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, inversion
43 lists, used in regular expression data structures, are scalars, each
44 consisting of an array of UVs which are accessed through PVs. But,
45 using it for non-strings requires care, as the underlying assumption of
46 much of the internals is that PVs are just for strings. Often, for
47 example, a trailing NUL is tacked on automatically. The non-string use
48 is documented only in this paragraph.)
50 The seven routines are:
55 SV* newSVpv(const char*, STRLEN);
56 SV* newSVpvn(const char*, STRLEN);
57 SV* newSVpvf(const char*, ...);
60 C<STRLEN> is an integer type (Size_t, usually defined as size_t in
61 F<config.h>) guaranteed to be large enough to represent the size of
62 any string that perl can handle.
64 In the unlikely case of a SV requiring more complex initialisation, you
65 can create an empty SV with newSV(len). If C<len> is 0 an empty SV of
66 type NULL is returned, else an SV of type PV is returned with len + 1 (for
67 the NUL) bytes of storage allocated, accessible via SvPVX. In both cases
68 the SV has the undef value.
70 SV *sv = newSV(0); /* no storage allocated */
71 SV *sv = newSV(10); /* 10 (+1) bytes of uninitialised storage
74 To change the value of an I<already-existing> SV, there are eight routines:
76 void sv_setiv(SV*, IV);
77 void sv_setuv(SV*, UV);
78 void sv_setnv(SV*, double);
79 void sv_setpv(SV*, const char*);
80 void sv_setpvn(SV*, const char*, STRLEN)
81 void sv_setpvf(SV*, const char*, ...);
82 void sv_vsetpvfn(SV*, const char*, STRLEN, va_list *,
84 void sv_setsv(SV*, SV*);
86 Notice that you can choose to specify the length of the string to be
87 assigned by using C<sv_setpvn>, C<newSVpvn>, or C<newSVpv>, or you may
88 allow Perl to calculate the length by using C<sv_setpv> or by specifying
89 0 as the second argument to C<newSVpv>. Be warned, though, that Perl will
90 determine the string's length by using C<strlen>, which depends on the
91 string terminating with a NUL character, and not otherwise containing
94 The arguments of C<sv_setpvf> are processed like C<sprintf>, and the
95 formatted output becomes the value.
97 C<sv_vsetpvfn> is an analogue of C<vsprintf>, but it allows you to specify
98 either a pointer to a variable argument list or the address and length of
99 an array of SVs. The last argument points to a boolean; on return, if that
100 boolean is true, then locale-specific information has been used to format
101 the string, and the string's contents are therefore untrustworthy (see
102 L<perlsec>). This pointer may be NULL if that information is not
103 important. Note that this function requires you to specify the length of
106 The C<sv_set*()> functions are not generic enough to operate on values
107 that have "magic". See L<Magic Virtual Tables> later in this document.
109 All SVs that contain strings should be terminated with a NUL character.
110 If it is not NUL-terminated there is a risk of
111 core dumps and corruptions from code which passes the string to C
112 functions or system calls which expect a NUL-terminated string.
113 Perl's own functions typically add a trailing NUL for this reason.
114 Nevertheless, you should be very careful when you pass a string stored
115 in an SV to a C function or system call.
117 To access the actual value that an SV points to, you can use the macros:
122 SvPV(SV*, STRLEN len)
125 which will automatically coerce the actual scalar type into an IV, UV, double,
128 In the C<SvPV> macro, the length of the string returned is placed into the
129 variable C<len> (this is a macro, so you do I<not> use C<&len>). If you do
130 not care what the length of the data is, use the C<SvPV_nolen> macro.
131 Historically the C<SvPV> macro with the global variable C<PL_na> has been
132 used in this case. But that can be quite inefficient because C<PL_na> must
133 be accessed in thread-local storage in threaded Perl. In any case, remember
134 that Perl allows arbitrary strings of data that may both contain NULs and
135 might not be terminated by a NUL.
137 Also remember that C doesn't allow you to safely say C<foo(SvPV(s, len),
138 len);>. It might work with your 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 NUL byte (perl's own string functions typically do
160 C<SvGROW(sv, len + 1)>).
162 If you have an SV and want to know what kind of data Perl thinks is stored
163 in it, you can use the following macros to check the type of SV you have.
169 You can get and set the current length of the string stored in an SV with
170 the following macros:
173 SvCUR_set(SV*, I32 val)
175 You can also get a pointer to the end of the string stored in the SV
180 But note that these last three macros are valid only if C<SvPOK()> is true.
182 If you want to append something to the end of string stored in an C<SV*>,
183 you can use the following functions:
185 void sv_catpv(SV*, const char*);
186 void sv_catpvn(SV*, const char*, STRLEN);
187 void sv_catpvf(SV*, const char*, ...);
188 void sv_vcatpvfn(SV*, const char*, STRLEN, va_list *, SV **,
190 void sv_catsv(SV*, SV*);
192 The first function calculates the length of the string to be appended by
193 using C<strlen>. In the second, you specify the length of the string
194 yourself. The third function processes its arguments like C<sprintf> and
195 appends the formatted output. The fourth function works like C<vsprintf>.
196 You can specify the address and length of an array of SVs instead of the
197 va_list argument. The fifth function extends the string stored in the first
198 SV with the string stored in the second SV. It also forces the second SV
199 to be interpreted as a string.
201 The C<sv_cat*()> functions are not generic enough to operate on values that
202 have "magic". See L<Magic Virtual Tables> later in this document.
204 If you know the name of a scalar variable, you can get a pointer to its SV
205 by using the following:
207 SV* get_sv("package::varname", 0);
209 This returns NULL if the variable does not exist.
211 If you want to know if this variable (or any other SV) is actually C<defined>,
216 The scalar C<undef> value is stored in an SV instance called C<PL_sv_undef>.
218 Its address can be used whenever an C<SV*> is needed. Make sure that
219 you don't try to compare a random sv with C<&PL_sv_undef>. For example
220 when interfacing Perl code, it'll work correctly for:
224 But won't work when called as:
229 So to repeat always use SvOK() to check whether an sv is defined.
231 Also you have to be careful when using C<&PL_sv_undef> as a value in
232 AVs or HVs (see L<AVs, HVs and undefined values>).
234 There are also the two values C<PL_sv_yes> and C<PL_sv_no>, which contain
235 boolean TRUE and FALSE values, respectively. Like C<PL_sv_undef>, their
236 addresses can be used whenever an C<SV*> is needed.
238 Do not be fooled into thinking that C<(SV *) 0> is the same as C<&PL_sv_undef>.
242 if (I-am-to-return-a-real-value) {
243 sv = sv_2mortal(newSViv(42));
247 This code tries to return a new SV (which contains the value 42) if it should
248 return a real value, or undef otherwise. Instead it has returned a NULL
249 pointer which, somewhere down the line, will cause a segmentation violation,
250 bus error, or just weird results. Change the zero to C<&PL_sv_undef> in the
251 first line and all will be well.
253 To free an SV that you've created, call C<SvREFCNT_dec(SV*)>. Normally this
254 call is not necessary (see L<Reference Counts and Mortality>).
258 Perl provides the function C<sv_chop> to efficiently remove characters
259 from the beginning of a string; you give it an SV and a pointer to
260 somewhere inside the PV, and it discards everything before the
261 pointer. The efficiency comes by means of a little hack: instead of
262 actually removing the characters, C<sv_chop> sets the flag C<OOK>
263 (offset OK) to signal to other functions that the offset hack is in
264 effect, and it puts the number of bytes chopped off into the IV field
265 of the SV. It then moves the PV pointer (called C<SvPVX>) forward that
266 many bytes, and adjusts C<SvCUR> and C<SvLEN>.
268 Hence, at this point, the start of the buffer that we allocated lives
269 at C<SvPVX(sv) - SvIV(sv)> in memory and the PV pointer is pointing
270 into the middle of this allocated storage.
272 This is best demonstrated by example:
274 % ./perl -Ilib -MDevel::Peek -le '$a="12345"; $a=~s/.//; Dump($a)'
275 SV = PVIV(0x8128450) at 0x81340f0
277 FLAGS = (POK,OOK,pPOK)
279 PV = 0x8135781 ( "1" . ) "2345"\0
283 Here the number of bytes chopped off (1) is put into IV, and
284 C<Devel::Peek::Dump> helpfully reminds us that this is an offset. The
285 portion of the string between the "real" and the "fake" beginnings is
286 shown in parentheses, and the values of C<SvCUR> and C<SvLEN> reflect
287 the fake beginning, not the real one.
289 Something similar to the offset hack is performed on AVs to enable
290 efficient shifting and splicing off the beginning of the array; while
291 C<AvARRAY> points to the first element in the array that is visible from
292 Perl, C<AvALLOC> points to the real start of the C array. These are
293 usually the same, but a C<shift> operation can be carried out by
294 increasing C<AvARRAY> by one and decreasing C<AvFILL> and C<AvMAX>.
295 Again, the location of the real start of the C array only comes into
296 play when freeing the array. See C<av_shift> in F<av.c>.
298 =head2 What's Really Stored in an SV?
300 Recall that the usual method of determining the type of scalar you have is
301 to use C<Sv*OK> macros. Because a scalar can be both a number and a string,
302 usually these macros will always return TRUE and calling the C<Sv*V>
303 macros will do the appropriate conversion of string to integer/double or
304 integer/double to string.
306 If you I<really> need to know if you have an integer, double, or string
307 pointer in an SV, you can use the following three macros instead:
313 These will tell you if you truly have an integer, double, or string pointer
314 stored in your SV. The "p" stands for private.
316 There are various ways in which the private and public flags may differ.
317 For example, a tied SV may have a valid underlying value in the IV slot
318 (so SvIOKp is true), but the data should be accessed via the FETCH
319 routine rather than directly, so SvIOK is false. Another is when
320 numeric conversion has occurred and precision has been lost: only the
321 private flag is set on 'lossy' values. So when an NV is converted to an
322 IV with loss, SvIOKp, SvNOKp and SvNOK will be set, while SvIOK wont be.
324 In general, though, it's best to use the C<Sv*V> macros.
326 =head2 Working with AVs
328 There are two ways to create and load an AV. The first method creates an
333 The second method both creates the AV and initially populates it with SVs:
335 AV* av_make(I32 num, SV **ptr);
337 The second argument points to an array containing C<num> C<SV*>'s. Once the
338 AV has been created, the SVs can be destroyed, if so desired.
340 Once the AV has been created, the following operations are possible on it:
342 void av_push(AV*, SV*);
345 void av_unshift(AV*, I32 num);
347 These should be familiar operations, with the exception of C<av_unshift>.
348 This routine adds C<num> elements at the front of the array with the C<undef>
349 value. You must then use C<av_store> (described below) to assign values
350 to these new elements.
352 Here are some other functions:
355 SV** av_fetch(AV*, I32 key, I32 lval);
356 SV** av_store(AV*, I32 key, SV* val);
358 The C<av_len> function returns the highest index value in an array (just
359 like $#array in Perl). If the array is empty, -1 is returned. The
360 C<av_fetch> function returns the value at index C<key>, but if C<lval>
361 is non-zero, then C<av_fetch> will store an undef value at that index.
362 The C<av_store> function stores the value C<val> at index C<key>, and does
363 not increment the reference count of C<val>. Thus the caller is responsible
364 for taking care of that, and if C<av_store> returns NULL, the caller will
365 have to decrement the reference count to avoid a memory leak. Note that
366 C<av_fetch> and C<av_store> both return C<SV**>'s, not C<SV*>'s as their
373 void av_extend(AV*, I32 key);
375 The C<av_clear> function deletes all the elements in the AV* array, but
376 does not actually delete the array itself. The C<av_undef> function will
377 delete all the elements in the array plus the array itself. The
378 C<av_extend> function extends the array so that it contains at least C<key+1>
379 elements. If C<key+1> is less than the currently allocated length of the array,
380 then nothing is done.
382 If you know the name of an array variable, you can get a pointer to its AV
383 by using the following:
385 AV* get_av("package::varname", 0);
387 This returns NULL if the variable does not exist.
389 See L<Understanding the Magic of Tied Hashes and Arrays> for more
390 information on how to use the array access functions on tied arrays.
392 =head2 Working with HVs
394 To create an HV, you use the following routine:
398 Once the HV has been created, the following operations are possible on it:
400 SV** hv_store(HV*, const char* key, U32 klen, SV* val, U32 hash);
401 SV** hv_fetch(HV*, const char* key, U32 klen, I32 lval);
403 The C<klen> parameter is the length of the key being passed in (Note that
404 you cannot pass 0 in as a value of C<klen> to tell Perl to measure the
405 length of the key). The C<val> argument contains the SV pointer to the
406 scalar being stored, and C<hash> is the precomputed hash value (zero if
407 you want C<hv_store> to calculate it for you). The C<lval> parameter
408 indicates whether this fetch is actually a part of a store operation, in
409 which case a new undefined value will be added to the HV with the supplied
410 key and C<hv_fetch> will return as if the value had already existed.
412 Remember that C<hv_store> and C<hv_fetch> return C<SV**>'s and not just
413 C<SV*>. To access the scalar value, you must first dereference the return
414 value. However, you should check to make sure that the return value is
415 not NULL before dereferencing it.
417 The first of these two functions checks if a hash table entry exists, and the
420 bool hv_exists(HV*, const char* key, U32 klen);
421 SV* hv_delete(HV*, const char* key, U32 klen, I32 flags);
423 If C<flags> does not include the C<G_DISCARD> flag then C<hv_delete> will
424 create and return a mortal copy of the deleted value.
426 And more miscellaneous functions:
431 Like their AV counterparts, C<hv_clear> deletes all the entries in the hash
432 table but does not actually delete the hash table. The C<hv_undef> deletes
433 both the entries and the hash table itself.
435 Perl keeps the actual data in a linked list of structures with a typedef of HE.
436 These contain the actual key and value pointers (plus extra administrative
437 overhead). The key is a string pointer; the value is an C<SV*>. However,
438 once you have an C<HE*>, to get the actual key and value, use the routines
441 I32 hv_iterinit(HV*);
442 /* Prepares starting point to traverse hash table */
443 HE* hv_iternext(HV*);
444 /* Get the next entry, and return a pointer to a
445 structure that has both the key and value */
446 char* hv_iterkey(HE* entry, I32* retlen);
447 /* Get the key from an HE structure and also return
448 the length of the key string */
449 SV* hv_iterval(HV*, HE* entry);
450 /* Return an SV pointer to the value of the HE
452 SV* hv_iternextsv(HV*, char** key, I32* retlen);
453 /* This convenience routine combines hv_iternext,
454 hv_iterkey, and hv_iterval. The key and retlen
455 arguments are return values for the key and its
456 length. The value is returned in the SV* argument */
458 If you know the name of a hash variable, you can get a pointer to its HV
459 by using the following:
461 HV* get_hv("package::varname", 0);
463 This returns NULL if the variable does not exist.
465 The hash algorithm is defined in the C<PERL_HASH> macro:
467 PERL_HASH(hash, key, klen)
469 The exact implementation of this macro varies by architecture and version
470 of perl, and the return value may change per invocation, so the value
471 is only valid for the duration of a single perl process.
473 See L<Understanding the Magic of Tied Hashes and Arrays> for more
474 information on how to use the hash access functions on tied hashes.
476 =head2 Hash API Extensions
478 Beginning with version 5.004, the following functions are also supported:
480 HE* hv_fetch_ent (HV* tb, SV* key, I32 lval, U32 hash);
481 HE* hv_store_ent (HV* tb, SV* key, SV* val, U32 hash);
483 bool hv_exists_ent (HV* tb, SV* key, U32 hash);
484 SV* hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
486 SV* hv_iterkeysv (HE* entry);
488 Note that these functions take C<SV*> keys, which simplifies writing
489 of extension code that deals with hash structures. These functions
490 also allow passing of C<SV*> keys to C<tie> functions without forcing
491 you to stringify the keys (unlike the previous set of functions).
493 They also return and accept whole hash entries (C<HE*>), making their
494 use more efficient (since the hash number for a particular string
495 doesn't have to be recomputed every time). See L<perlapi> for detailed
498 The following macros must always be used to access the contents of hash
499 entries. Note that the arguments to these macros must be simple
500 variables, since they may get evaluated more than once. See
501 L<perlapi> for detailed descriptions of these macros.
503 HePV(HE* he, STRLEN len)
507 HeSVKEY_force(HE* he)
508 HeSVKEY_set(HE* he, SV* sv)
510 These two lower level macros are defined, but must only be used when
511 dealing with keys that are not C<SV*>s:
516 Note that both C<hv_store> and C<hv_store_ent> do not increment the
517 reference count of the stored C<val>, which is the caller's responsibility.
518 If these functions return a NULL value, the caller will usually have to
519 decrement the reference count of C<val> to avoid a memory leak.
521 =head2 AVs, HVs and undefined values
523 Sometimes you have to store undefined values in AVs or HVs. Although
524 this may be a rare case, it can be tricky. That's because you're
525 used to using C<&PL_sv_undef> if you need an undefined SV.
527 For example, intuition tells you that this XS code:
530 av_store( av, 0, &PL_sv_undef );
532 is equivalent to this Perl code:
537 Unfortunately, this isn't true. AVs use C<&PL_sv_undef> as a marker
538 for indicating that an array element has not yet been initialized.
539 Thus, C<exists $av[0]> would be true for the above Perl code, but
540 false for the array generated by the XS code.
542 Other problems can occur when storing C<&PL_sv_undef> in HVs:
544 hv_store( hv, "key", 3, &PL_sv_undef, 0 );
546 This will indeed make the value C<undef>, but if you try to modify
547 the value of C<key>, you'll get the following error:
549 Modification of non-creatable hash value attempted
551 In perl 5.8.0, C<&PL_sv_undef> was also used to mark placeholders
552 in restricted hashes. This caused such hash entries not to appear
553 when iterating over the hash or when checking for the keys
554 with the C<hv_exists> function.
556 You can run into similar problems when you store C<&PL_sv_yes> or
557 C<&PL_sv_no> into AVs or HVs. Trying to modify such elements
558 will give you the following error:
560 Modification of a read-only value attempted
562 To make a long story short, you can use the special variables
563 C<&PL_sv_undef>, C<&PL_sv_yes> and C<&PL_sv_no> with AVs and
564 HVs, but you have to make sure you know what you're doing.
566 Generally, if you want to store an undefined value in an AV
567 or HV, you should not use C<&PL_sv_undef>, but rather create a
568 new undefined value using the C<newSV> function, for example:
570 av_store( av, 42, newSV(0) );
571 hv_store( hv, "foo", 3, newSV(0), 0 );
575 References are a special type of scalar that point to other data types
576 (including other references).
578 To create a reference, use either of the following functions:
580 SV* newRV_inc((SV*) thing);
581 SV* newRV_noinc((SV*) thing);
583 The C<thing> argument can be any of an C<SV*>, C<AV*>, or C<HV*>. The
584 functions are identical except that C<newRV_inc> increments the reference
585 count of the C<thing>, while C<newRV_noinc> does not. For historical
586 reasons, C<newRV> is a synonym for C<newRV_inc>.
588 Once you have a reference, you can use the following macro to dereference
593 then call the appropriate routines, casting the returned C<SV*> to either an
594 C<AV*> or C<HV*>, if required.
596 To determine if an SV is a reference, you can use the following macro:
600 To discover what type of value the reference refers to, use the following
601 macro and then check the return value.
605 The most useful types that will be returned are:
614 SVt_PVGV Glob (possibly a file handle)
615 SVt_PVMG Blessed or Magical Scalar
617 See the F<sv.h> header file for more details.
619 =head2 Blessed References and Class Objects
621 References are also used to support object-oriented programming. In perl's
622 OO lexicon, an object is simply a reference that has been blessed into a
623 package (or class). Once blessed, the programmer may now use the reference
624 to access the various methods in the class.
626 A reference can be blessed into a package with the following function:
628 SV* sv_bless(SV* sv, HV* stash);
630 The C<sv> argument must be a reference value. The C<stash> argument
631 specifies which class the reference will belong to. See
632 L<Stashes and Globs> for information on converting class names into stashes.
634 /* Still under construction */
636 The following function upgrades rv to reference if not already one.
637 Creates a new SV for rv to point to. If C<classname> is non-null, the SV
638 is blessed into the specified class. SV is returned.
640 SV* newSVrv(SV* rv, const char* classname);
642 The following three functions copy integer, unsigned integer or double
643 into an SV whose reference is C<rv>. SV is blessed if C<classname> is
646 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
647 SV* sv_setref_uv(SV* rv, const char* classname, UV uv);
648 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
650 The following function copies the pointer value (I<the address, not the
651 string!>) into an SV whose reference is rv. SV is blessed if C<classname>
654 SV* sv_setref_pv(SV* rv, const char* classname, void* pv);
656 The following function copies a string into an SV whose reference is C<rv>.
657 Set length to 0 to let Perl calculate the string length. SV is blessed if
658 C<classname> is non-null.
660 SV* sv_setref_pvn(SV* rv, const char* classname, char* pv,
663 The following function tests whether the SV is blessed into the specified
664 class. It does not check inheritance relationships.
666 int sv_isa(SV* sv, const char* name);
668 The following function tests whether the SV is a reference to a blessed object.
670 int sv_isobject(SV* sv);
672 The following function tests whether the SV is derived from the specified
673 class. SV can be either a reference to a blessed object or a string
674 containing a class name. This is the function implementing the
675 C<UNIVERSAL::isa> functionality.
677 bool sv_derived_from(SV* sv, const char* name);
679 To check if you've got an object derived from a specific class you have
682 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
684 =head2 Creating New Variables
686 To create a new Perl variable with an undef value which can be accessed from
687 your Perl script, use the following routines, depending on the variable type.
689 SV* get_sv("package::varname", GV_ADD);
690 AV* get_av("package::varname", GV_ADD);
691 HV* get_hv("package::varname", GV_ADD);
693 Notice the use of GV_ADD as the second parameter. The new variable can now
694 be set, using the routines appropriate to the data type.
696 There are additional macros whose values may be bitwise OR'ed with the
697 C<GV_ADD> argument to enable certain extra features. Those bits are:
703 Marks the variable as multiply defined, thus preventing the:
705 Name <varname> used only once: possible typo
713 Had to create <varname> unexpectedly
715 if the variable did not exist before the function was called.
719 If you do not specify a package name, the variable is created in the current
722 =head2 Reference Counts and Mortality
724 Perl uses a reference count-driven garbage collection mechanism. SVs,
725 AVs, or HVs (xV for short in the following) start their life with a
726 reference count of 1. If the reference count of an xV ever drops to 0,
727 then it will be destroyed and its memory made available for reuse.
729 This normally doesn't happen at the Perl level unless a variable is
730 undef'ed or the last variable holding a reference to it is changed or
731 overwritten. At the internal level, however, reference counts can be
732 manipulated with the following macros:
734 int SvREFCNT(SV* sv);
735 SV* SvREFCNT_inc(SV* sv);
736 void SvREFCNT_dec(SV* sv);
738 However, there is one other function which manipulates the reference
739 count of its argument. The C<newRV_inc> function, you will recall,
740 creates a reference to the specified argument. As a side effect,
741 it increments the argument's reference count. If this is not what
742 you want, use C<newRV_noinc> instead.
744 For example, imagine you want to return a reference from an XSUB function.
745 Inside the XSUB routine, you create an SV which initially has a reference
746 count of one. Then you call C<newRV_inc>, passing it the just-created SV.
747 This returns the reference as a new SV, but the reference count of the
748 SV you passed to C<newRV_inc> has been incremented to two. Now you
749 return the reference from the XSUB routine and forget about the SV.
750 But Perl hasn't! Whenever the returned reference is destroyed, the
751 reference count of the original SV is decreased to one and nothing happens.
752 The SV will hang around without any way to access it until Perl itself
753 terminates. This is a memory leak.
755 The correct procedure, then, is to use C<newRV_noinc> instead of
756 C<newRV_inc>. Then, if and when the last reference is destroyed,
757 the reference count of the SV will go to zero and it will be destroyed,
758 stopping any memory leak.
760 There are some convenience functions available that can help with the
761 destruction of xVs. These functions introduce the concept of "mortality".
762 An xV that is mortal has had its reference count marked to be decremented,
763 but not actually decremented, until "a short time later". Generally the
764 term "short time later" means a single Perl statement, such as a call to
765 an XSUB function. The actual determinant for when mortal xVs have their
766 reference count decremented depends on two macros, SAVETMPS and FREETMPS.
767 See L<perlcall> and L<perlxs> for more details on these macros.
769 "Mortalization" then is at its simplest a deferred C<SvREFCNT_dec>.
770 However, if you mortalize a variable twice, the reference count will
771 later be decremented twice.
773 "Mortal" SVs are mainly used for SVs that are placed on perl's stack.
774 For example an SV which is created just to pass a number to a called sub
775 is made mortal to have it cleaned up automatically when it's popped off
776 the stack. Similarly, results returned by XSUBs (which are pushed on the
777 stack) are often made mortal.
779 To create a mortal variable, use the functions:
783 SV* sv_mortalcopy(SV*)
785 The first call creates a mortal SV (with no value), the second converts an existing
786 SV to a mortal SV (and thus defers a call to C<SvREFCNT_dec>), and the
787 third creates a mortal copy of an existing SV.
788 Because C<sv_newmortal> gives the new SV no value, it must normally be given one
789 via C<sv_setpv>, C<sv_setiv>, etc. :
791 SV *tmp = sv_newmortal();
792 sv_setiv(tmp, an_integer);
794 As that is multiple C statements it is quite common so see this idiom instead:
796 SV *tmp = sv_2mortal(newSViv(an_integer));
799 You should be careful about creating mortal variables. Strange things
800 can happen if you make the same value mortal within multiple contexts,
801 or if you make a variable mortal multiple times. Thinking of "Mortalization"
802 as deferred C<SvREFCNT_dec> should help to minimize such problems.
803 For example if you are passing an SV which you I<know> has a high enough REFCNT
804 to survive its use on the stack you need not do any mortalization.
805 If you are not sure then doing an C<SvREFCNT_inc> and C<sv_2mortal>, or
806 making a C<sv_mortalcopy> is safer.
808 The mortal routines are not just for SVs; AVs and HVs can be
809 made mortal by passing their address (type-casted to C<SV*>) to the
810 C<sv_2mortal> or C<sv_mortalcopy> routines.
812 =head2 Stashes and Globs
814 A B<stash> is a hash that contains all variables that are defined
815 within a package. Each key of the stash is a symbol
816 name (shared by all the different types of objects that have the same
817 name), and each value in the hash table is a GV (Glob Value). This GV
818 in turn contains references to the various objects of that name,
819 including (but not limited to) the following:
828 There is a single stash called C<PL_defstash> that holds the items that exist
829 in the C<main> package. To get at the items in other packages, append the
830 string "::" to the package name. The items in the C<Foo> package are in
831 the stash C<Foo::> in PL_defstash. The items in the C<Bar::Baz> package are
832 in the stash C<Baz::> in C<Bar::>'s stash.
834 To get the stash pointer for a particular package, use the function:
836 HV* gv_stashpv(const char* name, I32 flags)
837 HV* gv_stashsv(SV*, I32 flags)
839 The first function takes a literal string, the second uses the string stored
840 in the SV. Remember that a stash is just a hash table, so you get back an
841 C<HV*>. The C<flags> flag will create a new package if it is set to GV_ADD.
843 The name that C<gv_stash*v> wants is the name of the package whose symbol table
844 you want. The default package is called C<main>. If you have multiply nested
845 packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
848 Alternately, if you have an SV that is a blessed reference, you can find
849 out the stash pointer by using:
851 HV* SvSTASH(SvRV(SV*));
853 then use the following to get the package name itself:
855 char* HvNAME(HV* stash);
857 If you need to bless or re-bless an object you can use the following
860 SV* sv_bless(SV*, HV* stash)
862 where the first argument, an C<SV*>, must be a reference, and the second
863 argument is a stash. The returned C<SV*> can now be used in the same way
866 For more information on references and blessings, consult L<perlref>.
868 =head2 Double-Typed SVs
870 Scalar variables normally contain only one type of value, an integer,
871 double, pointer, or reference. Perl will automatically convert the
872 actual scalar data from the stored type into the requested type.
874 Some scalar variables contain more than one type of scalar data. For
875 example, the variable C<$!> contains either the numeric value of C<errno>
876 or its string equivalent from either C<strerror> or C<sys_errlist[]>.
878 To force multiple data values into an SV, you must do two things: use the
879 C<sv_set*v> routines to add the additional scalar type, then set a flag
880 so that Perl will believe it contains more than one type of data. The
881 four macros to set the flags are:
888 The particular macro you must use depends on which C<sv_set*v> routine
889 you called first. This is because every C<sv_set*v> routine turns on
890 only the bit for the particular type of data being set, and turns off
893 For example, to create a new Perl variable called "dberror" that contains
894 both the numeric and descriptive string error values, you could use the
898 extern char *dberror_list;
900 SV* sv = get_sv("dberror", GV_ADD);
901 sv_setiv(sv, (IV) dberror);
902 sv_setpv(sv, dberror_list[dberror]);
905 If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
906 macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
908 =head2 Magic Variables
910 [This section still under construction. Ignore everything here. Post no
911 bills. Everything not permitted is forbidden.]
913 Any SV may be magical, that is, it has special features that a normal
914 SV does not have. These features are stored in the SV structure in a
915 linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
928 Note this is current as of patchlevel 0, and could change at any time.
930 =head2 Assigning Magic
932 Perl adds magic to an SV using the sv_magic function:
934 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
936 The C<sv> argument is a pointer to the SV that is to acquire a new magical
939 If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
940 convert C<sv> to type C<SVt_PVMG>. Perl then continues by adding new magic
941 to the beginning of the linked list of magical features. Any prior entry
942 of the same type of magic is deleted. Note that this can be overridden,
943 and multiple instances of the same type of magic can be associated with an
946 The C<name> and C<namlen> arguments are used to associate a string with
947 the magic, typically the name of a variable. C<namlen> is stored in the
948 C<mg_len> field and if C<name> is non-null then either a C<savepvn> copy of
949 C<name> or C<name> itself is stored in the C<mg_ptr> field, depending on
950 whether C<namlen> is greater than zero or equal to zero respectively. As a
951 special case, if C<(name && namlen == HEf_SVKEY)> then C<name> is assumed
952 to contain an C<SV*> and is stored as-is with its REFCNT incremented.
954 The sv_magic function uses C<how> to determine which, if any, predefined
955 "Magic Virtual Table" should be assigned to the C<mg_virtual> field.
956 See the L<Magic Virtual Tables> section below. The C<how> argument is also
957 stored in the C<mg_type> field. The value of C<how> should be chosen
958 from the set of macros C<PERL_MAGIC_foo> found in F<perl.h>. Note that before
959 these macros were added, Perl internals used to directly use character
960 literals, so you may occasionally come across old code or documentation
961 referring to 'U' magic rather than C<PERL_MAGIC_uvar> for example.
963 The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
964 structure. If it is not the same as the C<sv> argument, the reference
965 count of the C<obj> object is incremented. If it is the same, or if
966 the C<how> argument is C<PERL_MAGIC_arylen>, or if it is a NULL pointer,
967 then C<obj> is merely stored, without the reference count being incremented.
969 See also C<sv_magicext> in L<perlapi> for a more flexible way to add magic
972 There is also a function to add magic to an C<HV>:
974 void hv_magic(HV *hv, GV *gv, int how);
976 This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
978 To remove the magic from an SV, call the function sv_unmagic:
980 int sv_unmagic(SV *sv, int type);
982 The C<type> argument should be equal to the C<how> value when the C<SV>
983 was initially made magical.
985 However, note that C<sv_unmagic> removes all magic of a certain C<type> from the
986 C<SV>. If you want to remove only certain magic of a C<type> based on the magic
987 virtual table, use C<sv_unmagicext> instead:
989 int sv_unmagicext(SV *sv, int type, MGVTBL *vtbl);
991 =head2 Magic Virtual Tables
993 The C<mg_virtual> field in the C<MAGIC> structure is a pointer to an
994 C<MGVTBL>, which is a structure of function pointers and stands for
995 "Magic Virtual Table" to handle the various operations that might be
996 applied to that variable.
998 The C<MGVTBL> has five (or sometimes eight) pointers to the following
1001 int (*svt_get)(SV* sv, MAGIC* mg);
1002 int (*svt_set)(SV* sv, MAGIC* mg);
1003 U32 (*svt_len)(SV* sv, MAGIC* mg);
1004 int (*svt_clear)(SV* sv, MAGIC* mg);
1005 int (*svt_free)(SV* sv, MAGIC* mg);
1007 int (*svt_copy)(SV *sv, MAGIC* mg, SV *nsv,
1008 const char *name, I32 namlen);
1009 int (*svt_dup)(MAGIC *mg, CLONE_PARAMS *param);
1010 int (*svt_local)(SV *nsv, MAGIC *mg);
1013 This MGVTBL structure is set at compile-time in F<perl.h> and there are
1014 currently 32 types. These different structures contain pointers to various
1015 routines that perform additional actions depending on which function is
1018 Function pointer Action taken
1019 ---------------- ------------
1020 svt_get Do something before the value of the SV is
1022 svt_set Do something after the SV is assigned a value.
1023 svt_len Report on the SV's length.
1024 svt_clear Clear something the SV represents.
1025 svt_free Free any extra storage associated with the SV.
1027 svt_copy copy tied variable magic to a tied element
1028 svt_dup duplicate a magic structure during thread cloning
1029 svt_local copy magic to local value during 'local'
1031 For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
1032 to an C<mg_type> of C<PERL_MAGIC_sv>) contains:
1034 { magic_get, magic_set, magic_len, 0, 0 }
1036 Thus, when an SV is determined to be magical and of type C<PERL_MAGIC_sv>,
1037 if a get operation is being performed, the routine C<magic_get> is
1038 called. All the various routines for the various magical types begin
1039 with C<magic_>. NOTE: the magic routines are not considered part of
1040 the Perl API, and may not be exported by the Perl library.
1042 The last three slots are a recent addition, and for source code
1043 compatibility they are only checked for if one of the three flags
1044 MGf_COPY, MGf_DUP or MGf_LOCAL is set in mg_flags. This means that most
1045 code can continue declaring a vtable as a 5-element value. These three are
1046 currently used exclusively by the threading code, and are highly subject
1049 The current kinds of Magic Virtual Tables are:
1052 This table is generated by regen/mg_vtable.pl. Any changes made here
1055 =for mg_vtable.pl begin
1058 (old-style char and macro) MGVTBL Type of magic
1059 -------------------------- ------ -------------
1060 \0 PERL_MAGIC_sv vtbl_sv Special scalar variable
1061 # PERL_MAGIC_arylen vtbl_arylen Array length ($#ary)
1062 % PERL_MAGIC_rhash (none) extra data for restricted
1064 & PERL_MAGIC_proto (none) my sub prototype CV
1065 . PERL_MAGIC_pos vtbl_pos pos() lvalue
1066 : PERL_MAGIC_symtab (none) extra data for symbol
1068 < PERL_MAGIC_backref vtbl_backref for weak ref data
1069 @ PERL_MAGIC_arylen_p (none) to move arylen out of XPVAV
1070 B PERL_MAGIC_bm vtbl_regexp Boyer-Moore
1071 (fast string search)
1072 c PERL_MAGIC_overload_table vtbl_ovrld Holds overload table
1074 D PERL_MAGIC_regdata vtbl_regdata Regex match position data
1076 d PERL_MAGIC_regdatum vtbl_regdatum Regex match position data
1078 E PERL_MAGIC_env vtbl_env %ENV hash
1079 e PERL_MAGIC_envelem vtbl_envelem %ENV hash element
1080 f PERL_MAGIC_fm vtbl_regexp Formline
1082 g PERL_MAGIC_regex_global vtbl_mglob m//g target
1083 H PERL_MAGIC_hints vtbl_hints %^H hash
1084 h PERL_MAGIC_hintselem vtbl_hintselem %^H hash element
1085 I PERL_MAGIC_isa vtbl_isa @ISA array
1086 i PERL_MAGIC_isaelem vtbl_isaelem @ISA array element
1087 k PERL_MAGIC_nkeys vtbl_nkeys scalar(keys()) lvalue
1088 L PERL_MAGIC_dbfile (none) Debugger %_<filename
1089 l PERL_MAGIC_dbline vtbl_dbline Debugger %_<filename
1091 N PERL_MAGIC_shared (none) Shared between threads
1092 n PERL_MAGIC_shared_scalar (none) Shared between threads
1093 o PERL_MAGIC_collxfrm vtbl_collxfrm Locale transformation
1094 P PERL_MAGIC_tied vtbl_pack Tied array or hash
1095 p PERL_MAGIC_tiedelem vtbl_packelem Tied array or hash element
1096 q PERL_MAGIC_tiedscalar vtbl_packelem Tied scalar or handle
1097 r PERL_MAGIC_qr vtbl_regexp precompiled qr// regex
1098 S PERL_MAGIC_sig (none) %SIG hash
1099 s PERL_MAGIC_sigelem vtbl_sigelem %SIG hash element
1100 t PERL_MAGIC_taint vtbl_taint Taintedness
1101 U PERL_MAGIC_uvar vtbl_uvar Available for use by
1103 u PERL_MAGIC_uvar_elem (none) Reserved for use by
1105 V PERL_MAGIC_vstring (none) SV was vstring literal
1106 v PERL_MAGIC_vec vtbl_vec vec() lvalue
1107 w PERL_MAGIC_utf8 vtbl_utf8 Cached UTF-8 information
1108 x PERL_MAGIC_substr vtbl_substr substr() lvalue
1109 y PERL_MAGIC_defelem vtbl_defelem Shadow "foreach" iterator
1110 variable / smart parameter
1112 ] PERL_MAGIC_checkcall vtbl_checkcall inlining/mutation of call
1114 ~ PERL_MAGIC_ext (none) Available for use by
1117 =for mg_vtable.pl end
1119 When an uppercase and lowercase letter both exist in the table, then the
1120 uppercase letter is typically used to represent some kind of composite type
1121 (a list or a hash), and the lowercase letter is used to represent an element
1122 of that composite type. Some internals code makes use of this case
1123 relationship. However, 'v' and 'V' (vec and v-string) are in no way related.
1125 The C<PERL_MAGIC_ext> and C<PERL_MAGIC_uvar> magic types are defined
1126 specifically for use by extensions and will not be used by perl itself.
1127 Extensions can use C<PERL_MAGIC_ext> magic to 'attach' private information
1128 to variables (typically objects). This is especially useful because
1129 there is no way for normal perl code to corrupt this private information
1130 (unlike using extra elements of a hash object).
1132 Similarly, C<PERL_MAGIC_uvar> magic can be used much like tie() to call a
1133 C function any time a scalar's value is used or changed. The C<MAGIC>'s
1134 C<mg_ptr> field points to a C<ufuncs> structure:
1137 I32 (*uf_val)(pTHX_ IV, SV*);
1138 I32 (*uf_set)(pTHX_ IV, SV*);
1142 When the SV is read from or written to, the C<uf_val> or C<uf_set>
1143 function will be called with C<uf_index> as the first arg and a pointer to
1144 the SV as the second. A simple example of how to add C<PERL_MAGIC_uvar>
1145 magic is shown below. Note that the ufuncs structure is copied by
1146 sv_magic, so you can safely allocate it on the stack.
1154 uf.uf_val = &my_get_fn;
1155 uf.uf_set = &my_set_fn;
1157 sv_magic(sv, 0, PERL_MAGIC_uvar, (char*)&uf, sizeof(uf));
1159 Attaching C<PERL_MAGIC_uvar> to arrays is permissible but has no effect.
1161 For hashes there is a specialized hook that gives control over hash
1162 keys (but not values). This hook calls C<PERL_MAGIC_uvar> 'get' magic
1163 if the "set" function in the C<ufuncs> structure is NULL. The hook
1164 is activated whenever the hash is accessed with a key specified as
1165 an C<SV> through the functions C<hv_store_ent>, C<hv_fetch_ent>,
1166 C<hv_delete_ent>, and C<hv_exists_ent>. Accessing the key as a string
1167 through the functions without the C<..._ent> suffix circumvents the
1168 hook. See L<Hash::Util::FieldHash/GUTS> for a detailed description.
1170 Note that because multiple extensions may be using C<PERL_MAGIC_ext>
1171 or C<PERL_MAGIC_uvar> magic, it is important for extensions to take
1172 extra care to avoid conflict. Typically only using the magic on
1173 objects blessed into the same class as the extension is sufficient.
1174 For C<PERL_MAGIC_ext> magic, it is usually a good idea to define an
1175 C<MGVTBL>, even if all its fields will be C<0>, so that individual
1176 C<MAGIC> pointers can be identified as a particular kind of magic
1177 using their magic virtual table. C<mg_findext> provides an easy way
1180 STATIC MGVTBL my_vtbl = { 0, 0, 0, 0, 0, 0, 0, 0 };
1183 if ((mg = mg_findext(sv, PERL_MAGIC_ext, &my_vtbl))) {
1184 /* this is really ours, not another module's PERL_MAGIC_ext */
1185 my_priv_data_t *priv = (my_priv_data_t *)mg->mg_ptr;
1189 Also note that the C<sv_set*()> and C<sv_cat*()> functions described
1190 earlier do B<not> invoke 'set' magic on their targets. This must
1191 be done by the user either by calling the C<SvSETMAGIC()> macro after
1192 calling these functions, or by using one of the C<sv_set*_mg()> or
1193 C<sv_cat*_mg()> functions. Similarly, generic C code must call the
1194 C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
1195 obtained from external sources in functions that don't handle magic.
1196 See L<perlapi> for a description of these functions.
1197 For example, calls to the C<sv_cat*()> functions typically need to be
1198 followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
1199 since their implementation handles 'get' magic.
1201 =head2 Finding Magic
1203 MAGIC *mg_find(SV *sv, int type); /* Finds the magic pointer of that
1206 This routine returns a pointer to a C<MAGIC> structure stored in the SV.
1207 If the SV does not have that magical feature, C<NULL> is returned. If the
1208 SV has multiple instances of that magical feature, the first one will be
1209 returned. C<mg_findext> can be used to find a C<MAGIC> structure of an SV
1210 based on both its magic type and its magic virtual table:
1212 MAGIC *mg_findext(SV *sv, int type, MGVTBL *vtbl);
1214 Also, if the SV passed to C<mg_find> or C<mg_findext> is not of type
1215 SVt_PVMG, Perl may core dump.
1217 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
1219 This routine checks to see what types of magic C<sv> has. If the mg_type
1220 field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
1221 the mg_type field is changed to be the lowercase letter.
1223 =head2 Understanding the Magic of Tied Hashes and Arrays
1225 Tied hashes and arrays are magical beasts of the C<PERL_MAGIC_tied>
1228 WARNING: As of the 5.004 release, proper usage of the array and hash
1229 access functions requires understanding a few caveats. Some
1230 of these caveats are actually considered bugs in the API, to be fixed
1231 in later releases, and are bracketed with [MAYCHANGE] below. If
1232 you find yourself actually applying such information in this section, be
1233 aware that the behavior may change in the future, umm, without warning.
1235 The perl tie function associates a variable with an object that implements
1236 the various GET, SET, etc methods. To perform the equivalent of the perl
1237 tie function from an XSUB, you must mimic this behaviour. The code below
1238 carries out the necessary steps - firstly it creates a new hash, and then
1239 creates a second hash which it blesses into the class which will implement
1240 the tie methods. Lastly it ties the two hashes together, and returns a
1241 reference to the new tied hash. Note that the code below does NOT call the
1242 TIEHASH method in the MyTie class -
1243 see L<Calling Perl Routines from within C Programs> for details on how
1254 tie = newRV_noinc((SV*)newHV());
1255 stash = gv_stashpv("MyTie", GV_ADD);
1256 sv_bless(tie, stash);
1257 hv_magic(hash, (GV*)tie, PERL_MAGIC_tied);
1258 RETVAL = newRV_noinc(hash);
1262 The C<av_store> function, when given a tied array argument, merely
1263 copies the magic of the array onto the value to be "stored", using
1264 C<mg_copy>. It may also return NULL, indicating that the value did not
1265 actually need to be stored in the array. [MAYCHANGE] After a call to
1266 C<av_store> on a tied array, the caller will usually need to call
1267 C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1268 TIEARRAY object. If C<av_store> did return NULL, a call to
1269 C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1272 The previous paragraph is applicable verbatim to tied hash access using the
1273 C<hv_store> and C<hv_store_ent> functions as well.
1275 C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1276 C<hv_fetch_ent> actually return an undefined mortal value whose magic
1277 has been initialized using C<mg_copy>. Note the value so returned does not
1278 need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1279 need to call C<mg_get()> on the returned value in order to actually invoke
1280 the perl level "FETCH" method on the underlying TIE object. Similarly,
1281 you may also call C<mg_set()> on the return value after possibly assigning
1282 a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
1283 method on the TIE object. [/MAYCHANGE]
1286 In other words, the array or hash fetch/store functions don't really
1287 fetch and store actual values in the case of tied arrays and hashes. They
1288 merely call C<mg_copy> to attach magic to the values that were meant to be
1289 "stored" or "fetched". Later calls to C<mg_get> and C<mg_set> actually
1290 do the job of invoking the TIE methods on the underlying objects. Thus
1291 the magic mechanism currently implements a kind of lazy access to arrays
1294 Currently (as of perl version 5.004), use of the hash and array access
1295 functions requires the user to be aware of whether they are operating on
1296 "normal" hashes and arrays, or on their tied variants. The API may be
1297 changed to provide more transparent access to both tied and normal data
1298 types in future versions.
1301 You would do well to understand that the TIEARRAY and TIEHASH interfaces
1302 are mere sugar to invoke some perl method calls while using the uniform hash
1303 and array syntax. The use of this sugar imposes some overhead (typically
1304 about two to four extra opcodes per FETCH/STORE operation, in addition to
1305 the creation of all the mortal variables required to invoke the methods).
1306 This overhead will be comparatively small if the TIE methods are themselves
1307 substantial, but if they are only a few statements long, the overhead
1308 will not be insignificant.
1310 =head2 Localizing changes
1312 Perl has a very handy construction
1319 This construction is I<approximately> equivalent to
1328 The biggest difference is that the first construction would
1329 reinstate the initial value of $var, irrespective of how control exits
1330 the block: C<goto>, C<return>, C<die>/C<eval>, etc. It is a little bit
1331 more efficient as well.
1333 There is a way to achieve a similar task from C via Perl API: create a
1334 I<pseudo-block>, and arrange for some changes to be automatically
1335 undone at the end of it, either explicit, or via a non-local exit (via
1336 die()). A I<block>-like construct is created by a pair of
1337 C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1338 Such a construct may be created specially for some important localized
1339 task, or an existing one (like boundaries of enclosing Perl
1340 subroutine/block, or an existing pair for freeing TMPs) may be
1341 used. (In the second case the overhead of additional localization must
1342 be almost negligible.) Note that any XSUB is automatically enclosed in
1343 an C<ENTER>/C<LEAVE> pair.
1345 Inside such a I<pseudo-block> the following service is available:
1349 =item C<SAVEINT(int i)>
1351 =item C<SAVEIV(IV i)>
1353 =item C<SAVEI32(I32 i)>
1355 =item C<SAVELONG(long i)>
1357 These macros arrange things to restore the value of integer variable
1358 C<i> at the end of enclosing I<pseudo-block>.
1360 =item C<SAVESPTR(s)>
1362 =item C<SAVEPPTR(p)>
1364 These macros arrange things to restore the value of pointers C<s> and
1365 C<p>. C<s> must be a pointer of a type which survives conversion to
1366 C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1369 =item C<SAVEFREESV(SV *sv)>
1371 The refcount of C<sv> would be decremented at the end of
1372 I<pseudo-block>. This is similar to C<sv_2mortal> in that it is also a
1373 mechanism for doing a delayed C<SvREFCNT_dec>. However, while C<sv_2mortal>
1374 extends the lifetime of C<sv> until the beginning of the next statement,
1375 C<SAVEFREESV> extends it until the end of the enclosing scope. These
1376 lifetimes can be wildly different.
1378 Also compare C<SAVEMORTALIZESV>.
1380 =item C<SAVEMORTALIZESV(SV *sv)>
1382 Just like C<SAVEFREESV>, but mortalizes C<sv> at the end of the current
1383 scope instead of decrementing its reference count. This usually has the
1384 effect of keeping C<sv> alive until the statement that called the currently
1385 live scope has finished executing.
1387 =item C<SAVEFREEOP(OP *op)>
1389 The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1391 =item C<SAVEFREEPV(p)>
1393 The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1394 end of I<pseudo-block>.
1396 =item C<SAVECLEARSV(SV *sv)>
1398 Clears a slot in the current scratchpad which corresponds to C<sv> at
1399 the end of I<pseudo-block>.
1401 =item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1403 The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
1404 string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1405 short-lived storage, the corresponding string may be reallocated like
1408 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
1410 =item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
1412 At the end of I<pseudo-block> the function C<f> is called with the
1415 =item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1417 At the end of I<pseudo-block> the function C<f> is called with the
1418 implicit context argument (if any), and C<p>.
1420 =item C<SAVESTACK_POS()>
1422 The current offset on the Perl internal stack (cf. C<SP>) is restored
1423 at the end of I<pseudo-block>.
1427 The following API list contains functions, thus one needs to
1428 provide pointers to the modifiable data explicitly (either C pointers,
1429 or Perlish C<GV *>s). Where the above macros take C<int>, a similar
1430 function takes C<int *>.
1434 =item C<SV* save_scalar(GV *gv)>
1436 Equivalent to Perl code C<local $gv>.
1438 =item C<AV* save_ary(GV *gv)>
1440 =item C<HV* save_hash(GV *gv)>
1442 Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1444 =item C<void save_item(SV *item)>
1446 Duplicates the current value of C<SV>, on the exit from the current
1447 C<ENTER>/C<LEAVE> I<pseudo-block> will restore the value of C<SV>
1448 using the stored value. It doesn't handle magic. Use C<save_scalar> if
1451 =item C<void save_list(SV **sarg, I32 maxsarg)>
1453 A variant of C<save_item> which takes multiple arguments via an array
1454 C<sarg> of C<SV*> of length C<maxsarg>.
1456 =item C<SV* save_svref(SV **sptr)>
1458 Similar to C<save_scalar>, but will reinstate an C<SV *>.
1460 =item C<void save_aptr(AV **aptr)>
1462 =item C<void save_hptr(HV **hptr)>
1464 Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1468 The C<Alias> module implements localization of the basic types within the
1469 I<caller's scope>. People who are interested in how to localize things in
1470 the containing scope should take a look there too.
1474 =head2 XSUBs and the Argument Stack
1476 The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1477 An XSUB routine will have a stack that contains the arguments from the Perl
1478 program, and a way to map from the Perl data structures to a C equivalent.
1480 The stack arguments are accessible through the C<ST(n)> macro, which returns
1481 the C<n>'th stack argument. Argument 0 is the first argument passed in the
1482 Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1485 Most of the time, output from the C routine can be handled through use of
1486 the RETVAL and OUTPUT directives. However, there are some cases where the
1487 argument stack is not already long enough to handle all the return values.
1488 An example is the POSIX tzname() call, which takes no arguments, but returns
1489 two, the local time zone's standard and summer time abbreviations.
1491 To handle this situation, the PPCODE directive is used and the stack is
1492 extended using the macro:
1496 where C<SP> is the macro that represents the local copy of the stack pointer,
1497 and C<num> is the number of elements the stack should be extended by.
1499 Now that there is room on the stack, values can be pushed on it using C<PUSHs>
1500 macro. The pushed values will often need to be "mortal" (See
1501 L</Reference Counts and Mortality>):
1503 PUSHs(sv_2mortal(newSViv(an_integer)))
1504 PUSHs(sv_2mortal(newSVuv(an_unsigned_integer)))
1505 PUSHs(sv_2mortal(newSVnv(a_double)))
1506 PUSHs(sv_2mortal(newSVpv("Some String",0)))
1507 /* Although the last example is better written as the more
1509 PUSHs(newSVpvs_flags("Some String", SVs_TEMP))
1511 And now the Perl program calling C<tzname>, the two values will be assigned
1514 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1516 An alternate (and possibly simpler) method to pushing values on the stack is
1521 This macro automatically adjusts the stack for you, if needed. Thus, you
1522 do not need to call C<EXTEND> to extend the stack.
1524 Despite their suggestions in earlier versions of this document the macros
1525 C<(X)PUSH[iunp]> are I<not> suited to XSUBs which return multiple results.
1526 For that, either stick to the C<(X)PUSHs> macros shown above, or use the new
1527 C<m(X)PUSH[iunp]> macros instead; see L</Putting a C value on Perl stack>.
1529 For more information, consult L<perlxs> and L<perlxstut>.
1531 =head2 Autoloading with XSUBs
1533 If an AUTOLOAD routine is an XSUB, as with Perl subroutines, Perl puts the
1534 fully-qualified name of the autoloaded subroutine in the $AUTOLOAD variable
1535 of the XSUB's package.
1537 But it also puts the same information in certain fields of the XSUB itself:
1539 HV *stash = CvSTASH(cv);
1540 const char *subname = SvPVX(cv);
1541 STRLEN name_length = SvCUR(cv); /* in bytes */
1542 U32 is_utf8 = SvUTF8(cv);
1544 C<SvPVX(cv)> contains just the sub name itself, not including the package.
1545 For an AUTOLOAD routine in UNIVERSAL or one of its superclasses,
1546 C<CvSTASH(cv)> returns NULL during a method call on a nonexistent package.
1548 B<Note>: Setting $AUTOLOAD stopped working in 5.6.1, which did not support
1549 XS AUTOLOAD subs at all. Perl 5.8.0 introduced the use of fields in the
1550 XSUB itself. Perl 5.16.0 restored the setting of $AUTOLOAD. If you need
1551 to support 5.8-5.14, use the XSUB's fields.
1553 =head2 Calling Perl Routines from within C Programs
1555 There are four routines that can be used to call a Perl subroutine from
1556 within a C program. These four are:
1558 I32 call_sv(SV*, I32);
1559 I32 call_pv(const char*, I32);
1560 I32 call_method(const char*, I32);
1561 I32 call_argv(const char*, I32, register char**);
1563 The routine most often used is C<call_sv>. The C<SV*> argument
1564 contains either the name of the Perl subroutine to be called, or a
1565 reference to the subroutine. The second argument consists of flags
1566 that control the context in which the subroutine is called, whether
1567 or not the subroutine is being passed arguments, how errors should be
1568 trapped, and how to treat return values.
1570 All four routines return the number of arguments that the subroutine returned
1573 These routines used to be called C<perl_call_sv>, etc., before Perl v5.6.0,
1574 but those names are now deprecated; macros of the same name are provided for
1577 When using any of these routines (except C<call_argv>), the programmer
1578 must manipulate the Perl stack. These include the following macros and
1593 For a detailed description of calling conventions from C to Perl,
1594 consult L<perlcall>.
1596 =head2 Memory Allocation
1600 All memory meant to be used with the Perl API functions should be manipulated
1601 using the macros described in this section. The macros provide the necessary
1602 transparency between differences in the actual malloc implementation that is
1605 It is suggested that you enable the version of malloc that is distributed
1606 with Perl. It keeps pools of various sizes of unallocated memory in
1607 order to satisfy allocation requests more quickly. However, on some
1608 platforms, it may cause spurious malloc or free errors.
1610 The following three macros are used to initially allocate memory :
1612 Newx(pointer, number, type);
1613 Newxc(pointer, number, type, cast);
1614 Newxz(pointer, number, type);
1616 The first argument C<pointer> should be the name of a variable that will
1617 point to the newly allocated memory.
1619 The second and third arguments C<number> and C<type> specify how many of
1620 the specified type of data structure should be allocated. The argument
1621 C<type> is passed to C<sizeof>. The final argument to C<Newxc>, C<cast>,
1622 should be used if the C<pointer> argument is different from the C<type>
1625 Unlike the C<Newx> and C<Newxc> macros, the C<Newxz> macro calls C<memzero>
1626 to zero out all the newly allocated memory.
1630 Renew(pointer, number, type);
1631 Renewc(pointer, number, type, cast);
1634 These three macros are used to change a memory buffer size or to free a
1635 piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1636 match those of C<New> and C<Newc> with the exception of not needing the
1637 "magic cookie" argument.
1641 Move(source, dest, number, type);
1642 Copy(source, dest, number, type);
1643 Zero(dest, number, type);
1645 These three macros are used to move, copy, or zero out previously allocated
1646 memory. The C<source> and C<dest> arguments point to the source and
1647 destination starting points. Perl will move, copy, or zero out C<number>
1648 instances of the size of the C<type> data structure (using the C<sizeof>
1653 The most recent development releases of Perl have been experimenting with
1654 removing Perl's dependency on the "normal" standard I/O suite and allowing
1655 other stdio implementations to be used. This involves creating a new
1656 abstraction layer that then calls whichever implementation of stdio Perl
1657 was compiled with. All XSUBs should now use the functions in the PerlIO
1658 abstraction layer and not make any assumptions about what kind of stdio
1661 For a complete description of the PerlIO abstraction, consult L<perlapio>.
1663 =head2 Putting a C value on Perl stack
1665 A lot of opcodes (this is an elementary operation in the internal perl
1666 stack machine) put an SV* on the stack. However, as an optimization
1667 the corresponding SV is (usually) not recreated each time. The opcodes
1668 reuse specially assigned SVs (I<target>s) which are (as a corollary)
1669 not constantly freed/created.
1671 Each of the targets is created only once (but see
1672 L<Scratchpads and recursion> below), and when an opcode needs to put
1673 an integer, a double, or a string on stack, it just sets the
1674 corresponding parts of its I<target> and puts the I<target> on stack.
1676 The macro to put this target on stack is C<PUSHTARG>, and it is
1677 directly used in some opcodes, as well as indirectly in zillions of
1678 others, which use it via C<(X)PUSH[iunp]>.
1680 Because the target is reused, you must be careful when pushing multiple
1681 values on the stack. The following code will not do what you think:
1686 This translates as "set C<TARG> to 10, push a pointer to C<TARG> onto
1687 the stack; set C<TARG> to 20, push a pointer to C<TARG> onto the stack".
1688 At the end of the operation, the stack does not contain the values 10
1689 and 20, but actually contains two pointers to C<TARG>, which we have set
1692 If you need to push multiple different values then you should either use
1693 the C<(X)PUSHs> macros, or else use the new C<m(X)PUSH[iunp]> macros,
1694 none of which make use of C<TARG>. The C<(X)PUSHs> macros simply push an
1695 SV* on the stack, which, as noted under L</XSUBs and the Argument Stack>,
1696 will often need to be "mortal". The new C<m(X)PUSH[iunp]> macros make
1697 this a little easier to achieve by creating a new mortal for you (via
1698 C<(X)PUSHmortal>), pushing that onto the stack (extending it if necessary
1699 in the case of the C<mXPUSH[iunp]> macros), and then setting its value.
1700 Thus, instead of writing this to "fix" the example above:
1702 XPUSHs(sv_2mortal(newSViv(10)))
1703 XPUSHs(sv_2mortal(newSViv(20)))
1705 you can simply write:
1710 On a related note, if you do use C<(X)PUSH[iunp]>, then you're going to
1711 need a C<dTARG> in your variable declarations so that the C<*PUSH*>
1712 macros can make use of the local variable C<TARG>. See also C<dTARGET>
1717 The question remains on when the SVs which are I<target>s for opcodes
1718 are created. The answer is that they are created when the current
1719 unit--a subroutine or a file (for opcodes for statements outside of
1720 subroutines)--is compiled. During this time a special anonymous Perl
1721 array is created, which is called a scratchpad for the current unit.
1723 A scratchpad keeps SVs which are lexicals for the current unit and are
1724 targets for opcodes. One can deduce that an SV lives on a scratchpad
1725 by looking on its flags: lexicals have C<SVs_PADMY> set, and
1726 I<target>s have C<SVs_PADTMP> set.
1728 The correspondence between OPs and I<target>s is not 1-to-1. Different
1729 OPs in the compile tree of the unit can use the same target, if this
1730 would not conflict with the expected life of the temporary.
1732 =head2 Scratchpads and recursion
1734 In fact it is not 100% true that a compiled unit contains a pointer to
1735 the scratchpad AV. In fact it contains a pointer to an AV of
1736 (initially) one element, and this element is the scratchpad AV. Why do
1737 we need an extra level of indirection?
1739 The answer is B<recursion>, and maybe B<threads>. Both
1740 these can create several execution pointers going into the same
1741 subroutine. For the subroutine-child not write over the temporaries
1742 for the subroutine-parent (lifespan of which covers the call to the
1743 child), the parent and the child should have different
1744 scratchpads. (I<And> the lexicals should be separate anyway!)
1746 So each subroutine is born with an array of scratchpads (of length 1).
1747 On each entry to the subroutine it is checked that the current
1748 depth of the recursion is not more than the length of this array, and
1749 if it is, new scratchpad is created and pushed into the array.
1751 The I<target>s on this scratchpad are C<undef>s, but they are already
1752 marked with correct flags.
1754 =head1 Compiled code
1758 Here we describe the internal form your code is converted to by
1759 Perl. Start with a simple example:
1763 This is converted to a tree similar to this one:
1771 (but slightly more complicated). This tree reflects the way Perl
1772 parsed your code, but has nothing to do with the execution order.
1773 There is an additional "thread" going through the nodes of the tree
1774 which shows the order of execution of the nodes. In our simplified
1775 example above it looks like:
1777 $b ---> $c ---> + ---> $a ---> assign-to
1779 But with the actual compile tree for C<$a = $b + $c> it is different:
1780 some nodes I<optimized away>. As a corollary, though the actual tree
1781 contains more nodes than our simplified example, the execution order
1782 is the same as in our example.
1784 =head2 Examining the tree
1786 If you have your perl compiled for debugging (usually done with
1787 C<-DDEBUGGING> on the C<Configure> command line), you may examine the
1788 compiled tree by specifying C<-Dx> on the Perl command line. The
1789 output takes several lines per node, and for C<$b+$c> it looks like
1794 FLAGS = (SCALAR,KIDS)
1796 TYPE = null ===> (4)
1798 FLAGS = (SCALAR,KIDS)
1800 3 TYPE = gvsv ===> 4
1806 TYPE = null ===> (5)
1808 FLAGS = (SCALAR,KIDS)
1810 4 TYPE = gvsv ===> 5
1816 This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
1817 not optimized away (one per number in the left column). The immediate
1818 children of the given node correspond to C<{}> pairs on the same level
1819 of indentation, thus this listing corresponds to the tree:
1827 The execution order is indicated by C<===E<gt>> marks, thus it is C<3
1828 4 5 6> (node C<6> is not included into above listing), i.e.,
1829 C<gvsv gvsv add whatever>.
1831 Each of these nodes represents an op, a fundamental operation inside the
1832 Perl core. The code which implements each operation can be found in the
1833 F<pp*.c> files; the function which implements the op with type C<gvsv>
1834 is C<pp_gvsv>, and so on. As the tree above shows, different ops have
1835 different numbers of children: C<add> is a binary operator, as one would
1836 expect, and so has two children. To accommodate the various different
1837 numbers of children, there are various types of op data structure, and
1838 they link together in different ways.
1840 The simplest type of op structure is C<OP>: this has no children. Unary
1841 operators, C<UNOP>s, have one child, and this is pointed to by the
1842 C<op_first> field. Binary operators (C<BINOP>s) have not only an
1843 C<op_first> field but also an C<op_last> field. The most complex type of
1844 op is a C<LISTOP>, which has any number of children. In this case, the
1845 first child is pointed to by C<op_first> and the last child by
1846 C<op_last>. The children in between can be found by iteratively
1847 following the C<op_sibling> pointer from the first child to the last.
1849 There are also two other op types: a C<PMOP> holds a regular expression,
1850 and has no children, and a C<LOOP> may or may not have children. If the
1851 C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
1852 complicate matters, if a C<UNOP> is actually a C<null> op after
1853 optimization (see L</Compile pass 2: context propagation>) it will still
1854 have children in accordance with its former type.
1856 Another way to examine the tree is to use a compiler back-end module, such
1859 =head2 Compile pass 1: check routines
1861 The tree is created by the compiler while I<yacc> code feeds it
1862 the constructions it recognizes. Since I<yacc> works bottom-up, so does
1863 the first pass of perl compilation.
1865 What makes this pass interesting for perl developers is that some
1866 optimization may be performed on this pass. This is optimization by
1867 so-called "check routines". The correspondence between node names
1868 and corresponding check routines is described in F<opcode.pl> (do not
1869 forget to run C<make regen_headers> if you modify this file).
1871 A check routine is called when the node is fully constructed except
1872 for the execution-order thread. Since at this time there are no
1873 back-links to the currently constructed node, one can do most any
1874 operation to the top-level node, including freeing it and/or creating
1875 new nodes above/below it.
1877 The check routine returns the node which should be inserted into the
1878 tree (if the top-level node was not modified, check routine returns
1881 By convention, check routines have names C<ck_*>. They are usually
1882 called from C<new*OP> subroutines (or C<convert>) (which in turn are
1883 called from F<perly.y>).
1885 =head2 Compile pass 1a: constant folding
1887 Immediately after the check routine is called the returned node is
1888 checked for being compile-time executable. If it is (the value is
1889 judged to be constant) it is immediately executed, and a I<constant>
1890 node with the "return value" of the corresponding subtree is
1891 substituted instead. The subtree is deleted.
1893 If constant folding was not performed, the execution-order thread is
1896 =head2 Compile pass 2: context propagation
1898 When a context for a part of compile tree is known, it is propagated
1899 down through the tree. At this time the context can have 5 values
1900 (instead of 2 for runtime context): void, boolean, scalar, list, and
1901 lvalue. In contrast with the pass 1 this pass is processed from top
1902 to bottom: a node's context determines the context for its children.
1904 Additional context-dependent optimizations are performed at this time.
1905 Since at this moment the compile tree contains back-references (via
1906 "thread" pointers), nodes cannot be free()d now. To allow
1907 optimized-away nodes at this stage, such nodes are null()ified instead
1908 of free()ing (i.e. their type is changed to OP_NULL).
1910 =head2 Compile pass 3: peephole optimization
1912 After the compile tree for a subroutine (or for an C<eval> or a file)
1913 is created, an additional pass over the code is performed. This pass
1914 is neither top-down or bottom-up, but in the execution order (with
1915 additional complications for conditionals). Optimizations performed
1916 at this stage are subject to the same restrictions as in the pass 2.
1918 Peephole optimizations are done by calling the function pointed to
1919 by the global variable C<PL_peepp>. By default, C<PL_peepp> just
1920 calls the function pointed to by the global variable C<PL_rpeepp>.
1921 By default, that performs some basic op fixups and optimisations along
1922 the execution-order op chain, and recursively calls C<PL_rpeepp> for
1923 each side chain of ops (resulting from conditionals). Extensions may
1924 provide additional optimisations or fixups, hooking into either the
1925 per-subroutine or recursive stage, like this:
1927 static peep_t prev_peepp;
1928 static void my_peep(pTHX_ OP *o)
1930 /* custom per-subroutine optimisation goes here */
1932 /* custom per-subroutine optimisation may also go here */
1935 prev_peepp = PL_peepp;
1938 static peep_t prev_rpeepp;
1939 static void my_rpeep(pTHX_ OP *o)
1942 for(; o; o = o->op_next) {
1943 /* custom per-op optimisation goes here */
1945 prev_rpeepp(orig_o);
1948 prev_rpeepp = PL_rpeepp;
1949 PL_rpeepp = my_rpeep;
1951 =head2 Pluggable runops
1953 The compile tree is executed in a runops function. There are two runops
1954 functions, in F<run.c> and in F<dump.c>. C<Perl_runops_debug> is used
1955 with DEBUGGING and C<Perl_runops_standard> is used otherwise. For fine
1956 control over the execution of the compile tree it is possible to provide
1957 your own runops function.
1959 It's probably best to copy one of the existing runops functions and
1960 change it to suit your needs. Then, in the BOOT section of your XS
1963 PL_runops = my_runops;
1965 This function should be as efficient as possible to keep your programs
1966 running as fast as possible.
1968 =head2 Compile-time scope hooks
1970 As of perl 5.14 it is possible to hook into the compile-time lexical
1971 scope mechanism using C<Perl_blockhook_register>. This is used like
1974 STATIC void my_start_hook(pTHX_ int full);
1975 STATIC BHK my_hooks;
1978 BhkENTRY_set(&my_hooks, bhk_start, my_start_hook);
1979 Perl_blockhook_register(aTHX_ &my_hooks);
1981 This will arrange to have C<my_start_hook> called at the start of
1982 compiling every lexical scope. The available hooks are:
1986 =item C<void bhk_start(pTHX_ int full)>
1988 This is called just after starting a new lexical scope. Note that Perl
1993 creates two scopes: the first starts at the C<(> and has C<full == 1>,
1994 the second starts at the C<{> and has C<full == 0>. Both end at the
1995 C<}>, so calls to C<start> and C<pre/post_end> will match. Anything
1996 pushed onto the save stack by this hook will be popped just before the
1997 scope ends (between the C<pre_> and C<post_end> hooks, in fact).
1999 =item C<void bhk_pre_end(pTHX_ OP **o)>
2001 This is called at the end of a lexical scope, just before unwinding the
2002 stack. I<o> is the root of the optree representing the scope; it is a
2003 double pointer so you can replace the OP if you need to.
2005 =item C<void bhk_post_end(pTHX_ OP **o)>
2007 This is called at the end of a lexical scope, just after unwinding the
2008 stack. I<o> is as above. Note that it is possible for calls to C<pre_>
2009 and C<post_end> to nest, if there is something on the save stack that
2012 =item C<void bhk_eval(pTHX_ OP *const o)>
2014 This is called just before starting to compile an C<eval STRING>, C<do
2015 FILE>, C<require> or C<use>, after the eval has been set up. I<o> is the
2016 OP that requested the eval, and will normally be an C<OP_ENTEREVAL>,
2017 C<OP_DOFILE> or C<OP_REQUIRE>.
2021 Once you have your hook functions, you need a C<BHK> structure to put
2022 them in. It's best to allocate it statically, since there is no way to
2023 free it once it's registered. The function pointers should be inserted
2024 into this structure using the C<BhkENTRY_set> macro, which will also set
2025 flags indicating which entries are valid. If you do need to allocate
2026 your C<BHK> dynamically for some reason, be sure to zero it before you
2029 Once registered, there is no mechanism to switch these hooks off, so if
2030 that is necessary you will need to do this yourself. An entry in C<%^H>
2031 is probably the best way, so the effect is lexically scoped; however it
2032 is also possible to use the C<BhkDISABLE> and C<BhkENABLE> macros to
2033 temporarily switch entries on and off. You should also be aware that
2034 generally speaking at least one scope will have opened before your
2035 extension is loaded, so you will see some C<pre/post_end> pairs that
2036 didn't have a matching C<start>.
2038 =head1 Examining internal data structures with the C<dump> functions
2040 To aid debugging, the source file F<dump.c> contains a number of
2041 functions which produce formatted output of internal data structures.
2043 The most commonly used of these functions is C<Perl_sv_dump>; it's used
2044 for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
2045 C<sv_dump> to produce debugging output from Perl-space, so users of that
2046 module should already be familiar with its format.
2048 C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
2049 derivatives, and produces output similar to C<perl -Dx>; in fact,
2050 C<Perl_dump_eval> will dump the main root of the code being evaluated,
2051 exactly like C<-Dx>.
2053 Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
2054 op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
2055 subroutines in a package like so: (Thankfully, these are all xsubs, so
2056 there is no op tree)
2058 (gdb) print Perl_dump_packsubs(PL_defstash)
2060 SUB attributes::bootstrap = (xsub 0x811fedc 0)
2062 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
2064 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
2066 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
2068 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
2070 and C<Perl_dump_all>, which dumps all the subroutines in the stash and
2071 the op tree of the main root.
2073 =head1 How multiple interpreters and concurrency are supported
2075 =head2 Background and PERL_IMPLICIT_CONTEXT
2077 The Perl interpreter can be regarded as a closed box: it has an API
2078 for feeding it code or otherwise making it do things, but it also has
2079 functions for its own use. This smells a lot like an object, and
2080 there are ways for you to build Perl so that you can have multiple
2081 interpreters, with one interpreter represented either as a C structure,
2082 or inside a thread-specific structure. These structures contain all
2083 the context, the state of that interpreter.
2085 One macro controls the major Perl build flavor: MULTIPLICITY. The
2086 MULTIPLICITY build has a C structure that packages all the interpreter
2087 state. With multiplicity-enabled perls, PERL_IMPLICIT_CONTEXT is also
2088 normally defined, and enables the support for passing in a "hidden" first
2089 argument that represents all three data structures. MULTIPLICITY makes
2090 multi-threaded perls possible (with the ithreads threading model, related
2091 to the macro USE_ITHREADS.)
2093 Two other "encapsulation" macros are the PERL_GLOBAL_STRUCT and
2094 PERL_GLOBAL_STRUCT_PRIVATE (the latter turns on the former, and the
2095 former turns on MULTIPLICITY.) The PERL_GLOBAL_STRUCT causes all the
2096 internal variables of Perl to be wrapped inside a single global struct,
2097 struct perl_vars, accessible as (globals) &PL_Vars or PL_VarsPtr or
2098 the function Perl_GetVars(). The PERL_GLOBAL_STRUCT_PRIVATE goes
2099 one step further, there is still a single struct (allocated in main()
2100 either from heap or from stack) but there are no global data symbols
2101 pointing to it. In either case the global struct should be initialised
2102 as the very first thing in main() using Perl_init_global_struct() and
2103 correspondingly tear it down after perl_free() using Perl_free_global_struct(),
2104 please see F<miniperlmain.c> for usage details. You may also need
2105 to use C<dVAR> in your coding to "declare the global variables"
2106 when you are using them. dTHX does this for you automatically.
2108 To see whether you have non-const data you can use a BSD-compatible C<nm>:
2110 nm libperl.a | grep -v ' [TURtr] '
2112 If this displays any C<D> or C<d> symbols, you have non-const data.
2114 For backward compatibility reasons defining just PERL_GLOBAL_STRUCT
2115 doesn't actually hide all symbols inside a big global struct: some
2116 PerlIO_xxx vtables are left visible. The PERL_GLOBAL_STRUCT_PRIVATE
2117 then hides everything (see how the PERLIO_FUNCS_DECL is used).
2119 All this obviously requires a way for the Perl internal functions to be
2120 either subroutines taking some kind of structure as the first
2121 argument, or subroutines taking nothing as the first argument. To
2122 enable these two very different ways of building the interpreter,
2123 the Perl source (as it does in so many other situations) makes heavy
2124 use of macros and subroutine naming conventions.
2126 First problem: deciding which functions will be public API functions and
2127 which will be private. All functions whose names begin C<S_> are private
2128 (think "S" for "secret" or "static"). All other functions begin with
2129 "Perl_", but just because a function begins with "Perl_" does not mean it is
2130 part of the API. (See L</Internal Functions>.) The easiest way to be B<sure> a
2131 function is part of the API is to find its entry in L<perlapi>.
2132 If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
2133 think it should be (i.e., you need it for your extension), send mail via
2134 L<perlbug> explaining why you think it should be.
2136 Second problem: there must be a syntax so that the same subroutine
2137 declarations and calls can pass a structure as their first argument,
2138 or pass nothing. To solve this, the subroutines are named and
2139 declared in a particular way. Here's a typical start of a static
2140 function used within the Perl guts:
2143 S_incline(pTHX_ char *s)
2145 STATIC becomes "static" in C, and may be #define'd to nothing in some
2146 configurations in the future.
2148 A public function (i.e. part of the internal API, but not necessarily
2149 sanctioned for use in extensions) begins like this:
2152 Perl_sv_setiv(pTHX_ SV* dsv, IV num)
2154 C<pTHX_> is one of a number of macros (in F<perl.h>) that hide the
2155 details of the interpreter's context. THX stands for "thread", "this",
2156 or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
2157 The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
2158 or 'd' for B<d>eclaration, so we have C<pTHX>, C<aTHX> and C<dTHX>, and
2161 When Perl is built without options that set PERL_IMPLICIT_CONTEXT, there is no
2162 first argument containing the interpreter's context. The trailing underscore
2163 in the pTHX_ macro indicates that the macro expansion needs a comma
2164 after the context argument because other arguments follow it. If
2165 PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
2166 subroutine is not prototyped to take the extra argument. The form of the
2167 macro without the trailing underscore is used when there are no additional
2170 When a core function calls another, it must pass the context. This
2171 is normally hidden via macros. Consider C<sv_setiv>. It expands into
2172 something like this:
2174 #ifdef PERL_IMPLICIT_CONTEXT
2175 #define sv_setiv(a,b) Perl_sv_setiv(aTHX_ a, b)
2176 /* can't do this for vararg functions, see below */
2178 #define sv_setiv Perl_sv_setiv
2181 This works well, and means that XS authors can gleefully write:
2185 and still have it work under all the modes Perl could have been
2188 This doesn't work so cleanly for varargs functions, though, as macros
2189 imply that the number of arguments is known in advance. Instead we
2190 either need to spell them out fully, passing C<aTHX_> as the first
2191 argument (the Perl core tends to do this with functions like
2192 Perl_warner), or use a context-free version.
2194 The context-free version of Perl_warner is called
2195 Perl_warner_nocontext, and does not take the extra argument. Instead
2196 it does dTHX; to get the context from thread-local storage. We
2197 C<#define warner Perl_warner_nocontext> so that extensions get source
2198 compatibility at the expense of performance. (Passing an arg is
2199 cheaper than grabbing it from thread-local storage.)
2201 You can ignore [pad]THXx when browsing the Perl headers/sources.
2202 Those are strictly for use within the core. Extensions and embedders
2203 need only be aware of [pad]THX.
2205 =head2 So what happened to dTHR?
2207 C<dTHR> was introduced in perl 5.005 to support the older thread model.
2208 The older thread model now uses the C<THX> mechanism to pass context
2209 pointers around, so C<dTHR> is not useful any more. Perl 5.6.0 and
2210 later still have it for backward source compatibility, but it is defined
2213 =head2 How do I use all this in extensions?
2215 When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
2216 any functions in the Perl API will need to pass the initial context
2217 argument somehow. The kicker is that you will need to write it in
2218 such a way that the extension still compiles when Perl hasn't been
2219 built with PERL_IMPLICIT_CONTEXT enabled.
2221 There are three ways to do this. First, the easy but inefficient way,
2222 which is also the default, in order to maintain source compatibility
2223 with extensions: whenever F<XSUB.h> is #included, it redefines the aTHX
2224 and aTHX_ macros to call a function that will return the context.
2225 Thus, something like:
2229 in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
2232 Perl_sv_setiv(Perl_get_context(), sv, num);
2234 or to this otherwise:
2236 Perl_sv_setiv(sv, num);
2238 You don't have to do anything new in your extension to get this; since
2239 the Perl library provides Perl_get_context(), it will all just
2242 The second, more efficient way is to use the following template for
2245 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2250 STATIC void my_private_function(int arg1, int arg2);
2253 my_private_function(int arg1, int arg2)
2255 dTHX; /* fetch context */
2256 ... call many Perl API functions ...
2261 MODULE = Foo PACKAGE = Foo
2269 my_private_function(arg, 10);
2271 Note that the only two changes from the normal way of writing an
2272 extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
2273 including the Perl headers, followed by a C<dTHX;> declaration at
2274 the start of every function that will call the Perl API. (You'll
2275 know which functions need this, because the C compiler will complain
2276 that there's an undeclared identifier in those functions.) No changes
2277 are needed for the XSUBs themselves, because the XS() macro is
2278 correctly defined to pass in the implicit context if needed.
2280 The third, even more efficient way is to ape how it is done within
2284 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2289 /* pTHX_ only needed for functions that call Perl API */
2290 STATIC void my_private_function(pTHX_ int arg1, int arg2);
2293 my_private_function(pTHX_ int arg1, int arg2)
2295 /* dTHX; not needed here, because THX is an argument */
2296 ... call Perl API functions ...
2301 MODULE = Foo PACKAGE = Foo
2309 my_private_function(aTHX_ arg, 10);
2311 This implementation never has to fetch the context using a function
2312 call, since it is always passed as an extra argument. Depending on
2313 your needs for simplicity or efficiency, you may mix the previous
2314 two approaches freely.
2316 Never add a comma after C<pTHX> yourself--always use the form of the
2317 macro with the underscore for functions that take explicit arguments,
2318 or the form without the argument for functions with no explicit arguments.
2320 If one is compiling Perl with the C<-DPERL_GLOBAL_STRUCT> the C<dVAR>
2321 definition is needed if the Perl global variables (see F<perlvars.h>
2322 or F<globvar.sym>) are accessed in the function and C<dTHX> is not
2323 used (the C<dTHX> includes the C<dVAR> if necessary). One notices
2324 the need for C<dVAR> only with the said compile-time define, because
2325 otherwise the Perl global variables are visible as-is.
2327 =head2 Should I do anything special if I call perl from multiple threads?
2329 If you create interpreters in one thread and then proceed to call them in
2330 another, you need to make sure perl's own Thread Local Storage (TLS) slot is
2331 initialized correctly in each of those threads.
2333 The C<perl_alloc> and C<perl_clone> API functions will automatically set
2334 the TLS slot to the interpreter they created, so that there is no need to do
2335 anything special if the interpreter is always accessed in the same thread that
2336 created it, and that thread did not create or call any other interpreters
2337 afterwards. If that is not the case, you have to set the TLS slot of the
2338 thread before calling any functions in the Perl API on that particular
2339 interpreter. This is done by calling the C<PERL_SET_CONTEXT> macro in that
2340 thread as the first thing you do:
2342 /* do this before doing anything else with some_perl */
2343 PERL_SET_CONTEXT(some_perl);
2345 ... other Perl API calls on some_perl go here ...
2347 =head2 Future Plans and PERL_IMPLICIT_SYS
2349 Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
2350 that the interpreter knows about itself and pass it around, so too are
2351 there plans to allow the interpreter to bundle up everything it knows
2352 about the environment it's running on. This is enabled with the
2353 PERL_IMPLICIT_SYS macro. Currently it only works with USE_ITHREADS on
2356 This allows the ability to provide an extra pointer (called the "host"
2357 environment) for all the system calls. This makes it possible for
2358 all the system stuff to maintain their own state, broken down into
2359 seven C structures. These are thin wrappers around the usual system
2360 calls (see F<win32/perllib.c>) for the default perl executable, but for a
2361 more ambitious host (like the one that would do fork() emulation) all
2362 the extra work needed to pretend that different interpreters are
2363 actually different "processes", would be done here.
2365 The Perl engine/interpreter and the host are orthogonal entities.
2366 There could be one or more interpreters in a process, and one or
2367 more "hosts", with free association between them.
2369 =head1 Internal Functions
2371 All of Perl's internal functions which will be exposed to the outside
2372 world are prefixed by C<Perl_> so that they will not conflict with XS
2373 functions or functions used in a program in which Perl is embedded.
2374 Similarly, all global variables begin with C<PL_>. (By convention,
2375 static functions start with C<S_>.)
2377 Inside the Perl core (C<PERL_CORE> defined), you can get at the functions
2378 either with or without the C<Perl_> prefix, thanks to a bunch of defines
2379 that live in F<embed.h>. Note that extension code should I<not> set
2380 C<PERL_CORE>; this exposes the full perl internals, and is likely to cause
2381 breakage of the XS in each new perl release.
2383 The file F<embed.h> is generated automatically from
2384 F<embed.pl> and F<embed.fnc>. F<embed.pl> also creates the prototyping
2385 header files for the internal functions, generates the documentation
2386 and a lot of other bits and pieces. It's important that when you add
2387 a new function to the core or change an existing one, you change the
2388 data in the table in F<embed.fnc> as well. Here's a sample entry from
2391 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
2393 The second column is the return type, the third column the name. Columns
2394 after that are the arguments. The first column is a set of flags:
2400 This function is a part of the public API. All such functions should also
2401 have 'd', very few do not.
2405 This function has a C<Perl_> prefix; i.e. it is defined as
2410 This function has documentation using the C<apidoc> feature which we'll
2411 look at in a second. Some functions have 'd' but not 'A'; docs are good.
2415 Other available flags are:
2421 This is a static function and is defined as C<STATIC S_whatever>, and
2422 usually called within the sources as C<whatever(...)>.
2426 This does not need an interpreter context, so the definition has no
2427 C<pTHX>, and it follows that callers don't use C<aTHX>. (See
2428 L</Background and PERL_IMPLICIT_CONTEXT>.)
2432 This function never returns; C<croak>, C<exit> and friends.
2436 This function takes a variable number of arguments, C<printf> style.
2437 The argument list should end with C<...>, like this:
2439 Afprd |void |croak |const char* pat|...
2443 This function is part of the experimental development API, and may change
2444 or disappear without notice.
2448 This function should not have a compatibility macro to define, say,
2449 C<Perl_parse> to C<parse>. It must be called as C<Perl_parse>.
2453 This function isn't exported out of the Perl core.
2457 This is implemented as a macro.
2461 This function is explicitly exported.
2465 This function is visible to extensions included in the Perl core.
2469 Binary backward compatibility; this function is a macro but also has
2470 a C<Perl_> implementation (which is exported).
2474 See the comments at the top of C<embed.fnc> for others.
2478 If you edit F<embed.pl> or F<embed.fnc>, you will need to run
2479 C<make regen_headers> to force a rebuild of F<embed.h> and other
2480 auto-generated files.
2482 =head2 Formatted Printing of IVs, UVs, and NVs
2484 If you are printing IVs, UVs, or NVS instead of the stdio(3) style
2485 formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
2486 following macros for portability
2491 UVxf UV in hexadecimal
2496 These will take care of 64-bit integers and long doubles.
2499 printf("IV is %"IVdf"\n", iv);
2501 The IVdf will expand to whatever is the correct format for the IVs.
2503 If you are printing addresses of pointers, use UVxf combined
2504 with PTR2UV(), do not use %lx or %p.
2506 =head2 Pointer-To-Integer and Integer-To-Pointer
2508 Because pointer size does not necessarily equal integer size,
2509 use the follow macros to do it right.
2514 INT2PTR(pointertotype, integer)
2519 SV *sv = INT2PTR(SV*, iv);
2526 =head2 Exception Handling
2528 There are a couple of macros to do very basic exception handling in XS
2529 modules. You have to define C<NO_XSLOCKS> before including F<XSUB.h> to
2530 be able to use these macros:
2535 You can use these macros if you call code that may croak, but you need
2536 to do some cleanup before giving control back to Perl. For example:
2538 dXCPT; /* set up necessary variables */
2541 code_that_may_croak();
2546 /* do cleanup here */
2550 Note that you always have to rethrow an exception that has been
2551 caught. Using these macros, it is not possible to just catch the
2552 exception and ignore it. If you have to ignore the exception, you
2553 have to use the C<call_*> function.
2555 The advantage of using the above macros is that you don't have
2556 to setup an extra function for C<call_*>, and that using these
2557 macros is faster than using C<call_*>.
2559 =head2 Source Documentation
2561 There's an effort going on to document the internal functions and
2562 automatically produce reference manuals from them - L<perlapi> is one
2563 such manual which details all the functions which are available to XS
2564 writers. L<perlintern> is the autogenerated manual for the functions
2565 which are not part of the API and are supposedly for internal use only.
2567 Source documentation is created by putting POD comments into the C
2571 =for apidoc sv_setiv
2573 Copies an integer into the given SV. Does not handle 'set' magic. See
2579 Please try and supply some documentation if you add functions to the
2582 =head2 Backwards compatibility
2584 The Perl API changes over time. New functions are added or the interfaces
2585 of existing functions are changed. The C<Devel::PPPort> module tries to
2586 provide compatibility code for some of these changes, so XS writers don't
2587 have to code it themselves when supporting multiple versions of Perl.
2589 C<Devel::PPPort> generates a C header file F<ppport.h> that can also
2590 be run as a Perl script. To generate F<ppport.h>, run:
2592 perl -MDevel::PPPort -eDevel::PPPort::WriteFile
2594 Besides checking existing XS code, the script can also be used to retrieve
2595 compatibility information for various API calls using the C<--api-info>
2596 command line switch. For example:
2598 % perl ppport.h --api-info=sv_magicext
2600 For details, see C<perldoc ppport.h>.
2602 =head1 Unicode Support
2604 Perl 5.6.0 introduced Unicode support. It's important for porters and XS
2605 writers to understand this support and make sure that the code they
2606 write does not corrupt Unicode data.
2608 =head2 What B<is> Unicode, anyway?
2610 In the olden, less enlightened times, we all used to use ASCII. Most of
2611 us did, anyway. The big problem with ASCII is that it's American. Well,
2612 no, that's not actually the problem; the problem is that it's not
2613 particularly useful for people who don't use the Roman alphabet. What
2614 used to happen was that particular languages would stick their own
2615 alphabet in the upper range of the sequence, between 128 and 255. Of
2616 course, we then ended up with plenty of variants that weren't quite
2617 ASCII, and the whole point of it being a standard was lost.
2619 Worse still, if you've got a language like Chinese or
2620 Japanese that has hundreds or thousands of characters, then you really
2621 can't fit them into a mere 256, so they had to forget about ASCII
2622 altogether, and build their own systems using pairs of numbers to refer
2625 To fix this, some people formed Unicode, Inc. and
2626 produced a new character set containing all the characters you can
2627 possibly think of and more. There are several ways of representing these
2628 characters, and the one Perl uses is called UTF-8. UTF-8 uses
2629 a variable number of bytes to represent a character. You can learn more
2630 about Unicode and Perl's Unicode model in L<perlunicode>.
2632 =head2 How can I recognise a UTF-8 string?
2634 You can't. This is because UTF-8 data is stored in bytes just like
2635 non-UTF-8 data. The Unicode character 200, (C<0xC8> for you hex types)
2636 capital E with a grave accent, is represented by the two bytes
2637 C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
2638 has that byte sequence as well. So you can't tell just by looking - this
2639 is what makes Unicode input an interesting problem.
2641 In general, you either have to know what you're dealing with, or you
2642 have to guess. The API function C<is_utf8_string> can help; it'll tell
2643 you if a string contains only valid UTF-8 characters. However, it can't
2644 do the work for you. On a character-by-character basis,
2646 will tell you whether the current character in a string is valid UTF-8.
2648 =head2 How does UTF-8 represent Unicode characters?
2650 As mentioned above, UTF-8 uses a variable number of bytes to store a
2651 character. Characters with values 0...127 are stored in one byte, just
2652 like good ol' ASCII. Character 128 is stored as C<v194.128>; this
2653 continues up to character 191, which is C<v194.191>. Now we've run out of
2654 bits (191 is binary C<10111111>) so we move on; 192 is C<v195.128>. And
2655 so it goes on, moving to three bytes at character 2048.
2657 Assuming you know you're dealing with a UTF-8 string, you can find out
2658 how long the first character in it is with the C<UTF8SKIP> macro:
2660 char *utf = "\305\233\340\240\201";
2663 len = UTF8SKIP(utf); /* len is 2 here */
2665 len = UTF8SKIP(utf); /* len is 3 here */
2667 Another way to skip over characters in a UTF-8 string is to use
2668 C<utf8_hop>, which takes a string and a number of characters to skip
2669 over. You're on your own about bounds checking, though, so don't use it
2672 All bytes in a multi-byte UTF-8 character will have the high bit set,
2673 so you can test if you need to do something special with this
2674 character like this (the UTF8_IS_INVARIANT() is a macro that tests
2675 whether the byte can be encoded as a single byte even in UTF-8):
2678 U8 *utf_end; /* 1 beyond buffer pointed to by utf */
2679 UV uv; /* Note: a UV, not a U8, not a char */
2680 STRLEN len; /* length of character in bytes */
2682 if (!UTF8_IS_INVARIANT(*utf))
2683 /* Must treat this as UTF-8 */
2684 uv = utf8_to_uvchr_buf(utf, utf_end, &len);
2686 /* OK to treat this character as a byte */
2689 You can also see in that example that we use C<utf8_to_uvchr_buf> to get the
2690 value of the character; the inverse function C<uvchr_to_utf8> is available
2691 for putting a UV into UTF-8:
2693 if (!UTF8_IS_INVARIANT(uv))
2694 /* Must treat this as UTF8 */
2695 utf8 = uvchr_to_utf8(utf8, uv);
2697 /* OK to treat this character as a byte */
2700 You B<must> convert characters to UVs using the above functions if
2701 you're ever in a situation where you have to match UTF-8 and non-UTF-8
2702 characters. You may not skip over UTF-8 characters in this case. If you
2703 do this, you'll lose the ability to match hi-bit non-UTF-8 characters;
2704 for instance, if your UTF-8 string contains C<v196.172>, and you skip
2705 that character, you can never match a C<chr(200)> in a non-UTF-8 string.
2708 =head2 How does Perl store UTF-8 strings?
2710 Currently, Perl deals with Unicode strings and non-Unicode strings
2711 slightly differently. A flag in the SV, C<SVf_UTF8>, indicates that the
2712 string is internally encoded as UTF-8. Without it, the byte value is the
2713 codepoint number and vice versa (in other words, the string is encoded
2714 as iso-8859-1, but C<use feature 'unicode_strings'> is needed to get iso-8859-1
2715 semantics). You can check and manipulate this flag with the
2722 This flag has an important effect on Perl's treatment of the string: if
2723 Unicode data is not properly distinguished, regular expressions,
2724 C<length>, C<substr> and other string handling operations will have
2725 undesirable results.
2727 The problem comes when you have, for instance, a string that isn't
2728 flagged as UTF-8, and contains a byte sequence that could be UTF-8 -
2729 especially when combining non-UTF-8 and UTF-8 strings.
2731 Never forget that the C<SVf_UTF8> flag is separate to the PV value; you
2732 need be sure you don't accidentally knock it off while you're
2733 manipulating SVs. More specifically, you cannot expect to do this:
2742 nsv = newSVpvn(p, len);
2744 The C<char*> string does not tell you the whole story, and you can't
2745 copy or reconstruct an SV just by copying the string value. Check if the
2746 old SV has the UTF8 flag set, and act accordingly:
2750 nsv = newSVpvn(p, len);
2754 In fact, your C<frobnicate> function should be made aware of whether or
2755 not it's dealing with UTF-8 data, so that it can handle the string
2758 Since just passing an SV to an XS function and copying the data of
2759 the SV is not enough to copy the UTF8 flags, even less right is just
2760 passing a C<char *> to an XS function.
2762 =head2 How do I convert a string to UTF-8?
2764 If you're mixing UTF-8 and non-UTF-8 strings, it is necessary to upgrade
2765 one of the strings to UTF-8. If you've got an SV, the easiest way to do
2768 sv_utf8_upgrade(sv);
2770 However, you must not do this, for example:
2773 sv_utf8_upgrade(left);
2775 If you do this in a binary operator, you will actually change one of the
2776 strings that came into the operator, and, while it shouldn't be noticeable
2777 by the end user, it can cause problems in deficient code.
2779 Instead, C<bytes_to_utf8> will give you a UTF-8-encoded B<copy> of its
2780 string argument. This is useful for having the data available for
2781 comparisons and so on, without harming the original SV. There's also
2782 C<utf8_to_bytes> to go the other way, but naturally, this will fail if
2783 the string contains any characters above 255 that can't be represented
2786 =head2 Is there anything else I need to know?
2788 Not really. Just remember these things:
2794 There's no way to tell if a string is UTF-8 or not. You can tell if an SV
2795 is UTF-8 by looking at its C<SvUTF8> flag. Don't forget to set the flag if
2796 something should be UTF-8. Treat the flag as part of the PV, even though
2797 it's not - if you pass on the PV to somewhere, pass on the flag too.
2801 If a string is UTF-8, B<always> use C<utf8_to_uvchr_buf> to get at the value,
2802 unless C<UTF8_IS_INVARIANT(*s)> in which case you can use C<*s>.
2806 When writing a character C<uv> to a UTF-8 string, B<always> use
2807 C<uvchr_to_utf8>, unless C<UTF8_IS_INVARIANT(uv))> in which case
2808 you can use C<*s = uv>.
2812 Mixing UTF-8 and non-UTF-8 strings is tricky. Use C<bytes_to_utf8> to get
2813 a new string which is UTF-8 encoded, and then combine them.
2817 =head1 Custom Operators
2819 Custom operator support is a new experimental feature that allows you to
2820 define your own ops. This is primarily to allow the building of
2821 interpreters for other languages in the Perl core, but it also allows
2822 optimizations through the creation of "macro-ops" (ops which perform the
2823 functions of multiple ops which are usually executed together, such as
2824 C<gvsv, gvsv, add>.)
2826 This feature is implemented as a new op type, C<OP_CUSTOM>. The Perl
2827 core does not "know" anything special about this op type, and so it will
2828 not be involved in any optimizations. This also means that you can
2829 define your custom ops to be any op structure - unary, binary, list and
2832 It's important to know what custom operators won't do for you. They
2833 won't let you add new syntax to Perl, directly. They won't even let you
2834 add new keywords, directly. In fact, they won't change the way Perl
2835 compiles a program at all. You have to do those changes yourself, after
2836 Perl has compiled the program. You do this either by manipulating the op
2837 tree using a C<CHECK> block and the C<B::Generate> module, or by adding
2838 a custom peephole optimizer with the C<optimize> module.
2840 When you do this, you replace ordinary Perl ops with custom ops by
2841 creating ops with the type C<OP_CUSTOM> and the C<pp_addr> of your own
2842 PP function. This should be defined in XS code, and should look like
2843 the PP ops in C<pp_*.c>. You are responsible for ensuring that your op
2844 takes the appropriate number of values from the stack, and you are
2845 responsible for adding stack marks if necessary.
2847 You should also "register" your op with the Perl interpreter so that it
2848 can produce sensible error and warning messages. Since it is possible to
2849 have multiple custom ops within the one "logical" op type C<OP_CUSTOM>,
2850 Perl uses the value of C<< o->op_ppaddr >> to determine which custom op
2851 it is dealing with. You should create an C<XOP> structure for each
2852 ppaddr you use, set the properties of the custom op with
2853 C<XopENTRY_set>, and register the structure against the ppaddr using
2854 C<Perl_custom_op_register>. A trivial example might look like:
2857 static OP *my_pp(pTHX);
2860 XopENTRY_set(&my_xop, xop_name, "myxop");
2861 XopENTRY_set(&my_xop, xop_desc, "Useless custom op");
2862 Perl_custom_op_register(aTHX_ my_pp, &my_xop);
2864 The available fields in the structure are:
2870 A short name for your op. This will be included in some error messages,
2871 and will also be returned as C<< $op->name >> by the L<B|B> module, so
2872 it will appear in the output of module like L<B::Concise|B::Concise>.
2876 A short description of the function of the op.
2880 Which of the various C<*OP> structures this op uses. This should be one of
2881 the C<OA_*> constants from F<op.h>, namely
2901 =item OA_PVOP_OR_SVOP
2903 This should be interpreted as 'C<PVOP>' only. The C<_OR_SVOP> is because
2904 the only core C<PVOP>, C<OP_TRANS>, can sometimes be a C<SVOP> instead.
2912 The other C<OA_*> constants should not be used.
2916 This member is of type C<Perl_cpeep_t>, which expands to C<void
2917 (*Perl_cpeep_t)(aTHX_ OP *o, OP *oldop)>. If it is set, this function
2918 will be called from C<Perl_rpeep> when ops of this type are encountered
2919 by the peephole optimizer. I<o> is the OP that needs optimizing;
2920 I<oldop> is the previous OP optimized, whose C<op_next> points to I<o>.
2924 C<B::Generate> directly supports the creation of custom ops by name.
2928 Until May 1997, this document was maintained by Jeff Okamoto
2929 E<lt>okamoto@corp.hp.comE<gt>. It is now maintained as part of Perl
2930 itself by the Perl 5 Porters E<lt>perl5-porters@perl.orgE<gt>.
2932 With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
2933 Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
2934 Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
2935 Stephen McCamant, and Gurusamy Sarathy.
2939 L<perlapi>, L<perlintern>, L<perlxs>, L<perlembed>