3 perlguts - Introduction to the Perl API
7 This document attempts to describe how to use the Perl API, as well as
8 containing 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,
34 =head2 Working with SVs
36 An SV can be created and loaded with one command. There are four types of
37 values that can be loaded: an integer value (IV), a double (NV),
38 a string (PV), and another scalar (SV).
44 SV* newSVpv(const char*, int);
45 SV* newSVpvn(const char*, int);
46 SV* newSVpvf(const char*, ...);
49 To change the value of an *already-existing* SV, there are seven routines:
51 void sv_setiv(SV*, IV);
52 void sv_setuv(SV*, UV);
53 void sv_setnv(SV*, double);
54 void sv_setpv(SV*, const char*);
55 void sv_setpvn(SV*, const char*, int)
56 void sv_setpvf(SV*, const char*, ...);
57 void sv_setpvfn(SV*, const char*, STRLEN, va_list *, SV **, I32, bool);
58 void sv_setsv(SV*, SV*);
60 Notice that you can choose to specify the length of the string to be
61 assigned by using C<sv_setpvn>, C<newSVpvn>, or C<newSVpv>, or you may
62 allow Perl to calculate the length by using C<sv_setpv> or by specifying
63 0 as the second argument to C<newSVpv>. Be warned, though, that Perl will
64 determine the string's length by using C<strlen>, which depends on the
65 string terminating with a NUL character.
67 The arguments of C<sv_setpvf> are processed like C<sprintf>, and the
68 formatted output becomes the value.
70 C<sv_setpvfn> is an analogue of C<vsprintf>, but it allows you to specify
71 either a pointer to a variable argument list or the address and length of
72 an array of SVs. The last argument points to a boolean; on return, if that
73 boolean is true, then locale-specific information has been used to format
74 the string, and the string's contents are therefore untrustworthy (see
75 L<perlsec>). This pointer may be NULL if that information is not
76 important. Note that this function requires you to specify the length of
79 STRLEN is an integer type (Size_t, usually defined as size_t in
80 config.h) guaranteed to be large enough to represent the size of
81 any string that perl can handle.
83 The C<sv_set*()> functions are not generic enough to operate on values
84 that have "magic". See L<Magic Virtual Tables> later in this document.
86 All SVs that contain strings should be terminated with a NUL character.
87 If it is not NUL-terminated there is a risk of
88 core dumps and corruptions from code which passes the string to C
89 functions or system calls which expect a NUL-terminated string.
90 Perl's own functions typically add a trailing NUL for this reason.
91 Nevertheless, you should be very careful when you pass a string stored
92 in an SV to a C function or system call.
94 To access the actual value that an SV points to, you can use the macros:
102 which will automatically coerce the actual scalar type into an IV, UV, double,
105 In the C<SvPV> macro, the length of the string returned is placed into the
106 variable C<len> (this is a macro, so you do I<not> use C<&len>). If you do
107 not care what the length of the data is, use the C<SvPV_nolen> macro.
108 Historically the C<SvPV> macro with the global variable C<PL_na> has been
109 used in this case. But that can be quite inefficient because C<PL_na> must
110 be accessed in thread-local storage in threaded Perl. In any case, remember
111 that Perl allows arbitrary strings of data that may both contain NULs and
112 might not be terminated by a NUL.
114 Also remember that C doesn't allow you to safely say C<foo(SvPV(s, len),
115 len);>. It might work with your compiler, but it won't work for everyone.
116 Break this sort of statement up into separate assignments:
124 If you want to know if the scalar value is TRUE, you can use:
128 Although Perl will automatically grow strings for you, if you need to force
129 Perl to allocate more memory for your SV, you can use the macro
131 SvGROW(SV*, STRLEN newlen)
133 which will determine if more memory needs to be allocated. If so, it will
134 call the function C<sv_grow>. Note that C<SvGROW> can only increase, not
135 decrease, the allocated memory of an SV and that it does not automatically
136 add a byte for the a trailing NUL (perl's own string functions typically do
137 C<SvGROW(sv, len + 1)>).
139 If you have an SV and want to know what kind of data Perl thinks is stored
140 in it, you can use the following macros to check the type of SV you have.
146 You can get and set the current length of the string stored in an SV with
147 the following macros:
150 SvCUR_set(SV*, I32 val)
152 You can also get a pointer to the end of the string stored in the SV
157 But note that these last three macros are valid only if C<SvPOK()> is true.
159 If you want to append something to the end of string stored in an C<SV*>,
160 you can use the following functions:
162 void sv_catpv(SV*, const char*);
163 void sv_catpvn(SV*, const char*, STRLEN);
164 void sv_catpvf(SV*, const char*, ...);
165 void sv_catpvfn(SV*, const char*, STRLEN, va_list *, SV **, I32, bool);
166 void sv_catsv(SV*, SV*);
168 The first function calculates the length of the string to be appended by
169 using C<strlen>. In the second, you specify the length of the string
170 yourself. The third function processes its arguments like C<sprintf> and
171 appends the formatted output. The fourth function works like C<vsprintf>.
172 You can specify the address and length of an array of SVs instead of the
173 va_list argument. The fifth function extends the string stored in the first
174 SV with the string stored in the second SV. It also forces the second SV
175 to be interpreted as a string.
177 The C<sv_cat*()> functions are not generic enough to operate on values that
178 have "magic". See L<Magic Virtual Tables> later in this document.
180 If you know the name of a scalar variable, you can get a pointer to its SV
181 by using the following:
183 SV* get_sv("package::varname", FALSE);
185 This returns NULL if the variable does not exist.
187 If you want to know if this variable (or any other SV) is actually C<defined>,
192 The scalar C<undef> value is stored in an SV instance called C<PL_sv_undef>. Its
193 address can be used whenever an C<SV*> is needed.
195 There are also the two values C<PL_sv_yes> and C<PL_sv_no>, which contain Boolean
196 TRUE and FALSE values, respectively. Like C<PL_sv_undef>, their addresses can
197 be used whenever an C<SV*> is needed.
199 Do not be fooled into thinking that C<(SV *) 0> is the same as C<&PL_sv_undef>.
203 if (I-am-to-return-a-real-value) {
204 sv = sv_2mortal(newSViv(42));
208 This code tries to return a new SV (which contains the value 42) if it should
209 return a real value, or undef otherwise. Instead it has returned a NULL
210 pointer which, somewhere down the line, will cause a segmentation violation,
211 bus error, or just weird results. Change the zero to C<&PL_sv_undef> in the first
212 line and all will be well.
214 To free an SV that you've created, call C<SvREFCNT_dec(SV*)>. Normally this
215 call is not necessary (see L<Reference Counts and Mortality>).
219 Perl provides the function C<sv_chop> to efficiently remove characters
220 from the beginning of a string; you give it an SV and a pointer to
221 somewhere inside the the PV, and it discards everything before the
222 pointer. The efficiency comes by means of a little hack: instead of
223 actually removing the characters, C<sv_chop> sets the flag C<OOK>
224 (offset OK) to signal to other functions that the offset hack is in
225 effect, and it puts the number of bytes chopped off into the IV field
226 of the SV. It then moves the PV pointer (called C<SvPVX>) forward that
227 many bytes, and adjusts C<SvCUR> and C<SvLEN>.
229 Hence, at this point, the start of the buffer that we allocated lives
230 at C<SvPVX(sv) - SvIV(sv)> in memory and the PV pointer is pointing
231 into the middle of this allocated storage.
233 This is best demonstrated by example:
235 % ./perl -Ilib -MDevel::Peek -le '$a="12345"; $a=~s/.//; Dump($a)'
236 SV = PVIV(0x8128450) at 0x81340f0
238 FLAGS = (POK,OOK,pPOK)
240 PV = 0x8135781 ( "1" . ) "2345"\0
244 Here the number of bytes chopped off (1) is put into IV, and
245 C<Devel::Peek::Dump> helpfully reminds us that this is an offset. The
246 portion of the string between the "real" and the "fake" beginnings is
247 shown in parentheses, and the values of C<SvCUR> and C<SvLEN> reflect
248 the fake beginning, not the real one.
250 Something similar to the offset hack is perfomed on AVs to enable
251 efficient shifting and splicing off the beginning of the array; while
252 C<AvARRAY> points to the first element in the array that is visible from
253 Perl, C<AvALLOC> points to the real start of the C array. These are
254 usually the same, but a C<shift> operation can be carried out by
255 increasing C<AvARRAY> by one and decreasing C<AvFILL> and C<AvLEN>.
256 Again, the location of the real start of the C array only comes into
257 play when freeing the array. See C<av_shift> in F<av.c>.
259 =head2 What's Really Stored in an SV?
261 Recall that the usual method of determining the type of scalar you have is
262 to use C<Sv*OK> macros. Because a scalar can be both a number and a string,
263 usually these macros will always return TRUE and calling the C<Sv*V>
264 macros will do the appropriate conversion of string to integer/double or
265 integer/double to string.
267 If you I<really> need to know if you have an integer, double, or string
268 pointer in an SV, you can use the following three macros instead:
274 These will tell you if you truly have an integer, double, or string pointer
275 stored in your SV. The "p" stands for private.
277 In general, though, it's best to use the C<Sv*V> macros.
279 =head2 Working with AVs
281 There are two ways to create and load an AV. The first method creates an
286 The second method both creates the AV and initially populates it with SVs:
288 AV* av_make(I32 num, SV **ptr);
290 The second argument points to an array containing C<num> C<SV*>'s. Once the
291 AV has been created, the SVs can be destroyed, if so desired.
293 Once the AV has been created, the following operations are possible on AVs:
295 void av_push(AV*, SV*);
298 void av_unshift(AV*, I32 num);
300 These should be familiar operations, with the exception of C<av_unshift>.
301 This routine adds C<num> elements at the front of the array with the C<undef>
302 value. You must then use C<av_store> (described below) to assign values
303 to these new elements.
305 Here are some other functions:
308 SV** av_fetch(AV*, I32 key, I32 lval);
309 SV** av_store(AV*, I32 key, SV* val);
311 The C<av_len> function returns the highest index value in array (just
312 like $#array in Perl). If the array is empty, -1 is returned. The
313 C<av_fetch> function returns the value at index C<key>, but if C<lval>
314 is non-zero, then C<av_fetch> will store an undef value at that index.
315 The C<av_store> function stores the value C<val> at index C<key>, and does
316 not increment the reference count of C<val>. Thus the caller is responsible
317 for taking care of that, and if C<av_store> returns NULL, the caller will
318 have to decrement the reference count to avoid a memory leak. Note that
319 C<av_fetch> and C<av_store> both return C<SV**>'s, not C<SV*>'s as their
324 void av_extend(AV*, I32 key);
326 The C<av_clear> function deletes all the elements in the AV* array, but
327 does not actually delete the array itself. The C<av_undef> function will
328 delete all the elements in the array plus the array itself. The
329 C<av_extend> function extends the array so that it contains at least C<key+1>
330 elements. If C<key+1> is less than the currently allocated length of the array,
331 then nothing is done.
333 If you know the name of an array variable, you can get a pointer to its AV
334 by using the following:
336 AV* get_av("package::varname", FALSE);
338 This returns NULL if the variable does not exist.
340 See L<Understanding the Magic of Tied Hashes and Arrays> for more
341 information on how to use the array access functions on tied arrays.
343 =head2 Working with HVs
345 To create an HV, you use the following routine:
349 Once the HV has been created, the following operations are possible on HVs:
351 SV** hv_store(HV*, const char* key, U32 klen, SV* val, U32 hash);
352 SV** hv_fetch(HV*, const char* key, U32 klen, I32 lval);
354 The C<klen> parameter is the length of the key being passed in (Note that
355 you cannot pass 0 in as a value of C<klen> to tell Perl to measure the
356 length of the key). The C<val> argument contains the SV pointer to the
357 scalar being stored, and C<hash> is the precomputed hash value (zero if
358 you want C<hv_store> to calculate it for you). The C<lval> parameter
359 indicates whether this fetch is actually a part of a store operation, in
360 which case a new undefined value will be added to the HV with the supplied
361 key and C<hv_fetch> will return as if the value had already existed.
363 Remember that C<hv_store> and C<hv_fetch> return C<SV**>'s and not just
364 C<SV*>. To access the scalar value, you must first dereference the return
365 value. However, you should check to make sure that the return value is
366 not NULL before dereferencing it.
368 These two functions check if a hash table entry exists, and deletes it.
370 bool hv_exists(HV*, const char* key, U32 klen);
371 SV* hv_delete(HV*, const char* key, U32 klen, I32 flags);
373 If C<flags> does not include the C<G_DISCARD> flag then C<hv_delete> will
374 create and return a mortal copy of the deleted value.
376 And more miscellaneous functions:
381 Like their AV counterparts, C<hv_clear> deletes all the entries in the hash
382 table but does not actually delete the hash table. The C<hv_undef> deletes
383 both the entries and the hash table itself.
385 Perl keeps the actual data in linked list of structures with a typedef of HE.
386 These contain the actual key and value pointers (plus extra administrative
387 overhead). The key is a string pointer; the value is an C<SV*>. However,
388 once you have an C<HE*>, to get the actual key and value, use the routines
391 I32 hv_iterinit(HV*);
392 /* Prepares starting point to traverse hash table */
393 HE* hv_iternext(HV*);
394 /* Get the next entry, and return a pointer to a
395 structure that has both the key and value */
396 char* hv_iterkey(HE* entry, I32* retlen);
397 /* Get the key from an HE structure and also return
398 the length of the key string */
399 SV* hv_iterval(HV*, HE* entry);
400 /* Return a SV pointer to the value of the HE
402 SV* hv_iternextsv(HV*, char** key, I32* retlen);
403 /* This convenience routine combines hv_iternext,
404 hv_iterkey, and hv_iterval. The key and retlen
405 arguments are return values for the key and its
406 length. The value is returned in the SV* argument */
408 If you know the name of a hash variable, you can get a pointer to its HV
409 by using the following:
411 HV* get_hv("package::varname", FALSE);
413 This returns NULL if the variable does not exist.
415 The hash algorithm is defined in the C<PERL_HASH(hash, key, klen)> macro:
419 hash = (hash * 33) + *key++;
420 hash = hash + (hash >> 5); /* after 5.6 */
422 The last step was added in version 5.6 to improve distribution of
423 lower bits in the resulting hash value.
425 See L<Understanding the Magic of Tied Hashes and Arrays> for more
426 information on how to use the hash access functions on tied hashes.
428 =head2 Hash API Extensions
430 Beginning with version 5.004, the following functions are also supported:
432 HE* hv_fetch_ent (HV* tb, SV* key, I32 lval, U32 hash);
433 HE* hv_store_ent (HV* tb, SV* key, SV* val, U32 hash);
435 bool hv_exists_ent (HV* tb, SV* key, U32 hash);
436 SV* hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
438 SV* hv_iterkeysv (HE* entry);
440 Note that these functions take C<SV*> keys, which simplifies writing
441 of extension code that deals with hash structures. These functions
442 also allow passing of C<SV*> keys to C<tie> functions without forcing
443 you to stringify the keys (unlike the previous set of functions).
445 They also return and accept whole hash entries (C<HE*>), making their
446 use more efficient (since the hash number for a particular string
447 doesn't have to be recomputed every time). See L<perlapi> for detailed
450 The following macros must always be used to access the contents of hash
451 entries. Note that the arguments to these macros must be simple
452 variables, since they may get evaluated more than once. See
453 L<perlapi> for detailed descriptions of these macros.
455 HePV(HE* he, STRLEN len)
459 HeSVKEY_force(HE* he)
460 HeSVKEY_set(HE* he, SV* sv)
462 These two lower level macros are defined, but must only be used when
463 dealing with keys that are not C<SV*>s:
468 Note that both C<hv_store> and C<hv_store_ent> do not increment the
469 reference count of the stored C<val>, which is the caller's responsibility.
470 If these functions return a NULL value, the caller will usually have to
471 decrement the reference count of C<val> to avoid a memory leak.
475 References are a special type of scalar that point to other data types
476 (including references).
478 To create a reference, use either of the following functions:
480 SV* newRV_inc((SV*) thing);
481 SV* newRV_noinc((SV*) thing);
483 The C<thing> argument can be any of an C<SV*>, C<AV*>, or C<HV*>. The
484 functions are identical except that C<newRV_inc> increments the reference
485 count of the C<thing>, while C<newRV_noinc> does not. For historical
486 reasons, C<newRV> is a synonym for C<newRV_inc>.
488 Once you have a reference, you can use the following macro to dereference
493 then call the appropriate routines, casting the returned C<SV*> to either an
494 C<AV*> or C<HV*>, if required.
496 To determine if an SV is a reference, you can use the following macro:
500 To discover what type of value the reference refers to, use the following
501 macro and then check the return value.
505 The most useful types that will be returned are:
514 SVt_PVGV Glob (possible a file handle)
515 SVt_PVMG Blessed or Magical Scalar
517 See the sv.h header file for more details.
519 =head2 Blessed References and Class Objects
521 References are also used to support object-oriented programming. In the
522 OO lexicon, an object is simply a reference that has been blessed into a
523 package (or class). Once blessed, the programmer may now use the reference
524 to access the various methods in the class.
526 A reference can be blessed into a package with the following function:
528 SV* sv_bless(SV* sv, HV* stash);
530 The C<sv> argument must be a reference. The C<stash> argument specifies
531 which class the reference will belong to. See
532 L<Stashes and Globs> for information on converting class names into stashes.
534 /* Still under construction */
536 Upgrades rv to reference if not already one. Creates new SV for rv to
537 point to. If C<classname> is non-null, the SV is blessed into the specified
538 class. SV is returned.
540 SV* newSVrv(SV* rv, const char* classname);
542 Copies integer, unsigned integer or double into an SV whose reference is C<rv>. SV is blessed
543 if C<classname> is non-null.
545 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
546 SV* sv_setref_uv(SV* rv, const char* classname, UV uv);
547 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
549 Copies the pointer value (I<the address, not the string!>) into an SV whose
550 reference is rv. SV is blessed if C<classname> is non-null.
552 SV* sv_setref_pv(SV* rv, const char* classname, PV iv);
554 Copies string into an SV whose reference is C<rv>. Set length to 0 to let
555 Perl calculate the string length. SV is blessed if C<classname> is non-null.
557 SV* sv_setref_pvn(SV* rv, const char* classname, PV iv, STRLEN length);
559 Tests whether the SV is blessed into the specified class. It does not
560 check inheritance relationships.
562 int sv_isa(SV* sv, const char* name);
564 Tests whether the SV is a reference to a blessed object.
566 int sv_isobject(SV* sv);
568 Tests whether the SV is derived from the specified class. SV can be either
569 a reference to a blessed object or a string containing a class name. This
570 is the function implementing the C<UNIVERSAL::isa> functionality.
572 bool sv_derived_from(SV* sv, const char* name);
574 To check if you've got an object derived from a specific class you have
577 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
579 =head2 Creating New Variables
581 To create a new Perl variable with an undef value which can be accessed from
582 your Perl script, use the following routines, depending on the variable type.
584 SV* get_sv("package::varname", TRUE);
585 AV* get_av("package::varname", TRUE);
586 HV* get_hv("package::varname", TRUE);
588 Notice the use of TRUE as the second parameter. The new variable can now
589 be set, using the routines appropriate to the data type.
591 There are additional macros whose values may be bitwise OR'ed with the
592 C<TRUE> argument to enable certain extra features. Those bits are:
594 GV_ADDMULTI Marks the variable as multiply defined, thus preventing the
595 "Name <varname> used only once: possible typo" warning.
596 GV_ADDWARN Issues the warning "Had to create <varname> unexpectedly" if
597 the variable did not exist before the function was called.
599 If you do not specify a package name, the variable is created in the current
602 =head2 Reference Counts and Mortality
604 Perl uses an reference count-driven garbage collection mechanism. SVs,
605 AVs, or HVs (xV for short in the following) start their life with a
606 reference count of 1. If the reference count of an xV ever drops to 0,
607 then it will be destroyed and its memory made available for reuse.
609 This normally doesn't happen at the Perl level unless a variable is
610 undef'ed or the last variable holding a reference to it is changed or
611 overwritten. At the internal level, however, reference counts can be
612 manipulated with the following macros:
614 int SvREFCNT(SV* sv);
615 SV* SvREFCNT_inc(SV* sv);
616 void SvREFCNT_dec(SV* sv);
618 However, there is one other function which manipulates the reference
619 count of its argument. The C<newRV_inc> function, you will recall,
620 creates a reference to the specified argument. As a side effect,
621 it increments the argument's reference count. If this is not what
622 you want, use C<newRV_noinc> instead.
624 For example, imagine you want to return a reference from an XSUB function.
625 Inside the XSUB routine, you create an SV which initially has a reference
626 count of one. Then you call C<newRV_inc>, passing it the just-created SV.
627 This returns the reference as a new SV, but the reference count of the
628 SV you passed to C<newRV_inc> has been incremented to two. Now you
629 return the reference from the XSUB routine and forget about the SV.
630 But Perl hasn't! Whenever the returned reference is destroyed, the
631 reference count of the original SV is decreased to one and nothing happens.
632 The SV will hang around without any way to access it until Perl itself
633 terminates. This is a memory leak.
635 The correct procedure, then, is to use C<newRV_noinc> instead of
636 C<newRV_inc>. Then, if and when the last reference is destroyed,
637 the reference count of the SV will go to zero and it will be destroyed,
638 stopping any memory leak.
640 There are some convenience functions available that can help with the
641 destruction of xVs. These functions introduce the concept of "mortality".
642 An xV that is mortal has had its reference count marked to be decremented,
643 but not actually decremented, until "a short time later". Generally the
644 term "short time later" means a single Perl statement, such as a call to
645 an XSUB function. The actual determinant for when mortal xVs have their
646 reference count decremented depends on two macros, SAVETMPS and FREETMPS.
647 See L<perlcall> and L<perlxs> for more details on these macros.
649 "Mortalization" then is at its simplest a deferred C<SvREFCNT_dec>.
650 However, if you mortalize a variable twice, the reference count will
651 later be decremented twice.
653 You should be careful about creating mortal variables. Strange things
654 can happen if you make the same value mortal within multiple contexts,
655 or if you make a variable mortal multiple times.
657 To create a mortal variable, use the functions:
661 SV* sv_mortalcopy(SV*)
663 The first call creates a mortal SV, the second converts an existing
664 SV to a mortal SV (and thus defers a call to C<SvREFCNT_dec>), and the
665 third creates a mortal copy of an existing SV.
667 The mortal routines are not just for SVs -- AVs and HVs can be
668 made mortal by passing their address (type-casted to C<SV*>) to the
669 C<sv_2mortal> or C<sv_mortalcopy> routines.
671 =head2 Stashes and Globs
673 A "stash" is a hash that contains all of the different objects that
674 are contained within a package. Each key of the stash is a symbol
675 name (shared by all the different types of objects that have the same
676 name), and each value in the hash table is a GV (Glob Value). This GV
677 in turn contains references to the various objects of that name,
678 including (but not limited to) the following:
687 There is a single stash called "PL_defstash" that holds the items that exist
688 in the "main" package. To get at the items in other packages, append the
689 string "::" to the package name. The items in the "Foo" package are in
690 the stash "Foo::" in PL_defstash. The items in the "Bar::Baz" package are
691 in the stash "Baz::" in "Bar::"'s stash.
693 To get the stash pointer for a particular package, use the function:
695 HV* gv_stashpv(const char* name, I32 create)
696 HV* gv_stashsv(SV*, I32 create)
698 The first function takes a literal string, the second uses the string stored
699 in the SV. Remember that a stash is just a hash table, so you get back an
700 C<HV*>. The C<create> flag will create a new package if it is set.
702 The name that C<gv_stash*v> wants is the name of the package whose symbol table
703 you want. The default package is called C<main>. If you have multiply nested
704 packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
707 Alternately, if you have an SV that is a blessed reference, you can find
708 out the stash pointer by using:
710 HV* SvSTASH(SvRV(SV*));
712 then use the following to get the package name itself:
714 char* HvNAME(HV* stash);
716 If you need to bless or re-bless an object you can use the following
719 SV* sv_bless(SV*, HV* stash)
721 where the first argument, an C<SV*>, must be a reference, and the second
722 argument is a stash. The returned C<SV*> can now be used in the same way
725 For more information on references and blessings, consult L<perlref>.
727 =head2 Double-Typed SVs
729 Scalar variables normally contain only one type of value, an integer,
730 double, pointer, or reference. Perl will automatically convert the
731 actual scalar data from the stored type into the requested type.
733 Some scalar variables contain more than one type of scalar data. For
734 example, the variable C<$!> contains either the numeric value of C<errno>
735 or its string equivalent from either C<strerror> or C<sys_errlist[]>.
737 To force multiple data values into an SV, you must do two things: use the
738 C<sv_set*v> routines to add the additional scalar type, then set a flag
739 so that Perl will believe it contains more than one type of data. The
740 four macros to set the flags are:
747 The particular macro you must use depends on which C<sv_set*v> routine
748 you called first. This is because every C<sv_set*v> routine turns on
749 only the bit for the particular type of data being set, and turns off
752 For example, to create a new Perl variable called "dberror" that contains
753 both the numeric and descriptive string error values, you could use the
757 extern char *dberror_list;
759 SV* sv = get_sv("dberror", TRUE);
760 sv_setiv(sv, (IV) dberror);
761 sv_setpv(sv, dberror_list[dberror]);
764 If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
765 macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
767 =head2 Magic Variables
769 [This section still under construction. Ignore everything here. Post no
770 bills. Everything not permitted is forbidden.]
772 Any SV may be magical, that is, it has special features that a normal
773 SV does not have. These features are stored in the SV structure in a
774 linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
787 Note this is current as of patchlevel 0, and could change at any time.
789 =head2 Assigning Magic
791 Perl adds magic to an SV using the sv_magic function:
793 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
795 The C<sv> argument is a pointer to the SV that is to acquire a new magical
798 If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
799 set the C<SVt_PVMG> flag for the C<sv>. Perl then continues by adding
800 it to the beginning of the linked list of magical features. Any prior
801 entry of the same type of magic is deleted. Note that this can be
802 overridden, and multiple instances of the same type of magic can be
803 associated with an SV.
805 The C<name> and C<namlen> arguments are used to associate a string with
806 the magic, typically the name of a variable. C<namlen> is stored in the
807 C<mg_len> field and if C<name> is non-null and C<namlen> >= 0 a malloc'd
808 copy of the name is stored in C<mg_ptr> field.
810 The sv_magic function uses C<how> to determine which, if any, predefined
811 "Magic Virtual Table" should be assigned to the C<mg_virtual> field.
812 See the "Magic Virtual Table" section below. The C<how> argument is also
813 stored in the C<mg_type> field.
815 The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
816 structure. If it is not the same as the C<sv> argument, the reference
817 count of the C<obj> object is incremented. If it is the same, or if
818 the C<how> argument is "#", or if it is a NULL pointer, then C<obj> is
819 merely stored, without the reference count being incremented.
821 There is also a function to add magic to an C<HV>:
823 void hv_magic(HV *hv, GV *gv, int how);
825 This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
827 To remove the magic from an SV, call the function sv_unmagic:
829 void sv_unmagic(SV *sv, int type);
831 The C<type> argument should be equal to the C<how> value when the C<SV>
832 was initially made magical.
834 =head2 Magic Virtual Tables
836 The C<mg_virtual> field in the C<MAGIC> structure is a pointer to a
837 C<MGVTBL>, which is a structure of function pointers and stands for
838 "Magic Virtual Table" to handle the various operations that might be
839 applied to that variable.
841 The C<MGVTBL> has five pointers to the following routine types:
843 int (*svt_get)(SV* sv, MAGIC* mg);
844 int (*svt_set)(SV* sv, MAGIC* mg);
845 U32 (*svt_len)(SV* sv, MAGIC* mg);
846 int (*svt_clear)(SV* sv, MAGIC* mg);
847 int (*svt_free)(SV* sv, MAGIC* mg);
849 This MGVTBL structure is set at compile-time in C<perl.h> and there are
850 currently 19 types (or 21 with overloading turned on). These different
851 structures contain pointers to various routines that perform additional
852 actions depending on which function is being called.
854 Function pointer Action taken
855 ---------------- ------------
856 svt_get Do something after the value of the SV is retrieved.
857 svt_set Do something after the SV is assigned a value.
858 svt_len Report on the SV's length.
859 svt_clear Clear something the SV represents.
860 svt_free Free any extra storage associated with the SV.
862 For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
863 to an C<mg_type> of '\0') contains:
865 { magic_get, magic_set, magic_len, 0, 0 }
867 Thus, when an SV is determined to be magical and of type '\0', if a get
868 operation is being performed, the routine C<magic_get> is called. All
869 the various routines for the various magical types begin with C<magic_>.
870 NOTE: the magic routines are not considered part of the Perl API, and may
871 not be exported by the Perl library.
873 The current kinds of Magic Virtual Tables are:
875 mg_type MGVTBL Type of magic
876 ------- ------ ----------------------------
877 \0 vtbl_sv Special scalar variable
878 A vtbl_amagic %OVERLOAD hash
879 a vtbl_amagicelem %OVERLOAD hash element
880 c (none) Holds overload table (AMT) on stash
881 B vtbl_bm Boyer-Moore (fast string search)
882 D vtbl_regdata Regex match position data (@+ and @- vars)
883 d vtbl_regdatum Regex match position data element
885 e vtbl_envelem %ENV hash element
886 f vtbl_fm Formline ('compiled' format)
887 g vtbl_mglob m//g target / study()ed string
888 I vtbl_isa @ISA array
889 i vtbl_isaelem @ISA array element
890 k vtbl_nkeys scalar(keys()) lvalue
891 L (none) Debugger %_<filename
892 l vtbl_dbline Debugger %_<filename element
893 o vtbl_collxfrm Locale transformation
894 P vtbl_pack Tied array or hash
895 p vtbl_packelem Tied array or hash element
896 q vtbl_packelem Tied scalar or handle
898 s vtbl_sigelem %SIG hash element
899 t vtbl_taint Taintedness
900 U vtbl_uvar Available for use by extensions
901 v vtbl_vec vec() lvalue
902 x vtbl_substr substr() lvalue
903 y vtbl_defelem Shadow "foreach" iterator variable /
904 smart parameter vivification
905 * vtbl_glob GV (typeglob)
906 # vtbl_arylen Array length ($#ary)
907 . vtbl_pos pos() lvalue
908 ~ (none) Available for use by extensions
910 When an uppercase and lowercase letter both exist in the table, then the
911 uppercase letter is used to represent some kind of composite type (a list
912 or a hash), and the lowercase letter is used to represent an element of
915 The '~' and 'U' magic types are defined specifically for use by
916 extensions and will not be used by perl itself. Extensions can use
917 '~' magic to 'attach' private information to variables (typically
918 objects). This is especially useful because there is no way for
919 normal perl code to corrupt this private information (unlike using
920 extra elements of a hash object).
922 Similarly, 'U' magic can be used much like tie() to call a C function
923 any time a scalar's value is used or changed. The C<MAGIC>'s
924 C<mg_ptr> field points to a C<ufuncs> structure:
927 I32 (*uf_val)(IV, SV*);
928 I32 (*uf_set)(IV, SV*);
932 When the SV is read from or written to, the C<uf_val> or C<uf_set>
933 function will be called with C<uf_index> as the first arg and a
934 pointer to the SV as the second. A simple example of how to add 'U'
935 magic is shown below. Note that the ufuncs structure is copied by
936 sv_magic, so you can safely allocate it on the stack.
944 uf.uf_val = &my_get_fn;
945 uf.uf_set = &my_set_fn;
947 sv_magic(sv, 0, 'U', (char*)&uf, sizeof(uf));
949 Note that because multiple extensions may be using '~' or 'U' magic,
950 it is important for extensions to take extra care to avoid conflict.
951 Typically only using the magic on objects blessed into the same class
952 as the extension is sufficient. For '~' magic, it may also be
953 appropriate to add an I32 'signature' at the top of the private data
956 Also note that the C<sv_set*()> and C<sv_cat*()> functions described
957 earlier do B<not> invoke 'set' magic on their targets. This must
958 be done by the user either by calling the C<SvSETMAGIC()> macro after
959 calling these functions, or by using one of the C<sv_set*_mg()> or
960 C<sv_cat*_mg()> functions. Similarly, generic C code must call the
961 C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
962 obtained from external sources in functions that don't handle magic.
963 See L<perlapi> for a description of these functions.
964 For example, calls to the C<sv_cat*()> functions typically need to be
965 followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
966 since their implementation handles 'get' magic.
970 MAGIC* mg_find(SV*, int type); /* Finds the magic pointer of that type */
972 This routine returns a pointer to the C<MAGIC> structure stored in the SV.
973 If the SV does not have that magical feature, C<NULL> is returned. Also,
974 if the SV is not of type SVt_PVMG, Perl may core dump.
976 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
978 This routine checks to see what types of magic C<sv> has. If the mg_type
979 field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
980 the mg_type field is changed to be the lowercase letter.
982 =head2 Understanding the Magic of Tied Hashes and Arrays
984 Tied hashes and arrays are magical beasts of the 'P' magic type.
986 WARNING: As of the 5.004 release, proper usage of the array and hash
987 access functions requires understanding a few caveats. Some
988 of these caveats are actually considered bugs in the API, to be fixed
989 in later releases, and are bracketed with [MAYCHANGE] below. If
990 you find yourself actually applying such information in this section, be
991 aware that the behavior may change in the future, umm, without warning.
993 The perl tie function associates a variable with an object that implements
994 the various GET, SET etc methods. To perform the equivalent of the perl
995 tie function from an XSUB, you must mimic this behaviour. The code below
996 carries out the necessary steps - firstly it creates a new hash, and then
997 creates a second hash which it blesses into the class which will implement
998 the tie methods. Lastly it ties the two hashes together, and returns a
999 reference to the new tied hash. Note that the code below does NOT call the
1000 TIEHASH method in the MyTie class -
1001 see L<Calling Perl Routines from within C Programs> for details on how
1012 tie = newRV_noinc((SV*)newHV());
1013 stash = gv_stashpv("MyTie", TRUE);
1014 sv_bless(tie, stash);
1015 hv_magic(hash, tie, 'P');
1016 RETVAL = newRV_noinc(hash);
1020 The C<av_store> function, when given a tied array argument, merely
1021 copies the magic of the array onto the value to be "stored", using
1022 C<mg_copy>. It may also return NULL, indicating that the value did not
1023 actually need to be stored in the array. [MAYCHANGE] After a call to
1024 C<av_store> on a tied array, the caller will usually need to call
1025 C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1026 TIEARRAY object. If C<av_store> did return NULL, a call to
1027 C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1030 The previous paragraph is applicable verbatim to tied hash access using the
1031 C<hv_store> and C<hv_store_ent> functions as well.
1033 C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1034 C<hv_fetch_ent> actually return an undefined mortal value whose magic
1035 has been initialized using C<mg_copy>. Note the value so returned does not
1036 need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1037 need to call C<mg_get()> on the returned value in order to actually invoke
1038 the perl level "FETCH" method on the underlying TIE object. Similarly,
1039 you may also call C<mg_set()> on the return value after possibly assigning
1040 a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
1041 method on the TIE object. [/MAYCHANGE]
1044 In other words, the array or hash fetch/store functions don't really
1045 fetch and store actual values in the case of tied arrays and hashes. They
1046 merely call C<mg_copy> to attach magic to the values that were meant to be
1047 "stored" or "fetched". Later calls to C<mg_get> and C<mg_set> actually
1048 do the job of invoking the TIE methods on the underlying objects. Thus
1049 the magic mechanism currently implements a kind of lazy access to arrays
1052 Currently (as of perl version 5.004), use of the hash and array access
1053 functions requires the user to be aware of whether they are operating on
1054 "normal" hashes and arrays, or on their tied variants. The API may be
1055 changed to provide more transparent access to both tied and normal data
1056 types in future versions.
1059 You would do well to understand that the TIEARRAY and TIEHASH interfaces
1060 are mere sugar to invoke some perl method calls while using the uniform hash
1061 and array syntax. The use of this sugar imposes some overhead (typically
1062 about two to four extra opcodes per FETCH/STORE operation, in addition to
1063 the creation of all the mortal variables required to invoke the methods).
1064 This overhead will be comparatively small if the TIE methods are themselves
1065 substantial, but if they are only a few statements long, the overhead
1066 will not be insignificant.
1068 =head2 Localizing changes
1070 Perl has a very handy construction
1077 This construction is I<approximately> equivalent to
1086 The biggest difference is that the first construction would
1087 reinstate the initial value of $var, irrespective of how control exits
1088 the block: C<goto>, C<return>, C<die>/C<eval> etc. It is a little bit
1089 more efficient as well.
1091 There is a way to achieve a similar task from C via Perl API: create a
1092 I<pseudo-block>, and arrange for some changes to be automatically
1093 undone at the end of it, either explicit, or via a non-local exit (via
1094 die()). A I<block>-like construct is created by a pair of
1095 C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1096 Such a construct may be created specially for some important localized
1097 task, or an existing one (like boundaries of enclosing Perl
1098 subroutine/block, or an existing pair for freeing TMPs) may be
1099 used. (In the second case the overhead of additional localization must
1100 be almost negligible.) Note that any XSUB is automatically enclosed in
1101 an C<ENTER>/C<LEAVE> pair.
1103 Inside such a I<pseudo-block> the following service is available:
1107 =item C<SAVEINT(int i)>
1109 =item C<SAVEIV(IV i)>
1111 =item C<SAVEI32(I32 i)>
1113 =item C<SAVELONG(long i)>
1115 These macros arrange things to restore the value of integer variable
1116 C<i> at the end of enclosing I<pseudo-block>.
1118 =item C<SAVESPTR(s)>
1120 =item C<SAVEPPTR(p)>
1122 These macros arrange things to restore the value of pointers C<s> and
1123 C<p>. C<s> must be a pointer of a type which survives conversion to
1124 C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1127 =item C<SAVEFREESV(SV *sv)>
1129 The refcount of C<sv> would be decremented at the end of
1130 I<pseudo-block>. This is similar to C<sv_2mortal>, which should (?) be
1133 =item C<SAVEFREEOP(OP *op)>
1135 The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1137 =item C<SAVEFREEPV(p)>
1139 The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1140 end of I<pseudo-block>.
1142 =item C<SAVECLEARSV(SV *sv)>
1144 Clears a slot in the current scratchpad which corresponds to C<sv> at
1145 the end of I<pseudo-block>.
1147 =item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1149 The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
1150 string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1151 short-lived storage, the corresponding string may be reallocated like
1154 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
1156 =item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
1158 At the end of I<pseudo-block> the function C<f> is called with the
1161 =item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1163 At the end of I<pseudo-block> the function C<f> is called with the
1164 implicit context argument (if any), and C<p>.
1166 =item C<SAVESTACK_POS()>
1168 The current offset on the Perl internal stack (cf. C<SP>) is restored
1169 at the end of I<pseudo-block>.
1173 The following API list contains functions, thus one needs to
1174 provide pointers to the modifiable data explicitly (either C pointers,
1175 or Perlish C<GV *>s). Where the above macros take C<int>, a similar
1176 function takes C<int *>.
1180 =item C<SV* save_scalar(GV *gv)>
1182 Equivalent to Perl code C<local $gv>.
1184 =item C<AV* save_ary(GV *gv)>
1186 =item C<HV* save_hash(GV *gv)>
1188 Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1190 =item C<void save_item(SV *item)>
1192 Duplicates the current value of C<SV>, on the exit from the current
1193 C<ENTER>/C<LEAVE> I<pseudo-block> will restore the value of C<SV>
1194 using the stored value.
1196 =item C<void save_list(SV **sarg, I32 maxsarg)>
1198 A variant of C<save_item> which takes multiple arguments via an array
1199 C<sarg> of C<SV*> of length C<maxsarg>.
1201 =item C<SV* save_svref(SV **sptr)>
1203 Similar to C<save_scalar>, but will reinstate a C<SV *>.
1205 =item C<void save_aptr(AV **aptr)>
1207 =item C<void save_hptr(HV **hptr)>
1209 Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1213 The C<Alias> module implements localization of the basic types within the
1214 I<caller's scope>. People who are interested in how to localize things in
1215 the containing scope should take a look there too.
1219 =head2 XSUBs and the Argument Stack
1221 The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1222 An XSUB routine will have a stack that contains the arguments from the Perl
1223 program, and a way to map from the Perl data structures to a C equivalent.
1225 The stack arguments are accessible through the C<ST(n)> macro, which returns
1226 the C<n>'th stack argument. Argument 0 is the first argument passed in the
1227 Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1230 Most of the time, output from the C routine can be handled through use of
1231 the RETVAL and OUTPUT directives. However, there are some cases where the
1232 argument stack is not already long enough to handle all the return values.
1233 An example is the POSIX tzname() call, which takes no arguments, but returns
1234 two, the local time zone's standard and summer time abbreviations.
1236 To handle this situation, the PPCODE directive is used and the stack is
1237 extended using the macro:
1241 where C<SP> is the macro that represents the local copy of the stack pointer,
1242 and C<num> is the number of elements the stack should be extended by.
1244 Now that there is room on the stack, values can be pushed on it using the
1245 macros to push IVs, doubles, strings, and SV pointers respectively:
1252 And now the Perl program calling C<tzname>, the two values will be assigned
1255 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1257 An alternate (and possibly simpler) method to pushing values on the stack is
1265 These macros automatically adjust the stack for you, if needed. Thus, you
1266 do not need to call C<EXTEND> to extend the stack.
1267 However, see L</Putting a C value on Perl stack>
1269 For more information, consult L<perlxs> and L<perlxstut>.
1271 =head2 Calling Perl Routines from within C Programs
1273 There are four routines that can be used to call a Perl subroutine from
1274 within a C program. These four are:
1276 I32 call_sv(SV*, I32);
1277 I32 call_pv(const char*, I32);
1278 I32 call_method(const char*, I32);
1279 I32 call_argv(const char*, I32, register char**);
1281 The routine most often used is C<call_sv>. The C<SV*> argument
1282 contains either the name of the Perl subroutine to be called, or a
1283 reference to the subroutine. The second argument consists of flags
1284 that control the context in which the subroutine is called, whether
1285 or not the subroutine is being passed arguments, how errors should be
1286 trapped, and how to treat return values.
1288 All four routines return the number of arguments that the subroutine returned
1291 These routines used to be called C<perl_call_sv> etc., before Perl v5.6.0,
1292 but those names are now deprecated; macros of the same name are provided for
1295 When using any of these routines (except C<call_argv>), the programmer
1296 must manipulate the Perl stack. These include the following macros and
1311 For a detailed description of calling conventions from C to Perl,
1312 consult L<perlcall>.
1314 =head2 Memory Allocation
1316 All memory meant to be used with the Perl API functions should be manipulated
1317 using the macros described in this section. The macros provide the necessary
1318 transparency between differences in the actual malloc implementation that is
1321 It is suggested that you enable the version of malloc that is distributed
1322 with Perl. It keeps pools of various sizes of unallocated memory in
1323 order to satisfy allocation requests more quickly. However, on some
1324 platforms, it may cause spurious malloc or free errors.
1326 New(x, pointer, number, type);
1327 Newc(x, pointer, number, type, cast);
1328 Newz(x, pointer, number, type);
1330 These three macros are used to initially allocate memory.
1332 The first argument C<x> was a "magic cookie" that was used to keep track
1333 of who called the macro, to help when debugging memory problems. However,
1334 the current code makes no use of this feature (most Perl developers now
1335 use run-time memory checkers), so this argument can be any number.
1337 The second argument C<pointer> should be the name of a variable that will
1338 point to the newly allocated memory.
1340 The third and fourth arguments C<number> and C<type> specify how many of
1341 the specified type of data structure should be allocated. The argument
1342 C<type> is passed to C<sizeof>. The final argument to C<Newc>, C<cast>,
1343 should be used if the C<pointer> argument is different from the C<type>
1346 Unlike the C<New> and C<Newc> macros, the C<Newz> macro calls C<memzero>
1347 to zero out all the newly allocated memory.
1349 Renew(pointer, number, type);
1350 Renewc(pointer, number, type, cast);
1353 These three macros are used to change a memory buffer size or to free a
1354 piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1355 match those of C<New> and C<Newc> with the exception of not needing the
1356 "magic cookie" argument.
1358 Move(source, dest, number, type);
1359 Copy(source, dest, number, type);
1360 Zero(dest, number, type);
1362 These three macros are used to move, copy, or zero out previously allocated
1363 memory. The C<source> and C<dest> arguments point to the source and
1364 destination starting points. Perl will move, copy, or zero out C<number>
1365 instances of the size of the C<type> data structure (using the C<sizeof>
1370 The most recent development releases of Perl has been experimenting with
1371 removing Perl's dependency on the "normal" standard I/O suite and allowing
1372 other stdio implementations to be used. This involves creating a new
1373 abstraction layer that then calls whichever implementation of stdio Perl
1374 was compiled with. All XSUBs should now use the functions in the PerlIO
1375 abstraction layer and not make any assumptions about what kind of stdio
1378 For a complete description of the PerlIO abstraction, consult L<perlapio>.
1380 =head2 Putting a C value on Perl stack
1382 A lot of opcodes (this is an elementary operation in the internal perl
1383 stack machine) put an SV* on the stack. However, as an optimization
1384 the corresponding SV is (usually) not recreated each time. The opcodes
1385 reuse specially assigned SVs (I<target>s) which are (as a corollary)
1386 not constantly freed/created.
1388 Each of the targets is created only once (but see
1389 L<Scratchpads and recursion> below), and when an opcode needs to put
1390 an integer, a double, or a string on stack, it just sets the
1391 corresponding parts of its I<target> and puts the I<target> on stack.
1393 The macro to put this target on stack is C<PUSHTARG>, and it is
1394 directly used in some opcodes, as well as indirectly in zillions of
1395 others, which use it via C<(X)PUSH[pni]>.
1397 Because the target is reused, you must be careful when pushing multiple
1398 values on the stack. The following code will not do what you think:
1403 This translates as "set C<TARG> to 10, push a pointer to C<TARG> onto
1404 the stack; set C<TARG> to 20, push a pointer to C<TARG> onto the stack".
1405 At the end of the operation, the stack does not contain the values 10
1406 and 20, but actually contains two pointers to C<TARG>, which we have set
1407 to 20. If you need to push multiple different values, use C<XPUSHs>,
1408 which bypasses C<TARG>.
1410 On a related note, if you do use C<(X)PUSH[npi]>, then you're going to
1411 need a C<dTARG> in your variable declarations so that the C<*PUSH*>
1412 macros can make use of the local variable C<TARG>.
1416 The question remains on when the SVs which are I<target>s for opcodes
1417 are created. The answer is that they are created when the current unit --
1418 a subroutine or a file (for opcodes for statements outside of
1419 subroutines) -- is compiled. During this time a special anonymous Perl
1420 array is created, which is called a scratchpad for the current
1423 A scratchpad keeps SVs which are lexicals for the current unit and are
1424 targets for opcodes. One can deduce that an SV lives on a scratchpad
1425 by looking on its flags: lexicals have C<SVs_PADMY> set, and
1426 I<target>s have C<SVs_PADTMP> set.
1428 The correspondence between OPs and I<target>s is not 1-to-1. Different
1429 OPs in the compile tree of the unit can use the same target, if this
1430 would not conflict with the expected life of the temporary.
1432 =head2 Scratchpads and recursion
1434 In fact it is not 100% true that a compiled unit contains a pointer to
1435 the scratchpad AV. In fact it contains a pointer to an AV of
1436 (initially) one element, and this element is the scratchpad AV. Why do
1437 we need an extra level of indirection?
1439 The answer is B<recursion>, and maybe (sometime soon) B<threads>. Both
1440 these can create several execution pointers going into the same
1441 subroutine. For the subroutine-child not write over the temporaries
1442 for the subroutine-parent (lifespan of which covers the call to the
1443 child), the parent and the child should have different
1444 scratchpads. (I<And> the lexicals should be separate anyway!)
1446 So each subroutine is born with an array of scratchpads (of length 1).
1447 On each entry to the subroutine it is checked that the current
1448 depth of the recursion is not more than the length of this array, and
1449 if it is, new scratchpad is created and pushed into the array.
1451 The I<target>s on this scratchpad are C<undef>s, but they are already
1452 marked with correct flags.
1454 =head1 Compiled code
1458 Here we describe the internal form your code is converted to by
1459 Perl. Start with a simple example:
1463 This is converted to a tree similar to this one:
1471 (but slightly more complicated). This tree reflects the way Perl
1472 parsed your code, but has nothing to do with the execution order.
1473 There is an additional "thread" going through the nodes of the tree
1474 which shows the order of execution of the nodes. In our simplified
1475 example above it looks like:
1477 $b ---> $c ---> + ---> $a ---> assign-to
1479 But with the actual compile tree for C<$a = $b + $c> it is different:
1480 some nodes I<optimized away>. As a corollary, though the actual tree
1481 contains more nodes than our simplified example, the execution order
1482 is the same as in our example.
1484 =head2 Examining the tree
1486 If you have your perl compiled for debugging (usually done with C<-D
1487 optimize=-g> on C<Configure> command line), you may examine the
1488 compiled tree by specifying C<-Dx> on the Perl command line. The
1489 output takes several lines per node, and for C<$b+$c> it looks like
1494 FLAGS = (SCALAR,KIDS)
1496 TYPE = null ===> (4)
1498 FLAGS = (SCALAR,KIDS)
1500 3 TYPE = gvsv ===> 4
1506 TYPE = null ===> (5)
1508 FLAGS = (SCALAR,KIDS)
1510 4 TYPE = gvsv ===> 5
1516 This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
1517 not optimized away (one per number in the left column). The immediate
1518 children of the given node correspond to C<{}> pairs on the same level
1519 of indentation, thus this listing corresponds to the tree:
1527 The execution order is indicated by C<===E<gt>> marks, thus it is C<3
1528 4 5 6> (node C<6> is not included into above listing), i.e.,
1529 C<gvsv gvsv add whatever>.
1531 Each of these nodes represents an op, a fundamental operation inside the
1532 Perl core. The code which implements each operation can be found in the
1533 F<pp*.c> files; the function which implements the op with type C<gvsv>
1534 is C<pp_gvsv>, and so on. As the tree above shows, different ops have
1535 different numbers of children: C<add> is a binary operator, as one would
1536 expect, and so has two children. To accommodate the various different
1537 numbers of children, there are various types of op data structure, and
1538 they link together in different ways.
1540 The simplest type of op structure is C<OP>: this has no children. Unary
1541 operators, C<UNOP>s, have one child, and this is pointed to by the
1542 C<op_first> field. Binary operators (C<BINOP>s) have not only an
1543 C<op_first> field but also an C<op_last> field. The most complex type of
1544 op is a C<LISTOP>, which has any number of children. In this case, the
1545 first child is pointed to by C<op_first> and the last child by
1546 C<op_last>. The children in between can be found by iteratively
1547 following the C<op_sibling> pointer from the first child to the last.
1549 There are also two other op types: a C<PMOP> holds a regular expression,
1550 and has no children, and a C<LOOP> may or may not have children. If the
1551 C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
1552 complicate matters, if a C<UNOP> is actually a C<null> op after
1553 optimization (see L</Compile pass 2: context propagation>) it will still
1554 have children in accordance with its former type.
1556 =head2 Compile pass 1: check routines
1558 The tree is created by the compiler while I<yacc> code feeds it
1559 the constructions it recognizes. Since I<yacc> works bottom-up, so does
1560 the first pass of perl compilation.
1562 What makes this pass interesting for perl developers is that some
1563 optimization may be performed on this pass. This is optimization by
1564 so-called "check routines". The correspondence between node names
1565 and corresponding check routines is described in F<opcode.pl> (do not
1566 forget to run C<make regen_headers> if you modify this file).
1568 A check routine is called when the node is fully constructed except
1569 for the execution-order thread. Since at this time there are no
1570 back-links to the currently constructed node, one can do most any
1571 operation to the top-level node, including freeing it and/or creating
1572 new nodes above/below it.
1574 The check routine returns the node which should be inserted into the
1575 tree (if the top-level node was not modified, check routine returns
1578 By convention, check routines have names C<ck_*>. They are usually
1579 called from C<new*OP> subroutines (or C<convert>) (which in turn are
1580 called from F<perly.y>).
1582 =head2 Compile pass 1a: constant folding
1584 Immediately after the check routine is called the returned node is
1585 checked for being compile-time executable. If it is (the value is
1586 judged to be constant) it is immediately executed, and a I<constant>
1587 node with the "return value" of the corresponding subtree is
1588 substituted instead. The subtree is deleted.
1590 If constant folding was not performed, the execution-order thread is
1593 =head2 Compile pass 2: context propagation
1595 When a context for a part of compile tree is known, it is propagated
1596 down through the tree. At this time the context can have 5 values
1597 (instead of 2 for runtime context): void, boolean, scalar, list, and
1598 lvalue. In contrast with the pass 1 this pass is processed from top
1599 to bottom: a node's context determines the context for its children.
1601 Additional context-dependent optimizations are performed at this time.
1602 Since at this moment the compile tree contains back-references (via
1603 "thread" pointers), nodes cannot be free()d now. To allow
1604 optimized-away nodes at this stage, such nodes are null()ified instead
1605 of free()ing (i.e. their type is changed to OP_NULL).
1607 =head2 Compile pass 3: peephole optimization
1609 After the compile tree for a subroutine (or for an C<eval> or a file)
1610 is created, an additional pass over the code is performed. This pass
1611 is neither top-down or bottom-up, but in the execution order (with
1612 additional complications for conditionals). These optimizations are
1613 done in the subroutine peep(). Optimizations performed at this stage
1614 are subject to the same restrictions as in the pass 2.
1616 =head1 Examining internal data structures with the C<dump> functions
1618 To aid debugging, the source file F<dump.c> contains a number of
1619 functions which produce formatted output of internal data structures.
1621 The most commonly used of these functions is C<Perl_sv_dump>; it's used
1622 for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
1623 C<sv_dump> to produce debugging output from Perl-space, so users of that
1624 module should already be familiar with its format.
1626 C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
1627 derivatives, and produces output similiar to C<perl -Dx>; in fact,
1628 C<Perl_dump_eval> will dump the main root of the code being evaluated,
1629 exactly like C<-Dx>.
1631 Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
1632 op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
1633 subroutines in a package like so: (Thankfully, these are all xsubs, so
1634 there is no op tree)
1636 (gdb) print Perl_dump_packsubs(PL_defstash)
1638 SUB attributes::bootstrap = (xsub 0x811fedc 0)
1640 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
1642 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
1644 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
1646 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
1648 and C<Perl_dump_all>, which dumps all the subroutines in the stash and
1649 the op tree of the main root.
1651 =head1 How multiple interpreters and concurrency are supported
1653 =head2 Background and PERL_IMPLICIT_CONTEXT
1655 The Perl interpreter can be regarded as a closed box: it has an API
1656 for feeding it code or otherwise making it do things, but it also has
1657 functions for its own use. This smells a lot like an object, and
1658 there are ways for you to build Perl so that you can have multiple
1659 interpreters, with one interpreter represented either as a C++ object,
1660 a C structure, or inside a thread. The thread, the C structure, or
1661 the C++ object will contain all the context, the state of that
1664 Three macros control the major Perl build flavors: MULTIPLICITY,
1665 USE_THREADS and PERL_OBJECT. The MULTIPLICITY build has a C structure
1666 that packages all the interpreter state, there is a similar thread-specific
1667 data structure under USE_THREADS, and the (now deprecated) PERL_OBJECT
1668 build has a C++ class to maintain interpreter state. In all three cases,
1669 PERL_IMPLICIT_CONTEXT is also normally defined, and enables the
1670 support for passing in a "hidden" first argument that represents all three
1673 All this obviously requires a way for the Perl internal functions to be
1674 C++ methods, subroutines taking some kind of structure as the first
1675 argument, or subroutines taking nothing as the first argument. To
1676 enable these three very different ways of building the interpreter,
1677 the Perl source (as it does in so many other situations) makes heavy
1678 use of macros and subroutine naming conventions.
1680 First problem: deciding which functions will be public API functions and
1681 which will be private. All functions whose names begin C<S_> are private
1682 (think "S" for "secret" or "static"). All other functions begin with
1683 "Perl_", but just because a function begins with "Perl_" does not mean it is
1684 part of the API. (See L</Internal Functions>.) The easiest way to be B<sure> a
1685 function is part of the API is to find its entry in L<perlapi>.
1686 If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
1687 think it should be (i.e., you need it for your extension), send mail via
1688 L<perlbug> explaining why you think it should be.
1690 Second problem: there must be a syntax so that the same subroutine
1691 declarations and calls can pass a structure as their first argument,
1692 or pass nothing. To solve this, the subroutines are named and
1693 declared in a particular way. Here's a typical start of a static
1694 function used within the Perl guts:
1697 S_incline(pTHX_ char *s)
1699 STATIC becomes "static" in C, and is #define'd to nothing in C++.
1701 A public function (i.e. part of the internal API, but not necessarily
1702 sanctioned for use in extensions) begins like this:
1705 Perl_sv_setsv(pTHX_ SV* dsv, SV* ssv)
1707 C<pTHX_> is one of a number of macros (in perl.h) that hide the
1708 details of the interpreter's context. THX stands for "thread", "this",
1709 or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
1710 The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
1711 or 'd' for B<d>eclaration, so we have C<pTHX>, C<aTHX> and C<dTHX>, and
1714 When Perl is built without options that set PERL_IMPLICIT_CONTEXT, there is no
1715 first argument containing the interpreter's context. The trailing underscore
1716 in the pTHX_ macro indicates that the macro expansion needs a comma
1717 after the context argument because other arguments follow it. If
1718 PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
1719 subroutine is not prototyped to take the extra argument. The form of the
1720 macro without the trailing underscore is used when there are no additional
1723 When a core function calls another, it must pass the context. This
1724 is normally hidden via macros. Consider C<sv_setsv>. It expands into
1725 something like this:
1727 ifdef PERL_IMPLICIT_CONTEXT
1728 define sv_setsv(a,b) Perl_sv_setsv(aTHX_ a, b)
1729 /* can't do this for vararg functions, see below */
1731 define sv_setsv Perl_sv_setsv
1734 This works well, and means that XS authors can gleefully write:
1738 and still have it work under all the modes Perl could have been
1741 Under PERL_OBJECT in the core, that will translate to either:
1743 CPerlObj::Perl_sv_setsv(foo,bar); # in CPerlObj functions,
1744 # C++ takes care of 'this'
1747 pPerl->Perl_sv_setsv(foo,bar); # in truly static functions,
1750 Under PERL_OBJECT in extensions (aka PERL_CAPI), or under
1751 MULTIPLICITY/USE_THREADS with PERL_IMPLICIT_CONTEXT in both core
1752 and extensions, it will become:
1754 Perl_sv_setsv(aTHX_ foo, bar); # the canonical Perl "API"
1755 # for all build flavors
1757 This doesn't work so cleanly for varargs functions, though, as macros
1758 imply that the number of arguments is known in advance. Instead we
1759 either need to spell them out fully, passing C<aTHX_> as the first
1760 argument (the Perl core tends to do this with functions like
1761 Perl_warner), or use a context-free version.
1763 The context-free version of Perl_warner is called
1764 Perl_warner_nocontext, and does not take the extra argument. Instead
1765 it does dTHX; to get the context from thread-local storage. We
1766 C<#define warner Perl_warner_nocontext> so that extensions get source
1767 compatibility at the expense of performance. (Passing an arg is
1768 cheaper than grabbing it from thread-local storage.)
1770 You can ignore [pad]THX[xo] when browsing the Perl headers/sources.
1771 Those are strictly for use within the core. Extensions and embedders
1772 need only be aware of [pad]THX.
1774 =head2 So what happened to dTHR?
1776 C<dTHR> was introduced in perl 5.005 to support the older thread model.
1777 The older thread model now uses the C<THX> mechanism to pass context
1778 pointers around, so C<dTHR> is not useful any more. Perl 5.6.0 and
1779 later still have it for backward source compatibility, but it is defined
1782 =head2 How do I use all this in extensions?
1784 When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
1785 any functions in the Perl API will need to pass the initial context
1786 argument somehow. The kicker is that you will need to write it in
1787 such a way that the extension still compiles when Perl hasn't been
1788 built with PERL_IMPLICIT_CONTEXT enabled.
1790 There are three ways to do this. First, the easy but inefficient way,
1791 which is also the default, in order to maintain source compatibility
1792 with extensions: whenever XSUB.h is #included, it redefines the aTHX
1793 and aTHX_ macros to call a function that will return the context.
1794 Thus, something like:
1798 in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
1801 Perl_sv_setsv(Perl_get_context(), asv, bsv);
1803 or to this otherwise:
1805 Perl_sv_setsv(asv, bsv);
1807 You have to do nothing new in your extension to get this; since
1808 the Perl library provides Perl_get_context(), it will all just
1811 The second, more efficient way is to use the following template for
1814 #define PERL_NO_GET_CONTEXT /* we want efficiency */
1819 static my_private_function(int arg1, int arg2);
1822 my_private_function(int arg1, int arg2)
1824 dTHX; /* fetch context */
1825 ... call many Perl API functions ...
1830 MODULE = Foo PACKAGE = Foo
1838 my_private_function(arg, 10);
1840 Note that the only two changes from the normal way of writing an
1841 extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
1842 including the Perl headers, followed by a C<dTHX;> declaration at
1843 the start of every function that will call the Perl API. (You'll
1844 know which functions need this, because the C compiler will complain
1845 that there's an undeclared identifier in those functions.) No changes
1846 are needed for the XSUBs themselves, because the XS() macro is
1847 correctly defined to pass in the implicit context if needed.
1849 The third, even more efficient way is to ape how it is done within
1853 #define PERL_NO_GET_CONTEXT /* we want efficiency */
1858 /* pTHX_ only needed for functions that call Perl API */
1859 static my_private_function(pTHX_ int arg1, int arg2);
1862 my_private_function(pTHX_ int arg1, int arg2)
1864 /* dTHX; not needed here, because THX is an argument */
1865 ... call Perl API functions ...
1870 MODULE = Foo PACKAGE = Foo
1878 my_private_function(aTHX_ arg, 10);
1880 This implementation never has to fetch the context using a function
1881 call, since it is always passed as an extra argument. Depending on
1882 your needs for simplicity or efficiency, you may mix the previous
1883 two approaches freely.
1885 Never add a comma after C<pTHX> yourself--always use the form of the
1886 macro with the underscore for functions that take explicit arguments,
1887 or the form without the argument for functions with no explicit arguments.
1889 =head2 Should I do anything special if I call perl from multiple threads?
1891 If you create interpreters in one thread and then proceed to call them in
1892 another, you need to make sure perl's own Thread Local Storage (TLS) slot is
1893 initialized correctly in each of those threads.
1895 The C<perl_alloc> and C<perl_clone> API functions will automatically set
1896 the TLS slot to the interpreter they created, so that there is no need to do
1897 anything special if the interpreter is always accessed in the same thread that
1898 created it, and that thread did not create or call any other interpreters
1899 afterwards. If that is not the case, you have to set the TLS slot of the
1900 thread before calling any functions in the Perl API on that particular
1901 interpreter. This is done by calling the C<PERL_SET_CONTEXT> macro in that
1902 thread as the first thing you do:
1904 /* do this before doing anything else with some_perl */
1905 PERL_SET_CONTEXT(some_perl);
1907 ... other Perl API calls on some_perl go here ...
1909 =head2 Future Plans and PERL_IMPLICIT_SYS
1911 Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
1912 that the interpreter knows about itself and pass it around, so too are
1913 there plans to allow the interpreter to bundle up everything it knows
1914 about the environment it's running on. This is enabled with the
1915 PERL_IMPLICIT_SYS macro. Currently it only works with PERL_OBJECT
1916 and USE_THREADS on Windows (see inside iperlsys.h).
1918 This allows the ability to provide an extra pointer (called the "host"
1919 environment) for all the system calls. This makes it possible for
1920 all the system stuff to maintain their own state, broken down into
1921 seven C structures. These are thin wrappers around the usual system
1922 calls (see win32/perllib.c) for the default perl executable, but for a
1923 more ambitious host (like the one that would do fork() emulation) all
1924 the extra work needed to pretend that different interpreters are
1925 actually different "processes", would be done here.
1927 The Perl engine/interpreter and the host are orthogonal entities.
1928 There could be one or more interpreters in a process, and one or
1929 more "hosts", with free association between them.
1931 =head1 Internal Functions
1933 All of Perl's internal functions which will be exposed to the outside
1934 world are be prefixed by C<Perl_> so that they will not conflict with XS
1935 functions or functions used in a program in which Perl is embedded.
1936 Similarly, all global variables begin with C<PL_>. (By convention,
1937 static functions start with C<S_>)
1939 Inside the Perl core, you can get at the functions either with or
1940 without the C<Perl_> prefix, thanks to a bunch of defines that live in
1941 F<embed.h>. This header file is generated automatically from
1942 F<embed.pl>. F<embed.pl> also creates the prototyping header files for
1943 the internal functions, generates the documentation and a lot of other
1944 bits and pieces. It's important that when you add a new function to the
1945 core or change an existing one, you change the data in the table at the
1946 end of F<embed.pl> as well. Here's a sample entry from that table:
1948 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
1950 The second column is the return type, the third column the name. Columns
1951 after that are the arguments. The first column is a set of flags:
1957 This function is a part of the public API.
1961 This function has a C<Perl_> prefix; ie, it is defined as C<Perl_av_fetch>
1965 This function has documentation using the C<apidoc> feature which we'll
1966 look at in a second.
1970 Other available flags are:
1976 This is a static function and is defined as C<S_whatever>, and usually
1977 called within the sources as C<whatever(...)>.
1981 This does not use C<aTHX_> and C<pTHX> to pass interpreter context. (See
1982 L<perlguts/Background and PERL_IMPLICIT_CONTEXT>.)
1986 This function never returns; C<croak>, C<exit> and friends.
1990 This function takes a variable number of arguments, C<printf> style.
1991 The argument list should end with C<...>, like this:
1993 Afprd |void |croak |const char* pat|...
1997 This function is part of the experimental development API, and may change
1998 or disappear without notice.
2002 This function should not have a compatibility macro to define, say,
2003 C<Perl_parse> to C<parse>. It must be called as C<Perl_parse>.
2007 This function is not a member of C<CPerlObj>. If you don't know
2008 what this means, don't use it.
2012 This function isn't exported out of the Perl core.
2016 If you edit F<embed.pl>, you will need to run C<make regen_headers> to
2017 force a rebuild of F<embed.h> and other auto-generated files.
2019 =head2 Formatted Printing of IVs, UVs, and NVs
2021 If you are printing IVs, UVs, or NVS instead of the stdio(3) style
2022 formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
2023 following macros for portability
2028 UVxf UV in hexadecimal
2033 These will take care of 64-bit integers and long doubles.
2036 printf("IV is %"IVdf"\n", iv);
2038 The IVdf will expand to whatever is the correct format for the IVs.
2040 If you are printing addresses of pointers, use UVxf combined
2041 with PTR2UV(), do not use %lx or %p.
2043 =head2 Pointer-To-Integer and Integer-To-Pointer
2045 Because pointer size does not necessarily equal integer size,
2046 use the follow macros to do it right.
2051 INT2PTR(pointertotype, integer)
2056 SV *sv = INT2PTR(SV*, iv);
2063 =head2 Source Documentation
2065 There's an effort going on to document the internal functions and
2066 automatically produce reference manuals from them - L<perlapi> is one
2067 such manual which details all the functions which are available to XS
2068 writers. L<perlintern> is the autogenerated manual for the functions
2069 which are not part of the API and are supposedly for internal use only.
2071 Source documentation is created by putting POD comments into the C
2075 =for apidoc sv_setiv
2077 Copies an integer into the given SV. Does not handle 'set' magic. See
2083 Please try and supply some documentation if you add functions to the
2086 =head1 Unicode Support
2088 Perl 5.6.0 introduced Unicode support. It's important for porters and XS
2089 writers to understand this support and make sure that the code they
2090 write does not corrupt Unicode data.
2092 =head2 What B<is> Unicode, anyway?
2094 In the olden, less enlightened times, we all used to use ASCII. Most of
2095 us did, anyway. The big problem with ASCII is that it's American. Well,
2096 no, that's not actually the problem; the problem is that it's not
2097 particularly useful for people who don't use the Roman alphabet. What
2098 used to happen was that particular languages would stick their own
2099 alphabet in the upper range of the sequence, between 128 and 255. Of
2100 course, we then ended up with plenty of variants that weren't quite
2101 ASCII, and the whole point of it being a standard was lost.
2103 Worse still, if you've got a language like Chinese or
2104 Japanese that has hundreds or thousands of characters, then you really
2105 can't fit them into a mere 256, so they had to forget about ASCII
2106 altogether, and build their own systems using pairs of numbers to refer
2109 To fix this, some people formed Unicode, Inc. and
2110 produced a new character set containing all the characters you can
2111 possibly think of and more. There are several ways of representing these
2112 characters, and the one Perl uses is called UTF8. UTF8 uses
2113 a variable number of bytes to represent a character, instead of just
2114 one. You can learn more about Unicode at http://www.unicode.org/
2116 =head2 How can I recognise a UTF8 string?
2118 You can't. This is because UTF8 data is stored in bytes just like
2119 non-UTF8 data. The Unicode character 200, (C<0xC8> for you hex types)
2120 capital E with a grave accent, is represented by the two bytes
2121 C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
2122 has that byte sequence as well. So you can't tell just by looking - this
2123 is what makes Unicode input an interesting problem.
2125 The API function C<is_utf8_string> can help; it'll tell you if a string
2126 contains only valid UTF8 characters. However, it can't do the work for
2127 you. On a character-by-character basis, C<is_utf8_char> will tell you
2128 whether the current character in a string is valid UTF8.
2130 =head2 How does UTF8 represent Unicode characters?
2132 As mentioned above, UTF8 uses a variable number of bytes to store a
2133 character. Characters with values 1...128 are stored in one byte, just
2134 like good ol' ASCII. Character 129 is stored as C<v194.129>; this
2135 continues up to character 191, which is C<v194.191>. Now we've run out of
2136 bits (191 is binary C<10111111>) so we move on; 192 is C<v195.128>. And
2137 so it goes on, moving to three bytes at character 2048.
2139 Assuming you know you're dealing with a UTF8 string, you can find out
2140 how long the first character in it is with the C<UTF8SKIP> macro:
2142 char *utf = "\305\233\340\240\201";
2145 len = UTF8SKIP(utf); /* len is 2 here */
2147 len = UTF8SKIP(utf); /* len is 3 here */
2149 Another way to skip over characters in a UTF8 string is to use
2150 C<utf8_hop>, which takes a string and a number of characters to skip
2151 over. You're on your own about bounds checking, though, so don't use it
2154 All bytes in a multi-byte UTF8 character will have the high bit set, so
2155 you can test if you need to do something special with this character
2161 /* Must treat this as UTF8 */
2162 uv = utf8_to_uv(utf);
2164 /* OK to treat this character as a byte */
2167 You can also see in that example that we use C<utf8_to_uv> to get the
2168 value of the character; the inverse function C<uv_to_utf8> is available
2169 for putting a UV into UTF8:
2172 /* Must treat this as UTF8 */
2173 utf8 = uv_to_utf8(utf8, uv);
2175 /* OK to treat this character as a byte */
2178 You B<must> convert characters to UVs using the above functions if
2179 you're ever in a situation where you have to match UTF8 and non-UTF8
2180 characters. You may not skip over UTF8 characters in this case. If you
2181 do this, you'll lose the ability to match hi-bit non-UTF8 characters;
2182 for instance, if your UTF8 string contains C<v196.172>, and you skip
2183 that character, you can never match a C<chr(200)> in a non-UTF8 string.
2186 =head2 How does Perl store UTF8 strings?
2188 Currently, Perl deals with Unicode strings and non-Unicode strings
2189 slightly differently. If a string has been identified as being UTF-8
2190 encoded, Perl will set a flag in the SV, C<SVf_UTF8>. You can check and
2191 manipulate this flag with the following macros:
2197 This flag has an important effect on Perl's treatment of the string: if
2198 Unicode data is not properly distinguished, regular expressions,
2199 C<length>, C<substr> and other string handling operations will have
2200 undesirable results.
2202 The problem comes when you have, for instance, a string that isn't
2203 flagged is UTF8, and contains a byte sequence that could be UTF8 -
2204 especially when combining non-UTF8 and UTF8 strings.
2206 Never forget that the C<SVf_UTF8> flag is separate to the PV value; you
2207 need be sure you don't accidentally knock it off while you're
2208 manipulating SVs. More specifically, you cannot expect to do this:
2217 nsv = newSVpvn(p, len);
2219 The C<char*> string does not tell you the whole story, and you can't
2220 copy or reconstruct an SV just by copying the string value. Check if the
2221 old SV has the UTF8 flag set, and act accordingly:
2225 nsv = newSVpvn(p, len);
2229 In fact, your C<frobnicate> function should be made aware of whether or
2230 not it's dealing with UTF8 data, so that it can handle the string
2233 =head2 How do I convert a string to UTF8?
2235 If you're mixing UTF8 and non-UTF8 strings, you might find it necessary
2236 to upgrade one of the strings to UTF8. If you've got an SV, the easiest
2239 sv_utf8_upgrade(sv);
2241 However, you must not do this, for example:
2244 sv_utf8_upgrade(left);
2246 If you do this in a binary operator, you will actually change one of the
2247 strings that came into the operator, and, while it shouldn't be noticeable
2248 by the end user, it can cause problems.
2250 Instead, C<bytes_to_utf8> will give you a UTF8-encoded B<copy> of its
2251 string argument. This is useful for having the data available for
2252 comparisons and so on, without harming the original SV. There's also
2253 C<utf8_to_bytes> to go the other way, but naturally, this will fail if
2254 the string contains any characters above 255 that can't be represented
2257 =head2 Is there anything else I need to know?
2259 Not really. Just remember these things:
2265 There's no way to tell if a string is UTF8 or not. You can tell if an SV
2266 is UTF8 by looking at is C<SvUTF8> flag. Don't forget to set the flag if
2267 something should be UTF8. Treat the flag as part of the PV, even though
2268 it's not - if you pass on the PV to somewhere, pass on the flag too.
2272 If a string is UTF8, B<always> use C<utf8_to_uv> to get at the value,
2273 unless C<!(*s & 0x80)> in which case you can use C<*s>.
2277 When writing to a UTF8 string, B<always> use C<uv_to_utf8>, unless
2278 C<uv < 0x80> in which case you can use C<*s = uv>.
2282 Mixing UTF8 and non-UTF8 strings is tricky. Use C<bytes_to_utf8> to get
2283 a new string which is UTF8 encoded. There are tricks you can use to
2284 delay deciding whether you need to use a UTF8 string until you get to a
2285 high character - C<HALF_UPGRADE> is one of those.
2291 Until May 1997, this document was maintained by Jeff Okamoto
2292 <okamoto@corp.hp.com>. It is now maintained as part of Perl itself
2293 by the Perl 5 Porters <perl5-porters@perl.org>.
2295 With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
2296 Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
2297 Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
2298 Stephen McCamant, and Gurusamy Sarathy.
2300 API Listing originally by Dean Roehrich <roehrich@cray.com>.
2302 Modifications to autogenerate the API listing (L<perlapi>) by Benjamin
2307 perlapi(1), perlintern(1), perlxs(1), perlembed(1)