This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
overload int()
[perl5.git] / pod / perlguts.pod
CommitLineData
a0d0e21e
LW
1=head1 NAME
2
954c1994 3perlguts - Introduction to the Perl API
a0d0e21e
LW
4
5=head1 DESCRIPTION
6
b3b6085d
PP
7This document attempts to describe how to use the Perl API, as well as
8containing some info on the basic workings of the Perl core. It is far
9from complete and probably contains many errors. Please refer any
10questions or comments to the author below.
a0d0e21e 11
0a753a76
PP
12=head1 Variables
13
5f05dabc 14=head2 Datatypes
a0d0e21e
LW
15
16Perl has three typedefs that handle Perl's three main data types:
17
18 SV Scalar Value
19 AV Array Value
20 HV Hash Value
21
d1b91892 22Each typedef has specific routines that manipulate the various data types.
a0d0e21e
LW
23
24=head2 What is an "IV"?
25
954c1994 26Perl uses a special typedef IV which is a simple signed integer type that is
5f05dabc 27guaranteed to be large enough to hold a pointer (as well as an integer).
954c1994 28Additionally, there is the UV, which is simply an unsigned IV.
a0d0e21e 29
d1b91892 30Perl also uses two special typedefs, I32 and I16, which will always be at
954c1994
GS
31least 32-bits and 16-bits long, respectively. (Again, there are U32 and U16,
32as well.)
a0d0e21e 33
54310121 34=head2 Working with SVs
a0d0e21e
LW
35
36An SV can be created and loaded with one command. There are four types of
a7dfe00a
MS
37values that can be loaded: an integer value (IV), a double (NV),
38a string (PV), and another scalar (SV).
a0d0e21e 39
9da1e3b5 40The six routines are:
a0d0e21e
LW
41
42 SV* newSViv(IV);
43 SV* newSVnv(double);
08105a92
GS
44 SV* newSVpv(const char*, int);
45 SV* newSVpvn(const char*, int);
46fc3d4c 46 SV* newSVpvf(const char*, ...);
a0d0e21e
LW
47 SV* newSVsv(SV*);
48
deb3007b 49To change the value of an *already-existing* SV, there are seven routines:
a0d0e21e
LW
50
51 void sv_setiv(SV*, IV);
deb3007b 52 void sv_setuv(SV*, UV);
a0d0e21e 53 void sv_setnv(SV*, double);
08105a92
GS
54 void sv_setpv(SV*, const char*);
55 void sv_setpvn(SV*, const char*, int)
46fc3d4c 56 void sv_setpvf(SV*, const char*, ...);
9abd00ed 57 void sv_setpvfn(SV*, const char*, STRLEN, va_list *, SV **, I32, bool);
a0d0e21e
LW
58 void sv_setsv(SV*, SV*);
59
60Notice that you can choose to specify the length of the string to be
9da1e3b5
MUN
61assigned by using C<sv_setpvn>, C<newSVpvn>, or C<newSVpv>, or you may
62allow Perl to calculate the length by using C<sv_setpv> or by specifying
630 as the second argument to C<newSVpv>. Be warned, though, that Perl will
64determine the string's length by using C<strlen>, which depends on the
9abd00ed
GS
65string terminating with a NUL character.
66
67The arguments of C<sv_setpvf> are processed like C<sprintf>, and the
68formatted output becomes the value.
69
70C<sv_setpvfn> is an analogue of C<vsprintf>, but it allows you to specify
71either a pointer to a variable argument list or the address and length of
72an array of SVs. The last argument points to a boolean; on return, if that
73boolean is true, then locale-specific information has been used to format
c2611fb3 74the string, and the string's contents are therefore untrustworthy (see
9abd00ed
GS
75L<perlsec>). This pointer may be NULL if that information is not
76important. Note that this function requires you to specify the length of
77the format.
78
7c906a97
AD
79STRLEN is an integer type (Size_t, usually defined as size_t in
80config.h) guaranteed to be large enough to represent the size of
81any string that perl can handle.
82
9da1e3b5
MUN
83The C<sv_set*()> functions are not generic enough to operate on values
84that have "magic". See L<Magic Virtual Tables> later in this document.
a0d0e21e 85
a3cb178b
GS
86All SVs that contain strings should be terminated with a NUL character.
87If it is not NUL-terminated there is a risk of
5f05dabc
PP
88core dumps and corruptions from code which passes the string to C
89functions or system calls which expect a NUL-terminated string.
90Perl's own functions typically add a trailing NUL for this reason.
91Nevertheless, you should be very careful when you pass a string stored
92in an SV to a C function or system call.
93
a0d0e21e
LW
94To access the actual value that an SV points to, you can use the macros:
95
96 SvIV(SV*)
954c1994 97 SvUV(SV*)
a0d0e21e
LW
98 SvNV(SV*)
99 SvPV(SV*, STRLEN len)
1fa8b10d 100 SvPV_nolen(SV*)
a0d0e21e 101
954c1994 102which will automatically coerce the actual scalar type into an IV, UV, double,
a0d0e21e
LW
103or string.
104
105In the C<SvPV> macro, the length of the string returned is placed into the
1fa8b10d
JD
106variable C<len> (this is a macro, so you do I<not> use C<&len>). If you do
107not care what the length of the data is, use the C<SvPV_nolen> macro.
108Historically the C<SvPV> macro with the global variable C<PL_na> has been
109used in this case. But that can be quite inefficient because C<PL_na> must
110be accessed in thread-local storage in threaded Perl. In any case, remember
111that Perl allows arbitrary strings of data that may both contain NULs and
112might not be terminated by a NUL.
a0d0e21e 113
ce2f5d8f
KA
114Also remember that C doesn't allow you to safely say C<foo(SvPV(s, len),
115len);>. It might work with your compiler, but it won't work for everyone.
116Break this sort of statement up into separate assignments:
117
b2f5ed49 118 SV *s;
ce2f5d8f
KA
119 STRLEN len;
120 char * ptr;
b2f5ed49 121 ptr = SvPV(s, len);
ce2f5d8f
KA
122 foo(ptr, len);
123
07fa94a1 124If you want to know if the scalar value is TRUE, you can use:
a0d0e21e
LW
125
126 SvTRUE(SV*)
127
128Although Perl will automatically grow strings for you, if you need to force
129Perl to allocate more memory for your SV, you can use the macro
130
131 SvGROW(SV*, STRLEN newlen)
132
133which will determine if more memory needs to be allocated. If so, it will
134call the function C<sv_grow>. Note that C<SvGROW> can only increase, not
5f05dabc
PP
135decrease, the allocated memory of an SV and that it does not automatically
136add a byte for the a trailing NUL (perl's own string functions typically do
8ebc5c01 137C<SvGROW(sv, len + 1)>).
a0d0e21e
LW
138
139If you have an SV and want to know what kind of data Perl thinks is stored
140in it, you can use the following macros to check the type of SV you have.
141
142 SvIOK(SV*)
143 SvNOK(SV*)
144 SvPOK(SV*)
145
146You can get and set the current length of the string stored in an SV with
147the following macros:
148
149 SvCUR(SV*)
150 SvCUR_set(SV*, I32 val)
151
cb1a09d0
AD
152You can also get a pointer to the end of the string stored in the SV
153with the macro:
154
155 SvEND(SV*)
156
157But note that these last three macros are valid only if C<SvPOK()> is true.
a0d0e21e 158
d1b91892
AD
159If you want to append something to the end of string stored in an C<SV*>,
160you can use the following functions:
161
08105a92 162 void sv_catpv(SV*, const char*);
e65f3abd 163 void sv_catpvn(SV*, const char*, STRLEN);
46fc3d4c 164 void sv_catpvf(SV*, const char*, ...);
9abd00ed 165 void sv_catpvfn(SV*, const char*, STRLEN, va_list *, SV **, I32, bool);
d1b91892
AD
166 void sv_catsv(SV*, SV*);
167
168The first function calculates the length of the string to be appended by
169using C<strlen>. In the second, you specify the length of the string
46fc3d4c 170yourself. The third function processes its arguments like C<sprintf> and
9abd00ed
GS
171appends the formatted output. The fourth function works like C<vsprintf>.
172You can specify the address and length of an array of SVs instead of the
173va_list argument. The fifth function extends the string stored in the first
174SV with the string stored in the second SV. It also forces the second SV
175to be interpreted as a string.
176
177The C<sv_cat*()> functions are not generic enough to operate on values that
178have "magic". See L<Magic Virtual Tables> later in this document.
d1b91892 179
a0d0e21e
LW
180If you know the name of a scalar variable, you can get a pointer to its SV
181by using the following:
182
4929bf7b 183 SV* get_sv("package::varname", FALSE);
a0d0e21e
LW
184
185This returns NULL if the variable does not exist.
186
d1b91892 187If you want to know if this variable (or any other SV) is actually C<defined>,
a0d0e21e
LW
188you can call:
189
190 SvOK(SV*)
191
9cde0e7f 192The scalar C<undef> value is stored in an SV instance called C<PL_sv_undef>. Its
a0d0e21e
LW
193address can be used whenever an C<SV*> is needed.
194
9cde0e7f
GS
195There are also the two values C<PL_sv_yes> and C<PL_sv_no>, which contain Boolean
196TRUE and FALSE values, respectively. Like C<PL_sv_undef>, their addresses can
a0d0e21e
LW
197be used whenever an C<SV*> is needed.
198
9cde0e7f 199Do not be fooled into thinking that C<(SV *) 0> is the same as C<&PL_sv_undef>.
a0d0e21e
LW
200Take this code:
201
202 SV* sv = (SV*) 0;
203 if (I-am-to-return-a-real-value) {
204 sv = sv_2mortal(newSViv(42));
205 }
206 sv_setsv(ST(0), sv);
207
208This code tries to return a new SV (which contains the value 42) if it should
04343c6d 209return a real value, or undef otherwise. Instead it has returned a NULL
a0d0e21e 210pointer which, somewhere down the line, will cause a segmentation violation,
9cde0e7f 211bus error, or just weird results. Change the zero to C<&PL_sv_undef> in the first
5f05dabc 212line and all will be well.
a0d0e21e
LW
213
214To free an SV that you've created, call C<SvREFCNT_dec(SV*)>. Normally this
3fe9a6f1 215call is not necessary (see L<Reference Counts and Mortality>).
a0d0e21e 216
94dde4fb
SC
217=head2 Offsets
218
219Perl provides the function C<sv_chop> to efficiently remove characters
220from the beginning of a string; you give it an SV and a pointer to
221somewhere inside the the PV, and it discards everything before the
222pointer. The efficiency comes by means of a little hack: instead of
223actually 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
225effect, and it puts the number of bytes chopped off into the IV field
226of the SV. It then moves the PV pointer (called C<SvPVX>) forward that
227many bytes, and adjusts C<SvCUR> and C<SvLEN>.
228
229Hence, at this point, the start of the buffer that we allocated lives
230at C<SvPVX(sv) - SvIV(sv)> in memory and the PV pointer is pointing
231into the middle of this allocated storage.
232
233This is best demonstrated by example:
234
235 % ./perl -Ilib -MDevel::Peek -le '$a="12345"; $a=~s/.//; Dump($a)'
236 SV = PVIV(0x8128450) at 0x81340f0
237 REFCNT = 1
238 FLAGS = (POK,OOK,pPOK)
239 IV = 1 (OFFSET)
240 PV = 0x8135781 ( "1" . ) "2345"\0
241 CUR = 4
242 LEN = 5
243
244Here the number of bytes chopped off (1) is put into IV, and
245C<Devel::Peek::Dump> helpfully reminds us that this is an offset. The
246portion of the string between the "real" and the "fake" beginnings is
247shown in parentheses, and the values of C<SvCUR> and C<SvLEN> reflect
248the fake beginning, not the real one.
249
d1b91892 250=head2 What's Really Stored in an SV?
a0d0e21e
LW
251
252Recall that the usual method of determining the type of scalar you have is
5f05dabc 253to use C<Sv*OK> macros. Because a scalar can be both a number and a string,
d1b91892 254usually these macros will always return TRUE and calling the C<Sv*V>
a0d0e21e
LW
255macros will do the appropriate conversion of string to integer/double or
256integer/double to string.
257
258If you I<really> need to know if you have an integer, double, or string
259pointer in an SV, you can use the following three macros instead:
260
261 SvIOKp(SV*)
262 SvNOKp(SV*)
263 SvPOKp(SV*)
264
265These will tell you if you truly have an integer, double, or string pointer
d1b91892 266stored in your SV. The "p" stands for private.
a0d0e21e 267
07fa94a1 268In general, though, it's best to use the C<Sv*V> macros.
a0d0e21e 269
54310121 270=head2 Working with AVs
a0d0e21e 271
07fa94a1
JO
272There are two ways to create and load an AV. The first method creates an
273empty AV:
a0d0e21e
LW
274
275 AV* newAV();
276
54310121 277The second method both creates the AV and initially populates it with SVs:
a0d0e21e
LW
278
279 AV* av_make(I32 num, SV **ptr);
280
5f05dabc 281The second argument points to an array containing C<num> C<SV*>'s. Once the
54310121 282AV has been created, the SVs can be destroyed, if so desired.
a0d0e21e 283
54310121 284Once the AV has been created, the following operations are possible on AVs:
a0d0e21e
LW
285
286 void av_push(AV*, SV*);
287 SV* av_pop(AV*);
288 SV* av_shift(AV*);
289 void av_unshift(AV*, I32 num);
290
291These should be familiar operations, with the exception of C<av_unshift>.
292This routine adds C<num> elements at the front of the array with the C<undef>
293value. You must then use C<av_store> (described below) to assign values
294to these new elements.
295
296Here are some other functions:
297
5f05dabc 298 I32 av_len(AV*);
a0d0e21e 299 SV** av_fetch(AV*, I32 key, I32 lval);
a0d0e21e 300 SV** av_store(AV*, I32 key, SV* val);
a0d0e21e 301
5f05dabc
PP
302The C<av_len> function returns the highest index value in array (just
303like $#array in Perl). If the array is empty, -1 is returned. The
304C<av_fetch> function returns the value at index C<key>, but if C<lval>
305is non-zero, then C<av_fetch> will store an undef value at that index.
04343c6d
GS
306The C<av_store> function stores the value C<val> at index C<key>, and does
307not increment the reference count of C<val>. Thus the caller is responsible
308for taking care of that, and if C<av_store> returns NULL, the caller will
309have to decrement the reference count to avoid a memory leak. Note that
310C<av_fetch> and C<av_store> both return C<SV**>'s, not C<SV*>'s as their
311return value.
d1b91892 312
a0d0e21e 313 void av_clear(AV*);
a0d0e21e 314 void av_undef(AV*);
cb1a09d0 315 void av_extend(AV*, I32 key);
5f05dabc
PP
316
317The C<av_clear> function deletes all the elements in the AV* array, but
318does not actually delete the array itself. The C<av_undef> function will
319delete all the elements in the array plus the array itself. The
adc882cf
GS
320C<av_extend> function extends the array so that it contains at least C<key+1>
321elements. If C<key+1> is less than the currently allocated length of the array,
322then nothing is done.
a0d0e21e
LW
323
324If you know the name of an array variable, you can get a pointer to its AV
325by using the following:
326
4929bf7b 327 AV* get_av("package::varname", FALSE);
a0d0e21e
LW
328
329This returns NULL if the variable does not exist.
330
04343c6d
GS
331See L<Understanding the Magic of Tied Hashes and Arrays> for more
332information on how to use the array access functions on tied arrays.
333
54310121 334=head2 Working with HVs
a0d0e21e
LW
335
336To create an HV, you use the following routine:
337
338 HV* newHV();
339
54310121 340Once the HV has been created, the following operations are possible on HVs:
a0d0e21e 341
08105a92
GS
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);
a0d0e21e 344
5f05dabc
PP
345The C<klen> parameter is the length of the key being passed in (Note that
346you cannot pass 0 in as a value of C<klen> to tell Perl to measure the
347length of the key). The C<val> argument contains the SV pointer to the
54310121 348scalar being stored, and C<hash> is the precomputed hash value (zero if
5f05dabc
PP
349you want C<hv_store> to calculate it for you). The C<lval> parameter
350indicates whether this fetch is actually a part of a store operation, in
351which case a new undefined value will be added to the HV with the supplied
352key and C<hv_fetch> will return as if the value had already existed.
a0d0e21e 353
5f05dabc
PP
354Remember that C<hv_store> and C<hv_fetch> return C<SV**>'s and not just
355C<SV*>. To access the scalar value, you must first dereference the return
356value. However, you should check to make sure that the return value is
357not NULL before dereferencing it.
a0d0e21e
LW
358
359These two functions check if a hash table entry exists, and deletes it.
360
08105a92
GS
361 bool hv_exists(HV*, const char* key, U32 klen);
362 SV* hv_delete(HV*, const char* key, U32 klen, I32 flags);
a0d0e21e 363
5f05dabc
PP
364If C<flags> does not include the C<G_DISCARD> flag then C<hv_delete> will
365create and return a mortal copy of the deleted value.
366
a0d0e21e
LW
367And more miscellaneous functions:
368
369 void hv_clear(HV*);
a0d0e21e 370 void hv_undef(HV*);
5f05dabc
PP
371
372Like their AV counterparts, C<hv_clear> deletes all the entries in the hash
373table but does not actually delete the hash table. The C<hv_undef> deletes
374both the entries and the hash table itself.
a0d0e21e 375
d1b91892
AD
376Perl keeps the actual data in linked list of structures with a typedef of HE.
377These contain the actual key and value pointers (plus extra administrative
378overhead). The key is a string pointer; the value is an C<SV*>. However,
379once you have an C<HE*>, to get the actual key and value, use the routines
380specified below.
381
a0d0e21e
LW
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 */
cb1a09d0 390 SV* hv_iterval(HV*, HE* entry);
a0d0e21e
LW
391 /* Return a SV pointer to the value of the HE
392 structure */
cb1a09d0 393 SV* hv_iternextsv(HV*, char** key, I32* retlen);
d1b91892
AD
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 */
a0d0e21e
LW
398
399If you know the name of a hash variable, you can get a pointer to its HV
400by using the following:
401
4929bf7b 402 HV* get_hv("package::varname", FALSE);
a0d0e21e
LW
403
404This returns NULL if the variable does not exist.
405
8ebc5c01 406The hash algorithm is defined in the C<PERL_HASH(hash, key, klen)> macro:
a0d0e21e 407
a0d0e21e 408 hash = 0;
ab192400
GS
409 while (klen--)
410 hash = (hash * 33) + *key++;
87275199 411 hash = hash + (hash >> 5); /* after 5.6 */
ab192400 412
87275199 413The last step was added in version 5.6 to improve distribution of
ab192400 414lower bits in the resulting hash value.
a0d0e21e 415
04343c6d
GS
416See L<Understanding the Magic of Tied Hashes and Arrays> for more
417information on how to use the hash access functions on tied hashes.
418
1e422769
PP
419=head2 Hash API Extensions
420
421Beginning with version 5.004, the following functions are also supported:
422
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);
c47ff5f1 425
1e422769
PP
426 bool hv_exists_ent (HV* tb, SV* key, U32 hash);
427 SV* hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
c47ff5f1 428
1e422769
PP
429 SV* hv_iterkeysv (HE* entry);
430
431Note that these functions take C<SV*> keys, which simplifies writing
432of extension code that deals with hash structures. These functions
433also allow passing of C<SV*> keys to C<tie> functions without forcing
434you to stringify the keys (unlike the previous set of functions).
435
436They also return and accept whole hash entries (C<HE*>), making their
437use more efficient (since the hash number for a particular string
4a4eefd0
GS
438doesn't have to be recomputed every time). See L<perlapi> for detailed
439descriptions.
1e422769
PP
440
441The following macros must always be used to access the contents of hash
442entries. Note that the arguments to these macros must be simple
443variables, since they may get evaluated more than once. See
4a4eefd0 444L<perlapi> for detailed descriptions of these macros.
1e422769
PP
445
446 HePV(HE* he, STRLEN len)
447 HeVAL(HE* he)
448 HeHASH(HE* he)
449 HeSVKEY(HE* he)
450 HeSVKEY_force(HE* he)
451 HeSVKEY_set(HE* he, SV* sv)
452
453These two lower level macros are defined, but must only be used when
454dealing with keys that are not C<SV*>s:
455
456 HeKEY(HE* he)
457 HeKLEN(HE* he)
458
04343c6d
GS
459Note that both C<hv_store> and C<hv_store_ent> do not increment the
460reference count of the stored C<val>, which is the caller's responsibility.
461If these functions return a NULL value, the caller will usually have to
462decrement the reference count of C<val> to avoid a memory leak.
1e422769 463
a0d0e21e
LW
464=head2 References
465
d1b91892
AD
466References are a special type of scalar that point to other data types
467(including references).
a0d0e21e 468
07fa94a1 469To create a reference, use either of the following functions:
a0d0e21e 470
5f05dabc
PP
471 SV* newRV_inc((SV*) thing);
472 SV* newRV_noinc((SV*) thing);
a0d0e21e 473
5f05dabc 474The C<thing> argument can be any of an C<SV*>, C<AV*>, or C<HV*>. The
07fa94a1
JO
475functions are identical except that C<newRV_inc> increments the reference
476count of the C<thing>, while C<newRV_noinc> does not. For historical
477reasons, C<newRV> is a synonym for C<newRV_inc>.
478
479Once you have a reference, you can use the following macro to dereference
480the reference:
a0d0e21e
LW
481
482 SvRV(SV*)
483
484then call the appropriate routines, casting the returned C<SV*> to either an
d1b91892 485C<AV*> or C<HV*>, if required.
a0d0e21e 486
d1b91892 487To determine if an SV is a reference, you can use the following macro:
a0d0e21e
LW
488
489 SvROK(SV*)
490
07fa94a1
JO
491To discover what type of value the reference refers to, use the following
492macro and then check the return value.
d1b91892
AD
493
494 SvTYPE(SvRV(SV*))
495
496The most useful types that will be returned are:
497
498 SVt_IV Scalar
499 SVt_NV Scalar
500 SVt_PV Scalar
5f05dabc 501 SVt_RV Scalar
d1b91892
AD
502 SVt_PVAV Array
503 SVt_PVHV Hash
504 SVt_PVCV Code
5f05dabc
PP
505 SVt_PVGV Glob (possible a file handle)
506 SVt_PVMG Blessed or Magical Scalar
507
508 See the sv.h header file for more details.
d1b91892 509
cb1a09d0
AD
510=head2 Blessed References and Class Objects
511
512References are also used to support object-oriented programming. In the
513OO lexicon, an object is simply a reference that has been blessed into a
514package (or class). Once blessed, the programmer may now use the reference
515to access the various methods in the class.
516
517A reference can be blessed into a package with the following function:
518
519 SV* sv_bless(SV* sv, HV* stash);
520
521The C<sv> argument must be a reference. The C<stash> argument specifies
3fe9a6f1 522which class the reference will belong to. See
2ae324a7 523L<Stashes and Globs> for information on converting class names into stashes.
cb1a09d0
AD
524
525/* Still under construction */
526
527Upgrades rv to reference if not already one. Creates new SV for rv to
8ebc5c01
PP
528point to. If C<classname> is non-null, the SV is blessed into the specified
529class. SV is returned.
cb1a09d0 530
08105a92 531 SV* newSVrv(SV* rv, const char* classname);
cb1a09d0 532
8ebc5c01
PP
533Copies integer or double into an SV whose reference is C<rv>. SV is blessed
534if C<classname> is non-null.
cb1a09d0 535
08105a92
GS
536 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
537 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
cb1a09d0 538
5f05dabc 539Copies the pointer value (I<the address, not the string!>) into an SV whose
8ebc5c01 540reference is rv. SV is blessed if C<classname> is non-null.
cb1a09d0 541
08105a92 542 SV* sv_setref_pv(SV* rv, const char* classname, PV iv);
cb1a09d0 543
8ebc5c01
PP
544Copies string into an SV whose reference is C<rv>. Set length to 0 to let
545Perl calculate the string length. SV is blessed if C<classname> is non-null.
cb1a09d0 546
e65f3abd 547 SV* sv_setref_pvn(SV* rv, const char* classname, PV iv, STRLEN length);
cb1a09d0 548
9abd00ed
GS
549Tests whether the SV is blessed into the specified class. It does not
550check inheritance relationships.
551
08105a92 552 int sv_isa(SV* sv, const char* name);
9abd00ed
GS
553
554Tests whether the SV is a reference to a blessed object.
555
556 int sv_isobject(SV* sv);
557
558Tests whether the SV is derived from the specified class. SV can be either
559a reference to a blessed object or a string containing a class name. This
560is the function implementing the C<UNIVERSAL::isa> functionality.
561
08105a92 562 bool sv_derived_from(SV* sv, const char* name);
9abd00ed
GS
563
564To check if you've got an object derived from a specific class you have
565to write:
566
567 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
cb1a09d0 568
5f05dabc 569=head2 Creating New Variables
cb1a09d0 570
5f05dabc
PP
571To create a new Perl variable with an undef value which can be accessed from
572your Perl script, use the following routines, depending on the variable type.
cb1a09d0 573
4929bf7b
GS
574 SV* get_sv("package::varname", TRUE);
575 AV* get_av("package::varname", TRUE);
576 HV* get_hv("package::varname", TRUE);
cb1a09d0
AD
577
578Notice the use of TRUE as the second parameter. The new variable can now
579be set, using the routines appropriate to the data type.
580
5f05dabc
PP
581There are additional macros whose values may be bitwise OR'ed with the
582C<TRUE> argument to enable certain extra features. Those bits are:
cb1a09d0 583
5f05dabc 584 GV_ADDMULTI Marks the variable as multiply defined, thus preventing the
54310121 585 "Name <varname> used only once: possible typo" warning.
07fa94a1
JO
586 GV_ADDWARN Issues the warning "Had to create <varname> unexpectedly" if
587 the variable did not exist before the function was called.
cb1a09d0 588
07fa94a1
JO
589If you do not specify a package name, the variable is created in the current
590package.
cb1a09d0 591
5f05dabc 592=head2 Reference Counts and Mortality
a0d0e21e 593
54310121
PP
594Perl uses an reference count-driven garbage collection mechanism. SVs,
595AVs, or HVs (xV for short in the following) start their life with a
55497cff 596reference count of 1. If the reference count of an xV ever drops to 0,
07fa94a1 597then it will be destroyed and its memory made available for reuse.
55497cff
PP
598
599This normally doesn't happen at the Perl level unless a variable is
5f05dabc
PP
600undef'ed or the last variable holding a reference to it is changed or
601overwritten. At the internal level, however, reference counts can be
55497cff
PP
602manipulated with the following macros:
603
604 int SvREFCNT(SV* sv);
5f05dabc 605 SV* SvREFCNT_inc(SV* sv);
55497cff
PP
606 void SvREFCNT_dec(SV* sv);
607
608However, there is one other function which manipulates the reference
07fa94a1
JO
609count of its argument. The C<newRV_inc> function, you will recall,
610creates a reference to the specified argument. As a side effect,
611it increments the argument's reference count. If this is not what
612you want, use C<newRV_noinc> instead.
613
614For example, imagine you want to return a reference from an XSUB function.
615Inside the XSUB routine, you create an SV which initially has a reference
616count of one. Then you call C<newRV_inc>, passing it the just-created SV.
5f05dabc
PP
617This returns the reference as a new SV, but the reference count of the
618SV you passed to C<newRV_inc> has been incremented to two. Now you
07fa94a1
JO
619return the reference from the XSUB routine and forget about the SV.
620But Perl hasn't! Whenever the returned reference is destroyed, the
621reference count of the original SV is decreased to one and nothing happens.
622The SV will hang around without any way to access it until Perl itself
623terminates. This is a memory leak.
5f05dabc
PP
624
625The correct procedure, then, is to use C<newRV_noinc> instead of
faed5253
JO
626C<newRV_inc>. Then, if and when the last reference is destroyed,
627the reference count of the SV will go to zero and it will be destroyed,
07fa94a1 628stopping any memory leak.
55497cff 629
5f05dabc 630There are some convenience functions available that can help with the
54310121 631destruction of xVs. These functions introduce the concept of "mortality".
07fa94a1
JO
632An xV that is mortal has had its reference count marked to be decremented,
633but not actually decremented, until "a short time later". Generally the
634term "short time later" means a single Perl statement, such as a call to
54310121 635an XSUB function. The actual determinant for when mortal xVs have their
07fa94a1
JO
636reference count decremented depends on two macros, SAVETMPS and FREETMPS.
637See L<perlcall> and L<perlxs> for more details on these macros.
55497cff
PP
638
639"Mortalization" then is at its simplest a deferred C<SvREFCNT_dec>.
640However, if you mortalize a variable twice, the reference count will
641later be decremented twice.
642
643You should be careful about creating mortal variables. Strange things
644can happen if you make the same value mortal within multiple contexts,
5f05dabc 645or if you make a variable mortal multiple times.
a0d0e21e
LW
646
647To create a mortal variable, use the functions:
648
649 SV* sv_newmortal()
650 SV* sv_2mortal(SV*)
651 SV* sv_mortalcopy(SV*)
652
5f05dabc
PP
653The first call creates a mortal SV, the second converts an existing
654SV to a mortal SV (and thus defers a call to C<SvREFCNT_dec>), and the
655third creates a mortal copy of an existing SV.
a0d0e21e 656
54310121 657The mortal routines are not just for SVs -- AVs and HVs can be
faed5253 658made mortal by passing their address (type-casted to C<SV*>) to the
07fa94a1 659C<sv_2mortal> or C<sv_mortalcopy> routines.
a0d0e21e 660
5f05dabc 661=head2 Stashes and Globs
a0d0e21e 662
aa689395
PP
663A "stash" is a hash that contains all of the different objects that
664are contained within a package. Each key of the stash is a symbol
665name (shared by all the different types of objects that have the same
666name), and each value in the hash table is a GV (Glob Value). This GV
667in turn contains references to the various objects of that name,
668including (but not limited to) the following:
cb1a09d0 669
a0d0e21e
LW
670 Scalar Value
671 Array Value
672 Hash Value
a3cb178b 673 I/O Handle
a0d0e21e
LW
674 Format
675 Subroutine
676
9cde0e7f 677There is a single stash called "PL_defstash" that holds the items that exist
5f05dabc
PP
678in the "main" package. To get at the items in other packages, append the
679string "::" to the package name. The items in the "Foo" package are in
9cde0e7f 680the stash "Foo::" in PL_defstash. The items in the "Bar::Baz" package are
5f05dabc 681in the stash "Baz::" in "Bar::"'s stash.
a0d0e21e 682
d1b91892 683To get the stash pointer for a particular package, use the function:
a0d0e21e 684
08105a92 685 HV* gv_stashpv(const char* name, I32 create)
a0d0e21e
LW
686 HV* gv_stashsv(SV*, I32 create)
687
688The first function takes a literal string, the second uses the string stored
d1b91892 689in the SV. Remember that a stash is just a hash table, so you get back an
cb1a09d0 690C<HV*>. The C<create> flag will create a new package if it is set.
a0d0e21e
LW
691
692The name that C<gv_stash*v> wants is the name of the package whose symbol table
693you want. The default package is called C<main>. If you have multiply nested
d1b91892
AD
694packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
695language itself.
a0d0e21e
LW
696
697Alternately, if you have an SV that is a blessed reference, you can find
698out the stash pointer by using:
699
700 HV* SvSTASH(SvRV(SV*));
701
702then use the following to get the package name itself:
703
704 char* HvNAME(HV* stash);
705
5f05dabc
PP
706If you need to bless or re-bless an object you can use the following
707function:
a0d0e21e
LW
708
709 SV* sv_bless(SV*, HV* stash)
710
711where the first argument, an C<SV*>, must be a reference, and the second
712argument is a stash. The returned C<SV*> can now be used in the same way
713as any other SV.
714
d1b91892
AD
715For more information on references and blessings, consult L<perlref>.
716
54310121 717=head2 Double-Typed SVs
0a753a76
PP
718
719Scalar variables normally contain only one type of value, an integer,
720double, pointer, or reference. Perl will automatically convert the
721actual scalar data from the stored type into the requested type.
722
723Some scalar variables contain more than one type of scalar data. For
724example, the variable C<$!> contains either the numeric value of C<errno>
725or its string equivalent from either C<strerror> or C<sys_errlist[]>.
726
727To force multiple data values into an SV, you must do two things: use the
728C<sv_set*v> routines to add the additional scalar type, then set a flag
729so that Perl will believe it contains more than one type of data. The
730four macros to set the flags are:
731
732 SvIOK_on
733 SvNOK_on
734 SvPOK_on
735 SvROK_on
736
737The particular macro you must use depends on which C<sv_set*v> routine
738you called first. This is because every C<sv_set*v> routine turns on
739only the bit for the particular type of data being set, and turns off
740all the rest.
741
742For example, to create a new Perl variable called "dberror" that contains
743both the numeric and descriptive string error values, you could use the
744following code:
745
746 extern int dberror;
747 extern char *dberror_list;
748
4929bf7b 749 SV* sv = get_sv("dberror", TRUE);
0a753a76
PP
750 sv_setiv(sv, (IV) dberror);
751 sv_setpv(sv, dberror_list[dberror]);
752 SvIOK_on(sv);
753
754If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
755macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
756
757=head2 Magic Variables
a0d0e21e 758
d1b91892
AD
759[This section still under construction. Ignore everything here. Post no
760bills. Everything not permitted is forbidden.]
761
d1b91892
AD
762Any SV may be magical, that is, it has special features that a normal
763SV does not have. These features are stored in the SV structure in a
5f05dabc 764linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
d1b91892
AD
765
766 struct magic {
767 MAGIC* mg_moremagic;
768 MGVTBL* mg_virtual;
769 U16 mg_private;
770 char mg_type;
771 U8 mg_flags;
772 SV* mg_obj;
773 char* mg_ptr;
774 I32 mg_len;
775 };
776
777Note this is current as of patchlevel 0, and could change at any time.
778
779=head2 Assigning Magic
780
781Perl adds magic to an SV using the sv_magic function:
782
08105a92 783 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
d1b91892
AD
784
785The C<sv> argument is a pointer to the SV that is to acquire a new magical
786feature.
787
788If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
789set the C<SVt_PVMG> flag for the C<sv>. Perl then continues by adding
790it to the beginning of the linked list of magical features. Any prior
791entry of the same type of magic is deleted. Note that this can be
5fb8527f 792overridden, and multiple instances of the same type of magic can be
d1b91892
AD
793associated with an SV.
794
54310121
PP
795The C<name> and C<namlen> arguments are used to associate a string with
796the magic, typically the name of a variable. C<namlen> is stored in the
797C<mg_len> field and if C<name> is non-null and C<namlen> >= 0 a malloc'd
d1b91892
AD
798copy of the name is stored in C<mg_ptr> field.
799
800The 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.
cb1a09d0
AD
802See the "Magic Virtual Table" section below. The C<how> argument is also
803stored in the C<mg_type> field.
d1b91892
AD
804
805The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
806structure. If it is not the same as the C<sv> argument, the reference
807count of the C<obj> object is incremented. If it is the same, or if
04343c6d 808the C<how> argument is "#", or if it is a NULL pointer, then C<obj> is
d1b91892
AD
809merely stored, without the reference count being incremented.
810
cb1a09d0
AD
811There is also a function to add magic to an C<HV>:
812
813 void hv_magic(HV *hv, GV *gv, int how);
814
815This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
816
817To remove the magic from an SV, call the function sv_unmagic:
818
819 void sv_unmagic(SV *sv, int type);
820
821The C<type> argument should be equal to the C<how> value when the C<SV>
822was initially made magical.
823
d1b91892
AD
824=head2 Magic Virtual Tables
825
826The C<mg_virtual> field in the C<MAGIC> structure is a pointer to a
827C<MGVTBL>, which is a structure of function pointers and stands for
828"Magic Virtual Table" to handle the various operations that might be
829applied to that variable.
830
831The C<MGVTBL> has five pointers to the following routine types:
832
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);
838
839This MGVTBL structure is set at compile-time in C<perl.h> and there are
840currently 19 types (or 21 with overloading turned on). These different
841structures contain pointers to various routines that perform additional
842actions depending on which function is being called.
843
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.
851
852For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
853to an C<mg_type> of '\0') contains:
854
855 { magic_get, magic_set, magic_len, 0, 0 }
856
857Thus, when an SV is determined to be magical and of type '\0', if a get
858operation is being performed, the routine C<magic_get> is called. All
859the various routines for the various magical types begin with C<magic_>.
954c1994
GS
860NOTE: the magic routines are not considered part of the Perl API, and may
861not be exported by the Perl library.
d1b91892
AD
862
863The current kinds of Magic Virtual Tables are:
864
bdbeb323 865 mg_type MGVTBL Type of magic
5f05dabc 866 ------- ------ ----------------------------
bdbeb323
SM
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)
c2e66d9e
GS
872 D vtbl_regdata Regex match position data (@+ and @- vars)
873 d vtbl_regdatum Regex match position data element
d1b91892
AD
874 E vtbl_env %ENV hash
875 e vtbl_envelem %ENV hash element
bdbeb323
SM
876 f vtbl_fm Formline ('compiled' format)
877 g vtbl_mglob m//g target / study()ed string
d1b91892
AD
878 I vtbl_isa @ISA array
879 i vtbl_isaelem @ISA array element
bdbeb323
SM
880 k vtbl_nkeys scalar(keys()) lvalue
881 L (none) Debugger %_<filename
882 l vtbl_dbline Debugger %_<filename element
44a8e56a 883 o vtbl_collxfrm Locale transformation
bdbeb323
SM
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
887 S vtbl_sig %SIG hash
888 s vtbl_sigelem %SIG hash element
d1b91892 889 t vtbl_taint Taintedness
bdbeb323
SM
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
d1b91892 899
68dc0745
PP
900When an uppercase and lowercase letter both exist in the table, then the
901uppercase letter is used to represent some kind of composite type (a list
902or a hash), and the lowercase letter is used to represent an element of
d1b91892
AD
903that composite type.
904
bdbeb323
SM
905The '~' and 'U' magic types are defined specifically for use by
906extensions and will not be used by perl itself. Extensions can use
907'~' magic to 'attach' private information to variables (typically
908objects). This is especially useful because there is no way for
909normal perl code to corrupt this private information (unlike using
910extra elements of a hash object).
911
912Similarly, 'U' magic can be used much like tie() to call a C function
913any time a scalar's value is used or changed. The C<MAGIC>'s
914C<mg_ptr> field points to a C<ufuncs> structure:
915
916 struct ufuncs {
917 I32 (*uf_val)(IV, SV*);
918 I32 (*uf_set)(IV, SV*);
919 IV uf_index;
920 };
921
922When the SV is read from or written to, the C<uf_val> or C<uf_set>
923function will be called with C<uf_index> as the first arg and a
1526ead6
AB
924pointer to the SV as the second. A simple example of how to add 'U'
925magic is shown below. Note that the ufuncs structure is copied by
926sv_magic, so you can safely allocate it on the stack.
927
928 void
929 Umagic(sv)
930 SV *sv;
931 PREINIT:
932 struct ufuncs uf;
933 CODE:
934 uf.uf_val = &my_get_fn;
935 uf.uf_set = &my_set_fn;
936 uf.uf_index = 0;
937 sv_magic(sv, 0, 'U', (char*)&uf, sizeof(uf));
5f05dabc 938
bdbeb323
SM
939Note that because multiple extensions may be using '~' or 'U' magic,
940it is important for extensions to take extra care to avoid conflict.
941Typically only using the magic on objects blessed into the same class
942as the extension is sufficient. For '~' magic, it may also be
943appropriate to add an I32 'signature' at the top of the private data
944area and check that.
5f05dabc 945
ef50df4b
GS
946Also note that the C<sv_set*()> and C<sv_cat*()> functions described
947earlier do B<not> invoke 'set' magic on their targets. This must
948be done by the user either by calling the C<SvSETMAGIC()> macro after
949calling these functions, or by using one of the C<sv_set*_mg()> or
950C<sv_cat*_mg()> functions. Similarly, generic C code must call the
951C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
952obtained from external sources in functions that don't handle magic.
4a4eefd0 953See L<perlapi> for a description of these functions.
189b2af5
GS
954For example, calls to the C<sv_cat*()> functions typically need to be
955followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
956since their implementation handles 'get' magic.
957
d1b91892
AD
958=head2 Finding Magic
959
960 MAGIC* mg_find(SV*, int type); /* Finds the magic pointer of that type */
961
962This routine returns a pointer to the C<MAGIC> structure stored in the SV.
963If the SV does not have that magical feature, C<NULL> is returned. Also,
54310121 964if the SV is not of type SVt_PVMG, Perl may core dump.
d1b91892 965
08105a92 966 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
d1b91892
AD
967
968This routine checks to see what types of magic C<sv> has. If the mg_type
68dc0745
PP
969field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
970the mg_type field is changed to be the lowercase letter.
a0d0e21e 971
04343c6d
GS
972=head2 Understanding the Magic of Tied Hashes and Arrays
973
974Tied hashes and arrays are magical beasts of the 'P' magic type.
9edb2b46
GS
975
976WARNING: As of the 5.004 release, proper usage of the array and hash
977access functions requires understanding a few caveats. Some
978of these caveats are actually considered bugs in the API, to be fixed
979in later releases, and are bracketed with [MAYCHANGE] below. If
980you find yourself actually applying such information in this section, be
981aware that the behavior may change in the future, umm, without warning.
04343c6d 982
1526ead6
AB
983The perl tie function associates a variable with an object that implements
984the various GET, SET etc methods. To perform the equivalent of the perl
985tie function from an XSUB, you must mimic this behaviour. The code below
986carries out the necessary steps - firstly it creates a new hash, and then
987creates a second hash which it blesses into the class which will implement
988the tie methods. Lastly it ties the two hashes together, and returns a
989reference to the new tied hash. Note that the code below does NOT call the
990TIEHASH method in the MyTie class -
991see L<Calling Perl Routines from within C Programs> for details on how
992to do this.
993
994 SV*
995 mytie()
996 PREINIT:
997 HV *hash;
998 HV *stash;
999 SV *tie;
1000 CODE:
1001 hash = newHV();
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);
1007 OUTPUT:
1008 RETVAL
1009
04343c6d
GS
1010The C<av_store> function, when given a tied array argument, merely
1011copies the magic of the array onto the value to be "stored", using
1012C<mg_copy>. It may also return NULL, indicating that the value did not
9edb2b46
GS
1013actually need to be stored in the array. [MAYCHANGE] After a call to
1014C<av_store> on a tied array, the caller will usually need to call
1015C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1016TIEARRAY object. If C<av_store> did return NULL, a call to
1017C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1018leak. [/MAYCHANGE]
04343c6d
GS
1019
1020The previous paragraph is applicable verbatim to tied hash access using the
1021C<hv_store> and C<hv_store_ent> functions as well.
1022
1023C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1024C<hv_fetch_ent> actually return an undefined mortal value whose magic
1025has been initialized using C<mg_copy>. Note the value so returned does not
9edb2b46
GS
1026need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1027need to call C<mg_get()> on the returned value in order to actually invoke
1028the perl level "FETCH" method on the underlying TIE object. Similarly,
04343c6d
GS
1029you may also call C<mg_set()> on the return value after possibly assigning
1030a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
9edb2b46 1031method on the TIE object. [/MAYCHANGE]
04343c6d 1032
9edb2b46 1033[MAYCHANGE]
04343c6d
GS
1034In other words, the array or hash fetch/store functions don't really
1035fetch and store actual values in the case of tied arrays and hashes. They
1036merely 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
1038do the job of invoking the TIE methods on the underlying objects. Thus
9edb2b46 1039the magic mechanism currently implements a kind of lazy access to arrays
04343c6d
GS
1040and hashes.
1041
1042Currently (as of perl version 5.004), use of the hash and array access
1043functions requires the user to be aware of whether they are operating on
9edb2b46
GS
1044"normal" hashes and arrays, or on their tied variants. The API may be
1045changed to provide more transparent access to both tied and normal data
1046types in future versions.
1047[/MAYCHANGE]
04343c6d
GS
1048
1049You would do well to understand that the TIEARRAY and TIEHASH interfaces
1050are mere sugar to invoke some perl method calls while using the uniform hash
1051and array syntax. The use of this sugar imposes some overhead (typically
1052about two to four extra opcodes per FETCH/STORE operation, in addition to
1053the creation of all the mortal variables required to invoke the methods).
1054This overhead will be comparatively small if the TIE methods are themselves
1055substantial, but if they are only a few statements long, the overhead
1056will not be insignificant.
1057
d1c897a1
IZ
1058=head2 Localizing changes
1059
1060Perl has a very handy construction
1061
1062 {
1063 local $var = 2;
1064 ...
1065 }
1066
1067This construction is I<approximately> equivalent to
1068
1069 {
1070 my $oldvar = $var;
1071 $var = 2;
1072 ...
1073 $var = $oldvar;
1074 }
1075
1076The biggest difference is that the first construction would
1077reinstate the initial value of $var, irrespective of how control exits
1078the block: C<goto>, C<return>, C<die>/C<eval> etc. It is a little bit
1079more efficient as well.
1080
1081There is a way to achieve a similar task from C via Perl API: create a
1082I<pseudo-block>, and arrange for some changes to be automatically
1083undone at the end of it, either explicit, or via a non-local exit (via
1084die()). A I<block>-like construct is created by a pair of
b687b08b
TC
1085C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1086Such a construct may be created specially for some important localized
1087task, or an existing one (like boundaries of enclosing Perl
1088subroutine/block, or an existing pair for freeing TMPs) may be
1089used. (In the second case the overhead of additional localization must
1090be almost negligible.) Note that any XSUB is automatically enclosed in
1091an C<ENTER>/C<LEAVE> pair.
d1c897a1
IZ
1092
1093Inside such a I<pseudo-block> the following service is available:
1094
13a2d996 1095=over 4
d1c897a1
IZ
1096
1097=item C<SAVEINT(int i)>
1098
1099=item C<SAVEIV(IV i)>
1100
1101=item C<SAVEI32(I32 i)>
1102
1103=item C<SAVELONG(long i)>
1104
1105These macros arrange things to restore the value of integer variable
1106C<i> at the end of enclosing I<pseudo-block>.
1107
1108=item C<SAVESPTR(s)>
1109
1110=item C<SAVEPPTR(p)>
1111
1112These macros arrange things to restore the value of pointers C<s> and
1113C<p>. C<s> must be a pointer of a type which survives conversion to
1114C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1115and back.
1116
1117=item C<SAVEFREESV(SV *sv)>
1118
1119The refcount of C<sv> would be decremented at the end of
1120I<pseudo-block>. This is similar to C<sv_2mortal>, which should (?) be
1121used instead.
1122
1123=item C<SAVEFREEOP(OP *op)>
1124
1125The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1126
1127=item C<SAVEFREEPV(p)>
1128
1129The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1130end of I<pseudo-block>.
1131
1132=item C<SAVECLEARSV(SV *sv)>
1133
1134Clears a slot in the current scratchpad which corresponds to C<sv> at
1135the end of I<pseudo-block>.
1136
1137=item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1138
1139The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
1140string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1141short-lived storage, the corresponding string may be reallocated like
1142this:
1143
9cde0e7f 1144 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
d1c897a1 1145
c76ac1ee 1146=item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
d1c897a1
IZ
1147
1148At the end of I<pseudo-block> the function C<f> is called with the
c76ac1ee
GS
1149only argument C<p>.
1150
1151=item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1152
1153At the end of I<pseudo-block> the function C<f> is called with the
1154implicit context argument (if any), and C<p>.
d1c897a1
IZ
1155
1156=item C<SAVESTACK_POS()>
1157
1158The current offset on the Perl internal stack (cf. C<SP>) is restored
1159at the end of I<pseudo-block>.
1160
1161=back
1162
1163The following API list contains functions, thus one needs to
1164provide pointers to the modifiable data explicitly (either C pointers,
1165or Perlish C<GV *>s). Where the above macros take C<int>, a similar
1166function takes C<int *>.
1167
13a2d996 1168=over 4
d1c897a1
IZ
1169
1170=item C<SV* save_scalar(GV *gv)>
1171
1172Equivalent to Perl code C<local $gv>.
1173
1174=item C<AV* save_ary(GV *gv)>
1175
1176=item C<HV* save_hash(GV *gv)>
1177
1178Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1179
1180=item C<void save_item(SV *item)>
1181
1182Duplicates the current value of C<SV>, on the exit from the current
1183C<ENTER>/C<LEAVE> I<pseudo-block> will restore the value of C<SV>
1184using the stored value.
1185
1186=item C<void save_list(SV **sarg, I32 maxsarg)>
1187
1188A variant of C<save_item> which takes multiple arguments via an array
1189C<sarg> of C<SV*> of length C<maxsarg>.
1190
1191=item C<SV* save_svref(SV **sptr)>
1192
1193Similar to C<save_scalar>, but will reinstate a C<SV *>.
1194
1195=item C<void save_aptr(AV **aptr)>
1196
1197=item C<void save_hptr(HV **hptr)>
1198
1199Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1200
1201=back
1202
1203The C<Alias> module implements localization of the basic types within the
1204I<caller's scope>. People who are interested in how to localize things in
1205the containing scope should take a look there too.
1206
0a753a76 1207=head1 Subroutines
a0d0e21e 1208
68dc0745 1209=head2 XSUBs and the Argument Stack
5f05dabc
PP
1210
1211The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1212An XSUB routine will have a stack that contains the arguments from the Perl
1213program, and a way to map from the Perl data structures to a C equivalent.
1214
1215The stack arguments are accessible through the C<ST(n)> macro, which returns
1216the C<n>'th stack argument. Argument 0 is the first argument passed in the
1217Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1218an C<SV*> is used.
1219
1220Most of the time, output from the C routine can be handled through use of
1221the RETVAL and OUTPUT directives. However, there are some cases where the
1222argument stack is not already long enough to handle all the return values.
1223An example is the POSIX tzname() call, which takes no arguments, but returns
1224two, the local time zone's standard and summer time abbreviations.
1225
1226To handle this situation, the PPCODE directive is used and the stack is
1227extended using the macro:
1228
924508f0 1229 EXTEND(SP, num);
5f05dabc 1230
924508f0
GS
1231where C<SP> is the macro that represents the local copy of the stack pointer,
1232and C<num> is the number of elements the stack should be extended by.
5f05dabc
PP
1233
1234Now that there is room on the stack, values can be pushed on it using the
54310121 1235macros to push IVs, doubles, strings, and SV pointers respectively:
5f05dabc
PP
1236
1237 PUSHi(IV)
1238 PUSHn(double)
1239 PUSHp(char*, I32)
1240 PUSHs(SV*)
1241
1242And now the Perl program calling C<tzname>, the two values will be assigned
1243as in:
1244
1245 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1246
1247An alternate (and possibly simpler) method to pushing values on the stack is
1248to use the macros:
1249
1250 XPUSHi(IV)
1251 XPUSHn(double)
1252 XPUSHp(char*, I32)
1253 XPUSHs(SV*)
1254
1255These macros automatically adjust the stack for you, if needed. Thus, you
1256do not need to call C<EXTEND> to extend the stack.
1257
1258For more information, consult L<perlxs> and L<perlxstut>.
1259
1260=head2 Calling Perl Routines from within C Programs
a0d0e21e
LW
1261
1262There are four routines that can be used to call a Perl subroutine from
1263within a C program. These four are:
1264
954c1994
GS
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**);
a0d0e21e 1269
954c1994 1270The routine most often used is C<call_sv>. The C<SV*> argument
d1b91892
AD
1271contains either the name of the Perl subroutine to be called, or a
1272reference to the subroutine. The second argument consists of flags
1273that control the context in which the subroutine is called, whether
1274or not the subroutine is being passed arguments, how errors should be
1275trapped, and how to treat return values.
a0d0e21e
LW
1276
1277All four routines return the number of arguments that the subroutine returned
1278on the Perl stack.
1279
954c1994
GS
1280These routines used to be called C<perl_call_sv> etc., before Perl v5.6.0,
1281but those names are now deprecated; macros of the same name are provided for
1282compatibility.
1283
1284When using any of these routines (except C<call_argv>), the programmer
d1b91892
AD
1285must manipulate the Perl stack. These include the following macros and
1286functions:
a0d0e21e
LW
1287
1288 dSP
924508f0 1289 SP
a0d0e21e
LW
1290 PUSHMARK()
1291 PUTBACK
1292 SPAGAIN
1293 ENTER
1294 SAVETMPS
1295 FREETMPS
1296 LEAVE
1297 XPUSH*()
cb1a09d0 1298 POP*()
a0d0e21e 1299
5f05dabc
PP
1300For a detailed description of calling conventions from C to Perl,
1301consult L<perlcall>.
a0d0e21e 1302
5f05dabc 1303=head2 Memory Allocation
a0d0e21e 1304
86058a2d
GS
1305All memory meant to be used with the Perl API functions should be manipulated
1306using the macros described in this section. The macros provide the necessary
1307transparency between differences in the actual malloc implementation that is
1308used within perl.
1309
1310It is suggested that you enable the version of malloc that is distributed
5f05dabc 1311with Perl. It keeps pools of various sizes of unallocated memory in
07fa94a1
JO
1312order to satisfy allocation requests more quickly. However, on some
1313platforms, it may cause spurious malloc or free errors.
d1b91892
AD
1314
1315 New(x, pointer, number, type);
1316 Newc(x, pointer, number, type, cast);
1317 Newz(x, pointer, number, type);
1318
07fa94a1 1319These three macros are used to initially allocate memory.
5f05dabc
PP
1320
1321The first argument C<x> was a "magic cookie" that was used to keep track
1322of who called the macro, to help when debugging memory problems. However,
07fa94a1
JO
1323the current code makes no use of this feature (most Perl developers now
1324use run-time memory checkers), so this argument can be any number.
5f05dabc
PP
1325
1326The second argument C<pointer> should be the name of a variable that will
1327point to the newly allocated memory.
d1b91892 1328
d1b91892
AD
1329The third and fourth arguments C<number> and C<type> specify how many of
1330the specified type of data structure should be allocated. The argument
1331C<type> is passed to C<sizeof>. The final argument to C<Newc>, C<cast>,
1332should be used if the C<pointer> argument is different from the C<type>
1333argument.
1334
1335Unlike the C<New> and C<Newc> macros, the C<Newz> macro calls C<memzero>
1336to zero out all the newly allocated memory.
1337
1338 Renew(pointer, number, type);
1339 Renewc(pointer, number, type, cast);
1340 Safefree(pointer)
1341
1342These three macros are used to change a memory buffer size or to free a
1343piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1344match those of C<New> and C<Newc> with the exception of not needing the
1345"magic cookie" argument.
1346
1347 Move(source, dest, number, type);
1348 Copy(source, dest, number, type);
1349 Zero(dest, number, type);
1350
1351These three macros are used to move, copy, or zero out previously allocated
1352memory. The C<source> and C<dest> arguments point to the source and
1353destination starting points. Perl will move, copy, or zero out C<number>
1354instances of the size of the C<type> data structure (using the C<sizeof>
1355function).
a0d0e21e 1356
0cf5025f
SC
1357Here is a handy table of equivalents between ordinary C and Perl's
1358memory abstraction layer:
1359
8d1fbc99
JH
1360 Instead Of: Use:
1361
1362 t* p = malloc(n) New(id, p, n, t)
1363 t* p = calloc(n, s) Newz(id, p, n, t)
1364 p = realloc(p, n) Renew(p, n, t)
1365 memcpy(dst, src, n) Copy(src, dst, n, t)
1366 memmove(dst, src, n) Move(src, dst, n, t)
1367 free(p) Safefree(p)
1368 strdup(p) savepv(p)
1369 strndup(p, n) savepvn(p, n) (Hey, strndup doesn't exist!)
1370 memcpy/*(struct foo *) StructCopy(src, dst, t)
1371
1372 t type
1373 p pointer
1374 ck cookie for the memory region (now unused)
1375 n number of elements
1376 src source pointer
1377 dst destination pointer
1378
1379Notice the different order of arguments to C<Copy> and C<Move> than used
1380in C<memcpy> and C<memmove>.
0cf5025f 1381
5f05dabc 1382=head2 PerlIO
ce3d39e2 1383
5f05dabc
PP
1384The most recent development releases of Perl has been experimenting with
1385removing Perl's dependency on the "normal" standard I/O suite and allowing
1386other stdio implementations to be used. This involves creating a new
1387abstraction layer that then calls whichever implementation of stdio Perl
68dc0745 1388was compiled with. All XSUBs should now use the functions in the PerlIO
5f05dabc
PP
1389abstraction layer and not make any assumptions about what kind of stdio
1390is being used.
1391
1392For a complete description of the PerlIO abstraction, consult L<perlapio>.
1393
8ebc5c01 1394=head2 Putting a C value on Perl stack
ce3d39e2
IZ
1395
1396A lot of opcodes (this is an elementary operation in the internal perl
1397stack machine) put an SV* on the stack. However, as an optimization
1398the corresponding SV is (usually) not recreated each time. The opcodes
1399reuse specially assigned SVs (I<target>s) which are (as a corollary)
1400not constantly freed/created.
1401
0a753a76 1402Each of the targets is created only once (but see
ce3d39e2
IZ
1403L<Scratchpads and recursion> below), and when an opcode needs to put
1404an integer, a double, or a string on stack, it just sets the
1405corresponding parts of its I<target> and puts the I<target> on stack.
1406
1407The macro to put this target on stack is C<PUSHTARG>, and it is
1408directly used in some opcodes, as well as indirectly in zillions of
1409others, which use it via C<(X)PUSH[pni]>.
1410
8ebc5c01 1411=head2 Scratchpads
ce3d39e2 1412
54310121 1413The question remains on when the SVs which are I<target>s for opcodes
5f05dabc
PP
1414are created. The answer is that they are created when the current unit --
1415a subroutine or a file (for opcodes for statements outside of
1416subroutines) -- is compiled. During this time a special anonymous Perl
ce3d39e2
IZ
1417array is created, which is called a scratchpad for the current
1418unit.
1419
54310121 1420A scratchpad keeps SVs which are lexicals for the current unit and are
ce3d39e2
IZ
1421targets for opcodes. One can deduce that an SV lives on a scratchpad
1422by looking on its flags: lexicals have C<SVs_PADMY> set, and
1423I<target>s have C<SVs_PADTMP> set.
1424
54310121
PP
1425The correspondence between OPs and I<target>s is not 1-to-1. Different
1426OPs in the compile tree of the unit can use the same target, if this
ce3d39e2
IZ
1427would not conflict with the expected life of the temporary.
1428
2ae324a7 1429=head2 Scratchpads and recursion
ce3d39e2
IZ
1430
1431In fact it is not 100% true that a compiled unit contains a pointer to
1432the scratchpad AV. In fact it contains a pointer to an AV of
1433(initially) one element, and this element is the scratchpad AV. Why do
1434we need an extra level of indirection?
1435
1436The answer is B<recursion>, and maybe (sometime soon) B<threads>. Both
1437these can create several execution pointers going into the same
1438subroutine. For the subroutine-child not write over the temporaries
1439for the subroutine-parent (lifespan of which covers the call to the
1440child), the parent and the child should have different
1441scratchpads. (I<And> the lexicals should be separate anyway!)
1442
5f05dabc
PP
1443So each subroutine is born with an array of scratchpads (of length 1).
1444On each entry to the subroutine it is checked that the current
ce3d39e2
IZ
1445depth of the recursion is not more than the length of this array, and
1446if it is, new scratchpad is created and pushed into the array.
1447
1448The I<target>s on this scratchpad are C<undef>s, but they are already
1449marked with correct flags.
1450
0a753a76
PP
1451=head1 Compiled code
1452
1453=head2 Code tree
1454
1455Here we describe the internal form your code is converted to by
1456Perl. Start with a simple example:
1457
1458 $a = $b + $c;
1459
1460This is converted to a tree similar to this one:
1461
1462 assign-to
1463 / \
1464 + $a
1465 / \
1466 $b $c
1467
7b8d334a 1468(but slightly more complicated). This tree reflects the way Perl
0a753a76
PP
1469parsed your code, but has nothing to do with the execution order.
1470There is an additional "thread" going through the nodes of the tree
1471which shows the order of execution of the nodes. In our simplified
1472example above it looks like:
1473
1474 $b ---> $c ---> + ---> $a ---> assign-to
1475
1476But with the actual compile tree for C<$a = $b + $c> it is different:
1477some nodes I<optimized away>. As a corollary, though the actual tree
1478contains more nodes than our simplified example, the execution order
1479is the same as in our example.
1480
1481=head2 Examining the tree
1482
1483If you have your perl compiled for debugging (usually done with C<-D
1484optimize=-g> on C<Configure> command line), you may examine the
1485compiled tree by specifying C<-Dx> on the Perl command line. The
1486output takes several lines per node, and for C<$b+$c> it looks like
1487this:
1488
1489 5 TYPE = add ===> 6
1490 TARG = 1
1491 FLAGS = (SCALAR,KIDS)
1492 {
1493 TYPE = null ===> (4)
1494 (was rv2sv)
1495 FLAGS = (SCALAR,KIDS)
1496 {
1497 3 TYPE = gvsv ===> 4
1498 FLAGS = (SCALAR)
1499 GV = main::b
1500 }
1501 }
1502 {
1503 TYPE = null ===> (5)
1504 (was rv2sv)
1505 FLAGS = (SCALAR,KIDS)
1506 {
1507 4 TYPE = gvsv ===> 5
1508 FLAGS = (SCALAR)
1509 GV = main::c
1510 }
1511 }
1512
1513This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
1514not optimized away (one per number in the left column). The immediate
1515children of the given node correspond to C<{}> pairs on the same level
1516of indentation, thus this listing corresponds to the tree:
1517
1518 add
1519 / \
1520 null null
1521 | |
1522 gvsv gvsv
1523
1524The execution order is indicated by C<===E<gt>> marks, thus it is C<3
15254 5 6> (node C<6> is not included into above listing), i.e.,
1526C<gvsv gvsv add whatever>.
1527
9afa14e3
SC
1528Each of these nodes represents an op, a fundamental operation inside the
1529Perl core. The code which implements each operation can be found in the
1530F<pp*.c> files; the function which implements the op with type C<gvsv>
1531is C<pp_gvsv>, and so on. As the tree above shows, different ops have
1532different numbers of children: C<add> is a binary operator, as one would
1533expect, and so has two children. To accommodate the various different
1534numbers of children, there are various types of op data structure, and
1535they link together in different ways.
1536
1537The simplest type of op structure is C<OP>: this has no children. Unary
1538operators, C<UNOP>s, have one child, and this is pointed to by the
1539C<op_first> field. Binary operators (C<BINOP>s) have not only an
1540C<op_first> field but also an C<op_last> field. The most complex type of
1541op is a C<LISTOP>, which has any number of children. In this case, the
1542first child is pointed to by C<op_first> and the last child by
1543C<op_last>. The children in between can be found by iteratively
1544following the C<op_sibling> pointer from the first child to the last.
1545
1546There are also two other op types: a C<PMOP> holds a regular expression,
1547and has no children, and a C<LOOP> may or may not have children. If the
1548C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
1549complicate matters, if a C<UNOP> is actually a C<null> op after
1550optimization (see L</Compile pass 2: context propagation>) it will still
1551have children in accordance with its former type.
1552
0a753a76
PP
1553=head2 Compile pass 1: check routines
1554
8870b5c7
GS
1555The tree is created by the compiler while I<yacc> code feeds it
1556the constructions it recognizes. Since I<yacc> works bottom-up, so does
0a753a76
PP
1557the first pass of perl compilation.
1558
1559What makes this pass interesting for perl developers is that some
1560optimization may be performed on this pass. This is optimization by
8870b5c7 1561so-called "check routines". The correspondence between node names
0a753a76
PP
1562and corresponding check routines is described in F<opcode.pl> (do not
1563forget to run C<make regen_headers> if you modify this file).
1564
1565A check routine is called when the node is fully constructed except
7b8d334a 1566for the execution-order thread. Since at this time there are no
0a753a76
PP
1567back-links to the currently constructed node, one can do most any
1568operation to the top-level node, including freeing it and/or creating
1569new nodes above/below it.
1570
1571The check routine returns the node which should be inserted into the
1572tree (if the top-level node was not modified, check routine returns
1573its argument).
1574
1575By convention, check routines have names C<ck_*>. They are usually
1576called from C<new*OP> subroutines (or C<convert>) (which in turn are
1577called from F<perly.y>).
1578
1579=head2 Compile pass 1a: constant folding
1580
1581Immediately after the check routine is called the returned node is
1582checked for being compile-time executable. If it is (the value is
1583judged to be constant) it is immediately executed, and a I<constant>
1584node with the "return value" of the corresponding subtree is
1585substituted instead. The subtree is deleted.
1586
1587If constant folding was not performed, the execution-order thread is
1588created.
1589
1590=head2 Compile pass 2: context propagation
1591
1592When a context for a part of compile tree is known, it is propagated
a3cb178b 1593down through the tree. At this time the context can have 5 values
0a753a76
PP
1594(instead of 2 for runtime context): void, boolean, scalar, list, and
1595lvalue. In contrast with the pass 1 this pass is processed from top
1596to bottom: a node's context determines the context for its children.
1597
1598Additional context-dependent optimizations are performed at this time.
1599Since at this moment the compile tree contains back-references (via
1600"thread" pointers), nodes cannot be free()d now. To allow
1601optimized-away nodes at this stage, such nodes are null()ified instead
1602of free()ing (i.e. their type is changed to OP_NULL).
1603
1604=head2 Compile pass 3: peephole optimization
1605
1606After the compile tree for a subroutine (or for an C<eval> or a file)
1607is created, an additional pass over the code is performed. This pass
1608is neither top-down or bottom-up, but in the execution order (with
7b8d334a 1609additional complications for conditionals). These optimizations are
0a753a76
PP
1610done in the subroutine peep(). Optimizations performed at this stage
1611are subject to the same restrictions as in the pass 2.
1612
9afa14e3
SC
1613=head1 Examining internal data structures with the C<dump> functions
1614
1615To aid debugging, the source file F<dump.c> contains a number of
1616functions which produce formatted output of internal data structures.
1617
1618The most commonly used of these functions is C<Perl_sv_dump>; it's used
1619for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
1620C<sv_dump> to produce debugging output from Perl-space, so users of that
1621module should already be familiar with its format.
1622
1623C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
1624derivatives, and produces output similiar to C<perl -Dx>; in fact,
1625C<Perl_dump_eval> will dump the main root of the code being evaluated,
1626exactly like C<-Dx>.
1627
1628Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
1629op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
1630subroutines in a package like so: (Thankfully, these are all xsubs, so
1631there is no op tree)
1632
1633 (gdb) print Perl_dump_packsubs(PL_defstash)
1634
1635 SUB attributes::bootstrap = (xsub 0x811fedc 0)
1636
1637 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
1638
1639 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
1640
1641 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
1642
1643 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
1644
1645and C<Perl_dump_all>, which dumps all the subroutines in the stash and
1646the op tree of the main root.
1647
954c1994 1648=head1 How multiple interpreters and concurrency are supported
ee072b34 1649
ee072b34
GS
1650=head2 Background and PERL_IMPLICIT_CONTEXT
1651
1652The Perl interpreter can be regarded as a closed box: it has an API
1653for feeding it code or otherwise making it do things, but it also has
1654functions for its own use. This smells a lot like an object, and
1655there are ways for you to build Perl so that you can have multiple
1656interpreters, with one interpreter represented either as a C++ object,
1657a C structure, or inside a thread. The thread, the C structure, or
1658the C++ object will contain all the context, the state of that
1659interpreter.
1660
54aff467
GS
1661Three macros control the major Perl build flavors: MULTIPLICITY,
1662USE_THREADS and PERL_OBJECT. The MULTIPLICITY build has a C structure
1663that packages all the interpreter state, there is a similar thread-specific
1664data structure under USE_THREADS, and the PERL_OBJECT build has a C++
1665class to maintain interpreter state. In all three cases,
1666PERL_IMPLICIT_CONTEXT is also normally defined, and enables the
1667support for passing in a "hidden" first argument that represents all three
651a3225 1668data structures.
54aff467
GS
1669
1670All this obviously requires a way for the Perl internal functions to be
ee072b34
GS
1671C++ methods, subroutines taking some kind of structure as the first
1672argument, or subroutines taking nothing as the first argument. To
1673enable these three very different ways of building the interpreter,
1674the Perl source (as it does in so many other situations) makes heavy
1675use of macros and subroutine naming conventions.
1676
54aff467 1677First problem: deciding which functions will be public API functions and
954c1994
GS
1678which will be private. All functions whose names begin C<S_> are private
1679(think "S" for "secret" or "static"). All other functions begin with
1680"Perl_", but just because a function begins with "Perl_" does not mean it is
a422fd2d
SC
1681part of the API. (See L</Internal Functions>.) The easiest way to be B<sure> a
1682function is part of the API is to find its entry in L<perlapi>.
1683If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
1684think it should be (i.e., you need it for your extension), send mail via
1685L<perlbug> explaining why you think it should be.
ee072b34
GS
1686
1687Second problem: there must be a syntax so that the same subroutine
1688declarations and calls can pass a structure as their first argument,
1689or pass nothing. To solve this, the subroutines are named and
1690declared in a particular way. Here's a typical start of a static
1691function used within the Perl guts:
1692
1693 STATIC void
1694 S_incline(pTHX_ char *s)
1695
1696STATIC becomes "static" in C, and is #define'd to nothing in C++.
1697
651a3225
GS
1698A public function (i.e. part of the internal API, but not necessarily
1699sanctioned for use in extensions) begins like this:
ee072b34
GS
1700
1701 void
1702 Perl_sv_setsv(pTHX_ SV* dsv, SV* ssv)
1703
1704C<pTHX_> is one of a number of macros (in perl.h) that hide the
1705details of the interpreter's context. THX stands for "thread", "this",
1706or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
1707The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
1708or 'd' for B<d>eclaration.
1709
1710When Perl is built without PERL_IMPLICIT_CONTEXT, there is no first
1711argument containing the interpreter's context. The trailing underscore
1712in the pTHX_ macro indicates that the macro expansion needs a comma
1713after the context argument because other arguments follow it. If
1714PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
54aff467
GS
1715subroutine is not prototyped to take the extra argument. The form of the
1716macro without the trailing underscore is used when there are no additional
ee072b34
GS
1717explicit arguments.
1718
54aff467 1719When a core function calls another, it must pass the context. This
ee072b34
GS
1720is normally hidden via macros. Consider C<sv_setsv>. It expands
1721something like this:
1722
1723 ifdef PERL_IMPLICIT_CONTEXT
1724 define sv_setsv(a,b) Perl_sv_setsv(aTHX_ a, b)
1725 /* can't do this for vararg functions, see below */
1726 else
1727 define sv_setsv Perl_sv_setsv
1728 endif
1729
1730This works well, and means that XS authors can gleefully write:
1731
1732 sv_setsv(foo, bar);
1733
1734and still have it work under all the modes Perl could have been
1735compiled with.
1736
1737Under PERL_OBJECT in the core, that will translate to either:
1738
1739 CPerlObj::Perl_sv_setsv(foo,bar); # in CPerlObj functions,
1740 # C++ takes care of 'this'
1741 or
1742
1743 pPerl->Perl_sv_setsv(foo,bar); # in truly static functions,
1744 # see objXSUB.h
1745
1746Under PERL_OBJECT in extensions (aka PERL_CAPI), or under
1747MULTIPLICITY/USE_THREADS w/ PERL_IMPLICIT_CONTEXT in both core
1748and extensions, it will be:
1749
1750 Perl_sv_setsv(aTHX_ foo, bar); # the canonical Perl "API"
1751 # for all build flavors
1752
1753This doesn't work so cleanly for varargs functions, though, as macros
1754imply that the number of arguments is known in advance. Instead we
1755either need to spell them out fully, passing C<aTHX_> as the first
1756argument (the Perl core tends to do this with functions like
1757Perl_warner), or use a context-free version.
1758
1759The context-free version of Perl_warner is called
1760Perl_warner_nocontext, and does not take the extra argument. Instead
1761it does dTHX; to get the context from thread-local storage. We
1762C<#define warner Perl_warner_nocontext> so that extensions get source
1763compatibility at the expense of performance. (Passing an arg is
1764cheaper than grabbing it from thread-local storage.)
1765
1766You can ignore [pad]THX[xo] when browsing the Perl headers/sources.
1767Those are strictly for use within the core. Extensions and embedders
1768need only be aware of [pad]THX.
1769
1770=head2 How do I use all this in extensions?
1771
1772When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
1773any functions in the Perl API will need to pass the initial context
1774argument somehow. The kicker is that you will need to write it in
1775such a way that the extension still compiles when Perl hasn't been
1776built with PERL_IMPLICIT_CONTEXT enabled.
1777
1778There are three ways to do this. First, the easy but inefficient way,
1779which is also the default, in order to maintain source compatibility
1780with extensions: whenever XSUB.h is #included, it redefines the aTHX
1781and aTHX_ macros to call a function that will return the context.
1782Thus, something like:
1783
1784 sv_setsv(asv, bsv);
1785
4375e838 1786in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
54aff467 1787in effect:
ee072b34 1788
2fa86c13 1789 Perl_sv_setsv(Perl_get_context(), asv, bsv);
ee072b34 1790
54aff467 1791or to this otherwise:
ee072b34
GS
1792
1793 Perl_sv_setsv(asv, bsv);
1794
1795You have to do nothing new in your extension to get this; since
2fa86c13 1796the Perl library provides Perl_get_context(), it will all just
ee072b34
GS
1797work.
1798
1799The second, more efficient way is to use the following template for
1800your Foo.xs:
1801
1802 #define PERL_NO_GET_CONTEXT /* we want efficiency */
1803 #include "EXTERN.h"
1804 #include "perl.h"
1805 #include "XSUB.h"
1806
1807 static my_private_function(int arg1, int arg2);
1808
1809 static SV *
54aff467 1810 my_private_function(int arg1, int arg2)
ee072b34
GS
1811 {
1812 dTHX; /* fetch context */
1813 ... call many Perl API functions ...
1814 }
1815
1816 [... etc ...]
1817
1818 MODULE = Foo PACKAGE = Foo
1819
1820 /* typical XSUB */
1821
1822 void
1823 my_xsub(arg)
1824 int arg
1825 CODE:
1826 my_private_function(arg, 10);
1827
1828Note that the only two changes from the normal way of writing an
1829extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
1830including the Perl headers, followed by a C<dTHX;> declaration at
1831the start of every function that will call the Perl API. (You'll
1832know which functions need this, because the C compiler will complain
1833that there's an undeclared identifier in those functions.) No changes
1834are needed for the XSUBs themselves, because the XS() macro is
1835correctly defined to pass in the implicit context if needed.
1836
1837The third, even more efficient way is to ape how it is done within
1838the Perl guts:
1839
1840
1841 #define PERL_NO_GET_CONTEXT /* we want efficiency */
1842 #include "EXTERN.h"
1843 #include "perl.h"
1844 #include "XSUB.h"
1845
1846 /* pTHX_ only needed for functions that call Perl API */
1847 static my_private_function(pTHX_ int arg1, int arg2);
1848
1849 static SV *
1850 my_private_function(pTHX_ int arg1, int arg2)
1851 {
1852 /* dTHX; not needed here, because THX is an argument */
1853 ... call Perl API functions ...
1854 }
1855
1856 [... etc ...]
1857
1858 MODULE = Foo PACKAGE = Foo
1859
1860 /* typical XSUB */
1861
1862 void
1863 my_xsub(arg)
1864 int arg
1865 CODE:
1866 my_private_function(aTHX_ arg, 10);
1867
1868This implementation never has to fetch the context using a function
1869call, since it is always passed as an extra argument. Depending on
1870your needs for simplicity or efficiency, you may mix the previous
1871two approaches freely.
1872
651a3225
GS
1873Never add a comma after C<pTHX> yourself--always use the form of the
1874macro with the underscore for functions that take explicit arguments,
1875or the form without the argument for functions with no explicit arguments.
ee072b34
GS
1876
1877=head2 Future Plans and PERL_IMPLICIT_SYS
1878
1879Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
1880that the interpreter knows about itself and pass it around, so too are
1881there plans to allow the interpreter to bundle up everything it knows
1882about the environment it's running on. This is enabled with the
1883PERL_IMPLICIT_SYS macro. Currently it only works with PERL_OBJECT,
1884but is mostly there for MULTIPLICITY and USE_THREADS (see inside
1885iperlsys.h).
1886
1887This allows the ability to provide an extra pointer (called the "host"
1888environment) for all the system calls. This makes it possible for
1889all the system stuff to maintain their own state, broken down into
1890seven C structures. These are thin wrappers around the usual system
1891calls (see win32/perllib.c) for the default perl executable, but for a
1892more ambitious host (like the one that would do fork() emulation) all
1893the extra work needed to pretend that different interpreters are
1894actually different "processes", would be done here.
1895
1896The Perl engine/interpreter and the host are orthogonal entities.
1897There could be one or more interpreters in a process, and one or
1898more "hosts", with free association between them.
1899
a422fd2d
SC
1900=head1 Internal Functions
1901
1902All of Perl's internal functions which will be exposed to the outside
1903world are be prefixed by C<Perl_> so that they will not conflict with XS
1904functions or functions used in a program in which Perl is embedded.
1905Similarly, all global variables begin with C<PL_>. (By convention,
1906static functions start with C<S_>)
1907
1908Inside the Perl core, you can get at the functions either with or
1909without the C<Perl_> prefix, thanks to a bunch of defines that live in
1910F<embed.h>. This header file is generated automatically from
1911F<embed.pl>. F<embed.pl> also creates the prototyping header files for
1912the internal functions, generates the documentation and a lot of other
1913bits and pieces. It's important that when you add a new function to the
1914core or change an existing one, you change the data in the table at the
1915end of F<embed.pl> as well. Here's a sample entry from that table:
1916
1917 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
1918
1919The second column is the return type, the third column the name. Columns
1920after that are the arguments. The first column is a set of flags:
1921
1922=over 3
1923
1924=item A
1925
1926This function is a part of the public API.
1927
1928=item p
1929
1930This function has a C<Perl_> prefix; ie, it is defined as C<Perl_av_fetch>
1931
1932=item d
1933
1934This function has documentation using the C<apidoc> feature which we'll
1935look at in a second.
1936
1937=back
1938
1939Other available flags are:
1940
1941=over 3
1942
1943=item s
1944
1945This is a static function and is defined as C<S_whatever>.
1946
1947=item n
1948
1949This does not use C<aTHX_> and C<pTHX> to pass interpreter context. (See
1950L<perlguts/Background and PERL_IMPLICIT_CONTEXT>.)
1951
1952=item r
1953
1954This function never returns; C<croak>, C<exit> and friends.
1955
1956=item f
1957
1958This function takes a variable number of arguments, C<printf> style.
1959The argument list should end with C<...>, like this:
1960
1961 Afprd |void |croak |const char* pat|...
1962
1963=item m
1964
1965This function is part of the experimental development API, and may change
1966or disappear without notice.
1967
1968=item o
1969
1970This function should not have a compatibility macro to define, say,
1971C<Perl_parse> to C<parse>. It must be called as C<Perl_parse>.
1972
1973=item j
1974
1975This function is not a member of C<CPerlObj>. If you don't know
1976what this means, don't use it.
1977
1978=item x
1979
1980This function isn't exported out of the Perl core.
1981
1982=back
1983
1984If you edit F<embed.pl>, you will need to run C<make regen_headers> to
1985force a rebuild of F<embed.h> and other auto-generated files.
1986
6b4667fc 1987=head2 Formatted Printing of IVs, UVs, and NVs
9dd9db0b 1988
6b4667fc
A
1989If you are printing IVs, UVs, or NVS instead of the stdio(3) style
1990formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
1991following macros for portability
9dd9db0b
JH
1992
1993 IVdf IV in decimal
1994 UVuf UV in decimal
1995 UVof UV in octal
1996 UVxf UV in hexadecimal
6b4667fc
A
1997 NVef NV %e-like
1998 NVff NV %f-like
1999 NVgf NV %g-like
9dd9db0b 2000
6b4667fc
A
2001These will take care of 64-bit integers and long doubles.
2002For example:
2003
2004 printf("IV is %"IVdf"\n", iv);
2005
2006The IVdf will expand to whatever is the correct format for the IVs.
9dd9db0b 2007
8908e76d
JH
2008If you are printing addresses of pointers, use UVxf combined
2009with PTR2UV(), do not use %lx or %p.
2010
2011=head2 Pointer-To-Integer and Integer-To-Pointer
2012
2013Because pointer size does not necessarily equal integer size,
2014use the follow macros to do it right.
2015
2016 PTR2UV(pointer)
2017 PTR2IV(pointer)
2018 PTR2NV(pointer)
2019 INT2PTR(pointertotype, integer)
2020
2021For example:
2022
2023 IV iv = ...;
2024 SV *sv = INT2PTR(SV*, iv);
2025
2026and
2027
2028 AV *av = ...;
2029 UV uv = PTR2UV(av);
2030
a422fd2d
SC
2031=head2 Source Documentation
2032
2033There's an effort going on to document the internal functions and
2034automatically produce reference manuals from them - L<perlapi> is one
2035such manual which details all the functions which are available to XS
2036writers. L<perlintern> is the autogenerated manual for the functions
2037which are not part of the API and are supposedly for internal use only.
2038
2039Source documentation is created by putting POD comments into the C
2040source, like this:
2041
2042 /*
2043 =for apidoc sv_setiv
2044
2045 Copies an integer into the given SV. Does not handle 'set' magic. See
2046 C<sv_setiv_mg>.
2047
2048 =cut
2049 */
2050
2051Please try and supply some documentation if you add functions to the
2052Perl core.
2053
2054=head1 Unicode Support
2055
2056Perl 5.6.0 introduced Unicode support. It's important for porters and XS
2057writers to understand this support and make sure that the code they
2058write does not corrupt Unicode data.
2059
2060=head2 What B<is> Unicode, anyway?
2061
2062In the olden, less enlightened times, we all used to use ASCII. Most of
2063us did, anyway. The big problem with ASCII is that it's American. Well,
2064no, that's not actually the problem; the problem is that it's not
2065particularly useful for people who don't use the Roman alphabet. What
2066used to happen was that particular languages would stick their own
2067alphabet in the upper range of the sequence, between 128 and 255. Of
2068course, we then ended up with plenty of variants that weren't quite
2069ASCII, and the whole point of it being a standard was lost.
2070
2071Worse still, if you've got a language like Chinese or
2072Japanese that has hundreds or thousands of characters, then you really
2073can't fit them into a mere 256, so they had to forget about ASCII
2074altogether, and build their own systems using pairs of numbers to refer
2075to one character.
2076
2077To fix this, some people formed Unicode, Inc. and
2078produced a new character set containing all the characters you can
2079possibly think of and more. There are several ways of representing these
2080characters, and the one Perl uses is called UTF8. UTF8 uses
2081a variable number of bytes to represent a character, instead of just
b3b6085d 2082one. You can learn more about Unicode at http://www.unicode.org/
a422fd2d
SC
2083
2084=head2 How can I recognise a UTF8 string?
2085
2086You can't. This is because UTF8 data is stored in bytes just like
2087non-UTF8 data. The Unicode character 200, (C<0xC8> for you hex types)
2088capital E with a grave accent, is represented by the two bytes
2089C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
2090has that byte sequence as well. So you can't tell just by looking - this
2091is what makes Unicode input an interesting problem.
2092
2093The API function C<is_utf8_string> can help; it'll tell you if a string
2094contains only valid UTF8 characters. However, it can't do the work for
2095you. On a character-by-character basis, C<is_utf8_char> will tell you
2096whether the current character in a string is valid UTF8.
2097
2098=head2 How does UTF8 represent Unicode characters?
2099
2100As mentioned above, UTF8 uses a variable number of bytes to store a
2101character. Characters with values 1...128 are stored in one byte, just
2102like good ol' ASCII. Character 129 is stored as C<v194.129>; this
a31a806a 2103continues up to character 191, which is C<v194.191>. Now we've run out of
a422fd2d
SC
2104bits (191 is binary C<10111111>) so we move on; 192 is C<v195.128>. And
2105so it goes on, moving to three bytes at character 2048.
2106
2107Assuming you know you're dealing with a UTF8 string, you can find out
2108how long the first character in it is with the C<UTF8SKIP> macro:
2109
2110 char *utf = "\305\233\340\240\201";
2111 I32 len;
2112
2113 len = UTF8SKIP(utf); /* len is 2 here */
2114 utf += len;
2115 len = UTF8SKIP(utf); /* len is 3 here */
2116
2117Another way to skip over characters in a UTF8 string is to use
2118C<utf8_hop>, which takes a string and a number of characters to skip
2119over. You're on your own about bounds checking, though, so don't use it
2120lightly.
2121
2122All bytes in a multi-byte UTF8 character will have the high bit set, so
2123you can test if you need to do something special with this character
2124like this:
2125
2126 UV uv;
2127
2128 if (utf & 0x80)
2129 /* Must treat this as UTF8 */
2130 uv = utf8_to_uv(utf);
2131 else
2132 /* OK to treat this character as a byte */
2133 uv = *utf;
2134
2135You can also see in that example that we use C<utf8_to_uv> to get the
2136value of the character; the inverse function C<uv_to_utf8> is available
2137for putting a UV into UTF8:
2138
2139 if (uv > 0x80)
2140 /* Must treat this as UTF8 */
2141 utf8 = uv_to_utf8(utf8, uv);
2142 else
2143 /* OK to treat this character as a byte */
2144 *utf8++ = uv;
2145
2146You B<must> convert characters to UVs using the above functions if
2147you're ever in a situation where you have to match UTF8 and non-UTF8
2148characters. You may not skip over UTF8 characters in this case. If you
2149do this, you'll lose the ability to match hi-bit non-UTF8 characters;
2150for instance, if your UTF8 string contains C<v196.172>, and you skip
2151that character, you can never match a C<chr(200)> in a non-UTF8 string.
2152So don't do that!
2153
2154=head2 How does Perl store UTF8 strings?
2155
2156Currently, Perl deals with Unicode strings and non-Unicode strings
2157slightly differently. If a string has been identified as being UTF-8
2158encoded, Perl will set a flag in the SV, C<SVf_UTF8>. You can check and
2159manipulate this flag with the following macros:
2160
2161 SvUTF8(sv)
2162 SvUTF8_on(sv)
2163 SvUTF8_off(sv)
2164
2165This flag has an important effect on Perl's treatment of the string: if
2166Unicode data is not properly distinguished, regular expressions,
2167C<length>, C<substr> and other string handling operations will have
2168undesirable results.
2169
2170The problem comes when you have, for instance, a string that isn't
2171flagged is UTF8, and contains a byte sequence that could be UTF8 -
2172especially when combining non-UTF8 and UTF8 strings.
2173
2174Never forget that the C<SVf_UTF8> flag is separate to the PV value; you
2175need be sure you don't accidentally knock it off while you're
2176manipulating SVs. More specifically, you cannot expect to do this:
2177
2178 SV *sv;
2179 SV *nsv;
2180 STRLEN len;
2181 char *p;
2182
2183 p = SvPV(sv, len);
2184 frobnicate(p);
2185 nsv = newSVpvn(p, len);
2186
2187The C<char*> string does not tell you the whole story, and you can't
2188copy or reconstruct an SV just by copying the string value. Check if the
2189old SV has the UTF8 flag set, and act accordingly:
2190
2191 p = SvPV(sv, len);
2192 frobnicate(p);
2193 nsv = newSVpvn(p, len);
2194 if (SvUTF8(sv))
2195 SvUTF8_on(nsv);
2196
2197In fact, your C<frobnicate> function should be made aware of whether or
2198not it's dealing with UTF8 data, so that it can handle the string
2199appropriately.
2200
2201=head2 How do I convert a string to UTF8?
2202
2203If you're mixing UTF8 and non-UTF8 strings, you might find it necessary
2204to upgrade one of the strings to UTF8. If you've got an SV, the easiest
2205way to do this is:
2206
2207 sv_utf8_upgrade(sv);
2208
2209However, you must not do this, for example:
2210
2211 if (!SvUTF8(left))
2212 sv_utf8_upgrade(left);
2213
2214If you do this in a binary operator, you will actually change one of the
b1866b2d 2215strings that came into the operator, and, while it shouldn't be noticeable
a422fd2d
SC
2216by the end user, it can cause problems.
2217
2218Instead, C<bytes_to_utf8> will give you a UTF8-encoded B<copy> of its
2219string argument. This is useful for having the data available for
b1866b2d 2220comparisons and so on, without harming the original SV. There's also
a422fd2d
SC
2221C<utf8_to_bytes> to go the other way, but naturally, this will fail if
2222the string contains any characters above 255 that can't be represented
2223in a single byte.
2224
2225=head2 Is there anything else I need to know?
2226
2227Not really. Just remember these things:
2228
2229=over 3
2230
2231=item *
2232
2233There's no way to tell if a string is UTF8 or not. You can tell if an SV
2234is UTF8 by looking at is C<SvUTF8> flag. Don't forget to set the flag if
2235something should be UTF8. Treat the flag as part of the PV, even though
2236it's not - if you pass on the PV to somewhere, pass on the flag too.
2237
2238=item *
2239
2240If a string is UTF8, B<always> use C<utf8_to_uv> to get at the value,
2241unless C<!(*s & 0x80)> in which case you can use C<*s>.
2242
2243=item *
2244
2245When writing to a UTF8 string, B<always> use C<uv_to_utf8>, unless
2246C<uv < 0x80> in which case you can use C<*s = uv>.
2247
2248=item *
2249
2250Mixing UTF8 and non-UTF8 strings is tricky. Use C<bytes_to_utf8> to get
2251a new string which is UTF8 encoded. There are tricks you can use to
2252delay deciding whether you need to use a UTF8 string until you get to a
2253high character - C<HALF_UPGRADE> is one of those.
2254
2255=back
2256
954c1994 2257=head1 AUTHORS
e89caa19 2258
954c1994
GS
2259Until May 1997, this document was maintained by Jeff Okamoto
2260<okamoto@corp.hp.com>. It is now maintained as part of Perl itself
2261by the Perl 5 Porters <perl5-porters@perl.org>.
cb1a09d0 2262
954c1994
GS
2263With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
2264Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
2265Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
2266Stephen McCamant, and Gurusamy Sarathy.
cb1a09d0 2267
954c1994 2268API Listing originally by Dean Roehrich <roehrich@cray.com>.
cb1a09d0 2269
954c1994
GS
2270Modifications to autogenerate the API listing (L<perlapi>) by Benjamin
2271Stuhl.
cb1a09d0 2272
954c1994 2273=head1 SEE ALSO
cb1a09d0 2274
954c1994 2275perlapi(1), perlintern(1), perlxs(1), perlembed(1)