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 =head2 What's Really Stored in an SV?
252 Recall that the usual method of determining the type of scalar you have is
253 to use C<Sv*OK> macros. Because a scalar can be both a number and a string,
254 usually these macros will always return TRUE and calling the C<Sv*V>
255 macros will do the appropriate conversion of string to integer/double or
256 integer/double to string.
258 If you I<really> need to know if you have an integer, double, or string
259 pointer in an SV, you can use the following three macros instead:
265 These will tell you if you truly have an integer, double, or string pointer
266 stored in your SV. The "p" stands for private.
268 In general, though, it's best to use the C<Sv*V> macros.
270 =head2 Working with AVs
272 There are two ways to create and load an AV. The first method creates an
277 The second method both creates the AV and initially populates it with SVs:
279 AV* av_make(I32 num, SV **ptr);
281 The second argument points to an array containing C<num> C<SV*>'s. Once the
282 AV has been created, the SVs can be destroyed, if so desired.
284 Once the AV has been created, the following operations are possible on AVs:
286 void av_push(AV*, SV*);
289 void av_unshift(AV*, I32 num);
291 These should be familiar operations, with the exception of C<av_unshift>.
292 This routine adds C<num> elements at the front of the array with the C<undef>
293 value. You must then use C<av_store> (described below) to assign values
294 to these new elements.
296 Here are some other functions:
299 SV** av_fetch(AV*, I32 key, I32 lval);
300 SV** av_store(AV*, I32 key, SV* val);
302 The C<av_len> function returns the highest index value in array (just
303 like $#array in Perl). If the array is empty, -1 is returned. The
304 C<av_fetch> function returns the value at index C<key>, but if C<lval>
305 is non-zero, then C<av_fetch> will store an undef value at that index.
306 The C<av_store> function stores the value C<val> at index C<key>, and does
307 not increment the reference count of C<val>. Thus the caller is responsible
308 for taking care of that, and if C<av_store> returns NULL, the caller will
309 have to decrement the reference count to avoid a memory leak. Note that
310 C<av_fetch> and C<av_store> both return C<SV**>'s, not C<SV*>'s as their
315 void av_extend(AV*, I32 key);
317 The C<av_clear> function deletes all the elements in the AV* array, but
318 does not actually delete the array itself. The C<av_undef> function will
319 delete all the elements in the array plus the array itself. The
320 C<av_extend> function extends the array so that it contains at least C<key+1>
321 elements. If C<key+1> is less than the currently allocated length of the array,
322 then nothing is done.
324 If you know the name of an array variable, you can get a pointer to its AV
325 by using the following:
327 AV* get_av("package::varname", FALSE);
329 This returns NULL if the variable does not exist.
331 See L<Understanding the Magic of Tied Hashes and Arrays> for more
332 information on how to use the array access functions on tied arrays.
334 =head2 Working with HVs
336 To create an HV, you use the following routine:
340 Once the HV has been created, the following operations are possible on HVs:
342 SV** hv_store(HV*, const char* key, U32 klen, SV* val, U32 hash);
343 SV** hv_fetch(HV*, const char* key, U32 klen, I32 lval);
345 The C<klen> parameter is the length of the key being passed in (Note that
346 you cannot pass 0 in as a value of C<klen> to tell Perl to measure the
347 length of the key). The C<val> argument contains the SV pointer to the
348 scalar being stored, and C<hash> is the precomputed hash value (zero if
349 you want C<hv_store> to calculate it for you). The C<lval> parameter
350 indicates whether this fetch is actually a part of a store operation, in
351 which case a new undefined value will be added to the HV with the supplied
352 key and C<hv_fetch> will return as if the value had already existed.
354 Remember that C<hv_store> and C<hv_fetch> return C<SV**>'s and not just
355 C<SV*>. To access the scalar value, you must first dereference the return
356 value. However, you should check to make sure that the return value is
357 not NULL before dereferencing it.
359 These two functions check if a hash table entry exists, and deletes it.
361 bool hv_exists(HV*, const char* key, U32 klen);
362 SV* hv_delete(HV*, const char* key, U32 klen, I32 flags);
364 If C<flags> does not include the C<G_DISCARD> flag then C<hv_delete> will
365 create and return a mortal copy of the deleted value.
367 And more miscellaneous functions:
372 Like their AV counterparts, C<hv_clear> deletes all the entries in the hash
373 table but does not actually delete the hash table. The C<hv_undef> deletes
374 both the entries and the hash table itself.
376 Perl keeps the actual data in linked list of structures with a typedef of HE.
377 These contain the actual key and value pointers (plus extra administrative
378 overhead). The key is a string pointer; the value is an C<SV*>. However,
379 once you have an C<HE*>, to get the actual key and value, use the routines
382 I32 hv_iterinit(HV*);
383 /* Prepares starting point to traverse hash table */
384 HE* hv_iternext(HV*);
385 /* Get the next entry, and return a pointer to a
386 structure that has both the key and value */
387 char* hv_iterkey(HE* entry, I32* retlen);
388 /* Get the key from an HE structure and also return
389 the length of the key string */
390 SV* hv_iterval(HV*, HE* entry);
391 /* Return a SV pointer to the value of the HE
393 SV* hv_iternextsv(HV*, char** key, I32* retlen);
394 /* This convenience routine combines hv_iternext,
395 hv_iterkey, and hv_iterval. The key and retlen
396 arguments are return values for the key and its
397 length. The value is returned in the SV* argument */
399 If you know the name of a hash variable, you can get a pointer to its HV
400 by using the following:
402 HV* get_hv("package::varname", FALSE);
404 This returns NULL if the variable does not exist.
406 The hash algorithm is defined in the C<PERL_HASH(hash, key, klen)> macro:
410 hash = (hash * 33) + *key++;
411 hash = hash + (hash >> 5); /* after 5.6 */
413 The last step was added in version 5.6 to improve distribution of
414 lower bits in the resulting hash value.
416 See L<Understanding the Magic of Tied Hashes and Arrays> for more
417 information on how to use the hash access functions on tied hashes.
419 =head2 Hash API Extensions
421 Beginning with version 5.004, the following functions are also supported:
423 HE* hv_fetch_ent (HV* tb, SV* key, I32 lval, U32 hash);
424 HE* hv_store_ent (HV* tb, SV* key, SV* val, U32 hash);
426 bool hv_exists_ent (HV* tb, SV* key, U32 hash);
427 SV* hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
429 SV* hv_iterkeysv (HE* entry);
431 Note that these functions take C<SV*> keys, which simplifies writing
432 of extension code that deals with hash structures. These functions
433 also allow passing of C<SV*> keys to C<tie> functions without forcing
434 you to stringify the keys (unlike the previous set of functions).
436 They also return and accept whole hash entries (C<HE*>), making their
437 use more efficient (since the hash number for a particular string
438 doesn't have to be recomputed every time). See L<perlapi> for detailed
441 The following macros must always be used to access the contents of hash
442 entries. Note that the arguments to these macros must be simple
443 variables, since they may get evaluated more than once. See
444 L<perlapi> for detailed descriptions of these macros.
446 HePV(HE* he, STRLEN len)
450 HeSVKEY_force(HE* he)
451 HeSVKEY_set(HE* he, SV* sv)
453 These two lower level macros are defined, but must only be used when
454 dealing with keys that are not C<SV*>s:
459 Note that both C<hv_store> and C<hv_store_ent> do not increment the
460 reference count of the stored C<val>, which is the caller's responsibility.
461 If these functions return a NULL value, the caller will usually have to
462 decrement the reference count of C<val> to avoid a memory leak.
466 References are a special type of scalar that point to other data types
467 (including references).
469 To create a reference, use either of the following functions:
471 SV* newRV_inc((SV*) thing);
472 SV* newRV_noinc((SV*) thing);
474 The C<thing> argument can be any of an C<SV*>, C<AV*>, or C<HV*>. The
475 functions are identical except that C<newRV_inc> increments the reference
476 count of the C<thing>, while C<newRV_noinc> does not. For historical
477 reasons, C<newRV> is a synonym for C<newRV_inc>.
479 Once you have a reference, you can use the following macro to dereference
484 then call the appropriate routines, casting the returned C<SV*> to either an
485 C<AV*> or C<HV*>, if required.
487 To determine if an SV is a reference, you can use the following macro:
491 To discover what type of value the reference refers to, use the following
492 macro and then check the return value.
496 The most useful types that will be returned are:
505 SVt_PVGV Glob (possible a file handle)
506 SVt_PVMG Blessed or Magical Scalar
508 See the sv.h header file for more details.
510 =head2 Blessed References and Class Objects
512 References are also used to support object-oriented programming. In the
513 OO lexicon, an object is simply a reference that has been blessed into a
514 package (or class). Once blessed, the programmer may now use the reference
515 to access the various methods in the class.
517 A reference can be blessed into a package with the following function:
519 SV* sv_bless(SV* sv, HV* stash);
521 The C<sv> argument must be a reference. The C<stash> argument specifies
522 which class the reference will belong to. See
523 L<Stashes and Globs> for information on converting class names into stashes.
525 /* Still under construction */
527 Upgrades rv to reference if not already one. Creates new SV for rv to
528 point to. If C<classname> is non-null, the SV is blessed into the specified
529 class. SV is returned.
531 SV* newSVrv(SV* rv, const char* classname);
533 Copies integer or double into an SV whose reference is C<rv>. SV is blessed
534 if C<classname> is non-null.
536 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
537 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
539 Copies the pointer value (I<the address, not the string!>) into an SV whose
540 reference is rv. SV is blessed if C<classname> is non-null.
542 SV* sv_setref_pv(SV* rv, const char* classname, PV iv);
544 Copies string into an SV whose reference is C<rv>. Set length to 0 to let
545 Perl calculate the string length. SV is blessed if C<classname> is non-null.
547 SV* sv_setref_pvn(SV* rv, const char* classname, PV iv, STRLEN length);
549 Tests whether the SV is blessed into the specified class. It does not
550 check inheritance relationships.
552 int sv_isa(SV* sv, const char* name);
554 Tests whether the SV is a reference to a blessed object.
556 int sv_isobject(SV* sv);
558 Tests whether the SV is derived from the specified class. SV can be either
559 a reference to a blessed object or a string containing a class name. This
560 is the function implementing the C<UNIVERSAL::isa> functionality.
562 bool sv_derived_from(SV* sv, const char* name);
564 To check if you've got an object derived from a specific class you have
567 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
569 =head2 Creating New Variables
571 To create a new Perl variable with an undef value which can be accessed from
572 your Perl script, use the following routines, depending on the variable type.
574 SV* get_sv("package::varname", TRUE);
575 AV* get_av("package::varname", TRUE);
576 HV* get_hv("package::varname", TRUE);
578 Notice the use of TRUE as the second parameter. The new variable can now
579 be set, using the routines appropriate to the data type.
581 There are additional macros whose values may be bitwise OR'ed with the
582 C<TRUE> argument to enable certain extra features. Those bits are:
584 GV_ADDMULTI Marks the variable as multiply defined, thus preventing the
585 "Name <varname> used only once: possible typo" warning.
586 GV_ADDWARN Issues the warning "Had to create <varname> unexpectedly" if
587 the variable did not exist before the function was called.
589 If you do not specify a package name, the variable is created in the current
592 =head2 Reference Counts and Mortality
594 Perl uses an reference count-driven garbage collection mechanism. SVs,
595 AVs, or HVs (xV for short in the following) start their life with a
596 reference count of 1. If the reference count of an xV ever drops to 0,
597 then it will be destroyed and its memory made available for reuse.
599 This normally doesn't happen at the Perl level unless a variable is
600 undef'ed or the last variable holding a reference to it is changed or
601 overwritten. At the internal level, however, reference counts can be
602 manipulated with the following macros:
604 int SvREFCNT(SV* sv);
605 SV* SvREFCNT_inc(SV* sv);
606 void SvREFCNT_dec(SV* sv);
608 However, there is one other function which manipulates the reference
609 count of its argument. The C<newRV_inc> function, you will recall,
610 creates a reference to the specified argument. As a side effect,
611 it increments the argument's reference count. If this is not what
612 you want, use C<newRV_noinc> instead.
614 For example, imagine you want to return a reference from an XSUB function.
615 Inside the XSUB routine, you create an SV which initially has a reference
616 count of one. Then you call C<newRV_inc>, passing it the just-created SV.
617 This returns the reference as a new SV, but the reference count of the
618 SV you passed to C<newRV_inc> has been incremented to two. Now you
619 return the reference from the XSUB routine and forget about the SV.
620 But Perl hasn't! Whenever the returned reference is destroyed, the
621 reference count of the original SV is decreased to one and nothing happens.
622 The SV will hang around without any way to access it until Perl itself
623 terminates. This is a memory leak.
625 The correct procedure, then, is to use C<newRV_noinc> instead of
626 C<newRV_inc>. Then, if and when the last reference is destroyed,
627 the reference count of the SV will go to zero and it will be destroyed,
628 stopping any memory leak.
630 There are some convenience functions available that can help with the
631 destruction of xVs. These functions introduce the concept of "mortality".
632 An xV that is mortal has had its reference count marked to be decremented,
633 but not actually decremented, until "a short time later". Generally the
634 term "short time later" means a single Perl statement, such as a call to
635 an XSUB function. The actual determinant for when mortal xVs have their
636 reference count decremented depends on two macros, SAVETMPS and FREETMPS.
637 See L<perlcall> and L<perlxs> for more details on these macros.
639 "Mortalization" then is at its simplest a deferred C<SvREFCNT_dec>.
640 However, if you mortalize a variable twice, the reference count will
641 later be decremented twice.
643 You should be careful about creating mortal variables. Strange things
644 can happen if you make the same value mortal within multiple contexts,
645 or if you make a variable mortal multiple times.
647 To create a mortal variable, use the functions:
651 SV* sv_mortalcopy(SV*)
653 The first call creates a mortal SV, the second converts an existing
654 SV to a mortal SV (and thus defers a call to C<SvREFCNT_dec>), and the
655 third creates a mortal copy of an existing SV.
657 The mortal routines are not just for SVs -- AVs and HVs can be
658 made mortal by passing their address (type-casted to C<SV*>) to the
659 C<sv_2mortal> or C<sv_mortalcopy> routines.
661 =head2 Stashes and Globs
663 A "stash" is a hash that contains all of the different objects that
664 are contained within a package. Each key of the stash is a symbol
665 name (shared by all the different types of objects that have the same
666 name), and each value in the hash table is a GV (Glob Value). This GV
667 in turn contains references to the various objects of that name,
668 including (but not limited to) the following:
677 There is a single stash called "PL_defstash" that holds the items that exist
678 in the "main" package. To get at the items in other packages, append the
679 string "::" to the package name. The items in the "Foo" package are in
680 the stash "Foo::" in PL_defstash. The items in the "Bar::Baz" package are
681 in the stash "Baz::" in "Bar::"'s stash.
683 To get the stash pointer for a particular package, use the function:
685 HV* gv_stashpv(const char* name, I32 create)
686 HV* gv_stashsv(SV*, I32 create)
688 The first function takes a literal string, the second uses the string stored
689 in the SV. Remember that a stash is just a hash table, so you get back an
690 C<HV*>. The C<create> flag will create a new package if it is set.
692 The name that C<gv_stash*v> wants is the name of the package whose symbol table
693 you want. The default package is called C<main>. If you have multiply nested
694 packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
697 Alternately, if you have an SV that is a blessed reference, you can find
698 out the stash pointer by using:
700 HV* SvSTASH(SvRV(SV*));
702 then use the following to get the package name itself:
704 char* HvNAME(HV* stash);
706 If you need to bless or re-bless an object you can use the following
709 SV* sv_bless(SV*, HV* stash)
711 where the first argument, an C<SV*>, must be a reference, and the second
712 argument is a stash. The returned C<SV*> can now be used in the same way
715 For more information on references and blessings, consult L<perlref>.
717 =head2 Double-Typed SVs
719 Scalar variables normally contain only one type of value, an integer,
720 double, pointer, or reference. Perl will automatically convert the
721 actual scalar data from the stored type into the requested type.
723 Some scalar variables contain more than one type of scalar data. For
724 example, the variable C<$!> contains either the numeric value of C<errno>
725 or its string equivalent from either C<strerror> or C<sys_errlist[]>.
727 To force multiple data values into an SV, you must do two things: use the
728 C<sv_set*v> routines to add the additional scalar type, then set a flag
729 so that Perl will believe it contains more than one type of data. The
730 four macros to set the flags are:
737 The particular macro you must use depends on which C<sv_set*v> routine
738 you called first. This is because every C<sv_set*v> routine turns on
739 only the bit for the particular type of data being set, and turns off
742 For example, to create a new Perl variable called "dberror" that contains
743 both the numeric and descriptive string error values, you could use the
747 extern char *dberror_list;
749 SV* sv = get_sv("dberror", TRUE);
750 sv_setiv(sv, (IV) dberror);
751 sv_setpv(sv, dberror_list[dberror]);
754 If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
755 macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
757 =head2 Magic Variables
759 [This section still under construction. Ignore everything here. Post no
760 bills. Everything not permitted is forbidden.]
762 Any SV may be magical, that is, it has special features that a normal
763 SV does not have. These features are stored in the SV structure in a
764 linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
777 Note this is current as of patchlevel 0, and could change at any time.
779 =head2 Assigning Magic
781 Perl adds magic to an SV using the sv_magic function:
783 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
785 The C<sv> argument is a pointer to the SV that is to acquire a new magical
788 If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
789 set the C<SVt_PVMG> flag for the C<sv>. Perl then continues by adding
790 it to the beginning of the linked list of magical features. Any prior
791 entry of the same type of magic is deleted. Note that this can be
792 overridden, and multiple instances of the same type of magic can be
793 associated with an SV.
795 The C<name> and C<namlen> arguments are used to associate a string with
796 the magic, typically the name of a variable. C<namlen> is stored in the
797 C<mg_len> field and if C<name> is non-null and C<namlen> >= 0 a malloc'd
798 copy of the name is stored in C<mg_ptr> field.
800 The sv_magic function uses C<how> to determine which, if any, predefined
801 "Magic Virtual Table" should be assigned to the C<mg_virtual> field.
802 See the "Magic Virtual Table" section below. The C<how> argument is also
803 stored in the C<mg_type> field.
805 The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
806 structure. If it is not the same as the C<sv> argument, the reference
807 count of the C<obj> object is incremented. If it is the same, or if
808 the C<how> argument is "#", or if it is a NULL pointer, then C<obj> is
809 merely stored, without the reference count being incremented.
811 There is also a function to add magic to an C<HV>:
813 void hv_magic(HV *hv, GV *gv, int how);
815 This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
817 To remove the magic from an SV, call the function sv_unmagic:
819 void sv_unmagic(SV *sv, int type);
821 The C<type> argument should be equal to the C<how> value when the C<SV>
822 was initially made magical.
824 =head2 Magic Virtual Tables
826 The C<mg_virtual> field in the C<MAGIC> structure is a pointer to a
827 C<MGVTBL>, which is a structure of function pointers and stands for
828 "Magic Virtual Table" to handle the various operations that might be
829 applied to that variable.
831 The C<MGVTBL> has five pointers to the following routine types:
833 int (*svt_get)(SV* sv, MAGIC* mg);
834 int (*svt_set)(SV* sv, MAGIC* mg);
835 U32 (*svt_len)(SV* sv, MAGIC* mg);
836 int (*svt_clear)(SV* sv, MAGIC* mg);
837 int (*svt_free)(SV* sv, MAGIC* mg);
839 This MGVTBL structure is set at compile-time in C<perl.h> and there are
840 currently 19 types (or 21 with overloading turned on). These different
841 structures contain pointers to various routines that perform additional
842 actions depending on which function is being called.
844 Function pointer Action taken
845 ---------------- ------------
846 svt_get Do something after the value of the SV is retrieved.
847 svt_set Do something after the SV is assigned a value.
848 svt_len Report on the SV's length.
849 svt_clear Clear something the SV represents.
850 svt_free Free any extra storage associated with the SV.
852 For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
853 to an C<mg_type> of '\0') contains:
855 { magic_get, magic_set, magic_len, 0, 0 }
857 Thus, when an SV is determined to be magical and of type '\0', if a get
858 operation is being performed, the routine C<magic_get> is called. All
859 the various routines for the various magical types begin with C<magic_>.
860 NOTE: the magic routines are not considered part of the Perl API, and may
861 not be exported by the Perl library.
863 The current kinds of Magic Virtual Tables are:
865 mg_type MGVTBL Type of magic
866 ------- ------ ----------------------------
867 \0 vtbl_sv Special scalar variable
868 A vtbl_amagic %OVERLOAD hash
869 a vtbl_amagicelem %OVERLOAD hash element
870 c (none) Holds overload table (AMT) on stash
871 B vtbl_bm Boyer-Moore (fast string search)
872 D vtbl_regdata Regex match position data (@+ and @- vars)
873 d vtbl_regdatum Regex match position data element
875 e vtbl_envelem %ENV hash element
876 f vtbl_fm Formline ('compiled' format)
877 g vtbl_mglob m//g target / study()ed string
878 I vtbl_isa @ISA array
879 i vtbl_isaelem @ISA array element
880 k vtbl_nkeys scalar(keys()) lvalue
881 L (none) Debugger %_<filename
882 l vtbl_dbline Debugger %_<filename element
883 o vtbl_collxfrm Locale transformation
884 P vtbl_pack Tied array or hash
885 p vtbl_packelem Tied array or hash element
886 q vtbl_packelem Tied scalar or handle
888 s vtbl_sigelem %SIG hash element
889 t vtbl_taint Taintedness
890 U vtbl_uvar Available for use by extensions
891 v vtbl_vec vec() lvalue
892 x vtbl_substr substr() lvalue
893 y vtbl_defelem Shadow "foreach" iterator variable /
894 smart parameter vivification
895 * vtbl_glob GV (typeglob)
896 # vtbl_arylen Array length ($#ary)
897 . vtbl_pos pos() lvalue
898 ~ (none) Available for use by extensions
900 When an uppercase and lowercase letter both exist in the table, then the
901 uppercase letter is used to represent some kind of composite type (a list
902 or a hash), and the lowercase letter is used to represent an element of
905 The '~' and 'U' magic types are defined specifically for use by
906 extensions and will not be used by perl itself. Extensions can use
907 '~' magic to 'attach' private information to variables (typically
908 objects). This is especially useful because there is no way for
909 normal perl code to corrupt this private information (unlike using
910 extra elements of a hash object).
912 Similarly, 'U' magic can be used much like tie() to call a C function
913 any time a scalar's value is used or changed. The C<MAGIC>'s
914 C<mg_ptr> field points to a C<ufuncs> structure:
917 I32 (*uf_val)(IV, SV*);
918 I32 (*uf_set)(IV, SV*);
922 When the SV is read from or written to, the C<uf_val> or C<uf_set>
923 function will be called with C<uf_index> as the first arg and a
924 pointer to the SV as the second. A simple example of how to add 'U'
925 magic is shown below. Note that the ufuncs structure is copied by
926 sv_magic, so you can safely allocate it on the stack.
934 uf.uf_val = &my_get_fn;
935 uf.uf_set = &my_set_fn;
937 sv_magic(sv, 0, 'U', (char*)&uf, sizeof(uf));
939 Note that because multiple extensions may be using '~' or 'U' magic,
940 it is important for extensions to take extra care to avoid conflict.
941 Typically only using the magic on objects blessed into the same class
942 as the extension is sufficient. For '~' magic, it may also be
943 appropriate to add an I32 'signature' at the top of the private data
946 Also note that the C<sv_set*()> and C<sv_cat*()> functions described
947 earlier do B<not> invoke 'set' magic on their targets. This must
948 be done by the user either by calling the C<SvSETMAGIC()> macro after
949 calling these functions, or by using one of the C<sv_set*_mg()> or
950 C<sv_cat*_mg()> functions. Similarly, generic C code must call the
951 C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
952 obtained from external sources in functions that don't handle magic.
953 See L<perlapi> for a description of these functions.
954 For example, calls to the C<sv_cat*()> functions typically need to be
955 followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
956 since their implementation handles 'get' magic.
960 MAGIC* mg_find(SV*, int type); /* Finds the magic pointer of that type */
962 This routine returns a pointer to the C<MAGIC> structure stored in the SV.
963 If the SV does not have that magical feature, C<NULL> is returned. Also,
964 if the SV is not of type SVt_PVMG, Perl may core dump.
966 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
968 This routine checks to see what types of magic C<sv> has. If the mg_type
969 field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
970 the mg_type field is changed to be the lowercase letter.
972 =head2 Understanding the Magic of Tied Hashes and Arrays
974 Tied hashes and arrays are magical beasts of the 'P' magic type.
976 WARNING: As of the 5.004 release, proper usage of the array and hash
977 access functions requires understanding a few caveats. Some
978 of these caveats are actually considered bugs in the API, to be fixed
979 in later releases, and are bracketed with [MAYCHANGE] below. If
980 you find yourself actually applying such information in this section, be
981 aware that the behavior may change in the future, umm, without warning.
983 The perl tie function associates a variable with an object that implements
984 the various GET, SET etc methods. To perform the equivalent of the perl
985 tie function from an XSUB, you must mimic this behaviour. The code below
986 carries out the necessary steps - firstly it creates a new hash, and then
987 creates a second hash which it blesses into the class which will implement
988 the tie methods. Lastly it ties the two hashes together, and returns a
989 reference to the new tied hash. Note that the code below does NOT call the
990 TIEHASH method in the MyTie class -
991 see L<Calling Perl Routines from within C Programs> for details on how
1002 tie = newRV_noinc((SV*)newHV());
1003 stash = gv_stashpv("MyTie", TRUE);
1004 sv_bless(tie, stash);
1005 hv_magic(hash, tie, 'P');
1006 RETVAL = newRV_noinc(hash);
1010 The C<av_store> function, when given a tied array argument, merely
1011 copies the magic of the array onto the value to be "stored", using
1012 C<mg_copy>. It may also return NULL, indicating that the value did not
1013 actually need to be stored in the array. [MAYCHANGE] After a call to
1014 C<av_store> on a tied array, the caller will usually need to call
1015 C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1016 TIEARRAY object. If C<av_store> did return NULL, a call to
1017 C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1020 The previous paragraph is applicable verbatim to tied hash access using the
1021 C<hv_store> and C<hv_store_ent> functions as well.
1023 C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1024 C<hv_fetch_ent> actually return an undefined mortal value whose magic
1025 has been initialized using C<mg_copy>. Note the value so returned does not
1026 need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1027 need to call C<mg_get()> on the returned value in order to actually invoke
1028 the perl level "FETCH" method on the underlying TIE object. Similarly,
1029 you may also call C<mg_set()> on the return value after possibly assigning
1030 a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
1031 method on the TIE object. [/MAYCHANGE]
1034 In other words, the array or hash fetch/store functions don't really
1035 fetch and store actual values in the case of tied arrays and hashes. They
1036 merely call C<mg_copy> to attach magic to the values that were meant to be
1037 "stored" or "fetched". Later calls to C<mg_get> and C<mg_set> actually
1038 do the job of invoking the TIE methods on the underlying objects. Thus
1039 the magic mechanism currently implements a kind of lazy access to arrays
1042 Currently (as of perl version 5.004), use of the hash and array access
1043 functions requires the user to be aware of whether they are operating on
1044 "normal" hashes and arrays, or on their tied variants. The API may be
1045 changed to provide more transparent access to both tied and normal data
1046 types in future versions.
1049 You would do well to understand that the TIEARRAY and TIEHASH interfaces
1050 are mere sugar to invoke some perl method calls while using the uniform hash
1051 and array syntax. The use of this sugar imposes some overhead (typically
1052 about two to four extra opcodes per FETCH/STORE operation, in addition to
1053 the creation of all the mortal variables required to invoke the methods).
1054 This overhead will be comparatively small if the TIE methods are themselves
1055 substantial, but if they are only a few statements long, the overhead
1056 will not be insignificant.
1058 =head2 Localizing changes
1060 Perl has a very handy construction
1067 This construction is I<approximately> equivalent to
1076 The biggest difference is that the first construction would
1077 reinstate the initial value of $var, irrespective of how control exits
1078 the block: C<goto>, C<return>, C<die>/C<eval> etc. It is a little bit
1079 more efficient as well.
1081 There is a way to achieve a similar task from C via Perl API: create a
1082 I<pseudo-block>, and arrange for some changes to be automatically
1083 undone at the end of it, either explicit, or via a non-local exit (via
1084 die()). A I<block>-like construct is created by a pair of
1085 C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1086 Such a construct may be created specially for some important localized
1087 task, or an existing one (like boundaries of enclosing Perl
1088 subroutine/block, or an existing pair for freeing TMPs) may be
1089 used. (In the second case the overhead of additional localization must
1090 be almost negligible.) Note that any XSUB is automatically enclosed in
1091 an C<ENTER>/C<LEAVE> pair.
1093 Inside such a I<pseudo-block> the following service is available:
1097 =item C<SAVEINT(int i)>
1099 =item C<SAVEIV(IV i)>
1101 =item C<SAVEI32(I32 i)>
1103 =item C<SAVELONG(long i)>
1105 These macros arrange things to restore the value of integer variable
1106 C<i> at the end of enclosing I<pseudo-block>.
1108 =item C<SAVESPTR(s)>
1110 =item C<SAVEPPTR(p)>
1112 These macros arrange things to restore the value of pointers C<s> and
1113 C<p>. C<s> must be a pointer of a type which survives conversion to
1114 C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1117 =item C<SAVEFREESV(SV *sv)>
1119 The refcount of C<sv> would be decremented at the end of
1120 I<pseudo-block>. This is similar to C<sv_2mortal>, which should (?) be
1123 =item C<SAVEFREEOP(OP *op)>
1125 The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1127 =item C<SAVEFREEPV(p)>
1129 The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1130 end of I<pseudo-block>.
1132 =item C<SAVECLEARSV(SV *sv)>
1134 Clears a slot in the current scratchpad which corresponds to C<sv> at
1135 the end of I<pseudo-block>.
1137 =item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1139 The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
1140 string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1141 short-lived storage, the corresponding string may be reallocated like
1144 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
1146 =item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
1148 At the end of I<pseudo-block> the function C<f> is called with the
1151 =item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1153 At the end of I<pseudo-block> the function C<f> is called with the
1154 implicit context argument (if any), and C<p>.
1156 =item C<SAVESTACK_POS()>
1158 The current offset on the Perl internal stack (cf. C<SP>) is restored
1159 at the end of I<pseudo-block>.
1163 The following API list contains functions, thus one needs to
1164 provide pointers to the modifiable data explicitly (either C pointers,
1165 or Perlish C<GV *>s). Where the above macros take C<int>, a similar
1166 function takes C<int *>.
1170 =item C<SV* save_scalar(GV *gv)>
1172 Equivalent to Perl code C<local $gv>.
1174 =item C<AV* save_ary(GV *gv)>
1176 =item C<HV* save_hash(GV *gv)>
1178 Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1180 =item C<void save_item(SV *item)>
1182 Duplicates the current value of C<SV>, on the exit from the current
1183 C<ENTER>/C<LEAVE> I<pseudo-block> will restore the value of C<SV>
1184 using the stored value.
1186 =item C<void save_list(SV **sarg, I32 maxsarg)>
1188 A variant of C<save_item> which takes multiple arguments via an array
1189 C<sarg> of C<SV*> of length C<maxsarg>.
1191 =item C<SV* save_svref(SV **sptr)>
1193 Similar to C<save_scalar>, but will reinstate a C<SV *>.
1195 =item C<void save_aptr(AV **aptr)>
1197 =item C<void save_hptr(HV **hptr)>
1199 Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1203 The C<Alias> module implements localization of the basic types within the
1204 I<caller's scope>. People who are interested in how to localize things in
1205 the containing scope should take a look there too.
1209 =head2 XSUBs and the Argument Stack
1211 The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1212 An XSUB routine will have a stack that contains the arguments from the Perl
1213 program, and a way to map from the Perl data structures to a C equivalent.
1215 The stack arguments are accessible through the C<ST(n)> macro, which returns
1216 the C<n>'th stack argument. Argument 0 is the first argument passed in the
1217 Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1220 Most of the time, output from the C routine can be handled through use of
1221 the RETVAL and OUTPUT directives. However, there are some cases where the
1222 argument stack is not already long enough to handle all the return values.
1223 An example is the POSIX tzname() call, which takes no arguments, but returns
1224 two, the local time zone's standard and summer time abbreviations.
1226 To handle this situation, the PPCODE directive is used and the stack is
1227 extended using the macro:
1231 where C<SP> is the macro that represents the local copy of the stack pointer,
1232 and C<num> is the number of elements the stack should be extended by.
1234 Now that there is room on the stack, values can be pushed on it using the
1235 macros to push IVs, doubles, strings, and SV pointers respectively:
1242 And now the Perl program calling C<tzname>, the two values will be assigned
1245 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1247 An alternate (and possibly simpler) method to pushing values on the stack is
1255 These macros automatically adjust the stack for you, if needed. Thus, you
1256 do not need to call C<EXTEND> to extend the stack.
1258 For more information, consult L<perlxs> and L<perlxstut>.
1260 =head2 Calling Perl Routines from within C Programs
1262 There are four routines that can be used to call a Perl subroutine from
1263 within a C program. These four are:
1265 I32 call_sv(SV*, I32);
1266 I32 call_pv(const char*, I32);
1267 I32 call_method(const char*, I32);
1268 I32 call_argv(const char*, I32, register char**);
1270 The routine most often used is C<call_sv>. The C<SV*> argument
1271 contains either the name of the Perl subroutine to be called, or a
1272 reference to the subroutine. The second argument consists of flags
1273 that control the context in which the subroutine is called, whether
1274 or not the subroutine is being passed arguments, how errors should be
1275 trapped, and how to treat return values.
1277 All four routines return the number of arguments that the subroutine returned
1280 These routines used to be called C<perl_call_sv> etc., before Perl v5.6.0,
1281 but those names are now deprecated; macros of the same name are provided for
1284 When using any of these routines (except C<call_argv>), the programmer
1285 must manipulate the Perl stack. These include the following macros and
1300 For a detailed description of calling conventions from C to Perl,
1301 consult L<perlcall>.
1303 =head2 Memory Allocation
1305 All memory meant to be used with the Perl API functions should be manipulated
1306 using the macros described in this section. The macros provide the necessary
1307 transparency between differences in the actual malloc implementation that is
1310 It is suggested that you enable the version of malloc that is distributed
1311 with Perl. It keeps pools of various sizes of unallocated memory in
1312 order to satisfy allocation requests more quickly. However, on some
1313 platforms, it may cause spurious malloc or free errors.
1315 New(x, pointer, number, type);
1316 Newc(x, pointer, number, type, cast);
1317 Newz(x, pointer, number, type);
1319 These three macros are used to initially allocate memory.
1321 The first argument C<x> was a "magic cookie" that was used to keep track
1322 of who called the macro, to help when debugging memory problems. However,
1323 the current code makes no use of this feature (most Perl developers now
1324 use run-time memory checkers), so this argument can be any number.
1326 The second argument C<pointer> should be the name of a variable that will
1327 point to the newly allocated memory.
1329 The third and fourth arguments C<number> and C<type> specify how many of
1330 the specified type of data structure should be allocated. The argument
1331 C<type> is passed to C<sizeof>. The final argument to C<Newc>, C<cast>,
1332 should be used if the C<pointer> argument is different from the C<type>
1335 Unlike the C<New> and C<Newc> macros, the C<Newz> macro calls C<memzero>
1336 to zero out all the newly allocated memory.
1338 Renew(pointer, number, type);
1339 Renewc(pointer, number, type, cast);
1342 These three macros are used to change a memory buffer size or to free a
1343 piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1344 match those of C<New> and C<Newc> with the exception of not needing the
1345 "magic cookie" argument.
1347 Move(source, dest, number, type);
1348 Copy(source, dest, number, type);
1349 Zero(dest, number, type);
1351 These three macros are used to move, copy, or zero out previously allocated
1352 memory. The C<source> and C<dest> arguments point to the source and
1353 destination starting points. Perl will move, copy, or zero out C<number>
1354 instances of the size of the C<type> data structure (using the C<sizeof>
1359 The most recent development releases of Perl has been experimenting with
1360 removing Perl's dependency on the "normal" standard I/O suite and allowing
1361 other stdio implementations to be used. This involves creating a new
1362 abstraction layer that then calls whichever implementation of stdio Perl
1363 was compiled with. All XSUBs should now use the functions in the PerlIO
1364 abstraction layer and not make any assumptions about what kind of stdio
1367 For a complete description of the PerlIO abstraction, consult L<perlapio>.
1369 =head2 Putting a C value on Perl stack
1371 A lot of opcodes (this is an elementary operation in the internal perl
1372 stack machine) put an SV* on the stack. However, as an optimization
1373 the corresponding SV is (usually) not recreated each time. The opcodes
1374 reuse specially assigned SVs (I<target>s) which are (as a corollary)
1375 not constantly freed/created.
1377 Each of the targets is created only once (but see
1378 L<Scratchpads and recursion> below), and when an opcode needs to put
1379 an integer, a double, or a string on stack, it just sets the
1380 corresponding parts of its I<target> and puts the I<target> on stack.
1382 The macro to put this target on stack is C<PUSHTARG>, and it is
1383 directly used in some opcodes, as well as indirectly in zillions of
1384 others, which use it via C<(X)PUSH[pni]>.
1388 The question remains on when the SVs which are I<target>s for opcodes
1389 are created. The answer is that they are created when the current unit --
1390 a subroutine or a file (for opcodes for statements outside of
1391 subroutines) -- is compiled. During this time a special anonymous Perl
1392 array is created, which is called a scratchpad for the current
1395 A scratchpad keeps SVs which are lexicals for the current unit and are
1396 targets for opcodes. One can deduce that an SV lives on a scratchpad
1397 by looking on its flags: lexicals have C<SVs_PADMY> set, and
1398 I<target>s have C<SVs_PADTMP> set.
1400 The correspondence between OPs and I<target>s is not 1-to-1. Different
1401 OPs in the compile tree of the unit can use the same target, if this
1402 would not conflict with the expected life of the temporary.
1404 =head2 Scratchpads and recursion
1406 In fact it is not 100% true that a compiled unit contains a pointer to
1407 the scratchpad AV. In fact it contains a pointer to an AV of
1408 (initially) one element, and this element is the scratchpad AV. Why do
1409 we need an extra level of indirection?
1411 The answer is B<recursion>, and maybe (sometime soon) B<threads>. Both
1412 these can create several execution pointers going into the same
1413 subroutine. For the subroutine-child not write over the temporaries
1414 for the subroutine-parent (lifespan of which covers the call to the
1415 child), the parent and the child should have different
1416 scratchpads. (I<And> the lexicals should be separate anyway!)
1418 So each subroutine is born with an array of scratchpads (of length 1).
1419 On each entry to the subroutine it is checked that the current
1420 depth of the recursion is not more than the length of this array, and
1421 if it is, new scratchpad is created and pushed into the array.
1423 The I<target>s on this scratchpad are C<undef>s, but they are already
1424 marked with correct flags.
1426 =head1 Compiled code
1430 Here we describe the internal form your code is converted to by
1431 Perl. Start with a simple example:
1435 This is converted to a tree similar to this one:
1443 (but slightly more complicated). This tree reflects the way Perl
1444 parsed your code, but has nothing to do with the execution order.
1445 There is an additional "thread" going through the nodes of the tree
1446 which shows the order of execution of the nodes. In our simplified
1447 example above it looks like:
1449 $b ---> $c ---> + ---> $a ---> assign-to
1451 But with the actual compile tree for C<$a = $b + $c> it is different:
1452 some nodes I<optimized away>. As a corollary, though the actual tree
1453 contains more nodes than our simplified example, the execution order
1454 is the same as in our example.
1456 =head2 Examining the tree
1458 If you have your perl compiled for debugging (usually done with C<-D
1459 optimize=-g> on C<Configure> command line), you may examine the
1460 compiled tree by specifying C<-Dx> on the Perl command line. The
1461 output takes several lines per node, and for C<$b+$c> it looks like
1466 FLAGS = (SCALAR,KIDS)
1468 TYPE = null ===> (4)
1470 FLAGS = (SCALAR,KIDS)
1472 3 TYPE = gvsv ===> 4
1478 TYPE = null ===> (5)
1480 FLAGS = (SCALAR,KIDS)
1482 4 TYPE = gvsv ===> 5
1488 This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
1489 not optimized away (one per number in the left column). The immediate
1490 children of the given node correspond to C<{}> pairs on the same level
1491 of indentation, thus this listing corresponds to the tree:
1499 The execution order is indicated by C<===E<gt>> marks, thus it is C<3
1500 4 5 6> (node C<6> is not included into above listing), i.e.,
1501 C<gvsv gvsv add whatever>.
1503 Each of these nodes represents an op, a fundamental operation inside the
1504 Perl core. The code which implements each operation can be found in the
1505 F<pp*.c> files; the function which implements the op with type C<gvsv>
1506 is C<pp_gvsv>, and so on. As the tree above shows, different ops have
1507 different numbers of children: C<add> is a binary operator, as one would
1508 expect, and so has two children. To accommodate the various different
1509 numbers of children, there are various types of op data structure, and
1510 they link together in different ways.
1512 The simplest type of op structure is C<OP>: this has no children. Unary
1513 operators, C<UNOP>s, have one child, and this is pointed to by the
1514 C<op_first> field. Binary operators (C<BINOP>s) have not only an
1515 C<op_first> field but also an C<op_last> field. The most complex type of
1516 op is a C<LISTOP>, which has any number of children. In this case, the
1517 first child is pointed to by C<op_first> and the last child by
1518 C<op_last>. The children in between can be found by iteratively
1519 following the C<op_sibling> pointer from the first child to the last.
1521 There are also two other op types: a C<PMOP> holds a regular expression,
1522 and has no children, and a C<LOOP> may or may not have children. If the
1523 C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
1524 complicate matters, if a C<UNOP> is actually a C<null> op after
1525 optimization (see L</Compile pass 2: context propagation>) it will still
1526 have children in accordance with its former type.
1528 =head2 Compile pass 1: check routines
1530 The tree is created by the compiler while I<yacc> code feeds it
1531 the constructions it recognizes. Since I<yacc> works bottom-up, so does
1532 the first pass of perl compilation.
1534 What makes this pass interesting for perl developers is that some
1535 optimization may be performed on this pass. This is optimization by
1536 so-called "check routines". The correspondence between node names
1537 and corresponding check routines is described in F<opcode.pl> (do not
1538 forget to run C<make regen_headers> if you modify this file).
1540 A check routine is called when the node is fully constructed except
1541 for the execution-order thread. Since at this time there are no
1542 back-links to the currently constructed node, one can do most any
1543 operation to the top-level node, including freeing it and/or creating
1544 new nodes above/below it.
1546 The check routine returns the node which should be inserted into the
1547 tree (if the top-level node was not modified, check routine returns
1550 By convention, check routines have names C<ck_*>. They are usually
1551 called from C<new*OP> subroutines (or C<convert>) (which in turn are
1552 called from F<perly.y>).
1554 =head2 Compile pass 1a: constant folding
1556 Immediately after the check routine is called the returned node is
1557 checked for being compile-time executable. If it is (the value is
1558 judged to be constant) it is immediately executed, and a I<constant>
1559 node with the "return value" of the corresponding subtree is
1560 substituted instead. The subtree is deleted.
1562 If constant folding was not performed, the execution-order thread is
1565 =head2 Compile pass 2: context propagation
1567 When a context for a part of compile tree is known, it is propagated
1568 down through the tree. At this time the context can have 5 values
1569 (instead of 2 for runtime context): void, boolean, scalar, list, and
1570 lvalue. In contrast with the pass 1 this pass is processed from top
1571 to bottom: a node's context determines the context for its children.
1573 Additional context-dependent optimizations are performed at this time.
1574 Since at this moment the compile tree contains back-references (via
1575 "thread" pointers), nodes cannot be free()d now. To allow
1576 optimized-away nodes at this stage, such nodes are null()ified instead
1577 of free()ing (i.e. their type is changed to OP_NULL).
1579 =head2 Compile pass 3: peephole optimization
1581 After the compile tree for a subroutine (or for an C<eval> or a file)
1582 is created, an additional pass over the code is performed. This pass
1583 is neither top-down or bottom-up, but in the execution order (with
1584 additional complications for conditionals). These optimizations are
1585 done in the subroutine peep(). Optimizations performed at this stage
1586 are subject to the same restrictions as in the pass 2.
1588 =head1 Examining internal data structures with the C<dump> functions
1590 To aid debugging, the source file F<dump.c> contains a number of
1591 functions which produce formatted output of internal data structures.
1593 The most commonly used of these functions is C<Perl_sv_dump>; it's used
1594 for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
1595 C<sv_dump> to produce debugging output from Perl-space, so users of that
1596 module should already be familiar with its format.
1598 C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
1599 derivatives, and produces output similiar to C<perl -Dx>; in fact,
1600 C<Perl_dump_eval> will dump the main root of the code being evaluated,
1601 exactly like C<-Dx>.
1603 Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
1604 op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
1605 subroutines in a package like so: (Thankfully, these are all xsubs, so
1606 there is no op tree)
1608 (gdb) print Perl_dump_packsubs(PL_defstash)
1610 SUB attributes::bootstrap = (xsub 0x811fedc 0)
1612 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
1614 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
1616 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
1618 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
1620 and C<Perl_dump_all>, which dumps all the subroutines in the stash and
1621 the op tree of the main root.
1623 =head1 How multiple interpreters and concurrency are supported
1625 =head2 Background and PERL_IMPLICIT_CONTEXT
1627 The Perl interpreter can be regarded as a closed box: it has an API
1628 for feeding it code or otherwise making it do things, but it also has
1629 functions for its own use. This smells a lot like an object, and
1630 there are ways for you to build Perl so that you can have multiple
1631 interpreters, with one interpreter represented either as a C++ object,
1632 a C structure, or inside a thread. The thread, the C structure, or
1633 the C++ object will contain all the context, the state of that
1636 Three macros control the major Perl build flavors: MULTIPLICITY,
1637 USE_THREADS and PERL_OBJECT. The MULTIPLICITY build has a C structure
1638 that packages all the interpreter state, there is a similar thread-specific
1639 data structure under USE_THREADS, and the PERL_OBJECT build has a C++
1640 class to maintain interpreter state. In all three cases,
1641 PERL_IMPLICIT_CONTEXT is also normally defined, and enables the
1642 support for passing in a "hidden" first argument that represents all three
1645 All this obviously requires a way for the Perl internal functions to be
1646 C++ methods, subroutines taking some kind of structure as the first
1647 argument, or subroutines taking nothing as the first argument. To
1648 enable these three very different ways of building the interpreter,
1649 the Perl source (as it does in so many other situations) makes heavy
1650 use of macros and subroutine naming conventions.
1652 First problem: deciding which functions will be public API functions and
1653 which will be private. All functions whose names begin C<S_> are private
1654 (think "S" for "secret" or "static"). All other functions begin with
1655 "Perl_", but just because a function begins with "Perl_" does not mean it is
1656 part of the API. (See L</Internal Functions>.) The easiest way to be B<sure> a
1657 function is part of the API is to find its entry in L<perlapi>.
1658 If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
1659 think it should be (i.e., you need it for your extension), send mail via
1660 L<perlbug> explaining why you think it should be.
1662 Second problem: there must be a syntax so that the same subroutine
1663 declarations and calls can pass a structure as their first argument,
1664 or pass nothing. To solve this, the subroutines are named and
1665 declared in a particular way. Here's a typical start of a static
1666 function used within the Perl guts:
1669 S_incline(pTHX_ char *s)
1671 STATIC becomes "static" in C, and is #define'd to nothing in C++.
1673 A public function (i.e. part of the internal API, but not necessarily
1674 sanctioned for use in extensions) begins like this:
1677 Perl_sv_setsv(pTHX_ SV* dsv, SV* ssv)
1679 C<pTHX_> is one of a number of macros (in perl.h) that hide the
1680 details of the interpreter's context. THX stands for "thread", "this",
1681 or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
1682 The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
1683 or 'd' for B<d>eclaration.
1685 When Perl is built without PERL_IMPLICIT_CONTEXT, there is no first
1686 argument containing the interpreter's context. The trailing underscore
1687 in the pTHX_ macro indicates that the macro expansion needs a comma
1688 after the context argument because other arguments follow it. If
1689 PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
1690 subroutine is not prototyped to take the extra argument. The form of the
1691 macro without the trailing underscore is used when there are no additional
1694 When a core function calls another, it must pass the context. This
1695 is normally hidden via macros. Consider C<sv_setsv>. It expands
1696 something like this:
1698 ifdef PERL_IMPLICIT_CONTEXT
1699 define sv_setsv(a,b) Perl_sv_setsv(aTHX_ a, b)
1700 /* can't do this for vararg functions, see below */
1702 define sv_setsv Perl_sv_setsv
1705 This works well, and means that XS authors can gleefully write:
1709 and still have it work under all the modes Perl could have been
1712 Under PERL_OBJECT in the core, that will translate to either:
1714 CPerlObj::Perl_sv_setsv(foo,bar); # in CPerlObj functions,
1715 # C++ takes care of 'this'
1718 pPerl->Perl_sv_setsv(foo,bar); # in truly static functions,
1721 Under PERL_OBJECT in extensions (aka PERL_CAPI), or under
1722 MULTIPLICITY/USE_THREADS w/ PERL_IMPLICIT_CONTEXT in both core
1723 and extensions, it will be:
1725 Perl_sv_setsv(aTHX_ foo, bar); # the canonical Perl "API"
1726 # for all build flavors
1728 This doesn't work so cleanly for varargs functions, though, as macros
1729 imply that the number of arguments is known in advance. Instead we
1730 either need to spell them out fully, passing C<aTHX_> as the first
1731 argument (the Perl core tends to do this with functions like
1732 Perl_warner), or use a context-free version.
1734 The context-free version of Perl_warner is called
1735 Perl_warner_nocontext, and does not take the extra argument. Instead
1736 it does dTHX; to get the context from thread-local storage. We
1737 C<#define warner Perl_warner_nocontext> so that extensions get source
1738 compatibility at the expense of performance. (Passing an arg is
1739 cheaper than grabbing it from thread-local storage.)
1741 You can ignore [pad]THX[xo] when browsing the Perl headers/sources.
1742 Those are strictly for use within the core. Extensions and embedders
1743 need only be aware of [pad]THX.
1745 =head2 How do I use all this in extensions?
1747 When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
1748 any functions in the Perl API will need to pass the initial context
1749 argument somehow. The kicker is that you will need to write it in
1750 such a way that the extension still compiles when Perl hasn't been
1751 built with PERL_IMPLICIT_CONTEXT enabled.
1753 There are three ways to do this. First, the easy but inefficient way,
1754 which is also the default, in order to maintain source compatibility
1755 with extensions: whenever XSUB.h is #included, it redefines the aTHX
1756 and aTHX_ macros to call a function that will return the context.
1757 Thus, something like:
1761 in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
1764 Perl_sv_setsv(Perl_get_context(), asv, bsv);
1766 or to this otherwise:
1768 Perl_sv_setsv(asv, bsv);
1770 You have to do nothing new in your extension to get this; since
1771 the Perl library provides Perl_get_context(), it will all just
1774 The second, more efficient way is to use the following template for
1777 #define PERL_NO_GET_CONTEXT /* we want efficiency */
1782 static my_private_function(int arg1, int arg2);
1785 my_private_function(int arg1, int arg2)
1787 dTHX; /* fetch context */
1788 ... call many Perl API functions ...
1793 MODULE = Foo PACKAGE = Foo
1801 my_private_function(arg, 10);
1803 Note that the only two changes from the normal way of writing an
1804 extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
1805 including the Perl headers, followed by a C<dTHX;> declaration at
1806 the start of every function that will call the Perl API. (You'll
1807 know which functions need this, because the C compiler will complain
1808 that there's an undeclared identifier in those functions.) No changes
1809 are needed for the XSUBs themselves, because the XS() macro is
1810 correctly defined to pass in the implicit context if needed.
1812 The third, even more efficient way is to ape how it is done within
1816 #define PERL_NO_GET_CONTEXT /* we want efficiency */
1821 /* pTHX_ only needed for functions that call Perl API */
1822 static my_private_function(pTHX_ int arg1, int arg2);
1825 my_private_function(pTHX_ int arg1, int arg2)
1827 /* dTHX; not needed here, because THX is an argument */
1828 ... call Perl API functions ...
1833 MODULE = Foo PACKAGE = Foo
1841 my_private_function(aTHX_ arg, 10);
1843 This implementation never has to fetch the context using a function
1844 call, since it is always passed as an extra argument. Depending on
1845 your needs for simplicity or efficiency, you may mix the previous
1846 two approaches freely.
1848 Never add a comma after C<pTHX> yourself--always use the form of the
1849 macro with the underscore for functions that take explicit arguments,
1850 or the form without the argument for functions with no explicit arguments.
1852 =head2 Future Plans and PERL_IMPLICIT_SYS
1854 Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
1855 that the interpreter knows about itself and pass it around, so too are
1856 there plans to allow the interpreter to bundle up everything it knows
1857 about the environment it's running on. This is enabled with the
1858 PERL_IMPLICIT_SYS macro. Currently it only works with PERL_OBJECT,
1859 but is mostly there for MULTIPLICITY and USE_THREADS (see inside
1862 This allows the ability to provide an extra pointer (called the "host"
1863 environment) for all the system calls. This makes it possible for
1864 all the system stuff to maintain their own state, broken down into
1865 seven C structures. These are thin wrappers around the usual system
1866 calls (see win32/perllib.c) for the default perl executable, but for a
1867 more ambitious host (like the one that would do fork() emulation) all
1868 the extra work needed to pretend that different interpreters are
1869 actually different "processes", would be done here.
1871 The Perl engine/interpreter and the host are orthogonal entities.
1872 There could be one or more interpreters in a process, and one or
1873 more "hosts", with free association between them.
1875 =head1 Internal Functions
1877 All of Perl's internal functions which will be exposed to the outside
1878 world are be prefixed by C<Perl_> so that they will not conflict with XS
1879 functions or functions used in a program in which Perl is embedded.
1880 Similarly, all global variables begin with C<PL_>. (By convention,
1881 static functions start with C<S_>)
1883 Inside the Perl core, you can get at the functions either with or
1884 without the C<Perl_> prefix, thanks to a bunch of defines that live in
1885 F<embed.h>. This header file is generated automatically from
1886 F<embed.pl>. F<embed.pl> also creates the prototyping header files for
1887 the internal functions, generates the documentation and a lot of other
1888 bits and pieces. It's important that when you add a new function to the
1889 core or change an existing one, you change the data in the table at the
1890 end of F<embed.pl> as well. Here's a sample entry from that table:
1892 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
1894 The second column is the return type, the third column the name. Columns
1895 after that are the arguments. The first column is a set of flags:
1901 This function is a part of the public API.
1905 This function has a C<Perl_> prefix; ie, it is defined as C<Perl_av_fetch>
1909 This function has documentation using the C<apidoc> feature which we'll
1910 look at in a second.
1914 Other available flags are:
1920 This is a static function and is defined as C<S_whatever>.
1924 This does not use C<aTHX_> and C<pTHX> to pass interpreter context. (See
1925 L<perlguts/Background and PERL_IMPLICIT_CONTEXT>.)
1929 This function never returns; C<croak>, C<exit> and friends.
1933 This function takes a variable number of arguments, C<printf> style.
1934 The argument list should end with C<...>, like this:
1936 Afprd |void |croak |const char* pat|...
1940 This function is part of the experimental development API, and may change
1941 or disappear without notice.
1945 This function should not have a compatibility macro to define, say,
1946 C<Perl_parse> to C<parse>. It must be called as C<Perl_parse>.
1950 This function is not a member of C<CPerlObj>. If you don't know
1951 what this means, don't use it.
1955 This function isn't exported out of the Perl core.
1959 If you edit F<embed.pl>, you will need to run C<make regen_headers> to
1960 force a rebuild of F<embed.h> and other auto-generated files.
1962 =head2 Formatted Printing of IVs, UVs, and NVs
1964 If you are printing IVs, UVs, or NVS instead of the stdio(3) style
1965 formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
1966 following macros for portability
1971 UVxf UV in hexadecimal
1976 These will take care of 64-bit integers and long doubles.
1979 printf("IV is %"IVdf"\n", iv);
1981 The IVdf will expand to whatever is the correct format for the IVs.
1983 If you are printing addresses of pointers, use UVxf combined
1984 with PTR2UV(), do not use %lx or %p.
1986 =head2 Pointer-To-Integer and Integer-To-Pointer
1988 Because pointer size does not necessarily equal integer size,
1989 use the follow macros to do it right.
1994 INT2PTR(pointertotype, integer)
1999 SV *sv = INT2PTR(SV*, iv);
2006 =head2 Source Documentation
2008 There's an effort going on to document the internal functions and
2009 automatically produce reference manuals from them - L<perlapi> is one
2010 such manual which details all the functions which are available to XS
2011 writers. L<perlintern> is the autogenerated manual for the functions
2012 which are not part of the API and are supposedly for internal use only.
2014 Source documentation is created by putting POD comments into the C
2018 =for apidoc sv_setiv
2020 Copies an integer into the given SV. Does not handle 'set' magic. See
2026 Please try and supply some documentation if you add functions to the
2029 =head1 Unicode Support
2031 Perl 5.6.0 introduced Unicode support. It's important for porters and XS
2032 writers to understand this support and make sure that the code they
2033 write does not corrupt Unicode data.
2035 =head2 What B<is> Unicode, anyway?
2037 In the olden, less enlightened times, we all used to use ASCII. Most of
2038 us did, anyway. The big problem with ASCII is that it's American. Well,
2039 no, that's not actually the problem; the problem is that it's not
2040 particularly useful for people who don't use the Roman alphabet. What
2041 used to happen was that particular languages would stick their own
2042 alphabet in the upper range of the sequence, between 128 and 255. Of
2043 course, we then ended up with plenty of variants that weren't quite
2044 ASCII, and the whole point of it being a standard was lost.
2046 Worse still, if you've got a language like Chinese or
2047 Japanese that has hundreds or thousands of characters, then you really
2048 can't fit them into a mere 256, so they had to forget about ASCII
2049 altogether, and build their own systems using pairs of numbers to refer
2052 To fix this, some people formed Unicode, Inc. and
2053 produced a new character set containing all the characters you can
2054 possibly think of and more. There are several ways of representing these
2055 characters, and the one Perl uses is called UTF8. UTF8 uses
2056 a variable number of bytes to represent a character, instead of just
2057 one. You can learn more about Unicode at http://www.unicode.org/
2059 =head2 How can I recognise a UTF8 string?
2061 You can't. This is because UTF8 data is stored in bytes just like
2062 non-UTF8 data. The Unicode character 200, (C<0xC8> for you hex types)
2063 capital E with a grave accent, is represented by the two bytes
2064 C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
2065 has that byte sequence as well. So you can't tell just by looking - this
2066 is what makes Unicode input an interesting problem.
2068 The API function C<is_utf8_string> can help; it'll tell you if a string
2069 contains only valid UTF8 characters. However, it can't do the work for
2070 you. On a character-by-character basis, C<is_utf8_char> will tell you
2071 whether the current character in a string is valid UTF8.
2073 =head2 How does UTF8 represent Unicode characters?
2075 As mentioned above, UTF8 uses a variable number of bytes to store a
2076 character. Characters with values 1...128 are stored in one byte, just
2077 like good ol' ASCII. Character 129 is stored as C<v194.129>; this
2078 continues up to character 191, which is C<v194.191>. Now we've run out of
2079 bits (191 is binary C<10111111>) so we move on; 192 is C<v195.128>. And
2080 so it goes on, moving to three bytes at character 2048.
2082 Assuming you know you're dealing with a UTF8 string, you can find out
2083 how long the first character in it is with the C<UTF8SKIP> macro:
2085 char *utf = "\305\233\340\240\201";
2088 len = UTF8SKIP(utf); /* len is 2 here */
2090 len = UTF8SKIP(utf); /* len is 3 here */
2092 Another way to skip over characters in a UTF8 string is to use
2093 C<utf8_hop>, which takes a string and a number of characters to skip
2094 over. You're on your own about bounds checking, though, so don't use it
2097 All bytes in a multi-byte UTF8 character will have the high bit set, so
2098 you can test if you need to do something special with this character
2104 /* Must treat this as UTF8 */
2105 uv = utf8_to_uv(utf);
2107 /* OK to treat this character as a byte */
2110 You can also see in that example that we use C<utf8_to_uv> to get the
2111 value of the character; the inverse function C<uv_to_utf8> is available
2112 for putting a UV into UTF8:
2115 /* Must treat this as UTF8 */
2116 utf8 = uv_to_utf8(utf8, uv);
2118 /* OK to treat this character as a byte */
2121 You B<must> convert characters to UVs using the above functions if
2122 you're ever in a situation where you have to match UTF8 and non-UTF8
2123 characters. You may not skip over UTF8 characters in this case. If you
2124 do this, you'll lose the ability to match hi-bit non-UTF8 characters;
2125 for instance, if your UTF8 string contains C<v196.172>, and you skip
2126 that character, you can never match a C<chr(200)> in a non-UTF8 string.
2129 =head2 How does Perl store UTF8 strings?
2131 Currently, Perl deals with Unicode strings and non-Unicode strings
2132 slightly differently. If a string has been identified as being UTF-8
2133 encoded, Perl will set a flag in the SV, C<SVf_UTF8>. You can check and
2134 manipulate this flag with the following macros:
2140 This flag has an important effect on Perl's treatment of the string: if
2141 Unicode data is not properly distinguished, regular expressions,
2142 C<length>, C<substr> and other string handling operations will have
2143 undesirable results.
2145 The problem comes when you have, for instance, a string that isn't
2146 flagged is UTF8, and contains a byte sequence that could be UTF8 -
2147 especially when combining non-UTF8 and UTF8 strings.
2149 Never forget that the C<SVf_UTF8> flag is separate to the PV value; you
2150 need be sure you don't accidentally knock it off while you're
2151 manipulating SVs. More specifically, you cannot expect to do this:
2160 nsv = newSVpvn(p, len);
2162 The C<char*> string does not tell you the whole story, and you can't
2163 copy or reconstruct an SV just by copying the string value. Check if the
2164 old SV has the UTF8 flag set, and act accordingly:
2168 nsv = newSVpvn(p, len);
2172 In fact, your C<frobnicate> function should be made aware of whether or
2173 not it's dealing with UTF8 data, so that it can handle the string
2176 =head2 How do I convert a string to UTF8?
2178 If you're mixing UTF8 and non-UTF8 strings, you might find it necessary
2179 to upgrade one of the strings to UTF8. If you've got an SV, the easiest
2182 sv_utf8_upgrade(sv);
2184 However, you must not do this, for example:
2187 sv_utf8_upgrade(left);
2189 If you do this in a binary operator, you will actually change one of the
2190 strings that came into the operator, and, while it shouldn't be noticeable
2191 by the end user, it can cause problems.
2193 Instead, C<bytes_to_utf8> will give you a UTF8-encoded B<copy> of its
2194 string argument. This is useful for having the data available for
2195 comparisons and so on, without harming the original SV. There's also
2196 C<utf8_to_bytes> to go the other way, but naturally, this will fail if
2197 the string contains any characters above 255 that can't be represented
2200 =head2 Is there anything else I need to know?
2202 Not really. Just remember these things:
2208 There's no way to tell if a string is UTF8 or not. You can tell if an SV
2209 is UTF8 by looking at is C<SvUTF8> flag. Don't forget to set the flag if
2210 something should be UTF8. Treat the flag as part of the PV, even though
2211 it's not - if you pass on the PV to somewhere, pass on the flag too.
2215 If a string is UTF8, B<always> use C<utf8_to_uv> to get at the value,
2216 unless C<!(*s & 0x80)> in which case you can use C<*s>.
2220 When writing to a UTF8 string, B<always> use C<uv_to_utf8>, unless
2221 C<uv < 0x80> in which case you can use C<*s = uv>.
2225 Mixing UTF8 and non-UTF8 strings is tricky. Use C<bytes_to_utf8> to get
2226 a new string which is UTF8 encoded. There are tricks you can use to
2227 delay deciding whether you need to use a UTF8 string until you get to a
2228 high character - C<HALF_UPGRADE> is one of those.
2234 Until May 1997, this document was maintained by Jeff Okamoto
2235 <okamoto@corp.hp.com>. It is now maintained as part of Perl itself
2236 by the Perl 5 Porters <perl5-porters@perl.org>.
2238 With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
2239 Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
2240 Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
2241 Stephen McCamant, and Gurusamy Sarathy.
2243 API Listing originally by Dean Roehrich <roehrich@cray.com>.
2245 Modifications to autogenerate the API listing (L<perlapi>) by Benjamin
2250 perlapi(1), perlintern(1), perlxs(1), perlembed(1)