3 perlguts - Introduction to the Perl API
7 This document attempts to describe how to use the Perl API, as well as
8 to provide some info on the basic workings of the Perl core. It is far
9 from complete and probably contains many errors. Please refer any
10 questions or comments to the author below.
16 Perl has three typedefs that handle Perl's three main data types:
22 Each typedef has specific routines that manipulate the various data types.
24 =head2 What is an "IV"?
26 Perl uses a special typedef IV which is a simple signed integer type that is
27 guaranteed to be large enough to hold a pointer (as well as an integer).
28 Additionally, there is the UV, which is simply an unsigned IV.
30 Perl also uses two special typedefs, I32 and I16, which will always be at
31 least 32-bits and 16-bits long, respectively. (Again, there are U32 and U16,
32 as well.) They will usually be exactly 32 and 16 bits long, but on Crays
33 they will both be 64 bits.
35 =head2 Working with SVs
37 An SV can be created and loaded with one command. There are five types of
38 values that can be loaded: an integer value (IV), an unsigned integer
39 value (UV), a double (NV), a string (PV), and another scalar (SV).
40 ("PV" stands for "Pointer Value". You might think that it is misnamed
41 because it is described as pointing only to strings. However, it is
42 possible to have it point to other things. For example, it could point
43 to an array of UVs. But,
44 using it for non-strings requires care, as the underlying assumption of
45 much of the internals is that PVs are just for strings. Often, for
46 example, a trailing C<NUL> is tacked on automatically. The non-string use
47 is documented only in this paragraph.)
49 The seven routines are:
54 SV* newSVpv(const char*, STRLEN);
55 SV* newSVpvn(const char*, STRLEN);
56 SV* newSVpvf(const char*, ...);
59 C<STRLEN> is an integer type (Size_t, usually defined as size_t in
60 F<config.h>) guaranteed to be large enough to represent the size of
61 any string that perl can handle.
63 In the unlikely case of a SV requiring more complex initialization, you
64 can create an empty SV with newSV(len). If C<len> is 0 an empty SV of
65 type NULL is returned, else an SV of type PV is returned with len + 1 (for
66 the C<NUL>) bytes of storage allocated, accessible via SvPVX. In both cases
67 the SV has the undef value.
69 SV *sv = newSV(0); /* no storage allocated */
70 SV *sv = newSV(10); /* 10 (+1) bytes of uninitialised storage
73 To change the value of an I<already-existing> SV, there are eight routines:
75 void sv_setiv(SV*, IV);
76 void sv_setuv(SV*, UV);
77 void sv_setnv(SV*, double);
78 void sv_setpv(SV*, const char*);
79 void sv_setpvn(SV*, const char*, STRLEN)
80 void sv_setpvf(SV*, const char*, ...);
81 void sv_vsetpvfn(SV*, const char*, STRLEN, va_list *,
83 void sv_setsv(SV*, SV*);
85 Notice that you can choose to specify the length of the string to be
86 assigned by using C<sv_setpvn>, C<newSVpvn>, or C<newSVpv>, or you may
87 allow Perl to calculate the length by using C<sv_setpv> or by specifying
88 0 as the second argument to C<newSVpv>. Be warned, though, that Perl will
89 determine the string's length by using C<strlen>, which depends on the
90 string terminating with a C<NUL> character, and not otherwise containing
93 The arguments of C<sv_setpvf> are processed like C<sprintf>, and the
94 formatted output becomes the value.
96 C<sv_vsetpvfn> is an analogue of C<vsprintf>, but it allows you to specify
97 either a pointer to a variable argument list or the address and length of
98 an array of SVs. The last argument points to a boolean; on return, if that
99 boolean is true, then locale-specific information has been used to format
100 the string, and the string's contents are therefore untrustworthy (see
101 L<perlsec>). This pointer may be NULL if that information is not
102 important. Note that this function requires you to specify the length of
105 The C<sv_set*()> functions are not generic enough to operate on values
106 that have "magic". See L<Magic Virtual Tables> later in this document.
108 All SVs that contain strings should be terminated with a C<NUL> character.
109 If it is not C<NUL>-terminated there is a risk of
110 core dumps and corruptions from code which passes the string to C
111 functions or system calls which expect a C<NUL>-terminated string.
112 Perl's own functions typically add a trailing C<NUL> for this reason.
113 Nevertheless, you should be very careful when you pass a string stored
114 in an SV to a C function or system call.
116 To access the actual value that an SV points to, you can use the macros:
121 SvPV(SV*, STRLEN len)
124 which will automatically coerce the actual scalar type into an IV, UV, double,
127 In the C<SvPV> macro, the length of the string returned is placed into the
128 variable C<len> (this is a macro, so you do I<not> use C<&len>). If you do
129 not care what the length of the data is, use the C<SvPV_nolen> macro.
130 Historically the C<SvPV> macro with the global variable C<PL_na> has been
131 used in this case. But that can be quite inefficient because C<PL_na> must
132 be accessed in thread-local storage in threaded Perl. In any case, remember
133 that Perl allows arbitrary strings of data that may both contain NULs and
134 might not be terminated by a C<NUL>.
136 Also remember that C doesn't allow you to safely say C<foo(SvPV(s, len),
137 len);>. It might work with your
138 compiler, but it won't work for everyone.
139 Break this sort of statement up into separate assignments:
147 If you want to know if the scalar value is TRUE, you can use:
151 Although Perl will automatically grow strings for you, if you need to force
152 Perl to allocate more memory for your SV, you can use the macro
154 SvGROW(SV*, STRLEN newlen)
156 which will determine if more memory needs to be allocated. If so, it will
157 call the function C<sv_grow>. Note that C<SvGROW> can only increase, not
158 decrease, the allocated memory of an SV and that it does not automatically
159 add space for the trailing C<NUL> byte (perl's own string functions typically do
160 C<SvGROW(sv, len + 1)>).
162 If you want to write to an existing SV's buffer and set its value to a
163 string, use SvPV_force() or one of its variants to force the SV to be
164 a PV. This will remove any of various types of non-stringness from
165 the SV while preserving the content of the SV in the PV. This can be
166 used, for example, to append data from an API function to a buffer
167 without extra copying:
169 (void)SvPVbyte_force(sv, len);
170 s = SvGROW(sv, len + needlen + 1);
171 /* something that modifies up to needlen bytes at s+len, but
172 modifies newlen bytes
173 eg. newlen = read(fd, s + len, needlen);
174 ignoring errors for these examples
176 s[len + newlen] = '\0';
177 SvCUR_set(sv, len + newlen);
181 If you already have the data in memory or if you want to keep your
182 code simple, you can use one of the sv_cat*() variants, such as
183 sv_catpvn(). If you want to insert anywhere in the string you can use
184 sv_insert() or sv_insert_flags().
186 If you don't need the existing content of the SV, you can avoid some
189 sv_setpvn(sv, "", 0);
190 s = SvGROW(sv, needlen + 1);
191 /* something that modifies up to needlen bytes at s, but modifies
193 eg. newlen = read(fd, s. needlen);
196 SvCUR_set(sv, newlen);
197 SvPOK_only(sv); /* also clears SVf_UTF8 */
200 Again, if you already have the data in memory or want to avoid the
201 complexity of the above, you can use sv_setpvn().
203 If you have a buffer allocated with Newx() and want to set that as the
204 SV's value, you can use sv_usepvn_flags(). That has some requirements
205 if you want to avoid perl re-allocating the buffer to fit the trailing
208 Newx(buf, somesize+1, char);
209 /* ... fill in buf ... */
210 buf[somesize] = '\0';
211 sv_usepvn_flags(sv, buf, somesize, SV_SMAGIC | SV_HAS_TRAILING_NUL);
212 /* buf now belongs to perl, don't release it */
214 If you have an SV and want to know what kind of data Perl thinks is stored
215 in it, you can use the following macros to check the type of SV you have.
221 You can get and set the current length of the string stored in an SV with
222 the following macros:
225 SvCUR_set(SV*, I32 val)
227 You can also get a pointer to the end of the string stored in the SV
232 But note that these last three macros are valid only if C<SvPOK()> is true.
234 If you want to append something to the end of string stored in an C<SV*>,
235 you can use the following functions:
237 void sv_catpv(SV*, const char*);
238 void sv_catpvn(SV*, const char*, STRLEN);
239 void sv_catpvf(SV*, const char*, ...);
240 void sv_vcatpvfn(SV*, const char*, STRLEN, va_list *, SV **,
242 void sv_catsv(SV*, SV*);
244 The first function calculates the length of the string to be appended by
245 using C<strlen>. In the second, you specify the length of the string
246 yourself. The third function processes its arguments like C<sprintf> and
247 appends the formatted output. The fourth function works like C<vsprintf>.
248 You can specify the address and length of an array of SVs instead of the
249 va_list argument. The fifth function
250 extends the string stored in the first
251 SV with the string stored in the second SV. It also forces the second SV
252 to be interpreted as a string.
254 The C<sv_cat*()> functions are not generic enough to operate on values that
255 have "magic". See L<Magic Virtual Tables> later in this document.
257 If you know the name of a scalar variable, you can get a pointer to its SV
258 by using the following:
260 SV* get_sv("package::varname", 0);
262 This returns NULL if the variable does not exist.
264 If you want to know if this variable (or any other SV) is actually C<defined>,
269 The scalar C<undef> value is stored in an SV instance called C<PL_sv_undef>.
271 Its address can be used whenever an C<SV*> is needed. Make sure that
272 you don't try to compare a random sv with C<&PL_sv_undef>. For example
273 when interfacing Perl code, it'll work correctly for:
277 But won't work when called as:
282 So to repeat always use SvOK() to check whether an sv is defined.
284 Also you have to be careful when using C<&PL_sv_undef> as a value in
285 AVs or HVs (see L<AVs, HVs and undefined values>).
287 There are also the two values C<PL_sv_yes> and C<PL_sv_no>, which contain
288 boolean TRUE and FALSE values, respectively. Like C<PL_sv_undef>, their
289 addresses can be used whenever an C<SV*> is needed.
291 Do not be fooled into thinking that C<(SV *) 0> is the same as C<&PL_sv_undef>.
295 if (I-am-to-return-a-real-value) {
296 sv = sv_2mortal(newSViv(42));
300 This code tries to return a new SV (which contains the value 42) if it should
301 return a real value, or undef otherwise. Instead it has returned a NULL
302 pointer which, somewhere down the line, will cause a segmentation violation,
303 bus error, or just weird results. Change the zero to C<&PL_sv_undef> in the
304 first line and all will be well.
306 To free an SV that you've created, call C<SvREFCNT_dec(SV*)>. Normally this
307 call is not necessary (see L<Reference Counts and Mortality>).
311 Perl provides the function C<sv_chop> to efficiently remove characters
312 from the beginning of a string; you give it an SV and a pointer to
313 somewhere inside the PV, and it discards everything before the
314 pointer. The efficiency comes by means of a little hack: instead of
315 actually removing the characters, C<sv_chop> sets the flag C<OOK>
316 (offset OK) to signal to other functions that the offset hack is in
317 effect, and it moves the PV pointer (called C<SvPVX>) forward
318 by the number of bytes chopped off, and adjusts C<SvCUR> and C<SvLEN>
319 accordingly. (A portion of the space between the old and new PV
320 pointers is used to store the count of chopped bytes.)
322 Hence, at this point, the start of the buffer that we allocated lives
323 at C<SvPVX(sv) - SvIV(sv)> in memory and the PV pointer is pointing
324 into the middle of this allocated storage.
326 This is best demonstrated by example. 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"; $a=~s/.//; Dump($a)'
345 SV = PV(0x7ffa04008a70) at 0x7ffa04030390
347 FLAGS = (POK,OOK,pPOK)
349 PV = 0x7ffa03c05b61 ( "\1" . ) "23456789"\0
353 Here the number of bytes chopped off (1) is shown next as the OFFSET. The
354 portion of the string between the "real" and the "fake" beginnings is
355 shown in parentheses, and the values of C<SvCUR> and C<SvLEN> reflect
356 the fake beginning, not the real one. (The first character of the string
357 buffer happens to have changed to "\1" here, not "1", because the current
358 implementation stores the offset count in the string buffer. This is
361 Something similar to the offset hack is performed on AVs to enable
362 efficient shifting and splicing off the beginning of the array; while
363 C<AvARRAY> points to the first element in the array that is visible from
364 Perl, C<AvALLOC> points to the real start of the C array. These are
365 usually the same, but a C<shift> operation can be carried out by
366 increasing C<AvARRAY> by one and decreasing C<AvFILL> and C<AvMAX>.
367 Again, the location of the real start of the C array only comes into
368 play when freeing the array. See C<av_shift> in F<av.c>.
370 =head2 What's Really Stored in an SV?
372 Recall that the usual method of determining the type of scalar you have is
373 to use C<Sv*OK> macros. Because a scalar can be both a number and a string,
374 usually these macros will always return TRUE and calling the C<Sv*V>
375 macros will do the appropriate conversion of string to integer/double or
376 integer/double to string.
378 If you I<really> need to know if you have an integer, double, or string
379 pointer in an SV, you can use the following three macros instead:
385 These will tell you if you truly have an integer, double, or string pointer
386 stored in your SV. The "p" stands for private.
388 There are various ways in which the private and public flags may differ.
389 For example, in perl 5.16 and earlier a tied SV may have a valid
390 underlying value in the IV slot (so SvIOKp is true), but the data
391 should be accessed via the FETCH routine rather than directly,
392 so SvIOK is false. (In perl 5.18 onwards, tied scalars use
393 the flags the same way as untied scalars.) Another is when
394 numeric conversion has occurred and precision has been lost: only the
395 private flag is set on 'lossy' values. So when an NV is converted to an
396 IV with loss, SvIOKp, SvNOKp and SvNOK will be set, while SvIOK wont be.
398 In general, though, it's best to use the C<Sv*V> macros.
400 =head2 Working with AVs
402 There are two ways to create and load an AV. The first method creates an
407 The second method both creates the AV and initially populates it with SVs:
409 AV* av_make(SSize_t num, SV **ptr);
411 The second argument points to an array containing C<num> C<SV*>'s. Once the
412 AV has been created, the SVs can be destroyed, if so desired.
414 Once the AV has been created, the following operations are possible on it:
416 void av_push(AV*, SV*);
419 void av_unshift(AV*, SSize_t num);
421 These should be familiar operations, with the exception of C<av_unshift>.
422 This routine adds C<num> elements at the front of the array with the C<undef>
423 value. You must then use C<av_store> (described below) to assign values
424 to these new elements.
426 Here are some other functions:
428 SSize_t av_top_index(AV*);
429 SV** av_fetch(AV*, SSize_t key, I32 lval);
430 SV** av_store(AV*, SSize_t key, SV* val);
432 The C<av_top_index> function returns the highest index value in an array (just
433 like $#array in Perl). If the array is empty, -1 is returned. The
434 C<av_fetch> function returns the value at index C<key>, but if C<lval>
435 is non-zero, then C<av_fetch> will store an undef value at that index.
436 The C<av_store> function stores the value C<val> at index C<key>, and does
437 not increment the reference count of C<val>. Thus the caller is responsible
438 for taking care of that, and if C<av_store> returns NULL, the caller will
439 have to decrement the reference count to avoid a memory leak. Note that
440 C<av_fetch> and C<av_store> both return C<SV**>'s, not C<SV*>'s as their
447 void av_extend(AV*, SSize_t key);
449 The C<av_clear> function deletes all the elements in the AV* array, but
450 does not actually delete the array itself. The C<av_undef> function will
451 delete all the elements in the array plus the array itself. The
452 C<av_extend> function extends the array so that it contains at least C<key+1>
453 elements. If C<key+1> is less than the currently allocated length of the array,
454 then nothing is done.
456 If you know the name of an array variable, you can get a pointer to its AV
457 by using the following:
459 AV* get_av("package::varname", 0);
461 This returns NULL if the variable does not exist.
463 See L<Understanding the Magic of Tied Hashes and Arrays> for more
464 information on how to use the array access functions on tied arrays.
466 =head2 Working with HVs
468 To create an HV, you use the following routine:
472 Once the HV has been created, the following operations are possible on it:
474 SV** hv_store(HV*, const char* key, U32 klen, SV* val, U32 hash);
475 SV** hv_fetch(HV*, const char* key, U32 klen, I32 lval);
477 The C<klen> parameter is the length of the key being passed in (Note that
478 you cannot pass 0 in as a value of C<klen> to tell Perl to measure the
479 length of the key). The C<val> argument contains the SV pointer to the
480 scalar being stored, and C<hash> is the precomputed hash value (zero if
481 you want C<hv_store> to calculate it for you). The C<lval> parameter
482 indicates whether this fetch is actually a part of a store operation, in
483 which case a new undefined value will be added to the HV with the supplied
484 key and C<hv_fetch> will return as if the value had already existed.
486 Remember that C<hv_store> and C<hv_fetch> return C<SV**>'s and not just
487 C<SV*>. To access the scalar value, you must first dereference the return
488 value. However, you should check to make sure that the return value is
489 not NULL before dereferencing it.
491 The first of these two functions checks if a hash table entry exists, and the
494 bool hv_exists(HV*, const char* key, U32 klen);
495 SV* hv_delete(HV*, const char* key, U32 klen, I32 flags);
497 If C<flags> does not include the C<G_DISCARD> flag then C<hv_delete> will
498 create and return a mortal copy of the deleted value.
500 And more miscellaneous functions:
505 Like their AV counterparts, C<hv_clear> deletes all the entries in the hash
506 table but does not actually delete the hash table. The C<hv_undef> deletes
507 both the entries and the hash table itself.
509 Perl keeps the actual data in a linked list of structures with a typedef of HE.
510 These contain the actual key and value pointers (plus extra administrative
511 overhead). The key is a string pointer; the value is an C<SV*>. However,
512 once you have an C<HE*>, to get the actual key and value, use the routines
515 I32 hv_iterinit(HV*);
516 /* Prepares starting point to traverse hash table */
517 HE* hv_iternext(HV*);
518 /* Get the next entry, and return a pointer to a
519 structure that has both the key and value */
520 char* hv_iterkey(HE* entry, I32* retlen);
521 /* Get the key from an HE structure and also return
522 the length of the key string */
523 SV* hv_iterval(HV*, HE* entry);
524 /* Return an SV pointer to the value of the HE
526 SV* hv_iternextsv(HV*, char** key, I32* retlen);
527 /* This convenience routine combines hv_iternext,
528 hv_iterkey, and hv_iterval. The key and retlen
529 arguments are return values for the key and its
530 length. The value is returned in the SV* argument */
532 If you know the name of a hash variable, you can get a pointer to its HV
533 by using the following:
535 HV* get_hv("package::varname", 0);
537 This returns NULL if the variable does not exist.
539 The hash algorithm is defined in the C<PERL_HASH> macro:
541 PERL_HASH(hash, key, klen)
543 The exact implementation of this macro varies by architecture and version
544 of perl, and the return value may change per invocation, so the value
545 is only valid for the duration of a single perl process.
547 See L<Understanding the Magic of Tied Hashes and Arrays> for more
548 information on how to use the hash access functions on tied hashes.
550 =head2 Hash API Extensions
552 Beginning with version 5.004, the following functions are also supported:
554 HE* hv_fetch_ent (HV* tb, SV* key, I32 lval, U32 hash);
555 HE* hv_store_ent (HV* tb, SV* key, SV* val, U32 hash);
557 bool hv_exists_ent (HV* tb, SV* key, U32 hash);
558 SV* hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
560 SV* hv_iterkeysv (HE* entry);
562 Note that these functions take C<SV*> keys, which simplifies writing
563 of extension code that deals with hash structures. These functions
564 also allow passing of C<SV*> keys to C<tie> functions without forcing
565 you to stringify the keys (unlike the previous set of functions).
567 They also return and accept whole hash entries (C<HE*>), making their
568 use more efficient (since the hash number for a particular string
569 doesn't have to be recomputed every time). See L<perlapi> for detailed
572 The following macros must always be used to access the contents of hash
573 entries. Note that the arguments to these macros must be simple
574 variables, since they may get evaluated more than once. See
575 L<perlapi> for detailed descriptions of these macros.
577 HePV(HE* he, STRLEN len)
581 HeSVKEY_force(HE* he)
582 HeSVKEY_set(HE* he, SV* sv)
584 These two lower level macros are defined, but must only be used when
585 dealing with keys that are not C<SV*>s:
590 Note that both C<hv_store> and C<hv_store_ent> do not increment the
591 reference count of the stored C<val>, which is the caller's responsibility.
592 If these functions return a NULL value, the caller will usually have to
593 decrement the reference count of C<val> to avoid a memory leak.
595 =head2 AVs, HVs and undefined values
597 Sometimes you have to store undefined values in AVs or HVs. Although
598 this may be a rare case, it can be tricky. That's because you're
599 used to using C<&PL_sv_undef> if you need an undefined SV.
601 For example, intuition tells you that this XS code:
604 av_store( av, 0, &PL_sv_undef );
606 is equivalent to this Perl code:
611 Unfortunately, this isn't true. In perl 5.18 and earlier, AVs use C<&PL_sv_undef> as a marker
612 for indicating that an array element has not yet been initialized.
613 Thus, C<exists $av[0]> would be true for the above Perl code, but
614 false for the array generated by the XS code. In perl 5.20, storing
615 &PL_sv_undef will create a read-only element, because the scalar
616 &PL_sv_undef itself is stored, not a copy.
618 Similar problems can occur when storing C<&PL_sv_undef> in HVs:
620 hv_store( hv, "key", 3, &PL_sv_undef, 0 );
622 This will indeed make the value C<undef>, but if you try to modify
623 the value of C<key>, you'll get the following error:
625 Modification of non-creatable hash value attempted
627 In perl 5.8.0, C<&PL_sv_undef> was also used to mark placeholders
628 in restricted hashes. This caused such hash entries not to appear
629 when iterating over the hash or when checking for the keys
630 with the C<hv_exists> function.
632 You can run into similar problems when you store C<&PL_sv_yes> or
633 C<&PL_sv_no> into AVs or HVs. Trying to modify such elements
634 will give you the following error:
636 Modification of a read-only value attempted
638 To make a long story short, you can use the special variables
639 C<&PL_sv_undef>, C<&PL_sv_yes> and C<&PL_sv_no> with AVs and
640 HVs, but you have to make sure you know what you're doing.
642 Generally, if you want to store an undefined value in an AV
643 or HV, you should not use C<&PL_sv_undef>, but rather create a
644 new undefined value using the C<newSV> function, for example:
646 av_store( av, 42, newSV(0) );
647 hv_store( hv, "foo", 3, newSV(0), 0 );
651 References are a special type of scalar that point to other data types
652 (including other references).
654 To create a reference, use either of the following functions:
656 SV* newRV_inc((SV*) thing);
657 SV* newRV_noinc((SV*) thing);
659 The C<thing> argument can be any of an C<SV*>, C<AV*>, or C<HV*>. The
660 functions are identical except that C<newRV_inc> increments the reference
661 count of the C<thing>, while C<newRV_noinc> does not. For historical
662 reasons, C<newRV> is a synonym for C<newRV_inc>.
664 Once you have a reference, you can use the following macro to dereference
669 then call the appropriate routines, casting the returned C<SV*> to either an
670 C<AV*> or C<HV*>, if required.
672 To determine if an SV is a reference, you can use the following macro:
676 To discover what type of value the reference refers to, use the following
677 macro and then check the return value.
681 The most useful types that will be returned are:
687 SVt_PVGV Glob (possibly a file handle)
689 See L<perlapi/svtype> for more details.
691 =head2 Blessed References and Class Objects
693 References are also used to support object-oriented programming. In perl's
694 OO lexicon, an object is simply a reference that has been blessed into a
695 package (or class). Once blessed, the programmer may now use the reference
696 to access the various methods in the class.
698 A reference can be blessed into a package with the following function:
700 SV* sv_bless(SV* sv, HV* stash);
702 The C<sv> argument must be a reference value. The C<stash> argument
703 specifies which class the reference will belong to. See
704 L<Stashes and Globs> for information on converting class names into stashes.
706 /* Still under construction */
708 The following function upgrades rv to reference if not already one.
709 Creates a new SV for rv to point to. If C<classname> is non-null, the SV
710 is blessed into the specified class. SV is returned.
712 SV* newSVrv(SV* rv, const char* classname);
714 The following three functions copy integer, unsigned integer or double
715 into an SV whose reference is C<rv>. SV is blessed if C<classname> is
718 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
719 SV* sv_setref_uv(SV* rv, const char* classname, UV uv);
720 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
722 The following function copies the pointer value (I<the address, not the
723 string!>) into an SV whose reference is rv. SV is blessed if C<classname>
726 SV* sv_setref_pv(SV* rv, const char* classname, void* pv);
728 The following function copies a string into an SV whose reference is C<rv>.
729 Set length to 0 to let Perl calculate the string length. SV is blessed if
730 C<classname> is non-null.
732 SV* sv_setref_pvn(SV* rv, const char* classname, char* pv,
735 The following function tests whether the SV is blessed into the specified
736 class. It does not check inheritance relationships.
738 int sv_isa(SV* sv, const char* name);
740 The following function tests whether the SV is a reference to a blessed object.
742 int sv_isobject(SV* sv);
744 The following function tests whether the SV is derived from the specified
745 class. SV can be either a reference to a blessed object or a string
746 containing a class name. This is the function implementing the
747 C<UNIVERSAL::isa> functionality.
749 bool sv_derived_from(SV* sv, const char* name);
751 To check if you've got an object derived from a specific class you have
754 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
756 =head2 Creating New Variables
758 To create a new Perl variable with an undef value which can be accessed from
759 your Perl script, use the following routines, depending on the variable type.
761 SV* get_sv("package::varname", GV_ADD);
762 AV* get_av("package::varname", GV_ADD);
763 HV* get_hv("package::varname", GV_ADD);
765 Notice the use of GV_ADD as the second parameter. The new variable can now
766 be set, using the routines appropriate to the data type.
768 There are additional macros whose values may be bitwise OR'ed with the
769 C<GV_ADD> argument to enable certain extra features. Those bits are:
775 Marks the variable as multiply defined, thus preventing the:
777 Name <varname> used only once: possible typo
785 Had to create <varname> unexpectedly
787 if the variable did not exist before the function was called.
791 If you do not specify a package name, the variable is created in the current
794 =head2 Reference Counts and Mortality
796 Perl uses a reference count-driven garbage collection mechanism. SVs,
797 AVs, or HVs (xV for short in the following) start their life with a
798 reference count of 1. If the reference count of an xV ever drops to 0,
799 then it will be destroyed and its memory made available for reuse.
801 This normally doesn't happen at the Perl level unless a variable is
802 undef'ed or the last variable holding a reference to it is changed or
803 overwritten. At the internal level, however, reference counts can be
804 manipulated with the following macros:
806 int SvREFCNT(SV* sv);
807 SV* SvREFCNT_inc(SV* sv);
808 void SvREFCNT_dec(SV* sv);
810 However, there is one other function which manipulates the reference
811 count of its argument. The C<newRV_inc> function, you will recall,
812 creates a reference to the specified argument. As a side effect,
813 it increments the argument's reference count. If this is not what
814 you want, use C<newRV_noinc> instead.
816 For example, imagine you want to return a reference from an XSUB function.
817 Inside the XSUB routine, you create an SV which initially has a reference
818 count of one. Then you call C<newRV_inc>, passing it the just-created SV.
819 This returns the reference as a new SV, but the reference count of the
820 SV you passed to C<newRV_inc> has been incremented to two. Now you
821 return the reference from the XSUB routine and forget about the SV.
822 But Perl hasn't! Whenever the returned reference is destroyed, the
823 reference count of the original SV is decreased to one and nothing happens.
824 The SV will hang around without any way to access it until Perl itself
825 terminates. This is a memory leak.
827 The correct procedure, then, is to use C<newRV_noinc> instead of
828 C<newRV_inc>. Then, if and when the last reference is destroyed,
829 the reference count of the SV will go to zero and it will be destroyed,
830 stopping any memory leak.
832 There are some convenience functions available that can help with the
833 destruction of xVs. These functions introduce the concept of "mortality".
834 An xV that is mortal has had its reference count marked to be decremented,
835 but not actually decremented, until "a short time later". Generally the
836 term "short time later" means a single Perl statement, such as a call to
837 an XSUB function. The actual determinant for when mortal xVs have their
838 reference count decremented depends on two macros, SAVETMPS and FREETMPS.
839 See L<perlcall> and L<perlxs> for more details on these macros.
841 "Mortalization" then is at its simplest a deferred C<SvREFCNT_dec>.
842 However, if you mortalize a variable twice, the reference count will
843 later be decremented twice.
845 "Mortal" SVs are mainly used for SVs that are placed on perl's stack.
846 For example an SV which is created just to pass a number to a called sub
847 is made mortal to have it cleaned up automatically when it's popped off
848 the stack. Similarly, results returned by XSUBs (which are pushed on the
849 stack) are often made mortal.
851 To create a mortal variable, use the functions:
855 SV* sv_mortalcopy(SV*)
857 The first call creates a mortal SV (with no value), the second converts an existing
858 SV to a mortal SV (and thus defers a call to C<SvREFCNT_dec>), and the
859 third creates a mortal copy of an existing SV.
860 Because C<sv_newmortal> gives the new SV no value, it must normally be given one
861 via C<sv_setpv>, C<sv_setiv>, etc. :
863 SV *tmp = sv_newmortal();
864 sv_setiv(tmp, an_integer);
866 As that is multiple C statements it is quite common so see this idiom instead:
868 SV *tmp = sv_2mortal(newSViv(an_integer));
871 You should be careful about creating mortal variables. Strange things
872 can happen if you make the same value mortal within multiple contexts,
873 or if you make a variable mortal multiple
874 times. Thinking of "Mortalization"
875 as deferred C<SvREFCNT_dec> should help to minimize such problems.
876 For example if you are passing an SV which you I<know> has a high enough REFCNT
877 to survive its use on the stack you need not do any mortalization.
878 If you are not sure then doing an C<SvREFCNT_inc> and C<sv_2mortal>, or
879 making a C<sv_mortalcopy> is safer.
881 The mortal routines are not just for SVs; AVs and HVs can be
882 made mortal by passing their address (type-casted to C<SV*>) to the
883 C<sv_2mortal> or C<sv_mortalcopy> routines.
885 =head2 Stashes and Globs
887 A B<stash> is a hash that contains all variables that are defined
888 within a package. Each key of the stash is a symbol
889 name (shared by all the different types of objects that have the same
890 name), and each value in the hash table is a GV (Glob Value). This GV
891 in turn contains references to the various objects of that name,
892 including (but not limited to) the following:
901 There is a single stash called C<PL_defstash> that holds the items that exist
902 in the C<main> package. To get at the items in other packages, append the
903 string "::" to the package name. The items in the C<Foo> package are in
904 the stash C<Foo::> in PL_defstash. The items in the C<Bar::Baz> package are
905 in the stash C<Baz::> in C<Bar::>'s stash.
907 To get the stash pointer for a particular package, use the function:
909 HV* gv_stashpv(const char* name, I32 flags)
910 HV* gv_stashsv(SV*, I32 flags)
912 The first function takes a literal string, the second uses the string stored
913 in the SV. Remember that a stash is just a hash table, so you get back an
914 C<HV*>. The C<flags> flag will create a new package if it is set to GV_ADD.
916 The name that C<gv_stash*v> wants is the name of the package whose symbol table
917 you want. The default package is called C<main>. If you have multiply nested
918 packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
921 Alternately, if you have an SV that is a blessed reference, you can find
922 out the stash pointer by using:
924 HV* SvSTASH(SvRV(SV*));
926 then use the following to get the package name itself:
928 char* HvNAME(HV* stash);
930 If you need to bless or re-bless an object you can use the following
933 SV* sv_bless(SV*, HV* stash)
935 where the first argument, an C<SV*>, must be a reference, and the second
936 argument is a stash. The returned C<SV*> can now be used in the same way
939 For more information on references and blessings, consult L<perlref>.
941 =head2 Double-Typed SVs
943 Scalar variables normally contain only one type of value, an integer,
944 double, pointer, or reference. Perl will automatically convert the
945 actual scalar data from the stored type into the requested type.
947 Some scalar variables contain more than one type of scalar data. For
948 example, the variable C<$!> contains either the numeric value of C<errno>
949 or its string equivalent from either C<strerror> or C<sys_errlist[]>.
951 To force multiple data values into an SV, you must do two things: use the
952 C<sv_set*v> routines to add the additional scalar type, then set a flag
953 so that Perl will believe it contains more than one type of data. The
954 four macros to set the flags are:
961 The particular macro you must use depends on which C<sv_set*v> routine
962 you called first. This is because every C<sv_set*v> routine turns on
963 only the bit for the particular type of data being set, and turns off
966 For example, to create a new Perl variable called "dberror" that contains
967 both the numeric and descriptive string error values, you could use the
971 extern char *dberror_list;
973 SV* sv = get_sv("dberror", GV_ADD);
974 sv_setiv(sv, (IV) dberror);
975 sv_setpv(sv, dberror_list[dberror]);
978 If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
979 macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
981 =head2 Read-Only Values
983 In Perl 5.16 and earlier, copy-on-write (see the next section) shared a
984 flag bit with read-only scalars. So the only way to test whether
985 C<sv_setsv>, etc., will raise a "Modification of a read-only value" error
986 in those versions is:
988 SvREADONLY(sv) && !SvIsCOW(sv)
990 Under Perl 5.18 and later, SvREADONLY only applies to read-only variables,
991 and, under 5.20, copy-on-write scalars can also be read-only, so the above
992 check is incorrect. You just want:
996 If you need to do this check often, define your own macro like this:
998 #if PERL_VERSION >= 18
999 # define SvTRULYREADONLY(sv) SvREADONLY(sv)
1001 # define SvTRULYREADONLY(sv) (SvREADONLY(sv) && !SvIsCOW(sv))
1004 =head2 Copy on Write
1006 Perl implements a copy-on-write (COW) mechanism for scalars, in which
1007 string copies are not immediately made when requested, but are deferred
1008 until made necessary by one or the other scalar changing. This is mostly
1009 transparent, but one must take care not to modify string buffers that are
1010 shared by multiple SVs.
1012 You can test whether an SV is using copy-on-write with C<SvIsCOW(sv)>.
1014 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).
1016 If you want to make the SV drop its string buffer, use
1017 C<sv_force_normal_flags(sv, SV_COW_DROP_PV)> or simply
1018 C<sv_setsv(sv, NULL)>.
1020 All of these functions will croak on read-only scalars (see the previous
1021 section for more on those).
1023 To test that your code is behaving correctly and not modifying COW buffers,
1024 on systems that support L<mmap(2)> (i.e., Unix) you can configure perl with
1025 C<-Accflags=-DPERL_DEBUG_READONLY_COW> and it will turn buffer violations
1026 into crashes. You will find it to be marvellously slow, so you may want to
1027 skip perl's own tests.
1029 =head2 Magic Variables
1031 [This section still under construction. Ignore everything here. Post no
1032 bills. Everything not permitted is forbidden.]
1034 Any SV may be magical, that is, it has special features that a normal
1035 SV does not have. These features are stored in the SV structure in a
1036 linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
1039 MAGIC* mg_moremagic;
1049 Note this is current as of patchlevel 0, and could change at any time.
1051 =head2 Assigning Magic
1053 Perl adds magic to an SV using the sv_magic function:
1055 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
1057 The C<sv> argument is a pointer to the SV that is to acquire a new magical
1060 If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
1061 convert C<sv> to type C<SVt_PVMG>.
1062 Perl then continues by adding new magic
1063 to the beginning of the linked list of magical features. Any prior entry
1064 of the same type of magic is deleted. Note that this can be overridden,
1065 and multiple instances of the same type of magic can be associated with an
1068 The C<name> and C<namlen> arguments are used to associate a string with
1069 the magic, typically the name of a variable. C<namlen> is stored in the
1070 C<mg_len> field and if C<name> is non-null then either a C<savepvn> copy of
1071 C<name> or C<name> itself is stored in the C<mg_ptr> field, depending on
1072 whether C<namlen> is greater than zero or equal to zero respectively. As a
1073 special case, if C<(name && namlen == HEf_SVKEY)> then C<name> is assumed
1074 to contain an C<SV*> and is stored as-is with its REFCNT incremented.
1076 The sv_magic function uses C<how> to determine which, if any, predefined
1077 "Magic Virtual Table" should be assigned to the C<mg_virtual> field.
1078 See the L<Magic Virtual Tables> section below. The C<how> argument is also
1079 stored in the C<mg_type> field. The value of
1080 C<how> should be chosen from the set of macros
1081 C<PERL_MAGIC_foo> found in F<perl.h>. Note that before
1082 these macros were added, Perl internals used to directly use character
1083 literals, so you may occasionally come across old code or documentation
1084 referring to 'U' magic rather than C<PERL_MAGIC_uvar> for example.
1086 The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
1087 structure. If it is not the same as the C<sv> argument, the reference
1088 count of the C<obj> object is incremented. If it is the same, or if
1089 the C<how> argument is C<PERL_MAGIC_arylen>, or if it is a NULL pointer,
1090 then C<obj> is merely stored, without the reference count being incremented.
1092 See also C<sv_magicext> in L<perlapi> for a more flexible way to add magic
1095 There is also a function to add magic to an C<HV>:
1097 void hv_magic(HV *hv, GV *gv, int how);
1099 This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
1101 To remove the magic from an SV, call the function sv_unmagic:
1103 int sv_unmagic(SV *sv, int type);
1105 The C<type> argument should be equal to the C<how> value when the C<SV>
1106 was initially made magical.
1108 However, note that C<sv_unmagic> removes all magic of a certain C<type> from the
1109 C<SV>. If you want to remove only certain
1110 magic of a C<type> based on the magic
1111 virtual table, use C<sv_unmagicext> instead:
1113 int sv_unmagicext(SV *sv, int type, MGVTBL *vtbl);
1115 =head2 Magic Virtual Tables
1117 The C<mg_virtual> field in the C<MAGIC> structure is a pointer to an
1118 C<MGVTBL>, which is a structure of function pointers and stands for
1119 "Magic Virtual Table" to handle the various operations that might be
1120 applied to that variable.
1122 The C<MGVTBL> has five (or sometimes eight) pointers to the following
1125 int (*svt_get)(SV* sv, MAGIC* mg);
1126 int (*svt_set)(SV* sv, MAGIC* mg);
1127 U32 (*svt_len)(SV* sv, MAGIC* mg);
1128 int (*svt_clear)(SV* sv, MAGIC* mg);
1129 int (*svt_free)(SV* sv, MAGIC* mg);
1131 int (*svt_copy)(SV *sv, MAGIC* mg, SV *nsv,
1132 const char *name, I32 namlen);
1133 int (*svt_dup)(MAGIC *mg, CLONE_PARAMS *param);
1134 int (*svt_local)(SV *nsv, MAGIC *mg);
1137 This MGVTBL structure is set at compile-time in F<perl.h> and there are
1138 currently 32 types. These different structures contain pointers to various
1139 routines that perform additional actions depending on which function is
1142 Function pointer Action taken
1143 ---------------- ------------
1144 svt_get Do something before the value of the SV is
1146 svt_set Do something after the SV is assigned a value.
1147 svt_len Report on the SV's length.
1148 svt_clear Clear something the SV represents.
1149 svt_free Free any extra storage associated with the SV.
1151 svt_copy copy tied variable magic to a tied element
1152 svt_dup duplicate a magic structure during thread cloning
1153 svt_local copy magic to local value during 'local'
1155 For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
1156 to an C<mg_type> of C<PERL_MAGIC_sv>) contains:
1158 { magic_get, magic_set, magic_len, 0, 0 }
1160 Thus, when an SV is determined to be magical and of type C<PERL_MAGIC_sv>,
1161 if a get operation is being performed, the routine C<magic_get> is
1162 called. All the various routines for the various magical types begin
1163 with C<magic_>. NOTE: the magic routines are not considered part of
1164 the Perl API, and may not be exported by the Perl library.
1166 The last three slots are a recent addition, and for source code
1167 compatibility they are only checked for if one of the three flags
1168 MGf_COPY, MGf_DUP or MGf_LOCAL is set in mg_flags.
1169 This means that most code can continue declaring
1170 a vtable as a 5-element value. These three are
1171 currently used exclusively by the threading code, and are highly subject
1174 The current kinds of Magic Virtual Tables are:
1177 This table is generated by regen/mg_vtable.pl. Any changes made here
1180 =for mg_vtable.pl begin
1183 (old-style char and macro) MGVTBL Type of magic
1184 -------------------------- ------ -------------
1185 \0 PERL_MAGIC_sv vtbl_sv Special scalar variable
1186 # PERL_MAGIC_arylen vtbl_arylen Array length ($#ary)
1187 % PERL_MAGIC_rhash (none) Extra data for restricted
1189 * PERL_MAGIC_debugvar vtbl_debugvar $DB::single, signal, trace
1191 . PERL_MAGIC_pos vtbl_pos pos() lvalue
1192 : PERL_MAGIC_symtab (none) Extra data for symbol
1194 < PERL_MAGIC_backref vtbl_backref For weak ref data
1195 @ PERL_MAGIC_arylen_p (none) To move arylen out of XPVAV
1196 B PERL_MAGIC_bm vtbl_regexp Boyer-Moore
1197 (fast string search)
1198 c PERL_MAGIC_overload_table vtbl_ovrld Holds overload table
1200 D PERL_MAGIC_regdata vtbl_regdata Regex match position data
1202 d PERL_MAGIC_regdatum vtbl_regdatum Regex match position data
1204 E PERL_MAGIC_env vtbl_env %ENV hash
1205 e PERL_MAGIC_envelem vtbl_envelem %ENV hash element
1206 f PERL_MAGIC_fm vtbl_regexp Formline
1208 g PERL_MAGIC_regex_global vtbl_mglob m//g target
1209 H PERL_MAGIC_hints vtbl_hints %^H hash
1210 h PERL_MAGIC_hintselem vtbl_hintselem %^H hash element
1211 I PERL_MAGIC_isa vtbl_isa @ISA array
1212 i PERL_MAGIC_isaelem vtbl_isaelem @ISA array element
1213 k PERL_MAGIC_nkeys vtbl_nkeys scalar(keys()) lvalue
1214 L PERL_MAGIC_dbfile (none) Debugger %_<filename
1215 l PERL_MAGIC_dbline vtbl_dbline Debugger %_<filename
1217 N PERL_MAGIC_shared (none) Shared between threads
1218 n PERL_MAGIC_shared_scalar (none) Shared between threads
1219 o PERL_MAGIC_collxfrm vtbl_collxfrm Locale transformation
1220 P PERL_MAGIC_tied vtbl_pack Tied array or hash
1221 p PERL_MAGIC_tiedelem vtbl_packelem Tied array or hash element
1222 q PERL_MAGIC_tiedscalar vtbl_packelem Tied scalar or handle
1223 r PERL_MAGIC_qr vtbl_regexp Precompiled qr// regex
1224 S PERL_MAGIC_sig (none) %SIG hash
1225 s PERL_MAGIC_sigelem vtbl_sigelem %SIG hash element
1226 t PERL_MAGIC_taint vtbl_taint Taintedness
1227 U PERL_MAGIC_uvar vtbl_uvar Available for use by
1229 u PERL_MAGIC_uvar_elem (none) Reserved for use by
1231 V PERL_MAGIC_vstring (none) SV was vstring literal
1232 v PERL_MAGIC_vec vtbl_vec vec() lvalue
1233 w PERL_MAGIC_utf8 vtbl_utf8 Cached UTF-8 information
1234 x PERL_MAGIC_substr vtbl_substr substr() lvalue
1235 y PERL_MAGIC_defelem vtbl_defelem Shadow "foreach" iterator
1236 variable / smart parameter
1238 \ PERL_MAGIC_lvref vtbl_lvref Lvalue reference
1240 ] PERL_MAGIC_checkcall vtbl_checkcall Inlining/mutation of call
1242 ~ PERL_MAGIC_ext (none) Available for use by
1245 =for mg_vtable.pl end
1247 When an uppercase and lowercase letter both exist in the table, then the
1248 uppercase letter is typically used to represent some kind of composite type
1249 (a list or a hash), and the lowercase letter is used to represent an element
1250 of that composite type. Some internals code makes use of this case
1251 relationship. However, 'v' and 'V' (vec and v-string) are in no way related.
1253 The C<PERL_MAGIC_ext> and C<PERL_MAGIC_uvar> magic types are defined
1254 specifically for use by extensions and will not be used by perl itself.
1255 Extensions can use C<PERL_MAGIC_ext> magic to 'attach' private information
1256 to variables (typically objects). This is especially useful because
1257 there is no way for normal perl code to corrupt this private information
1258 (unlike using extra elements of a hash object).
1260 Similarly, C<PERL_MAGIC_uvar> magic can be used much like tie() to call a
1261 C function any time a scalar's value is used or changed. The C<MAGIC>'s
1262 C<mg_ptr> field points to a C<ufuncs> structure:
1265 I32 (*uf_val)(pTHX_ IV, SV*);
1266 I32 (*uf_set)(pTHX_ IV, SV*);
1270 When the SV is read from or written to, the C<uf_val> or C<uf_set>
1271 function will be called with C<uf_index> as the first arg and a pointer to
1272 the SV as the second. A simple example of how to add C<PERL_MAGIC_uvar>
1273 magic is shown below. Note that the ufuncs structure is copied by
1274 sv_magic, so you can safely allocate it on the stack.
1282 uf.uf_val = &my_get_fn;
1283 uf.uf_set = &my_set_fn;
1285 sv_magic(sv, 0, PERL_MAGIC_uvar, (char*)&uf, sizeof(uf));
1287 Attaching C<PERL_MAGIC_uvar> to arrays is permissible but has no effect.
1289 For hashes there is a specialized hook that gives control over hash
1290 keys (but not values). This hook calls C<PERL_MAGIC_uvar> 'get' magic
1291 if the "set" function in the C<ufuncs> structure is NULL. The hook
1292 is activated whenever the hash is accessed with a key specified as
1293 an C<SV> through the functions C<hv_store_ent>, C<hv_fetch_ent>,
1294 C<hv_delete_ent>, and C<hv_exists_ent>. Accessing the key as a string
1295 through the functions without the C<..._ent> suffix circumvents the
1296 hook. See L<Hash::Util::FieldHash/GUTS> for a detailed description.
1298 Note that because multiple extensions may be using C<PERL_MAGIC_ext>
1299 or C<PERL_MAGIC_uvar> magic, it is important for extensions to take
1300 extra care to avoid conflict. Typically only using the magic on
1301 objects blessed into the same class as the extension is sufficient.
1302 For C<PERL_MAGIC_ext> magic, it is usually a good idea to define an
1303 C<MGVTBL>, even if all its fields will be C<0>, so that individual
1304 C<MAGIC> pointers can be identified as a particular kind of magic
1305 using their magic virtual table. C<mg_findext> provides an easy way
1308 STATIC MGVTBL my_vtbl = { 0, 0, 0, 0, 0, 0, 0, 0 };
1311 if ((mg = mg_findext(sv, PERL_MAGIC_ext, &my_vtbl))) {
1312 /* this is really ours, not another module's PERL_MAGIC_ext */
1313 my_priv_data_t *priv = (my_priv_data_t *)mg->mg_ptr;
1317 Also note that the C<sv_set*()> and C<sv_cat*()> functions described
1318 earlier do B<not> invoke 'set' magic on their targets. This must
1319 be done by the user either by calling the C<SvSETMAGIC()> macro after
1320 calling these functions, or by using one of the C<sv_set*_mg()> or
1321 C<sv_cat*_mg()> functions. Similarly, generic C code must call the
1322 C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
1323 obtained from external sources in functions that don't handle magic.
1324 See L<perlapi> for a description of these functions.
1325 For example, calls to the C<sv_cat*()> functions typically need to be
1326 followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
1327 since their implementation handles 'get' magic.
1329 =head2 Finding Magic
1331 MAGIC *mg_find(SV *sv, int type); /* Finds the magic pointer of that
1334 This routine returns a pointer to a C<MAGIC> structure stored in the SV.
1335 If the SV does not have that magical
1336 feature, C<NULL> is returned. If the
1337 SV has multiple instances of that magical feature, the first one will be
1338 returned. C<mg_findext> can be used
1339 to find a C<MAGIC> structure of an SV
1340 based on both its magic type and its magic virtual table:
1342 MAGIC *mg_findext(SV *sv, int type, MGVTBL *vtbl);
1344 Also, if the SV passed to C<mg_find> or C<mg_findext> is not of type
1345 SVt_PVMG, Perl may core dump.
1347 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
1349 This routine checks to see what types of magic C<sv> has. If the mg_type
1350 field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
1351 the mg_type field is changed to be the lowercase letter.
1353 =head2 Understanding the Magic of Tied Hashes and Arrays
1355 Tied hashes and arrays are magical beasts of the C<PERL_MAGIC_tied>
1358 WARNING: As of the 5.004 release, proper usage of the array and hash
1359 access functions requires understanding a few caveats. Some
1360 of these caveats are actually considered bugs in the API, to be fixed
1361 in later releases, and are bracketed with [MAYCHANGE] below. If
1362 you find yourself actually applying such information in this section, be
1363 aware that the behavior may change in the future, umm, without warning.
1365 The perl tie function associates a variable with an object that implements
1366 the various GET, SET, etc methods. To perform the equivalent of the perl
1367 tie function from an XSUB, you must mimic this behaviour. The code below
1368 carries out the necessary steps -- firstly it creates a new hash, and then
1369 creates a second hash which it blesses into the class which will implement
1370 the tie methods. Lastly it ties the two hashes together, and returns a
1371 reference to the new tied hash. Note that the code below does NOT call the
1372 TIEHASH method in the MyTie class -
1373 see L<Calling Perl Routines from within C Programs> for details on how
1384 tie = newRV_noinc((SV*)newHV());
1385 stash = gv_stashpv("MyTie", GV_ADD);
1386 sv_bless(tie, stash);
1387 hv_magic(hash, (GV*)tie, PERL_MAGIC_tied);
1388 RETVAL = newRV_noinc(hash);
1392 The C<av_store> function, when given a tied array argument, merely
1393 copies the magic of the array onto the value to be "stored", using
1394 C<mg_copy>. It may also return NULL, indicating that the value did not
1395 actually need to be stored in the array. [MAYCHANGE] After a call to
1396 C<av_store> on a tied array, the caller will usually need to call
1397 C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1398 TIEARRAY object. If C<av_store> did return NULL, a call to
1399 C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1402 The previous paragraph is applicable verbatim to tied hash access using the
1403 C<hv_store> and C<hv_store_ent> functions as well.
1405 C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1406 C<hv_fetch_ent> actually return an undefined mortal value whose magic
1407 has been initialized using C<mg_copy>. Note the value so returned does not
1408 need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1409 need to call C<mg_get()> on the returned value in order to actually invoke
1410 the perl level "FETCH" method on the underlying TIE object. Similarly,
1411 you may also call C<mg_set()> on the return value after possibly assigning
1412 a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
1413 method on the TIE object. [/MAYCHANGE]
1416 In other words, the array or hash fetch/store functions don't really
1417 fetch and store actual values in the case of tied arrays and hashes. They
1418 merely call C<mg_copy> to attach magic to the values that were meant to be
1419 "stored" or "fetched". Later calls to C<mg_get> and C<mg_set> actually
1420 do the job of invoking the TIE methods on the underlying objects. Thus
1421 the magic mechanism currently implements a kind of lazy access to arrays
1424 Currently (as of perl version 5.004), use of the hash and array access
1425 functions requires the user to be aware of whether they are operating on
1426 "normal" hashes and arrays, or on their tied variants. The API may be
1427 changed to provide more transparent access to both tied and normal data
1428 types in future versions.
1431 You would do well to understand that the TIEARRAY and TIEHASH interfaces
1432 are mere sugar to invoke some perl method calls while using the uniform hash
1433 and array syntax. The use of this sugar imposes some overhead (typically
1434 about two to four extra opcodes per FETCH/STORE operation, in addition to
1435 the creation of all the mortal variables required to invoke the methods).
1436 This overhead will be comparatively small if the TIE methods are themselves
1437 substantial, but if they are only a few statements long, the overhead
1438 will not be insignificant.
1440 =head2 Localizing changes
1442 Perl has a very handy construction
1449 This construction is I<approximately> equivalent to
1458 The biggest difference is that the first construction would
1459 reinstate the initial value of $var, irrespective of how control exits
1460 the block: C<goto>, C<return>, C<die>/C<eval>, etc. It is a little bit
1461 more efficient as well.
1463 There is a way to achieve a similar task from C via Perl API: create a
1464 I<pseudo-block>, and arrange for some changes to be automatically
1465 undone at the end of it, either explicit, or via a non-local exit (via
1466 die()). A I<block>-like construct is created by a pair of
1467 C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1468 Such a construct may be created specially for some important localized
1469 task, or an existing one (like boundaries of enclosing Perl
1470 subroutine/block, or an existing pair for freeing TMPs) may be
1471 used. (In the second case the overhead of additional localization must
1472 be almost negligible.) Note that any XSUB is automatically enclosed in
1473 an C<ENTER>/C<LEAVE> pair.
1475 Inside such a I<pseudo-block> the following service is available:
1479 =item C<SAVEINT(int i)>
1481 =item C<SAVEIV(IV i)>
1483 =item C<SAVEI32(I32 i)>
1485 =item C<SAVELONG(long i)>
1487 These macros arrange things to restore the value of integer variable
1488 C<i> at the end of enclosing I<pseudo-block>.
1490 =item C<SAVESPTR(s)>
1492 =item C<SAVEPPTR(p)>
1494 These macros arrange things to restore the value of pointers C<s> and
1495 C<p>. C<s> must be a pointer of a type which survives conversion to
1496 C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1499 =item C<SAVEFREESV(SV *sv)>
1501 The refcount of C<sv> would be decremented at the end of
1502 I<pseudo-block>. This is similar to C<sv_2mortal> in that it is also a
1503 mechanism for doing a delayed C<SvREFCNT_dec>. However, while C<sv_2mortal>
1504 extends the lifetime of C<sv> until the beginning of the next statement,
1505 C<SAVEFREESV> extends it until the end of the enclosing scope. These
1506 lifetimes can be wildly different.
1508 Also compare C<SAVEMORTALIZESV>.
1510 =item C<SAVEMORTALIZESV(SV *sv)>
1512 Just like C<SAVEFREESV>, but mortalizes C<sv> at the end of the current
1513 scope instead of decrementing its reference count. This usually has the
1514 effect of keeping C<sv> alive until the statement that called the currently
1515 live scope has finished executing.
1517 =item C<SAVEFREEOP(OP *op)>
1519 The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1521 =item C<SAVEFREEPV(p)>
1523 The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1524 end of I<pseudo-block>.
1526 =item C<SAVECLEARSV(SV *sv)>
1528 Clears a slot in the current scratchpad which corresponds to C<sv> at
1529 the end of I<pseudo-block>.
1531 =item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1533 The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
1534 string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1535 short-lived storage, the corresponding string may be reallocated like
1538 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
1540 =item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
1542 At the end of I<pseudo-block> the function C<f> is called with the
1545 =item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1547 At the end of I<pseudo-block> the function C<f> is called with the
1548 implicit context argument (if any), and C<p>.
1550 =item C<SAVESTACK_POS()>
1552 The current offset on the Perl internal stack (cf. C<SP>) is restored
1553 at the end of I<pseudo-block>.
1557 The following API list contains functions, thus one needs to
1558 provide pointers to the modifiable data explicitly (either C pointers,
1559 or Perlish C<GV *>s). Where the above macros take C<int>, a similar
1560 function takes C<int *>.
1564 =item C<SV* save_scalar(GV *gv)>
1566 Equivalent to Perl code C<local $gv>.
1568 =item C<AV* save_ary(GV *gv)>
1570 =item C<HV* save_hash(GV *gv)>
1572 Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1574 =item C<void save_item(SV *item)>
1576 Duplicates the current value of C<SV>, on the exit from the current
1577 C<ENTER>/C<LEAVE> I<pseudo-block> will restore the value of C<SV>
1578 using the stored value. It doesn't handle magic. Use C<save_scalar> if
1581 =item C<void save_list(SV **sarg, I32 maxsarg)>
1583 A variant of C<save_item> which takes multiple arguments via an array
1584 C<sarg> of C<SV*> of length C<maxsarg>.
1586 =item C<SV* save_svref(SV **sptr)>
1588 Similar to C<save_scalar>, but will reinstate an C<SV *>.
1590 =item C<void save_aptr(AV **aptr)>
1592 =item C<void save_hptr(HV **hptr)>
1594 Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1598 The C<Alias> module implements localization of the basic types within the
1599 I<caller's scope>. People who are interested in how to localize things in
1600 the containing scope should take a look there too.
1604 =head2 XSUBs and the Argument Stack
1606 The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1607 An XSUB routine will have a stack that contains the arguments from the Perl
1608 program, and a way to map from the Perl data structures to a C equivalent.
1610 The stack arguments are accessible through the C<ST(n)> macro, which returns
1611 the C<n>'th stack argument. Argument 0 is the first argument passed in the
1612 Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1615 Most of the time, output from the C routine can be handled through use of
1616 the RETVAL and OUTPUT directives. However, there are some cases where the
1617 argument stack is not already long enough to handle all the return values.
1618 An example is the POSIX tzname() call, which takes no arguments, but returns
1619 two, the local time zone's standard and summer time abbreviations.
1621 To handle this situation, the PPCODE directive is used and the stack is
1622 extended using the macro:
1626 where C<SP> is the macro that represents the local copy of the stack pointer,
1627 and C<num> is the number of elements the stack should be extended by.
1629 Now that there is room on the stack, values can be pushed on it using C<PUSHs>
1630 macro. The pushed values will often need to be "mortal" (See
1631 L</Reference Counts and Mortality>):
1633 PUSHs(sv_2mortal(newSViv(an_integer)))
1634 PUSHs(sv_2mortal(newSVuv(an_unsigned_integer)))
1635 PUSHs(sv_2mortal(newSVnv(a_double)))
1636 PUSHs(sv_2mortal(newSVpv("Some String",0)))
1637 /* Although the last example is better written as the more
1639 PUSHs(newSVpvs_flags("Some String", SVs_TEMP))
1641 And now the Perl program calling C<tzname>, the two values will be assigned
1644 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1646 An alternate (and possibly simpler) method to pushing values on the stack is
1651 This macro automatically adjusts the stack for you, if needed. Thus, you
1652 do not need to call C<EXTEND> to extend the stack.
1654 Despite their suggestions in earlier versions of this document the macros
1655 C<(X)PUSH[iunp]> are I<not> suited to XSUBs which return multiple results.
1656 For that, either stick to the C<(X)PUSHs> macros shown above, or use the new
1657 C<m(X)PUSH[iunp]> macros instead; see L</Putting a C value on Perl stack>.
1659 For more information, consult L<perlxs> and L<perlxstut>.
1661 =head2 Autoloading with XSUBs
1663 If an AUTOLOAD routine is an XSUB, as with Perl subroutines, Perl puts the
1664 fully-qualified name of the autoloaded subroutine in the $AUTOLOAD variable
1665 of the XSUB's package.
1667 But it also puts the same information in certain fields of the XSUB itself:
1669 HV *stash = CvSTASH(cv);
1670 const char *subname = SvPVX(cv);
1671 STRLEN name_length = SvCUR(cv); /* in bytes */
1672 U32 is_utf8 = SvUTF8(cv);
1674 C<SvPVX(cv)> contains just the sub name itself, not including the package.
1675 For an AUTOLOAD routine in UNIVERSAL or one of its superclasses,
1676 C<CvSTASH(cv)> returns NULL during a method call on a nonexistent package.
1678 B<Note>: Setting $AUTOLOAD stopped working in 5.6.1, which did not support
1679 XS AUTOLOAD subs at all. Perl 5.8.0 introduced the use of fields in the
1680 XSUB itself. Perl 5.16.0 restored the setting of $AUTOLOAD. If you need
1681 to support 5.8-5.14, use the XSUB's fields.
1683 =head2 Calling Perl Routines from within C Programs
1685 There are four routines that can be used to call a Perl subroutine from
1686 within a C program. These four are:
1688 I32 call_sv(SV*, I32);
1689 I32 call_pv(const char*, I32);
1690 I32 call_method(const char*, I32);
1691 I32 call_argv(const char*, I32, char**);
1693 The routine most often used is C<call_sv>. The C<SV*> argument
1694 contains either the name of the Perl subroutine to be called, or a
1695 reference to the subroutine. The second argument consists of flags
1696 that control the context in which the subroutine is called, whether
1697 or not the subroutine is being passed arguments, how errors should be
1698 trapped, and how to treat return values.
1700 All four routines return the number of arguments that the subroutine returned
1703 These routines used to be called C<perl_call_sv>, etc., before Perl v5.6.0,
1704 but those names are now deprecated; macros of the same name are provided for
1707 When using any of these routines (except C<call_argv>), the programmer
1708 must manipulate the Perl stack. These include the following macros and
1723 For a detailed description of calling conventions from C to Perl,
1724 consult L<perlcall>.
1726 =head2 Putting a C value on Perl stack
1728 A lot of opcodes (this is an elementary operation in the internal perl
1729 stack machine) put an SV* on the stack. However, as an optimization
1730 the corresponding SV is (usually) not recreated each time. The opcodes
1731 reuse specially assigned SVs (I<target>s) which are (as a corollary)
1732 not constantly freed/created.
1734 Each of the targets is created only once (but see
1735 L<Scratchpads and recursion> below), and when an opcode needs to put
1736 an integer, a double, or a string on stack, it just sets the
1737 corresponding parts of its I<target> and puts the I<target> on stack.
1739 The macro to put this target on stack is C<PUSHTARG>, and it is
1740 directly used in some opcodes, as well as indirectly in zillions of
1741 others, which use it via C<(X)PUSH[iunp]>.
1743 Because the target is reused, you must be careful when pushing multiple
1744 values on the stack. The following code will not do what you think:
1749 This translates as "set C<TARG> to 10, push a pointer to C<TARG> onto
1750 the stack; set C<TARG> to 20, push a pointer to C<TARG> onto the stack".
1751 At the end of the operation, the stack does not contain the values 10
1752 and 20, but actually contains two pointers to C<TARG>, which we have set
1755 If you need to push multiple different values then you should either use
1756 the C<(X)PUSHs> macros, or else use the new C<m(X)PUSH[iunp]> macros,
1757 none of which make use of C<TARG>. The C<(X)PUSHs> macros simply push an
1758 SV* on the stack, which, as noted under L</XSUBs and the Argument Stack>,
1759 will often need to be "mortal". The new C<m(X)PUSH[iunp]> macros make
1760 this a little easier to achieve by creating a new mortal for you (via
1761 C<(X)PUSHmortal>), pushing that onto the stack (extending it if necessary
1762 in the case of the C<mXPUSH[iunp]> macros), and then setting its value.
1763 Thus, instead of writing this to "fix" the example above:
1765 XPUSHs(sv_2mortal(newSViv(10)))
1766 XPUSHs(sv_2mortal(newSViv(20)))
1768 you can simply write:
1773 On a related note, if you do use C<(X)PUSH[iunp]>, then you're going to
1774 need a C<dTARG> in your variable declarations so that the C<*PUSH*>
1775 macros can make use of the local variable C<TARG>. See also C<dTARGET>
1780 The question remains on when the SVs which are I<target>s for opcodes
1781 are created. The answer is that they are created when the current
1782 unit--a subroutine or a file (for opcodes for statements outside of
1783 subroutines)--is compiled. During this time a special anonymous Perl
1784 array is created, which is called a scratchpad for the current unit.
1786 A scratchpad keeps SVs which are lexicals for the current unit and are
1787 targets for opcodes. A previous version of this document
1788 stated that one can deduce that an SV lives on a scratchpad
1789 by looking on its flags: lexicals have C<SVs_PADMY> set, and
1790 I<target>s have C<SVs_PADTMP> set. But this has never been fully true.
1791 C<SVs_PADMY> could be set on a variable that no longer resides in any pad.
1792 While I<target>s do have C<SVs_PADTMP> set, it can also be set on variables
1793 that have never resided in a pad, but nonetheless act like I<target>s. As
1794 of perl 5.21.5, the C<SVs_PADMY> flag is no longer used and is defined as
1795 0. C<SvPADMY()> now returns true for anything without C<SVs_PADTMP>.
1797 The correspondence between OPs and I<target>s is not 1-to-1. Different
1798 OPs in the compile tree of the unit can use the same target, if this
1799 would not conflict with the expected life of the temporary.
1801 =head2 Scratchpads and recursion
1803 In fact it is not 100% true that a compiled unit contains a pointer to
1804 the scratchpad AV. In fact it contains a pointer to an AV of
1805 (initially) one element, and this element is the scratchpad AV. Why do
1806 we need an extra level of indirection?
1808 The answer is B<recursion>, and maybe B<threads>. Both
1809 these can create several execution pointers going into the same
1810 subroutine. For the subroutine-child not write over the temporaries
1811 for the subroutine-parent (lifespan of which covers the call to the
1812 child), the parent and the child should have different
1813 scratchpads. (I<And> the lexicals should be separate anyway!)
1815 So each subroutine is born with an array of scratchpads (of length 1).
1816 On each entry to the subroutine it is checked that the current
1817 depth of the recursion is not more than the length of this array, and
1818 if it is, new scratchpad is created and pushed into the array.
1820 The I<target>s on this scratchpad are C<undef>s, but they are already
1821 marked with correct flags.
1823 =head1 Memory Allocation
1827 All memory meant to be used with the Perl API functions should be manipulated
1828 using the macros described in this section. The macros provide the necessary
1829 transparency between differences in the actual malloc implementation that is
1832 It is suggested that you enable the version of malloc that is distributed
1833 with Perl. It keeps pools of various sizes of unallocated memory in
1834 order to satisfy allocation requests more quickly. However, on some
1835 platforms, it may cause spurious malloc or free errors.
1837 The following three macros are used to initially allocate memory :
1839 Newx(pointer, number, type);
1840 Newxc(pointer, number, type, cast);
1841 Newxz(pointer, number, type);
1843 The first argument C<pointer> should be the name of a variable that will
1844 point to the newly allocated memory.
1846 The second and third arguments C<number> and C<type> specify how many of
1847 the specified type of data structure should be allocated. The argument
1848 C<type> is passed to C<sizeof>. The final argument to C<Newxc>, C<cast>,
1849 should be used if the C<pointer> argument is different from the C<type>
1852 Unlike the C<Newx> and C<Newxc> macros, the C<Newxz> macro calls C<memzero>
1853 to zero out all the newly allocated memory.
1857 Renew(pointer, number, type);
1858 Renewc(pointer, number, type, cast);
1861 These three macros are used to change a memory buffer size or to free a
1862 piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1863 match those of C<New> and C<Newc> with the exception of not needing the
1864 "magic cookie" argument.
1868 Move(source, dest, number, type);
1869 Copy(source, dest, number, type);
1870 Zero(dest, number, type);
1872 These three macros are used to move, copy, or zero out previously allocated
1873 memory. The C<source> and C<dest> arguments point to the source and
1874 destination starting points. Perl will move, copy, or zero out C<number>
1875 instances of the size of the C<type> data structure (using the C<sizeof>
1880 The most recent development releases of Perl have been experimenting with
1881 removing Perl's dependency on the "normal" standard I/O suite and allowing
1882 other stdio implementations to be used. This involves creating a new
1883 abstraction layer that then calls whichever implementation of stdio Perl
1884 was compiled with. All XSUBs should now use the functions in the PerlIO
1885 abstraction layer and not make any assumptions about what kind of stdio
1888 For a complete description of the PerlIO abstraction, consult L<perlapio>.
1890 =head1 Compiled code
1894 Here we describe the internal form your code is converted to by
1895 Perl. Start with a simple example:
1899 This is converted to a tree similar to this one:
1907 (but slightly more complicated). This tree reflects the way Perl
1908 parsed your code, but has nothing to do with the execution order.
1909 There is an additional "thread" going through the nodes of the tree
1910 which shows the order of execution of the nodes. In our simplified
1911 example above it looks like:
1913 $b ---> $c ---> + ---> $a ---> assign-to
1915 But with the actual compile tree for C<$a = $b + $c> it is different:
1916 some nodes I<optimized away>. As a corollary, though the actual tree
1917 contains more nodes than our simplified example, the execution order
1918 is the same as in our example.
1920 =head2 Examining the tree
1922 If you have your perl compiled for debugging (usually done with
1923 C<-DDEBUGGING> on the C<Configure> command line), you may examine the
1924 compiled tree by specifying C<-Dx> on the Perl command line. The
1925 output takes several lines per node, and for C<$b+$c> it looks like
1930 FLAGS = (SCALAR,KIDS)
1932 TYPE = null ===> (4)
1934 FLAGS = (SCALAR,KIDS)
1936 3 TYPE = gvsv ===> 4
1942 TYPE = null ===> (5)
1944 FLAGS = (SCALAR,KIDS)
1946 4 TYPE = gvsv ===> 5
1952 This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
1953 not optimized away (one per number in the left column). The immediate
1954 children of the given node correspond to C<{}> pairs on the same level
1955 of indentation, thus this listing corresponds to the tree:
1963 The execution order is indicated by C<===E<gt>> marks, thus it is C<3
1964 4 5 6> (node C<6> is not included into above listing), i.e.,
1965 C<gvsv gvsv add whatever>.
1967 Each of these nodes represents an op, a fundamental operation inside the
1968 Perl core. The code which implements each operation can be found in the
1969 F<pp*.c> files; the function which implements the op with type C<gvsv>
1970 is C<pp_gvsv>, and so on. As the tree above shows, different ops have
1971 different numbers of children: C<add> is a binary operator, as one would
1972 expect, and so has two children. To accommodate the various different
1973 numbers of children, there are various types of op data structure, and
1974 they link together in different ways.
1976 The simplest type of op structure is C<OP>: this has no children. Unary
1977 operators, C<UNOP>s, have one child, and this is pointed to by the
1978 C<op_first> field. Binary operators (C<BINOP>s) have not only an
1979 C<op_first> field but also an C<op_last> field. The most complex type of
1980 op is a C<LISTOP>, which has any number of children. In this case, the
1981 first child is pointed to by C<op_first> and the last child by
1982 C<op_last>. The children in between can be found by iteratively
1983 following the C<OpSIBLING> pointer from the first child to the last (but
1986 There are also some other op types: a C<PMOP> holds a regular expression,
1987 and has no children, and a C<LOOP> may or may not have children. If the
1988 C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
1989 complicate matters, if a C<UNOP> is actually a C<null> op after
1990 optimization (see L</Compile pass 2: context propagation>) it will still
1991 have children in accordance with its former type.
1993 Finally, there is a C<LOGOP>, or logic op. Like a C<LISTOP>, this has one
1994 or more children, but it doesn't have an C<op_last> field: so you have to
1995 follow C<op_first> and then the C<OpSIBLING> chain itself to find the
1996 last child. Instead it has an C<op_other> field, which is comparable to
1997 the C<op_next> field described below, and represents an alternate
1998 execution path. Operators like C<and>, C<or> and C<?> are C<LOGOP>s. Note
1999 that in general, C<op_other> may not point to any of the direct children
2002 Starting in version 5.21.2, perls built with the experimental
2003 define C<-DPERL_OP_PARENT> add an extra boolean flag for each op,
2004 C<op_moresib>. When not set, this indicates that this is the last op in an
2005 C<OpSIBLING> chain. This frees up the C<op_sibling> field on the last
2006 sibling to point back to the parent op. Under this build, that field is
2007 also renamed C<op_sibparent> to reflect its joint role. The macro
2008 C<OpSIBLING(o)> wraps this special behaviour, and always returns NULL on
2009 the last sibling. With this build the C<op_parent(o)> function can be
2010 used to find the parent of any op. Thus for forward compatibility, you
2011 should always use the C<OpSIBLING(o)> macro rather than accessing
2012 C<op_sibling> directly.
2014 Another way to examine the tree is to use a compiler back-end module, such
2017 =head2 Compile pass 1: check routines
2019 The tree is created by the compiler while I<yacc> code feeds it
2020 the constructions it recognizes. Since I<yacc> works bottom-up, so does
2021 the first pass of perl compilation.
2023 What makes this pass interesting for perl developers is that some
2024 optimization may be performed on this pass. This is optimization by
2025 so-called "check routines". The correspondence between node names
2026 and corresponding check routines is described in F<opcode.pl> (do not
2027 forget to run C<make regen_headers> if you modify this file).
2029 A check routine is called when the node is fully constructed except
2030 for the execution-order thread. Since at this time there are no
2031 back-links to the currently constructed node, one can do most any
2032 operation to the top-level node, including freeing it and/or creating
2033 new nodes above/below it.
2035 The check routine returns the node which should be inserted into the
2036 tree (if the top-level node was not modified, check routine returns
2039 By convention, check routines have names C<ck_*>. They are usually
2040 called from C<new*OP> subroutines (or C<convert>) (which in turn are
2041 called from F<perly.y>).
2043 =head2 Compile pass 1a: constant folding
2045 Immediately after the check routine is called the returned node is
2046 checked for being compile-time executable. If it is (the value is
2047 judged to be constant) it is immediately executed, and a I<constant>
2048 node with the "return value" of the corresponding subtree is
2049 substituted instead. The subtree is deleted.
2051 If constant folding was not performed, the execution-order thread is
2054 =head2 Compile pass 2: context propagation
2056 When a context for a part of compile tree is known, it is propagated
2057 down through the tree. At this time the context can have 5 values
2058 (instead of 2 for runtime context): void, boolean, scalar, list, and
2059 lvalue. In contrast with the pass 1 this pass is processed from top
2060 to bottom: a node's context determines the context for its children.
2062 Additional context-dependent optimizations are performed at this time.
2063 Since at this moment the compile tree contains back-references (via
2064 "thread" pointers), nodes cannot be free()d now. To allow
2065 optimized-away nodes at this stage, such nodes are null()ified instead
2066 of free()ing (i.e. their type is changed to OP_NULL).
2068 =head2 Compile pass 3: peephole optimization
2070 After the compile tree for a subroutine (or for an C<eval> or a file)
2071 is created, an additional pass over the code is performed. This pass
2072 is neither top-down or bottom-up, but in the execution order (with
2073 additional complications for conditionals). Optimizations performed
2074 at this stage are subject to the same restrictions as in the pass 2.
2076 Peephole optimizations are done by calling the function pointed to
2077 by the global variable C<PL_peepp>. By default, C<PL_peepp> just
2078 calls the function pointed to by the global variable C<PL_rpeepp>.
2079 By default, that performs some basic op fixups and optimisations along
2080 the execution-order op chain, and recursively calls C<PL_rpeepp> for
2081 each side chain of ops (resulting from conditionals). Extensions may
2082 provide additional optimisations or fixups, hooking into either the
2083 per-subroutine or recursive stage, like this:
2085 static peep_t prev_peepp;
2086 static void my_peep(pTHX_ OP *o)
2088 /* custom per-subroutine optimisation goes here */
2089 prev_peepp(aTHX_ o);
2090 /* custom per-subroutine optimisation may also go here */
2093 prev_peepp = PL_peepp;
2096 static peep_t prev_rpeepp;
2097 static void my_rpeep(pTHX_ OP *o)
2100 for(; o; o = o->op_next) {
2101 /* custom per-op optimisation goes here */
2103 prev_rpeepp(aTHX_ orig_o);
2106 prev_rpeepp = PL_rpeepp;
2107 PL_rpeepp = my_rpeep;
2109 =head2 Pluggable runops
2111 The compile tree is executed in a runops function. There are two runops
2112 functions, in F<run.c> and in F<dump.c>. C<Perl_runops_debug> is used
2113 with DEBUGGING and C<Perl_runops_standard> is used otherwise. For fine
2114 control over the execution of the compile tree it is possible to provide
2115 your own runops function.
2117 It's probably best to copy one of the existing runops functions and
2118 change it to suit your needs. Then, in the BOOT section of your XS
2121 PL_runops = my_runops;
2123 This function should be as efficient as possible to keep your programs
2124 running as fast as possible.
2126 =head2 Compile-time scope hooks
2128 As of perl 5.14 it is possible to hook into the compile-time lexical
2129 scope mechanism using C<Perl_blockhook_register>. This is used like
2132 STATIC void my_start_hook(pTHX_ int full);
2133 STATIC BHK my_hooks;
2136 BhkENTRY_set(&my_hooks, bhk_start, my_start_hook);
2137 Perl_blockhook_register(aTHX_ &my_hooks);
2139 This will arrange to have C<my_start_hook> called at the start of
2140 compiling every lexical scope. The available hooks are:
2144 =item C<void bhk_start(pTHX_ int full)>
2146 This is called just after starting a new lexical scope. Note that Perl
2151 creates two scopes: the first starts at the C<(> and has C<full == 1>,
2152 the second starts at the C<{> and has C<full == 0>. Both end at the
2153 C<}>, so calls to C<start> and C<pre>/C<post_end> will match. Anything
2154 pushed onto the save stack by this hook will be popped just before the
2155 scope ends (between the C<pre_> and C<post_end> hooks, in fact).
2157 =item C<void bhk_pre_end(pTHX_ OP **o)>
2159 This is called at the end of a lexical scope, just before unwinding the
2160 stack. I<o> is the root of the optree representing the scope; it is a
2161 double pointer so you can replace the OP if you need to.
2163 =item C<void bhk_post_end(pTHX_ OP **o)>
2165 This is called at the end of a lexical scope, just after unwinding the
2166 stack. I<o> is as above. Note that it is possible for calls to C<pre_>
2167 and C<post_end> to nest, if there is something on the save stack that
2170 =item C<void bhk_eval(pTHX_ OP *const o)>
2172 This is called just before starting to compile an C<eval STRING>, C<do
2173 FILE>, C<require> or C<use>, after the eval has been set up. I<o> is the
2174 OP that requested the eval, and will normally be an C<OP_ENTEREVAL>,
2175 C<OP_DOFILE> or C<OP_REQUIRE>.
2179 Once you have your hook functions, you need a C<BHK> structure to put
2180 them in. It's best to allocate it statically, since there is no way to
2181 free it once it's registered. The function pointers should be inserted
2182 into this structure using the C<BhkENTRY_set> macro, which will also set
2183 flags indicating which entries are valid. If you do need to allocate
2184 your C<BHK> dynamically for some reason, be sure to zero it before you
2187 Once registered, there is no mechanism to switch these hooks off, so if
2188 that is necessary you will need to do this yourself. An entry in C<%^H>
2189 is probably the best way, so the effect is lexically scoped; however it
2190 is also possible to use the C<BhkDISABLE> and C<BhkENABLE> macros to
2191 temporarily switch entries on and off. You should also be aware that
2192 generally speaking at least one scope will have opened before your
2193 extension is loaded, so you will see some C<pre>/C<post_end> pairs that
2194 didn't have a matching C<start>.
2196 =head1 Examining internal data structures with the C<dump> functions
2198 To aid debugging, the source file F<dump.c> contains a number of
2199 functions which produce formatted output of internal data structures.
2201 The most commonly used of these functions is C<Perl_sv_dump>; it's used
2202 for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
2203 C<sv_dump> to produce debugging output from Perl-space, so users of that
2204 module should already be familiar with its format.
2206 C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
2207 derivatives, and produces output similar to C<perl -Dx>; in fact,
2208 C<Perl_dump_eval> will dump the main root of the code being evaluated,
2209 exactly like C<-Dx>.
2211 Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
2212 op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
2213 subroutines in a package like so: (Thankfully, these are all xsubs, so
2214 there is no op tree)
2216 (gdb) print Perl_dump_packsubs(PL_defstash)
2218 SUB attributes::bootstrap = (xsub 0x811fedc 0)
2220 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
2222 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
2224 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
2226 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
2228 and C<Perl_dump_all>, which dumps all the subroutines in the stash and
2229 the op tree of the main root.
2231 =head1 How multiple interpreters and concurrency are supported
2233 =head2 Background and PERL_IMPLICIT_CONTEXT
2235 The Perl interpreter can be regarded as a closed box: it has an API
2236 for feeding it code or otherwise making it do things, but it also has
2237 functions for its own use. This smells a lot like an object, and
2238 there are ways for you to build Perl so that you can have multiple
2239 interpreters, with one interpreter represented either as a C structure,
2240 or inside a thread-specific structure. These structures contain all
2241 the context, the state of that interpreter.
2243 One macro controls the major Perl build flavor: MULTIPLICITY. The
2244 MULTIPLICITY build has a C structure that packages all the interpreter
2245 state. With multiplicity-enabled perls, PERL_IMPLICIT_CONTEXT is also
2246 normally defined, and enables the support for passing in a "hidden" first
2247 argument that represents all three data structures. MULTIPLICITY makes
2248 multi-threaded perls possible (with the ithreads threading model, related
2249 to the macro USE_ITHREADS.)
2251 Two other "encapsulation" macros are the PERL_GLOBAL_STRUCT and
2252 PERL_GLOBAL_STRUCT_PRIVATE (the latter turns on the former, and the
2253 former turns on MULTIPLICITY.) The PERL_GLOBAL_STRUCT causes all the
2254 internal variables of Perl to be wrapped inside a single global struct,
2255 struct perl_vars, accessible as (globals) &PL_Vars or PL_VarsPtr or
2256 the function Perl_GetVars(). The PERL_GLOBAL_STRUCT_PRIVATE goes
2257 one step further, there is still a single struct (allocated in main()
2258 either from heap or from stack) but there are no global data symbols
2259 pointing to it. In either case the global struct should be initialized
2260 as the very first thing in main() using Perl_init_global_struct() and
2261 correspondingly tear it down after perl_free() using Perl_free_global_struct(),
2262 please see F<miniperlmain.c> for usage details. You may also need
2263 to use C<dVAR> in your coding to "declare the global variables"
2264 when you are using them. dTHX does this for you automatically.
2266 To see whether you have non-const data you can use a BSD (or GNU)
2269 nm libperl.a | grep -v ' [TURtr] '
2271 If this displays any C<D> or C<d> symbols (or possibly C<C> or C<c>),
2272 you have non-const data. The symbols the C<grep> removed are as follows:
2273 C<Tt> are I<text>, or code, the C<Rr> are I<read-only> (const) data,
2274 and the C<U> is <undefined>, external symbols referred to.
2276 The test F<t/porting/libperl.t> does this kind of symbol sanity
2277 checking on C<libperl.a>.
2279 For backward compatibility reasons defining just PERL_GLOBAL_STRUCT
2280 doesn't actually hide all symbols inside a big global struct: some
2281 PerlIO_xxx vtables are left visible. The PERL_GLOBAL_STRUCT_PRIVATE
2282 then hides everything (see how the PERLIO_FUNCS_DECL is used).
2284 All this obviously requires a way for the Perl internal functions to be
2285 either subroutines taking some kind of structure as the first
2286 argument, or subroutines taking nothing as the first argument. To
2287 enable these two very different ways of building the interpreter,
2288 the Perl source (as it does in so many other situations) makes heavy
2289 use of macros and subroutine naming conventions.
2291 First problem: deciding which functions will be public API functions and
2292 which will be private. All functions whose names begin C<S_> are private
2293 (think "S" for "secret" or "static"). All other functions begin with
2294 "Perl_", but just because a function begins with "Perl_" does not mean it is
2295 part of the API. (See L</Internal
2296 Functions>.) The easiest way to be B<sure> a
2297 function is part of the API is to find its entry in L<perlapi>.
2298 If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
2299 think it should be (i.e., you need it for your extension), send mail via
2300 L<perlbug> explaining why you think it should be.
2302 Second problem: there must be a syntax so that the same subroutine
2303 declarations and calls can pass a structure as their first argument,
2304 or pass nothing. To solve this, the subroutines are named and
2305 declared in a particular way. Here's a typical start of a static
2306 function used within the Perl guts:
2309 S_incline(pTHX_ char *s)
2311 STATIC becomes "static" in C, and may be #define'd to nothing in some
2312 configurations in the future.
2314 A public function (i.e. part of the internal API, but not necessarily
2315 sanctioned for use in extensions) begins like this:
2318 Perl_sv_setiv(pTHX_ SV* dsv, IV num)
2320 C<pTHX_> is one of a number of macros (in F<perl.h>) that hide the
2321 details of the interpreter's context. THX stands for "thread", "this",
2322 or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
2323 The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
2324 or 'd' for B<d>eclaration, so we have C<pTHX>, C<aTHX> and C<dTHX>, and
2327 When Perl is built without options that set PERL_IMPLICIT_CONTEXT, there is no
2328 first argument containing the interpreter's context. The trailing underscore
2329 in the pTHX_ macro indicates that the macro expansion needs a comma
2330 after the context argument because other arguments follow it. If
2331 PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
2332 subroutine is not prototyped to take the extra argument. The form of the
2333 macro without the trailing underscore is used when there are no additional
2336 When a core function calls another, it must pass the context. This
2337 is normally hidden via macros. Consider C<sv_setiv>. It expands into
2338 something like this:
2340 #ifdef PERL_IMPLICIT_CONTEXT
2341 #define sv_setiv(a,b) Perl_sv_setiv(aTHX_ a, b)
2342 /* can't do this for vararg functions, see below */
2344 #define sv_setiv Perl_sv_setiv
2347 This works well, and means that XS authors can gleefully write:
2351 and still have it work under all the modes Perl could have been
2354 This doesn't work so cleanly for varargs functions, though, as macros
2355 imply that the number of arguments is known in advance. Instead we
2356 either need to spell them out fully, passing C<aTHX_> as the first
2357 argument (the Perl core tends to do this with functions like
2358 Perl_warner), or use a context-free version.
2360 The context-free version of Perl_warner is called
2361 Perl_warner_nocontext, and does not take the extra argument. Instead
2362 it does dTHX; to get the context from thread-local storage. We
2363 C<#define warner Perl_warner_nocontext> so that extensions get source
2364 compatibility at the expense of performance. (Passing an arg is
2365 cheaper than grabbing it from thread-local storage.)
2367 You can ignore [pad]THXx when browsing the Perl headers/sources.
2368 Those are strictly for use within the core. Extensions and embedders
2369 need only be aware of [pad]THX.
2371 =head2 So what happened to dTHR?
2373 C<dTHR> was introduced in perl 5.005 to support the older thread model.
2374 The older thread model now uses the C<THX> mechanism to pass context
2375 pointers around, so C<dTHR> is not useful any more. Perl 5.6.0 and
2376 later still have it for backward source compatibility, but it is defined
2379 =head2 How do I use all this in extensions?
2381 When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
2382 any functions in the Perl API will need to pass the initial context
2383 argument somehow. The kicker is that you will need to write it in
2384 such a way that the extension still compiles when Perl hasn't been
2385 built with PERL_IMPLICIT_CONTEXT enabled.
2387 There are three ways to do this. First, the easy but inefficient way,
2388 which is also the default, in order to maintain source compatibility
2389 with extensions: whenever F<XSUB.h> is #included, it redefines the aTHX
2390 and aTHX_ macros to call a function that will return the context.
2391 Thus, something like:
2395 in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
2398 Perl_sv_setiv(Perl_get_context(), sv, num);
2400 or to this otherwise:
2402 Perl_sv_setiv(sv, num);
2404 You don't have to do anything new in your extension to get this; since
2405 the Perl library provides Perl_get_context(), it will all just
2408 The second, more efficient way is to use the following template for
2411 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2416 STATIC void my_private_function(int arg1, int arg2);
2419 my_private_function(int arg1, int arg2)
2421 dTHX; /* fetch context */
2422 ... call many Perl API functions ...
2427 MODULE = Foo PACKAGE = Foo
2435 my_private_function(arg, 10);
2437 Note that the only two changes from the normal way of writing an
2438 extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
2439 including the Perl headers, followed by a C<dTHX;> declaration at
2440 the start of every function that will call the Perl API. (You'll
2441 know which functions need this, because the C compiler will complain
2442 that there's an undeclared identifier in those functions.) No changes
2443 are needed for the XSUBs themselves, because the XS() macro is
2444 correctly defined to pass in the implicit context if needed.
2446 The third, even more efficient way is to ape how it is done within
2450 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2455 /* pTHX_ only needed for functions that call Perl API */
2456 STATIC void my_private_function(pTHX_ int arg1, int arg2);
2459 my_private_function(pTHX_ int arg1, int arg2)
2461 /* dTHX; not needed here, because THX is an argument */
2462 ... call Perl API functions ...
2467 MODULE = Foo PACKAGE = Foo
2475 my_private_function(aTHX_ arg, 10);
2477 This implementation never has to fetch the context using a function
2478 call, since it is always passed as an extra argument. Depending on
2479 your needs for simplicity or efficiency, you may mix the previous
2480 two approaches freely.
2482 Never add a comma after C<pTHX> yourself--always use the form of the
2483 macro with the underscore for functions that take explicit arguments,
2484 or the form without the argument for functions with no explicit arguments.
2486 If one is compiling Perl with the C<-DPERL_GLOBAL_STRUCT> the C<dVAR>
2487 definition is needed if the Perl global variables (see F<perlvars.h>
2488 or F<globvar.sym>) are accessed in the function and C<dTHX> is not
2489 used (the C<dTHX> includes the C<dVAR> if necessary). One notices
2490 the need for C<dVAR> only with the said compile-time define, because
2491 otherwise the Perl global variables are visible as-is.
2493 =head2 Should I do anything special if I call perl from multiple threads?
2495 If you create interpreters in one thread and then proceed to call them in
2496 another, you need to make sure perl's own Thread Local Storage (TLS) slot is
2497 initialized correctly in each of those threads.
2499 The C<perl_alloc> and C<perl_clone> API functions will automatically set
2500 the TLS slot to the interpreter they created, so that there is no need to do
2501 anything special if the interpreter is always accessed in the same thread that
2502 created it, and that thread did not create or call any other interpreters
2503 afterwards. If that is not the case, you have to set the TLS slot of the
2504 thread before calling any functions in the Perl API on that particular
2505 interpreter. This is done by calling the C<PERL_SET_CONTEXT> macro in that
2506 thread as the first thing you do:
2508 /* do this before doing anything else with some_perl */
2509 PERL_SET_CONTEXT(some_perl);
2511 ... other Perl API calls on some_perl go here ...
2513 =head2 Future Plans and PERL_IMPLICIT_SYS
2515 Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
2516 that the interpreter knows about itself and pass it around, so too are
2517 there plans to allow the interpreter to bundle up everything it knows
2518 about the environment it's running on. This is enabled with the
2519 PERL_IMPLICIT_SYS macro. Currently it only works with USE_ITHREADS on
2522 This allows the ability to provide an extra pointer (called the "host"
2523 environment) for all the system calls. This makes it possible for
2524 all the system stuff to maintain their own state, broken down into
2525 seven C structures. These are thin wrappers around the usual system
2526 calls (see F<win32/perllib.c>) for the default perl executable, but for a
2527 more ambitious host (like the one that would do fork() emulation) all
2528 the extra work needed to pretend that different interpreters are
2529 actually different "processes", would be done here.
2531 The Perl engine/interpreter and the host are orthogonal entities.
2532 There could be one or more interpreters in a process, and one or
2533 more "hosts", with free association between them.
2535 =head1 Internal Functions
2537 All of Perl's internal functions which will be exposed to the outside
2538 world are prefixed by C<Perl_> so that they will not conflict with XS
2539 functions or functions used in a program in which Perl is embedded.
2540 Similarly, all global variables begin with C<PL_>. (By convention,
2541 static functions start with C<S_>.)
2543 Inside the Perl core (C<PERL_CORE> defined), you can get at the functions
2544 either with or without the C<Perl_> prefix, thanks to a bunch of defines
2545 that live in F<embed.h>. Note that extension code should I<not> set
2546 C<PERL_CORE>; this exposes the full perl internals, and is likely to cause
2547 breakage of the XS in each new perl release.
2549 The file F<embed.h> is generated automatically from
2550 F<embed.pl> and F<embed.fnc>. F<embed.pl> also creates the prototyping
2551 header files for the internal functions, generates the documentation
2552 and a lot of other bits and pieces. It's important that when you add
2553 a new function to the core or change an existing one, you change the
2554 data in the table in F<embed.fnc> as well. Here's a sample entry from
2557 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
2559 The second column is the return type, the third column the name. Columns
2560 after that are the arguments. The first column is a set of flags:
2566 This function is a part of the public
2567 API. All such functions should also
2568 have 'd', very few do not.
2572 This function has a C<Perl_> prefix; i.e. it is defined as
2577 This function has documentation using the C<apidoc> feature which we'll
2578 look at in a second. Some functions have 'd' but not 'A'; docs are good.
2582 Other available flags are:
2588 This is a static function and is defined as C<STATIC S_whatever>, and
2589 usually called within the sources as C<whatever(...)>.
2593 This does not need an interpreter context, so the definition has no
2594 C<pTHX>, and it follows that callers don't use C<aTHX>. (See
2595 L</Background and PERL_IMPLICIT_CONTEXT>.)
2599 This function never returns; C<croak>, C<exit> and friends.
2603 This function takes a variable number of arguments, C<printf> style.
2604 The argument list should end with C<...>, like this:
2606 Afprd |void |croak |const char* pat|...
2610 This function is part of the experimental development API, and may change
2611 or disappear without notice.
2615 This function should not have a compatibility macro to define, say,
2616 C<Perl_parse> to C<parse>. It must be called as C<Perl_parse>.
2620 This function isn't exported out of the Perl core.
2624 This is implemented as a macro.
2628 This function is explicitly exported.
2632 This function is visible to extensions included in the Perl core.
2636 Binary backward compatibility; this function is a macro but also has
2637 a C<Perl_> implementation (which is exported).
2641 See the comments at the top of C<embed.fnc> for others.
2645 If you edit F<embed.pl> or F<embed.fnc>, you will need to run
2646 C<make regen_headers> to force a rebuild of F<embed.h> and other
2647 auto-generated files.
2649 =head2 Formatted Printing of IVs, UVs, and NVs
2651 If you are printing IVs, UVs, or NVS instead of the stdio(3) style
2652 formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
2653 following macros for portability
2658 UVxf UV in hexadecimal
2663 These will take care of 64-bit integers and long doubles.
2666 printf("IV is %"IVdf"\n", iv);
2668 The IVdf will expand to whatever is the correct format for the IVs.
2670 Note that there are different "long doubles": Perl will use
2671 whatever the compiler has.
2673 If you are printing addresses of pointers, use UVxf combined
2674 with PTR2UV(), do not use %lx or %p.
2676 =head2 Pointer-To-Integer and Integer-To-Pointer
2678 Because pointer size does not necessarily equal integer size,
2679 use the follow macros to do it right.
2684 INT2PTR(pointertotype, integer)
2689 SV *sv = INT2PTR(SV*, iv);
2696 =head2 Exception Handling
2698 There are a couple of macros to do very basic exception handling in XS
2699 modules. You have to define C<NO_XSLOCKS> before including F<XSUB.h> to
2700 be able to use these macros:
2705 You can use these macros if you call code that may croak, but you need
2706 to do some cleanup before giving control back to Perl. For example:
2708 dXCPT; /* set up necessary variables */
2711 code_that_may_croak();
2716 /* do cleanup here */
2720 Note that you always have to rethrow an exception that has been
2721 caught. Using these macros, it is not possible to just catch the
2722 exception and ignore it. If you have to ignore the exception, you
2723 have to use the C<call_*> function.
2725 The advantage of using the above macros is that you don't have
2726 to setup an extra function for C<call_*>, and that using these
2727 macros is faster than using C<call_*>.
2729 =head2 Source Documentation
2731 There's an effort going on to document the internal functions and
2732 automatically produce reference manuals from them -- L<perlapi> is one
2733 such manual which details all the functions which are available to XS
2734 writers. L<perlintern> is the autogenerated manual for the functions
2735 which are not part of the API and are supposedly for internal use only.
2737 Source documentation is created by putting POD comments into the C
2741 =for apidoc sv_setiv
2743 Copies an integer into the given SV. Does not handle 'set' magic. See
2749 Please try and supply some documentation if you add functions to the
2752 =head2 Backwards compatibility
2754 The Perl API changes over time. New functions are
2755 added or the interfaces of existing functions are
2756 changed. The C<Devel::PPPort> module tries to
2757 provide compatibility code for some of these changes, so XS writers don't
2758 have to code it themselves when supporting multiple versions of Perl.
2760 C<Devel::PPPort> generates a C header file F<ppport.h> that can also
2761 be run as a Perl script. To generate F<ppport.h>, run:
2763 perl -MDevel::PPPort -eDevel::PPPort::WriteFile
2765 Besides checking existing XS code, the script can also be used to retrieve
2766 compatibility information for various API calls using the C<--api-info>
2767 command line switch. For example:
2769 % perl ppport.h --api-info=sv_magicext
2771 For details, see C<perldoc ppport.h>.
2773 =head1 Unicode Support
2775 Perl 5.6.0 introduced Unicode support. It's important for porters and XS
2776 writers to understand this support and make sure that the code they
2777 write does not corrupt Unicode data.
2779 =head2 What B<is> Unicode, anyway?
2781 In the olden, less enlightened times, we all used to use ASCII. Most of
2782 us did, anyway. The big problem with ASCII is that it's American. Well,
2783 no, that's not actually the problem; the problem is that it's not
2784 particularly useful for people who don't use the Roman alphabet. What
2785 used to happen was that particular languages would stick their own
2786 alphabet in the upper range of the sequence, between 128 and 255. Of
2787 course, we then ended up with plenty of variants that weren't quite
2788 ASCII, and the whole point of it being a standard was lost.
2790 Worse still, if you've got a language like Chinese or
2791 Japanese that has hundreds or thousands of characters, then you really
2792 can't fit them into a mere 256, so they had to forget about ASCII
2793 altogether, and build their own systems using pairs of numbers to refer
2796 To fix this, some people formed Unicode, Inc. and
2797 produced a new character set containing all the characters you can
2798 possibly think of and more. There are several ways of representing these
2799 characters, and the one Perl uses is called UTF-8. UTF-8 uses
2800 a variable number of bytes to represent a character. You can learn more
2801 about Unicode and Perl's Unicode model in L<perlunicode>.
2803 (On EBCDIC platforms, Perl uses instead UTF-EBCDIC, which is a form of
2804 UTF-8 adapted for EBCDIC platforms. Below, we just talk about UTF-8.
2805 UTF-EBCDIC is like UTF-8, but the details are different. The macros
2806 hide the differences from you, just remember that the particular numbers
2807 and bit patterns presented below will differ in UTF-EBCDIC.)
2809 =head2 How can I recognise a UTF-8 string?
2811 You can't. This is because UTF-8 data is stored in bytes just like
2812 non-UTF-8 data. The Unicode character 200, (C<0xC8> for you hex types)
2813 capital E with a grave accent, is represented by the two bytes
2814 C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
2815 has that byte sequence as well. So you can't tell just by looking -- this
2816 is what makes Unicode input an interesting problem.
2818 In general, you either have to know what you're dealing with, or you
2819 have to guess. The API function C<is_utf8_string> can help; it'll tell
2820 you if a string contains only valid UTF-8 characters, and the chances
2821 of a non-UTF-8 string looking like valid UTF-8 become very small very
2822 quickly with increasing string length. On a character-by-character
2823 basis, C<isUTF8_CHAR>
2824 will tell you whether the current character in a string is valid UTF-8.
2826 =head2 How does UTF-8 represent Unicode characters?
2828 As mentioned above, UTF-8 uses a variable number of bytes to store a
2829 character. Characters with values 0...127 are stored in one
2830 byte, just like good ol' ASCII. Character 128 is stored as
2831 C<v194.128>; this continues up to character 191, which is
2832 C<v194.191>. Now we've run out of bits (191 is binary
2833 C<10111111>) so we move on; character 192 is C<v195.128>. And
2834 so it goes on, moving to three bytes at character 2048.
2835 L<perlunicode/Unicode Encodings> has pictures of how this works.
2837 Assuming you know you're dealing with a UTF-8 string, you can find out
2838 how long the first character in it is with the C<UTF8SKIP> macro:
2840 char *utf = "\305\233\340\240\201";
2843 len = UTF8SKIP(utf); /* len is 2 here */
2845 len = UTF8SKIP(utf); /* len is 3 here */
2847 Another way to skip over characters in a UTF-8 string is to use
2848 C<utf8_hop>, which takes a string and a number of characters to skip
2849 over. You're on your own about bounds checking, though, so don't use it
2852 All bytes in a multi-byte UTF-8 character will have the high bit set,
2853 so you can test if you need to do something special with this
2854 character like this (the C<UTF8_IS_INVARIANT()> is a macro that tests
2855 whether the byte is encoded as a single byte even in UTF-8):
2858 U8 *utf_end; /* 1 beyond buffer pointed to by utf */
2859 UV uv; /* Note: a UV, not a U8, not a char */
2860 STRLEN len; /* length of character in bytes */
2862 if (!UTF8_IS_INVARIANT(*utf))
2863 /* Must treat this as UTF-8 */
2864 uv = utf8_to_uvchr_buf(utf, utf_end, &len);
2866 /* OK to treat this character as a byte */
2869 You can also see in that example that we use C<utf8_to_uvchr_buf> to get the
2870 value of the character; the inverse function C<uvchr_to_utf8> is available
2871 for putting a UV into UTF-8:
2873 if (!UVCHR_IS_INVARIANT(uv))
2874 /* Must treat this as UTF8 */
2875 utf8 = uvchr_to_utf8(utf8, uv);
2877 /* OK to treat this character as a byte */
2880 You B<must> convert characters to UVs using the above functions if
2881 you're ever in a situation where you have to match UTF-8 and non-UTF-8
2882 characters. You may not skip over UTF-8 characters in this case. If you
2883 do this, you'll lose the ability to match hi-bit non-UTF-8 characters;
2884 for instance, if your UTF-8 string contains C<v196.172>, and you skip
2885 that character, you can never match a C<chr(200)> in a non-UTF-8 string.
2888 (Note that we don't have to test for invariant characters in the
2889 examples above. The functions work on any well-formed UTF-8 input.
2890 It's just that its faster to avoid the function overhead when it's not
2893 =head2 How does Perl store UTF-8 strings?
2895 Currently, Perl deals with UTF-8 strings and non-UTF-8 strings
2896 slightly differently. A flag in the SV, C<SVf_UTF8>, indicates that the
2897 string is internally encoded as UTF-8. Without it, the byte value is the
2898 codepoint number and vice versa. This flag is only meaningful if the SV
2899 is C<SvPOK> or immediately after stringification via C<SvPV> or a
2900 similar macro. You can check and manipulate this flag with the
2907 This flag has an important effect on Perl's treatment of the string: if
2908 UTF-8 data is not properly distinguished, regular expressions,
2909 C<length>, C<substr> and other string handling operations will have
2910 undesirable (wrong) results.
2912 The problem comes when you have, for instance, a string that isn't
2913 flagged as UTF-8, and contains a byte sequence that could be UTF-8 --
2914 especially when combining non-UTF-8 and UTF-8 strings.
2916 Never forget that the C<SVf_UTF8> flag is separate from the PV value; you
2917 need to be sure you don't accidentally knock it off while you're
2918 manipulating SVs. More specifically, you cannot expect to do this:
2927 nsv = newSVpvn(p, len);
2929 The C<char*> string does not tell you the whole story, and you can't
2930 copy or reconstruct an SV just by copying the string value. Check if the
2931 old SV has the UTF8 flag set (I<after> the C<SvPV> call), and act
2935 is_utf8 = SvUTF8(sv);
2936 frobnicate(p, is_utf8);
2937 nsv = newSVpvn(p, len);
2941 In the above, your C<frobnicate> function has been changed to be made
2942 aware of whether or not it's dealing with UTF-8 data, so that it can
2943 handle the string appropriately.
2945 Since just passing an SV to an XS function and copying the data of
2946 the SV is not enough to copy the UTF8 flags, even less right is just
2947 passing a S<C<char *>> to an XS function.
2949 For full generality, use the L<C<DO_UTF8>|perlapi/DO_UTF8> macro to see if the
2950 string in an SV is to be I<treated> as UTF-8. This takes into account
2951 if the call to the XS function is being made from within the scope of
2952 L<S<C<use bytes>>|bytes>. If so, the underlying bytes that comprise the
2953 UTF-8 string are to be exposed, rather than the character they
2954 represent. But this pragma should only really be used for debugging and
2955 perhaps low-level testing at the byte level. Hence most XS code need
2956 not concern itself with this, but various areas of the perl core do need
2959 And this isn't the whole story. Starting in Perl v5.12, strings that
2960 aren't encoded in UTF-8 may also be treated as Unicode under various
2961 conditions (see L<perlunicode/ASCII Rules versus Unicode Rules>).
2962 This is only really a problem for characters whose ordinals are between
2963 128 and 255, and their behavior varies under ASCII versus Unicode rules
2964 in ways that your code cares about (see L<perlunicode/The "Unicode Bug">).
2965 There is no published API for dealing with this, as it is subject to
2966 change, but you can look at the code for C<pp_lc> in F<pp.c> for an
2967 example as to how it's currently done.
2969 =head2 How do I convert a string to UTF-8?
2971 If you're mixing UTF-8 and non-UTF-8 strings, it is necessary to upgrade
2972 the non-UTF-8 strings to UTF-8. If you've got an SV, the easiest way to do
2975 sv_utf8_upgrade(sv);
2977 However, you must not do this, for example:
2980 sv_utf8_upgrade(left);
2982 If you do this in a binary operator, you will actually change one of the
2983 strings that came into the operator, and, while it shouldn't be noticeable
2984 by the end user, it can cause problems in deficient code.
2986 Instead, C<bytes_to_utf8> will give you a UTF-8-encoded B<copy> of its
2987 string argument. This is useful for having the data available for
2988 comparisons and so on, without harming the original SV. There's also
2989 C<utf8_to_bytes> to go the other way, but naturally, this will fail if
2990 the string contains any characters above 255 that can't be represented
2993 =head2 How do I compare strings?
2995 L<perlapi/sv_cmp> and L<perlapi/sv_cmp_flags> do a lexigraphic
2996 comparison of two SV's, and handle UTF-8ness properly. Note, however,
2997 that Unicode specifies a much fancier mechanism for collation, available
2998 via the L<Unicode::Collate> module.
3000 To just compare two strings for equality/non-equality, you can just use
3001 L<C<memEQ()>|perlapi/memEQ> and L<C<memNE()>|perlapi/memEQ> as usual,
3002 except the strings must be both UTF-8 or not UTF-8 encoded.
3004 To compare two strings case-insensitively, use
3005 L<C<foldEQ_utf8()>|perlapi/foldEQ_utf8> (the strings don't have to have
3006 the same UTF-8ness).
3008 =head2 Is there anything else I need to know?
3010 Not really. Just remember these things:
3016 There's no way to tell if a S<C<char *>> or S<C<U8 *>> string is UTF-8
3017 or not. But you can tell if an SV is to be treated as UTF-8 by calling
3018 C<DO_UTF8> on it, after stringifying it with C<SvPV> or a similar
3019 macro. And, you can tell if SV is actually UTF-8 (even if it is not to
3020 be treated as such) by looking at its C<SvUTF8> flag (again after
3021 stringifying it). Don't forget to set the flag if something should be
3023 Treat the flag as part of the PV, even though it's not -- if you pass on
3024 the PV to somewhere, pass on the flag too.
3028 If a string is UTF-8, B<always> use C<utf8_to_uvchr_buf> to get at the value,
3029 unless C<UTF8_IS_INVARIANT(*s)> in which case you can use C<*s>.
3033 When writing a character UV to a UTF-8 string, B<always> use
3034 C<uvchr_to_utf8>, unless C<UVCHR_IS_INVARIANT(uv))> in which case
3035 you can use C<*s = uv>.
3039 Mixing UTF-8 and non-UTF-8 strings is
3040 tricky. Use C<bytes_to_utf8> to get
3041 a new string which is UTF-8 encoded, and then combine them.
3045 =head1 Custom Operators
3047 Custom operator support is an experimental feature that allows you to
3048 define your own ops. This is primarily to allow the building of
3049 interpreters for other languages in the Perl core, but it also allows
3050 optimizations through the creation of "macro-ops" (ops which perform the
3051 functions of multiple ops which are usually executed together, such as
3052 C<gvsv, gvsv, add>.)
3054 This feature is implemented as a new op type, C<OP_CUSTOM>. The Perl
3055 core does not "know" anything special about this op type, and so it will
3056 not be involved in any optimizations. This also means that you can
3057 define your custom ops to be any op structure -- unary, binary, list and
3060 It's important to know what custom operators won't do for you. They
3061 won't let you add new syntax to Perl, directly. They won't even let you
3062 add new keywords, directly. In fact, they won't change the way Perl
3063 compiles a program at all. You have to do those changes yourself, after
3064 Perl has compiled the program. You do this either by manipulating the op
3065 tree using a C<CHECK> block and the C<B::Generate> module, or by adding
3066 a custom peephole optimizer with the C<optimize> module.
3068 When you do this, you replace ordinary Perl ops with custom ops by
3069 creating ops with the type C<OP_CUSTOM> and the C<op_ppaddr> of your own
3070 PP function. This should be defined in XS code, and should look like
3071 the PP ops in C<pp_*.c>. You are responsible for ensuring that your op
3072 takes the appropriate number of values from the stack, and you are
3073 responsible for adding stack marks if necessary.
3075 You should also "register" your op with the Perl interpreter so that it
3076 can produce sensible error and warning messages. Since it is possible to
3077 have multiple custom ops within the one "logical" op type C<OP_CUSTOM>,
3078 Perl uses the value of C<< o->op_ppaddr >> to determine which custom op
3079 it is dealing with. You should create an C<XOP> structure for each
3080 ppaddr you use, set the properties of the custom op with
3081 C<XopENTRY_set>, and register the structure against the ppaddr using
3082 C<Perl_custom_op_register>. A trivial example might look like:
3085 static OP *my_pp(pTHX);
3088 XopENTRY_set(&my_xop, xop_name, "myxop");
3089 XopENTRY_set(&my_xop, xop_desc, "Useless custom op");
3090 Perl_custom_op_register(aTHX_ my_pp, &my_xop);
3092 The available fields in the structure are:
3098 A short name for your op. This will be included in some error messages,
3099 and will also be returned as C<< $op->name >> by the L<B|B> module, so
3100 it will appear in the output of module like L<B::Concise|B::Concise>.
3104 A short description of the function of the op.
3108 Which of the various C<*OP> structures this op uses. This should be one of
3109 the C<OA_*> constants from F<op.h>, namely
3129 =item OA_PVOP_OR_SVOP
3131 This should be interpreted as 'C<PVOP>' only. The C<_OR_SVOP> is because
3132 the only core C<PVOP>, C<OP_TRANS>, can sometimes be a C<SVOP> instead.
3140 The other C<OA_*> constants should not be used.
3144 This member is of type C<Perl_cpeep_t>, which expands to C<void
3145 (*Perl_cpeep_t)(aTHX_ OP *o, OP *oldop)>. If it is set, this function
3146 will be called from C<Perl_rpeep> when ops of this type are encountered
3147 by the peephole optimizer. I<o> is the OP that needs optimizing;
3148 I<oldop> is the previous OP optimized, whose C<op_next> points to I<o>.
3152 C<B::Generate> directly supports the creation of custom ops by name.
3156 Until May 1997, this document was maintained by Jeff Okamoto
3157 E<lt>okamoto@corp.hp.comE<gt>. It is now maintained as part of Perl
3158 itself by the Perl 5 Porters E<lt>perl5-porters@perl.orgE<gt>.
3160 With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
3161 Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
3162 Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
3163 Stephen McCamant, and Gurusamy Sarathy.
3167 L<perlapi>, L<perlintern>, L<perlxs>, L<perlembed>