_SYS
2558 Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
2559 that the interpreter knows about itself and pass it around, so too are
2560 there plans to allow the interpreter to bundle up everything it knows
2561 about the environment it's running on. This is enabled with the
2562 PERL_IMPLICIT_SYS macro. Currently it only works with USE_ITHREADS on
2565 This allows the ability to provide an extra pointer (called the "host"
2566 environment) for all the system calls. This makes it possible for
2567 all the system stuff to maintain their own state, broken down into
2568 seven C structures. These are thin wrappers around the usual system
2569 calls (see F<win32/perllib.c>) for the default perl executable, but for a
2570 more ambitious host (like the one that would do fork() emulation) all
2571 the extra work needed to pretend that different interpreters are
2572 actually different "processes", would be done here.
2574 The Perl engine/interpreter and the host are orthogonal entities.
2575 There could be one or more interpreters in a process, and one or
2576 more "hosts", with free association between them.
2578 =head1 Internal Functions
2580 All of Perl's internal functions which will be exposed to the outside
2581 world are prefixed by C<Perl_> so that they will not conflict with XS
2582 functions or functions used in a program in which Perl is embedded.
2583 Similarly, all global variables begin with C<PL_>. (By convention,
2584 static functions start with C<S_>.)
2586 Inside the Perl core (C<PERL_CORE> defined), you can get at the functions
2587 either with or without the C<Perl_> prefix, thanks to a bunch of defines
2588 that live in F<embed.h>. Note that extension code should I<not> set
2589 C<PERL_CORE>; this exposes the full perl internals, and is likely to cause
2590 breakage of the XS in each new perl release.
2592 The file F<embed.h> is generated automatically from
2593 F<embed.pl> and F<embed.fnc>. F<embed.pl> also creates the prototyping
2594 header files for the internal functions, generates the documentation
2595 and a lot of other bits and pieces. It's important that when you add
2596 a new function to the core or change an existing one, you change the
2597 data in the table in F<embed.fnc> as well. Here's a sample entry from
2600 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
2602 The second column is the return type, the third column the name. Columns
2603 after that are the arguments. The first column is a set of flags:
2609 This function is a part of the public
2610 API. All such functions should also
2611 have 'd', very few do not.
2615 This function has a C<Perl_> prefix; i.e. it is defined as
2620 This function has documentation using the C<apidoc> feature which we'll
2621 look at in a second. Some functions have 'd' but not 'A'; docs are good.
2625 Other available flags are:
2631 This is a static function and is defined as C<STATIC S_whatever>, and
2632 usually called within the sources as C<whatever(...)>.
2636 This does not need an interpreter context, so the definition has no
2637 C<pTHX>, and it follows that callers don't use C<aTHX>. (See
2638 L</Background and PERL_IMPLICIT_CONTEXT>.)
2642 This function never returns; C<croak>, C<exit> and friends.
2646 This function takes a variable number of arguments, C<printf> style.
2647 The argument list should end with C<...>, like this:
2649 Afprd |void |croak |const char* pat|...
2653 This function is part of the experimental development API, and may change
2654 or disappear without notice.
2658 This function should not have a compatibility macro to define, say,
2659 C<Perl_parse> to C<parse>. It must be called as C<Perl_parse>.
2663 This function isn't exported out of the Perl core.
2667 This is implemented as a macro.
2671 This function is explicitly exported.
2675 This function is visible to extensions included in the Perl core.
2679 Binary backward compatibility; this function is a macro but also has
2680 a C<Perl_> implementation (which is exported).
2684 See the comments at the top of C<embed.fnc> for others.
2688 If you edit F<embed.pl> or F<embed.fnc>, you will need to run
2689 C<make regen_headers> to force a rebuild of F<embed.h> and other
2690 auto-generated files.
2692 =head2 Formatted Printing of IVs, UVs, and NVs
2694 If you are printing IVs, UVs, or NVS instead of the stdio(3) style
2695 formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
2696 following macros for portability
2701 UVxf UV in hexadecimal
2706 These will take care of 64-bit integers and long doubles.
2709 printf("IV is %"IVdf"\n", iv);
2711 The IVdf will expand to whatever is the correct format for the IVs.
2713 Note that there are different "long doubles": Perl will use
2714 whatever the compiler has.
2716 If you are printing addresses of pointers, use UVxf combined
2717 with PTR2UV(), do not use %lx or %p.
2719 =head2 Formatted Printing of C<Size_t> and C<SSize_t>
2721 The most general way to do this is to cast them to a UV or IV, and
2723 L<previous section|/Formatted Printing of IVs, UVs, and NVs>.
2725 But if you're using C<PerlIO_printf()>, it's less typing and visual
2726 clutter to use the C<"%z"> length modifier (for I<siZe>):
2728 PerlIO_printf("STRLEN is %zu\n", len);
2730 This modifier is not portable, so its use should be restricted to
2733 =head2 Pointer-To-Integer and Integer-To-Pointer
2735 Because pointer size does not necessarily equal integer size,
2736 use the follow macros to do it right.
2741 INT2PTR(pointertotype, integer)
2746 SV *sv = INT2PTR(SV*, iv);
2753 =head2 Exception Handling
2755 There are a couple of macros to do very basic exception handling in XS
2756 modules. You have to define C<NO_XSLOCKS> before including F<XSUB.h> to
2757 be able to use these macros:
2762 You can use these macros if you call code that may croak, but you need
2763 to do some cleanup before giving control back to Perl. For example:
2765 dXCPT; /* set up necessary variables */
2768 code_that_may_croak();
2773 /* do cleanup here */
2777 Note that you always have to rethrow an exception that has been
2778 caught. Using these macros, it is not possible to just catch the
2779 exception and ignore it. If you have to ignore the exception, you
2780 have to use the C<call_*> function.
2782 The advantage of using the above macros is that you don't have
2783 to setup an extra function for C<call_*>, and that using these
2784 macros is faster than using C<call_*>.
2786 =head2 Source Documentation
2788 There's an effort going on to document the internal functions and
2789 automatically produce reference manuals from them -- L<perlapi> is one
2790 such manual which details all the functions which are available to XS
2791 writers. L<perlintern> is the autogenerated manual for the functions
2792 which are not part of the API and are supposedly for internal use only.
2794 Source documentation is created by putting POD comments into the C
2798 =for apidoc sv_setiv
2800 Copies an integer into the given SV. Does not handle 'set' magic. See
2801 L<perlapi/sv_setiv_mg>.
2806 Please try and supply some documentation if you add functions to the
2809 =head2 Backwards compatibility
2811 The Perl API changes over time. New functions are
2812 added or the interfaces of existing functions are
2813 changed. The C<Devel::PPPort> module tries to
2814 provide compatibility code for some of these changes, so XS writers don't
2815 have to code it themselves when supporting multiple versions of Perl.
2817 C<Devel::PPPort> generates a C header file F<ppport.h> that can also
2818 be run as a Perl script. To generate F<ppport.h>, run:
2820 perl -MDevel::PPPort -eDevel::PPPort::WriteFile
2822 Besides checking existing XS code, the script can also be used to retrieve
2823 compatibility information for various API calls using the C<--api-info>
2824 command line switch. For example:
2826 % perl ppport.h --api-info=sv_magicext
2828 For details, see C<perldoc ppport.h>.
2830 =head1 Unicode Support
2832 Perl 5.6.0 introduced Unicode support. It's important for porters and XS
2833 writers to understand this support and make sure that the code they
2834 write does not corrupt Unicode data.
2836 =head2 What B<is> Unicode, anyway?
2838 In the olden, less enlightened times, we all used to use ASCII. Most of
2839 us did, anyway. The big problem with ASCII is that it's American. Well,
2840 no, that's not actually the problem; the problem is that it's not
2841 particularly useful for people who don't use the Roman alphabet. What
2842 used to happen was that particular languages would stick their own
2843 alphabet in the upper range of the sequence, between 128 and 255. Of
2844 course, we then ended up with plenty of variants that weren't quite
2845 ASCII, and the whole point of it being a standard was lost.
2847 Worse still, if you've got a language like Chinese or
2848 Japanese that has hundreds or thousands of characters, then you really
2849 can't fit them into a mere 256, so they had to forget about ASCII
2850 altogether, and build their own systems using pairs of numbers to refer
2853 To fix this, some people formed Unicode, Inc. and
2854 produced a new character set containing all the characters you can
2855 possibly think of and more. There are several ways of representing these
2856 characters, and the one Perl uses is called UTF-8. UTF-8 uses
2857 a variable number of bytes to represent a character. You can learn more
2858 about Unicode and Perl's Unicode model in L<perlunicode>.
2860 (On EBCDIC platforms, Perl uses instead UTF-EBCDIC, which is a form of
2861 UTF-8 adapted for EBCDIC platforms. Below, we just talk about UTF-8.
2862 UTF-EBCDIC is like UTF-8, but the details are different. The macros
2863 hide the differences from you, just remember that the particular numbers
2864 and bit patterns presented below will differ in UTF-EBCDIC.)
2866 =head2 How can I recognise a UTF-8 string?
2868 You can't. This is because UTF-8 data is stored in bytes just like
2869 non-UTF-8 data. The Unicode character 200, (C<0xC8> for you hex types)
2870 capital E with a grave accent, is represented by the two bytes
2871 C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
2872 has that byte sequence as well. So you can't tell just by looking -- this
2873 is what makes Unicode input an interesting problem.
2875 In general, you either have to know what you're dealing with, or you
2876 have to guess. The API function C<is_utf8_string> can help; it'll tell
2877 you if a string contains only valid UTF-8 characters, and the chances
2878 of a non-UTF-8 string looking like valid UTF-8 become very small very
2879 quickly with increasing string length. On a character-by-character
2880 basis, C<isUTF8_CHAR>
2881 will tell you whether the current character in a string is valid UTF-8.
2883 =head2 How does UTF-8 represent Unicode characters?
2885 As mentioned above, UTF-8 uses a variable number of bytes to store a
2886 character. Characters with values 0...127 are stored in one
2887 byte, just like good ol' ASCII. Character 128 is stored as
2888 C<v194.128>; this continues up to character 191, which is
2889 C<v194.191>. Now we've run out of bits (191 is binary
2890 C<10111111>) so we move on; character 192 is C<v195.128>. And
2891 so it goes on, moving to three bytes at character 2048.
2892 L<perlunicode/Unicode Encodings> has pictures of how this works.
2894 Assuming you know you're dealing with a UTF-8 string, you can find out
2895 how long the first character in it is with the C<UTF8SKIP> macro:
2897 char *utf = "\305\233\340\240\201";
2900 len = UTF8SKIP(utf); /* len is 2 here */
2902 len = UTF8SKIP(utf); /* len is 3 here */
2904 Another way to skip over characters in a UTF-8 string is to use
2905 C<utf8_hop>, which takes a string and a number of characters to skip
2906 over. You're on your own about bounds checking, though, so don't use it
2909 All bytes in a multi-byte UTF-8 character will have the high bit set,
2910 so you can test if you need to do something special with this
2911 character like this (the C<UTF8_IS_INVARIANT()> is a macro that tests
2912 whether the byte is encoded as a single byte even in UTF-8):
2914 U8 *utf; /* Initialize this to point to the beginning of the
2915 sequence to convert */
2916 U8 *utf_end; /* Initialize this to 1 beyond the end of the sequence
2917 pointed to by 'utf' */
2918 UV uv; /* Returned code point; note: a UV, not a U8, not a
2920 STRLEN len; /* Returned length of character in bytes */
2922 if (!UTF8_IS_INVARIANT(*utf))
2923 /* Must treat this as UTF-8 */
2924 uv = utf8_to_uvchr_buf(utf, utf_end, &len);
2926 /* OK to treat this character as a byte */
2929 You can also see in that example that we use C<utf8_to_uvchr_buf> to get the
2930 value of the character; the inverse function C<uvchr_to_utf8> is available
2931 for putting a UV into UTF-8:
2933 if (!UVCHR_IS_INVARIANT(uv))
2934 /* Must treat this as UTF8 */
2935 utf8 = uvchr_to_utf8(utf8, uv);
2937 /* OK to treat this character as a byte */
2940 You B<must> convert characters to UVs using the above functions if
2941 you're ever in a situation where you have to match UTF-8 and non-UTF-8
2942 characters. You may not skip over UTF-8 characters in this case. If you
2943 do this, you'll lose the ability to match hi-bit non-UTF-8 characters;
2944 for instance, if your UTF-8 string contains C<v196.172>, and you skip
2945 that character, you can never match a C<chr(200)> in a non-UTF-8 string.
2948 (Note that we don't have to test for invariant characters in the
2949 examples above. The functions work on any well-formed UTF-8 input.
2950 It's just that its faster to avoid the function overhead when it's not
2953 =head2 How does Perl store UTF-8 strings?
2955 Currently, Perl deals with UTF-8 strings and non-UTF-8 strings
2956 slightly differently. A flag in the SV, C<SVf_UTF8>, indicates that the
2957 string is internally encoded as UTF-8. Without it, the byte value is the
2958 codepoint number and vice versa. This flag is only meaningful if the SV
2959 is C<SvPOK> or immediately after stringification via C<SvPV> or a
2960 similar macro. You can check and manipulate this flag with the
2967 This flag has an important effect on Perl's treatment of the string: if
2968 UTF-8 data is not properly distinguished, regular expressions,
2969 C<length>, C<substr> and other string handling operations will have
2970 undesirable (wrong) results.
2972 The problem comes when you have, for instance, a string that isn't
2973 flagged as UTF-8, and contains a byte sequence that could be UTF-8 --
2974 especially when combining non-UTF-8 and UTF-8 strings.
2976 Never forget that the C<SVf_UTF8> flag is separate from the PV value; you
2977 need to be sure you don't accidentally knock it off while you're
2978 manipulating SVs. More specifically, you cannot expect to do this:
2987 nsv = newSVpvn(p, len);
2989 The C<char*> string does not tell you the whole story, and you can't
2990 copy or reconstruct an SV just by copying the string value. Check if the
2991 old SV has the UTF8 flag set (I<after> the C<SvPV> call), and act
2995 is_utf8 = SvUTF8(sv);
2996 frobnicate(p, is_utf8);
2997 nsv = newSVpvn(p, len);
3001 In the above, your C<frobnicate> function has been changed to be made
3002 aware of whether or not it's dealing with UTF-8 data, so that it can
3003 handle the string appropriately.
3005 Since just passing an SV to an XS function and copying the data of
3006 the SV is not enough to copy the UTF8 flags, even less right is just
3007 passing a S<C<char *>> to an XS function.
3009 For full generality, use the L<C<DO_UTF8>|perlapi/DO_UTF8> macro to see if the
3010 string in an SV is to be I<treated> as UTF-8. This takes into account
3011 if the call to the XS function is being made from within the scope of
3012 L<S<C<use bytes>>|bytes>. If so, the underlying bytes that comprise the
3013 UTF-8 string are to be exposed, rather than the character they
3014 represent. But this pragma should only really be used for debugging and
3015 perhaps low-level testing at the byte level. Hence most XS code need
3016 not concern itself with this, but various areas of the perl core do need
3019 And this isn't the whole story. Starting in Perl v5.12, strings that
3020 aren't encoded in UTF-8 may also be treated as Unicode under various
3021 conditions (see L<perlunicode/ASCII Rules versus Unicode Rules>).
3022 This is only really a problem for characters whose ordinals are between
3023 128 and 255, and their behavior varies under ASCII versus Unicode rules
3024 in ways that your code cares about (see L<perlunicode/The "Unicode Bug">).
3025 There is no published API for dealing with this, as it is subject to
3026 change, but you can look at the code for C<pp_lc> in F<pp.c> for an
3027 example as to how it's currently done.
3029 =head2 How do I convert a string to UTF-8?
3031 If you're mixing UTF-8 and non-UTF-8 strings, it is necessary to upgrade
3032 the non-UTF-8 strings to UTF-8. If you've got an SV, the easiest way to do
3035 sv_utf8_upgrade(sv);
3037 However, you must not do this, for example:
3040 sv_utf8_upgrade(left);
3042 If you do this in a binary operator, you will actually change one of the
3043 strings that came into the operator, and, while it shouldn't be noticeable
3044 by the end user, it can cause problems in deficient code.
3046 Instead, C<bytes_to_utf8> will give you a UTF-8-encoded B<copy> of its
3047 string argument. This is useful for having the data available for
3048 comparisons and so on, without harming the original SV. There's also
3049 C<utf8_to_bytes> to go the other way, but naturally, this will fail if
3050 the string contains any characters above 255 that can't be represented
3053 =head2 How do I compare strings?
3055 L<perlapi/sv_cmp> and L<perlapi/sv_cmp_flags> do a lexigraphic
3056 comparison of two SV's, and handle UTF-8ness properly. Note, however,
3057 that Unicode specifies a much fancier mechanism for collation, available
3058 via the L<Unicode::Collate> module.
3060 To just compare two strings for equality/non-equality, you can just use
3061 L<C<memEQ()>|perlapi/memEQ> and L<C<memNE()>|perlapi/memEQ> as usual,
3062 except the strings must be both UTF-8 or not UTF-8 encoded.
3064 To compare two strings case-insensitively, use
3065 L<C<foldEQ_utf8()>|perlapi/foldEQ_utf8> (the strings don't have to have
3066 the same UTF-8ness).
3068 =head2 Is there anything else I need to know?
3070 Not really. Just remember these things:
3076 There's no way to tell if a S<C<char *>> or S<C<U8 *>> string is UTF-8
3077 or not. But you can tell if an SV is to be treated as UTF-8 by calling
3078 C<DO_UTF8> on it, after stringifying it with C<SvPV> or a similar
3079 macro. And, you can tell if SV is actually UTF-8 (even if it is not to
3080 be treated as such) by looking at its C<SvUTF8> flag (again after
3081 stringifying it). Don't forget to set the flag if something should be
3083 Treat the flag as part of the PV, even though it's not -- if you pass on
3084 the PV to somewhere, pass on the flag too.
3088 If a string is UTF-8, B<always> use C<utf8_to_uvchr_buf> to get at the value,
3089 unless C<UTF8_IS_INVARIANT(*s)> in which case you can use C<*s>.
3093 When writing a character UV to a UTF-8 string, B<always> use
3094 C<uvchr_to_utf8>, unless C<UVCHR_IS_INVARIANT(uv))> in which case
3095 you can use C<*s = uv>.
3099 Mixing UTF-8 and non-UTF-8 strings is
3100 tricky. Use C<bytes_to_utf8> to get
3101 a new string which is UTF-8 encoded, and then combine them.
3105 =head1 Custom Operators
3107 Custom operator support is an experimental feature that allows you to
3108 define your own ops. This is primarily to allow the building of
3109 interpreters for other languages in the Perl core, but it also allows
3110 optimizations through the creation of "macro-ops" (ops which perform the
3111 functions of multiple ops which are usually executed together, such as
3112 C<gvsv, gvsv, add>.)
3114 This feature is implemented as a new op type, C<OP_CUSTOM>. The Perl
3115 core does not "know" anything special about this op type, and so it will
3116 not be involved in any optimizations. This also means that you can
3117 define your custom ops to be any op structure -- unary, binary, list and
3120 It's important to know what custom operators won't do for you. They
3121 won't let you add new syntax to Perl, directly. They won't even let you
3122 add new keywords, directly. In fact, they won't change the way Perl
3123 compiles a program at all. You have to do those changes yourself, after
3124 Perl has compiled the program. You do this either by manipulating the op
3125 tree using a C<CHECK> block and the C<B::Generate> module, or by adding
3126 a custom peephole optimizer with the C<optimize> module.
3128 When you do this, you replace ordinary Perl ops with custom ops by
3129 creating ops with the type C<OP_CUSTOM> and the C<op_ppaddr> of your own
3130 PP function. This should be defined in XS code, and should look like
3131 the PP ops in C<pp_*.c>. You are responsible for ensuring that your op
3132 takes the appropriate number of values from the stack, and you are
3133 responsible for adding stack marks if necessary.
3135 You should also "register" your op with the Perl interpreter so that it
3136 can produce sensible error and warning messages. Since it is possible to
3137 have multiple custom ops within the one "logical" op type C<OP_CUSTOM>,
3138 Perl uses the value of C<< o->op_ppaddr >> to determine which custom op
3139 it is dealing with. You should create an C<XOP> structure for each
3140 ppaddr you use, set the properties of the custom op with
3141 C<XopENTRY_set>, and register the structure against the ppaddr using
3142 C<Perl_custom_op_register>. A trivial example might look like:
3145 static OP *my_pp(pTHX);
3148 XopENTRY_set(&my_xop, xop_name, "myxop");
3149 XopENTRY_set(&my_xop, xop_desc, "Useless custom op");
3150 Perl_custom_op_register(aTHX_ my_pp, &my_xop);
3152 The available fields in the structure are:
3158 A short name for your op. This will be included in some error messages,
3159 and will also be returned as C<< $op->name >> by the L<B|B> module, so
3160 it will appear in the output of module like L<B::Concise|B::Concise>.
3164 A short description of the function of the op.
3168 Which of the various C<*OP> structures this op uses. This should be one of
3169 the C<OA_*> constants from F<op.h>, namely
3189 =item OA_PVOP_OR_SVOP
3191 This should be interpreted as 'C<PVOP>' only. The C<_OR_SVOP> is because
3192 the only core C<PVOP>, C<OP_TRANS>, can sometimes be a C<SVOP> instead.
3200 The other C<OA_*> constants should not be used.
3204 This member is of type C<Perl_cpeep_t>, which expands to C<void
3205 (*Perl_cpeep_t)(aTHX_ OP *o, OP *oldop)>. If it is set, this function
3206 will be called from C<Perl_rpeep> when ops of this type are encountered
3207 by the peephole optimizer. I<o> is the OP that needs optimizing;
3208 I<oldop> is the previous OP optimized, whose C<op_next> points to I<o>.
3212 C<B::Generate> directly supports the creation of custom ops by name.
3215 =head1 Dynamic Scope and the Context Stack
3217 B<Note:> this section describes a non-public internal API that is subject
3218 to change without notice.
3220 =head2 Introduction to the context stack
3222 In Perl, dynamic scoping refers to the runtime nesting of things like
3223 subroutine calls, evals etc, as well as the entering and exiting of block
3224 scopes. For example, the restoring of a C<local>ised variable is
3225 determined by the dynamic scope.
3227 Perl tracks the dynamic scope by a data structure called the context
3228 stack, which is an array of C<PERL_CONTEXT> structures, and which is
3229 itself a big union for all the types of context. Whenever a new scope is
3230 entered (such as a block, a C<for> loop, or a subroutine call), a new
3231 context entry is pushed onto the stack. Similarly when leaving a block or
3232 returning from a subroutine call etc. a context is popped. Since the
3233 context stack represents the current dynamic scope, it can be searched.
3234 For example, C<next LABEL> searches back through the stack looking for a
3235 loop context that matches the label; C<return> pops contexts until it
3236 finds a sub or eval context or similar; C<caller> examines sub contexts on
3239 Each context entry is labelled with a context type, C<cx_type>. Typical
3240 context types are C<CXt_SUB>, C<CXt_EVAL> etc., as well as C<CXt_BLOCK>
3241 and C<CXt_NULL> which represent a basic scope (as pushed by C<pp_enter>)
3242 and a sort block. The type determines which part of the context union are
3245 The main division in the context struct is between a substitution scope
3246 (C<CXt_SUBST>) and block scopes, which are everything else. The former is
3247 just used while executing C<s///e>, and won't be discussed further
3250 All the block scope types share a common base, which corresponds to
3251 C<CXt_BLOCK>. This stores the old values of various scope-related
3252 variables like C<PL_curpm>, as well as information about the current
3253 scope, such as C<gimme>. On scope exit, the old variables are restored.
3255 Particular block scope types store extra per-type information. For
3256 example, C<CXt_SUB> stores the currently executing CV, while the various
3257 for loop types might hold the original loop variable SV. On scope exit,
3258 the per-type data is processed; for example the CV has its reference count
3259 decremented, and the original loop variable is restored.
3261 The macro C<cxstack> returns the base of the current context stack, while
3262 C<cxstack_ix> is the index of the current frame within that stack.
3264 In fact, the context stack is actually part of a stack-of-stacks system;
3265 whenever something unusual is done such as calling a C<DESTROY> or tie
3266 handler, a new stack is pushed, then popped at the end.
3268 Note that the API described here changed considerably in perl 5.24; prior
3269 to that, big macros like C<PUSHBLOCK> and C<POPSUB> were used; in 5.24
3270 they were replaced by the inline static functions described below. In
3271 addition, the ordering and detail of how these macros/function work
3272 changed in many ways, often subtly. In particular they didn't handle
3273 saving the savestack and temps stack positions, and required additional
3274 C<ENTER>, C<SAVETMPS> and C<LEAVE> compared to the new functions. The
3275 old-style macros will not be described further.
3278 =head2 Pushing contexts
3280 For pushing a new context, the two basic functions are
3281 C<cx = cx_pushblock()>, which pushes a new basic context block and returns
3282 its address, and a family of similar functions with names like
3283 C<cx_pushsub(cx)> which populate the additional type-dependent fields in
3284 the C<cx> struct. Note that C<CXt_NULL> and C<CXt_BLOCK> don't have their
3285 own push functions, as they don't store any data beyond that pushed by
3288 The fields of the context struct and the arguments to the C<cx_*>
3289 functions are subject to change between perl releases, representing
3290 whatever is convenient or efficient for that release.
3292 A typical context stack pushing can be found in C<pp_entersub>; the
3293 following shows a simplified and stripped-down example of a non-XS call,
3294 along with comments showing roughly what each function does.
3298 bool hasargs = cBOOL(PL_op->op_flags & OPf_STACKED);
3299 OP *retop = PL_op->op_next;
3300 I32 old_ss_ix = PL_savestack_ix;
3303 /* ... make mortal copies of stack args which are PADTMPs here ... */
3305 /* ... do any additional savestack pushes here ... */
3307 /* Now push a new context entry of type 'CXt_SUB'; initially just
3308 * doing the actions common to all block types: */
3310 cx = cx_pushblock(CXt_SUB, gimme, MARK, old_ss_ix);
3312 /* this does (approximately):
3313 CXINC; /* cxstack_ix++ (grow if necessary) */
3314 cx = CX_CUR(); /* and get the address of new frame */
3315 cx->cx_type = CXt_SUB;
3316 cx->blk_gimme = gimme;
3317 cx->blk_oldsp = MARK - PL_stack_base;
3318 cx->blk_oldsaveix = old_ss_ix;
3319 cx->blk_oldcop = PL_curcop;
3320 cx->blk_oldmarksp = PL_markstack_ptr - PL_markstack;
3321 cx->blk_oldscopesp = PL_scopestack_ix;
3322 cx->blk_oldpm = PL_curpm;
3323 cx->blk_old_tmpsfloor = PL_tmps_floor;
3325 PL_tmps_floor = PL_tmps_ix;
3329 /* then update the new context frame with subroutine-specific info,
3330 * such as the CV about to be executed: */
3332 cx_pushsub(cx, cv, retop, hasargs);
3334 /* this does (approximately):
3335 cx->blk_sub.cv = cv;
3336 cx->blk_sub.olddepth = CvDEPTH(cv);
3337 cx->blk_sub.prevcomppad = PL_comppad;
3338 cx->cx_type |= (hasargs) ? CXp_HASARGS : 0;
3339 cx->blk_sub.retop = retop;
3340 SvREFCNT_inc_simple_void_NN(cv);
3343 Note that C<cx_pushblock()> sets two new floors: for the args stack (to
3344 C<MARK>) and the temps stack (to C<PL_tmps_ix>). While executing at this
3345 scope level, every C<nextstate> (amongst others) will reset the args and
3346 tmps stack levels to these floors. Note that since C<cx_pushblock> uses
3347 the current value of C<PL_tmps_ix> rather than it being passed as an arg,
3348 this dictates at what point C<cx_pushblock> should be called. In
3349 particular, any new mortals which should be freed only on scope exit
3350 (rather than at the next C<nextstate>) should be created first.
3352 Most callers of C<cx_pushblock> simply set the new args stack floor to the
3353 top of the previous stack frame, but for C<CXt_LOOP_LIST> it stores the
3354 items being iterated over on the stack, and so sets C<blk_oldsp> to the
3355 top of these items instead. Note that, contrary to its name, C<blk_oldsp>
3356 doesn't always represent the value to restore C<PL_stack_sp> to on scope
3359 Note the early capture of C<PL_savestack_ix> to C<old_ss_ix>, which is
3360 later passed as an arg to C<cx_pushblock>. In the case of C<pp_entersub>,
3361 this is because, although most values needing saving are stored in fields
3362 of the context struct, an extra value needs saving only when the debugger
3363 is running, and it doesn't make sense to bloat the struct for this rare
3364 case. So instead it is saved on the savestack. Since this value gets
3365 calculated and saved before the context is pushed, it is necessary to pass
3366 the old value of C<PL_savestack_ix> to C<cx_pushblock>, to ensure that the
3367 saved value gets freed during scope exit. For most users of
3368 C<cx_pushblock>, where nothing needs pushing on the save stack,
3369 C<PL_savestack_ix> is just passed directly as an arg to C<cx_pushblock>.
3371 Note that where possible, values should be saved in the context struct
3372 rather than on the save stack; it's much faster that way.
3374 Normally C<cx_pushblock> should be immediately followed by the appropriate
3375 C<cx_pushfoo>, with nothing between them; this is because if code
3376 in-between could die (e.g. a warning upgraded to fatal), then the context
3377 stack unwinding code in C<dounwind> would see (in the example above) a
3378 C<CXt_SUB> context frame, but without all the subroutine-specific fields
3379 set, and crashes would soon ensue.
3381 Where the two must be separate, initially set the type to C<CXt_NULL> or
3382 C<CXt_BLOCK>, and later change it to C<CXt_foo> when doing the
3383 C<cx_pushfoo>. This is exactly what C<pp_enteriter> does, once it's
3384 determined which type of loop it's pushing.
3386 =head2 Popping contexts
3388 Contexts are popped using C<cx_popsub()> etc. and C<cx_popblock()>. Note
3389 however, that unlike C<cx_pushblock>, neither of these functions actually
3390 decrement the current context stack index; this is done separately using
3393 There are two main ways that contexts are popped. During normal execution
3394 as scopes are exited, functions like C<pp_leave>, C<pp_leaveloop> and
3395 C<pp_leavesub> process and pop just one context using C<cx_popfoo> and
3396 C<cx_popblock>. On the other hand, things like C<pp_return> and C<next>
3397 may have to pop back several scopes until a sub or loop context is found,
3398 and exceptions (such as C<die>) need to pop back contexts until an eval
3399 context is found. Both of these are accomplished by C<dounwind()>, which
3400 is capable of processing and popping all contexts above the target one.
3402 Here is a typical example of context popping, as found in C<pp_leavesub>
3403 (simplified slightly):
3412 gimme = cx->blk_gimme;
3413 oldsp = PL_stack_base + cx->blk_oldsp; /* last arg of previous frame */
3415 if (gimme == G_VOID)
3416 PL_stack_sp = oldsp;
3418 leave_adjust_stacks(oldsp, oldsp, gimme, 0);
3423 retop = cx->blk_sub.retop;
3428 The steps above are in a very specific order, designed to be the reverse
3429 order of when the context was pushed. The first thing to do is to copy
3430 and/or protect any any return arguments and free any temps in the current
3431 scope. Scope exits like an rvalue sub normally return a mortal copy of
3432 their return args (as opposed to lvalue subs). It is important to make
3433 this copy before the save stack is popped or variables are restored, or
3434 bad things like the following can happen:
3436 sub f { my $x =...; $x } # $x freed before we get to copy it
3437 sub f { /(...)/; $1 } # PL_curpm restored before $1 copied
3439 Although we wish to free any temps at the same time, we have to be careful
3440 not to free any temps which are keeping return args alive; nor to free the
3441 temps we have just created while mortal copying return args. Fortunately,
3442 C<leave_adjust_stacks()> is capable of making mortal copies of return args,
3443 shifting args down the stack, and only processing those entries on the
3444 temps stack that are safe to do so.
3446 In void context no args are returned, so it's more efficient to skip
3447 calling C<leave_adjust_stacks()>. Also in void context, a C<nextstate> op
3448 is likely to be imminently called which will do a C<FREETMPS>, so there's
3449 no need to do that either.
3451 The next step is to pop savestack entries: C<CX_LEAVE_SCOPE(cx)> is just
3452 defined as C<< LEAVE_SCOPE(cx->blk_oldsaveix) >>. Note that during the
3453 popping, it's possible for perl to call destructors, call C<STORE> to undo
3454 localisations of tied vars, and so on. Any of these can die or call
3455 C<exit()>. In this case, C<dounwind()> will be called, and the current
3456 context stack frame will be re-processed. Thus it is vital that all steps
3457 in popping a context are done in such a way to support reentrancy. The
3458 other alternative, of decrementing C<cxstack_ix> I<before> processing the
3459 frame, would lead to leaks and the like if something died halfway through,
3460 or overwriting of the current frame.
3462 C<CX_LEAVE_SCOPE> itself is safely re-entrant: if only half the savestack
3463 items have been popped before dying and getting trapped by eval, then the
3464 C<CX_LEAVE_SCOPE>s in C<dounwind> or C<pp_leaveeval> will continue where
3465 the first one left off.
3467 The next step is the type-specific context processing; in this case
3468 C<cx_popsub>. In part, this looks like:
3470 cv = cx->blk_sub.cv;
3471 CvDEPTH(cv) = cx->blk_sub.olddepth;
3472 cx->blk_sub.cv = NULL;
3475 where its processing the just-executed CV. Note that before it decrements
3476 the CV's reference count, it nulls the C<blk_sub.cv>. This means that if
3477 it re-enters, the CV won't be freed twice. It also means that you can't
3478 rely on such type-specific fields having useful values after the return
3481 Next, C<cx_popblock> restores all the various interpreter vars to their
3482 previous values or previous high water marks; it expands to:
3484 PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
3485 PL_scopestack_ix = cx->blk_oldscopesp;
3486 PL_curpm = cx->blk_oldpm;
3487 PL_curcop = cx->blk_oldcop;
3488 PL_tmps_floor = cx->blk_old_tmpsfloor;
3490 Note that it I<doesn't> restore C<PL_stack_sp>; as mentioned earlier,
3491 which value to restore it to depends on the context type (specifically
3492 C<for (list) {}>), and what args (if any) it returns; and that will
3493 already have been sorted out earlier by C<leave_adjust_stacks()>.
3495 Finally, the context stack pointer is actually decremented by C<CX_POP(cx)>.
3496 After this point, it's possible that that the current context frame could
3497 be overwritten by other contexts being pushed. Although things like ties
3498 and C<DESTROY> are supposed to work within a new context stack, it's best
3499 not to assume this. Indeed on debugging builds, C<CX_POP(cx)> deliberately
3500 sets C<cx> to null to detect code that is still relying on the field
3501 values in that context frame. Note in the C<pp_leavesub()> example above,
3502 we grab C<blk_sub.retop> I<before> calling C<CX_POP>.
3504 =head2 Redoing contexts
3506 Finally, there is C<cx_topblock(cx)>, which acts like a super-C<nextstate>
3507 as regards to resetting various vars to their base values. It is used in
3508 places like C<pp_next>, C<pp_redo> and C<pp_goto> where rather than
3509 exiting a scope, we want to re-initialise the scope. As well as resetting
3510 C<PL_stack_sp> like C<nextstate>, it also resets C<PL_markstack_ptr>,
3511 C<PL_scopestack_ix> and C<PL_curpm>. Note that it doesn't do a
3517 Until May 1997, this document was maintained by Jeff Okamoto
3518 E<lt>okamoto@corp.hp.comE<gt>. It is now maintained as part of Perl
3519 itself by the Perl 5 Porters E<lt>perl5-porters@perl.orgE<gt>.
3521 With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
3522 Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
3523 Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
3524 Stephen McCamant, and Gurusamy Sarathy.
3528 L<perlapi>, L<perlintern>, L<perlxs>, L<perlembed>