This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Merge wrap_keyword_plugin() into blead
[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 7This document attempts to describe how to use the Perl API, as well as
10e2eb10
FC
8to provide some info on the basic workings of the Perl core. It is far
9from complete and probably contains many errors. Please refer any
b3b6085d 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
10e2eb10 31least 32-bits and 16-bits long, respectively. (Again, there are U32 and U16,
20dbd849
NC
32as well.) They will usually be exactly 32 and 16 bits long, but on Crays
33they will both be 64 bits.
a0d0e21e 34
54310121 35=head2 Working with SVs
a0d0e21e 36
20dbd849
NC
37An SV can be created and loaded with one command. There are five types of
38values that can be loaded: an integer value (IV), an unsigned integer
39value (UV), a double (NV), a string (PV), and another scalar (SV).
61984ee1
KW
40("PV" stands for "Pointer Value". You might think that it is misnamed
41because it is described as pointing only to strings. However, it is
3ee1a09c 42possible to have it point to other things. For example, it could point
d6605d24 43to an array of UVs. But,
61984ee1
KW
44using it for non-strings requires care, as the underlying assumption of
45much of the internals is that PVs are just for strings. Often, for
6602b933 46example, a trailing C<NUL> is tacked on automatically. The non-string use
61984ee1 47is documented only in this paragraph.)
a0d0e21e 48
20dbd849 49The seven routines are:
a0d0e21e
LW
50
51 SV* newSViv(IV);
20dbd849 52 SV* newSVuv(UV);
a0d0e21e 53 SV* newSVnv(double);
06f6df17
RGS
54 SV* newSVpv(const char*, STRLEN);
55 SV* newSVpvn(const char*, STRLEN);
46fc3d4c 56 SV* newSVpvf(const char*, ...);
a0d0e21e
LW
57 SV* newSVsv(SV*);
58
e613617c 59C<STRLEN> is an integer type (C<Size_t>, usually defined as C<size_t> in
06f6df17
RGS
60F<config.h>) guaranteed to be large enough to represent the size of
61any string that perl can handle.
62
3bf17896 63In the unlikely case of a SV requiring more complex initialization, you
06f6df17
RGS
64can create an empty SV with newSV(len). If C<len> is 0 an empty SV of
65type NULL is returned, else an SV of type PV is returned with len + 1 (for
6602b933 66the C<NUL>) bytes of storage allocated, accessible via SvPVX. In both cases
da8c5729 67the SV has the undef value.
20dbd849 68
06f6df17 69 SV *sv = newSV(0); /* no storage allocated */
a9b0660e
KW
70 SV *sv = newSV(10); /* 10 (+1) bytes of uninitialised storage
71 * allocated */
20dbd849 72
06f6df17 73To change the value of an I<already-existing> SV, there are eight routines:
a0d0e21e
LW
74
75 void sv_setiv(SV*, IV);
deb3007b 76 void sv_setuv(SV*, UV);
a0d0e21e 77 void sv_setnv(SV*, double);
08105a92 78 void sv_setpv(SV*, const char*);
06f6df17 79 void sv_setpvn(SV*, const char*, STRLEN)
46fc3d4c 80 void sv_setpvf(SV*, const char*, ...);
a9b0660e 81 void sv_vsetpvfn(SV*, const char*, STRLEN, va_list *,
03a22d83 82 SV **, Size_t, bool *);
a0d0e21e
LW
83 void sv_setsv(SV*, SV*);
84
85Notice that you can choose to specify the length of the string to be
9da1e3b5
MUN
86assigned by using C<sv_setpvn>, C<newSVpvn>, or C<newSVpv>, or you may
87allow Perl to calculate the length by using C<sv_setpv> or by specifying
880 as the second argument to C<newSVpv>. Be warned, though, that Perl will
89determine the string's length by using C<strlen>, which depends on the
6602b933 90string terminating with a C<NUL> character, and not otherwise containing
a9b0660e 91NULs.
9abd00ed
GS
92
93The arguments of C<sv_setpvf> are processed like C<sprintf>, and the
94formatted output becomes the value.
95
328bf373 96C<sv_vsetpvfn> is an analogue of C<vsprintf>, but it allows you to specify
9abd00ed
GS
97either a pointer to a variable argument list or the address and length of
98an array of SVs. The last argument points to a boolean; on return, if that
99boolean is true, then locale-specific information has been used to format
c2611fb3 100the string, and the string's contents are therefore untrustworthy (see
9abd00ed
GS
101L<perlsec>). This pointer may be NULL if that information is not
102important. Note that this function requires you to specify the length of
103the format.
104
9da1e3b5 105The C<sv_set*()> functions are not generic enough to operate on values
5a0de581 106that have "magic". See L</Magic Virtual Tables> later in this document.
a0d0e21e 107
6602b933
KW
108All SVs that contain strings should be terminated with a C<NUL> character.
109If it is not C<NUL>-terminated there is a risk of
5f05dabc 110core dumps and corruptions from code which passes the string to C
6602b933
KW
111functions or system calls which expect a C<NUL>-terminated string.
112Perl's own functions typically add a trailing C<NUL> for this reason.
5f05dabc
PP
113Nevertheless, you should be very careful when you pass a string stored
114in an SV to a C function or system call.
115
a0d0e21e
LW
116To access the actual value that an SV points to, you can use the macros:
117
118 SvIV(SV*)
954c1994 119 SvUV(SV*)
a0d0e21e
LW
120 SvNV(SV*)
121 SvPV(SV*, STRLEN len)
1fa8b10d 122 SvPV_nolen(SV*)
a0d0e21e 123
954c1994 124which will automatically coerce the actual scalar type into an IV, UV, double,
a0d0e21e
LW
125or string.
126
127In the C<SvPV> macro, the length of the string returned is placed into the
1fa8b10d
JD
128variable C<len> (this is a macro, so you do I<not> use C<&len>). If you do
129not care what the length of the data is, use the C<SvPV_nolen> macro.
130Historically the C<SvPV> macro with the global variable C<PL_na> has been
131used in this case. But that can be quite inefficient because C<PL_na> must
132be accessed in thread-local storage in threaded Perl. In any case, remember
133that Perl allows arbitrary strings of data that may both contain NULs and
6602b933 134might not be terminated by a C<NUL>.
a0d0e21e 135
ce2f5d8f 136Also remember that C doesn't allow you to safely say C<foo(SvPV(s, len),
10e2eb10
FC
137len);>. It might work with your
138compiler, but it won't work for everyone.
ce2f5d8f
KA
139Break this sort of statement up into separate assignments:
140
1aa6ea50
JC
141 SV *s;
142 STRLEN len;
61955433 143 char *ptr;
1aa6ea50
JC
144 ptr = SvPV(s, len);
145 foo(ptr, len);
ce2f5d8f 146
07fa94a1 147If you want to know if the scalar value is TRUE, you can use:
a0d0e21e
LW
148
149 SvTRUE(SV*)
150
151Although Perl will automatically grow strings for you, if you need to force
152Perl to allocate more memory for your SV, you can use the macro
153
154 SvGROW(SV*, STRLEN newlen)
155
156which will determine if more memory needs to be allocated. If so, it will
157call the function C<sv_grow>. Note that C<SvGROW> can only increase, not
5f05dabc 158decrease, the allocated memory of an SV and that it does not automatically
6602b933 159add space for the trailing C<NUL> byte (perl's own string functions typically do
8ebc5c01 160C<SvGROW(sv, len + 1)>).
a0d0e21e 161
21134f66
TC
162If you want to write to an existing SV's buffer and set its value to a
163string, use SvPV_force() or one of its variants to force the SV to be
164a PV. This will remove any of various types of non-stringness from
165the SV while preserving the content of the SV in the PV. This can be
166used, for example, to append data from an API function to a buffer
167without extra copying:
168
169 (void)SvPVbyte_force(sv, len);
170 s = SvGROW(sv, len + needlen + 1);
171 /* something that modifies up to needlen bytes at s+len, but
172 modifies newlen bytes
173 eg. newlen = read(fd, s + len, needlen);
174 ignoring errors for these examples
175 */
176 s[len + newlen] = '\0';
177 SvCUR_set(sv, len + newlen);
178 SvUTF8_off(sv);
179 SvSETMAGIC(sv);
180
181If you already have the data in memory or if you want to keep your
182code simple, you can use one of the sv_cat*() variants, such as
183sv_catpvn(). If you want to insert anywhere in the string you can use
184sv_insert() or sv_insert_flags().
185
186If you don't need the existing content of the SV, you can avoid some
187copying with:
188
5b1fede8 189 SvPVCLEAR(sv);
21134f66
TC
190 s = SvGROW(sv, needlen + 1);
191 /* something that modifies up to needlen bytes at s, but modifies
192 newlen bytes
193 eg. newlen = read(fd, s. needlen);
194 */
195 s[newlen] = '\0';
196 SvCUR_set(sv, newlen);
197 SvPOK_only(sv); /* also clears SVf_UTF8 */
198 SvSETMAGIC(sv);
199
200Again, if you already have the data in memory or want to avoid the
201complexity of the above, you can use sv_setpvn().
202
203If you have a buffer allocated with Newx() and want to set that as the
204SV's value, you can use sv_usepvn_flags(). That has some requirements
205if you want to avoid perl re-allocating the buffer to fit the trailing
206NUL:
207
208 Newx(buf, somesize+1, char);
209 /* ... fill in buf ... */
210 buf[somesize] = '\0';
211 sv_usepvn_flags(sv, buf, somesize, SV_SMAGIC | SV_HAS_TRAILING_NUL);
212 /* buf now belongs to perl, don't release it */
213
a0d0e21e
LW
214If you have an SV and want to know what kind of data Perl thinks is stored
215in it, you can use the following macros to check the type of SV you have.
216
217 SvIOK(SV*)
218 SvNOK(SV*)
219 SvPOK(SV*)
220
221You can get and set the current length of the string stored in an SV with
222the following macros:
223
224 SvCUR(SV*)
225 SvCUR_set(SV*, I32 val)
226
cb1a09d0
AD
227You can also get a pointer to the end of the string stored in the SV
228with the macro:
229
230 SvEND(SV*)
231
232But note that these last three macros are valid only if C<SvPOK()> is true.
a0d0e21e 233
d1b91892
AD
234If you want to append something to the end of string stored in an C<SV*>,
235you can use the following functions:
236
08105a92 237 void sv_catpv(SV*, const char*);
e65f3abd 238 void sv_catpvn(SV*, const char*, STRLEN);
46fc3d4c 239 void sv_catpvf(SV*, const char*, ...);
a9b0660e
KW
240 void sv_vcatpvfn(SV*, const char*, STRLEN, va_list *, SV **,
241 I32, bool);
d1b91892
AD
242 void sv_catsv(SV*, SV*);
243
244The first function calculates the length of the string to be appended by
245using C<strlen>. In the second, you specify the length of the string
46fc3d4c 246yourself. The third function processes its arguments like C<sprintf> and
9abd00ed
GS
247appends the formatted output. The fourth function works like C<vsprintf>.
248You can specify the address and length of an array of SVs instead of the
10e2eb10
FC
249va_list argument. The fifth function
250extends the string stored in the first
9abd00ed
GS
251SV with the string stored in the second SV. It also forces the second SV
252to be interpreted as a string.
253
254The C<sv_cat*()> functions are not generic enough to operate on values that
5a0de581 255have "magic". See L</Magic Virtual Tables> later in this document.
d1b91892 256
a0d0e21e
LW
257If you know the name of a scalar variable, you can get a pointer to its SV
258by using the following:
259
64ace3f8 260 SV* get_sv("package::varname", 0);
a0d0e21e
LW
261
262This returns NULL if the variable does not exist.
263
d1b91892 264If you want to know if this variable (or any other SV) is actually C<defined>,
a0d0e21e
LW
265you can call:
266
267 SvOK(SV*)
268
06f6df17 269The scalar C<undef> value is stored in an SV instance called C<PL_sv_undef>.
9adebda4 270
10e2eb10
FC
271Its address can be used whenever an C<SV*> is needed. Make sure that
272you don't try to compare a random sv with C<&PL_sv_undef>. For example
9adebda4
SB
273when interfacing Perl code, it'll work correctly for:
274
275 foo(undef);
276
277But won't work when called as:
278
279 $x = undef;
280 foo($x);
281
282So to repeat always use SvOK() to check whether an sv is defined.
283
284Also you have to be careful when using C<&PL_sv_undef> as a value in
5a0de581 285AVs or HVs (see L</AVs, HVs and undefined values>).
a0d0e21e 286
06f6df17
RGS
287There are also the two values C<PL_sv_yes> and C<PL_sv_no>, which contain
288boolean TRUE and FALSE values, respectively. Like C<PL_sv_undef>, their
289addresses can be used whenever an C<SV*> is needed.
a0d0e21e 290
9cde0e7f 291Do not be fooled into thinking that C<(SV *) 0> is the same as C<&PL_sv_undef>.
a0d0e21e
LW
292Take this code:
293
294 SV* sv = (SV*) 0;
295 if (I-am-to-return-a-real-value) {
296 sv = sv_2mortal(newSViv(42));
297 }
298 sv_setsv(ST(0), sv);
299
300This code tries to return a new SV (which contains the value 42) if it should
04343c6d 301return a real value, or undef otherwise. Instead it has returned a NULL
a0d0e21e 302pointer which, somewhere down the line, will cause a segmentation violation,
06f6df17
RGS
303bus error, or just weird results. Change the zero to C<&PL_sv_undef> in the
304first line and all will be well.
a0d0e21e
LW
305
306To free an SV that you've created, call C<SvREFCNT_dec(SV*)>. Normally this
5a0de581 307call is not necessary (see L</Reference Counts and Mortality>).
a0d0e21e 308
94dde4fb
SC
309=head2 Offsets
310
311Perl provides the function C<sv_chop> to efficiently remove characters
312from the beginning of a string; you give it an SV and a pointer to
da75cd15 313somewhere inside the PV, and it discards everything before the
10e2eb10 314pointer. The efficiency comes by means of a little hack: instead of
94dde4fb
SC
315actually removing the characters, C<sv_chop> sets the flag C<OOK>
316(offset OK) to signal to other functions that the offset hack is in
883bb8c0
KW
317effect, and it moves the PV pointer (called C<SvPVX>) forward
318by the number of bytes chopped off, and adjusts C<SvCUR> and C<SvLEN>
319accordingly. (A portion of the space between the old and new PV
320pointers is used to store the count of chopped bytes.)
94dde4fb
SC
321
322Hence, at this point, the start of the buffer that we allocated lives
323at C<SvPVX(sv) - SvIV(sv)> in memory and the PV pointer is pointing
324into the middle of this allocated storage.
325
f942a0df
FC
326This is best demonstrated by example. Normally copy-on-write will prevent
327the substitution from operator from using this hack, but if you can craft a
328string for which copy-on-write is not possible, you can see it in play. In
329the current implementation, the final byte of a string buffer is used as a
330copy-on-write reference count. If the buffer is not big enough, then
331copy-on-write is skipped. First have a look at an empty string:
332
333 % ./perl -Ilib -MDevel::Peek -le '$a=""; $a .= ""; Dump $a'
334 SV = PV(0x7ffb7c008a70) at 0x7ffb7c030390
335 REFCNT = 1
336 FLAGS = (POK,pPOK)
337 PV = 0x7ffb7bc05b50 ""\0
338 CUR = 0
339 LEN = 10
340
341Notice here the LEN is 10. (It may differ on your platform.) Extend the
342length of the string to one less than 10, and do a substitution:
94dde4fb 343
e46aa1dd
KW
344 % ./perl -Ilib -MDevel::Peek -le '$a=""; $a.="123456789"; $a=~s/.//; \
345 Dump($a)'
346 SV = PV(0x7ffa04008a70) at 0x7ffa04030390
347 REFCNT = 1
348 FLAGS = (POK,OOK,pPOK)
349 OFFSET = 1
350 PV = 0x7ffa03c05b61 ( "\1" . ) "23456789"\0
351 CUR = 8
352 LEN = 9
94dde4fb 353
f942a0df 354Here the number of bytes chopped off (1) is shown next as the OFFSET. The
94dde4fb
SC
355portion of the string between the "real" and the "fake" beginnings is
356shown in parentheses, and the values of C<SvCUR> and C<SvLEN> reflect
f942a0df
FC
357the fake beginning, not the real one. (The first character of the string
358buffer happens to have changed to "\1" here, not "1", because the current
359implementation stores the offset count in the string buffer. This is
360subject to change.)
94dde4fb 361
fe854a6f 362Something similar to the offset hack is performed on AVs to enable
319cef53
SC
363efficient shifting and splicing off the beginning of the array; while
364C<AvARRAY> points to the first element in the array that is visible from
10e2eb10 365Perl, C<AvALLOC> points to the real start of the C array. These are
319cef53 366usually the same, but a C<shift> operation can be carried out by
6de131f0 367increasing C<AvARRAY> by one and decreasing C<AvFILL> and C<AvMAX>.
319cef53 368Again, the location of the real start of the C array only comes into
10e2eb10 369play when freeing the array. See C<av_shift> in F<av.c>.
319cef53 370
d1b91892 371=head2 What's Really Stored in an SV?
a0d0e21e
LW
372
373Recall that the usual method of determining the type of scalar you have is
5f05dabc 374to use C<Sv*OK> macros. Because a scalar can be both a number and a string,
d1b91892 375usually these macros will always return TRUE and calling the C<Sv*V>
a0d0e21e
LW
376macros will do the appropriate conversion of string to integer/double or
377integer/double to string.
378
379If you I<really> need to know if you have an integer, double, or string
380pointer in an SV, you can use the following three macros instead:
381
382 SvIOKp(SV*)
383 SvNOKp(SV*)
384 SvPOKp(SV*)
385
386These will tell you if you truly have an integer, double, or string pointer
d1b91892 387stored in your SV. The "p" stands for private.
a0d0e21e 388
da8c5729 389There are various ways in which the private and public flags may differ.
9090718a
FC
390For example, in perl 5.16 and earlier a tied SV may have a valid
391underlying value in the IV slot (so SvIOKp is true), but the data
392should be accessed via the FETCH routine rather than directly,
393so SvIOK is false. (In perl 5.18 onwards, tied scalars use
394the flags the same way as untied scalars.) Another is when
d7f8936a 395numeric conversion has occurred and precision has been lost: only the
10e2eb10 396private flag is set on 'lossy' values. So when an NV is converted to an
9e9796d6
JH
397IV with loss, SvIOKp, SvNOKp and SvNOK will be set, while SvIOK wont be.
398
07fa94a1 399In general, though, it's best to use the C<Sv*V> macros.
a0d0e21e 400
54310121 401=head2 Working with AVs
a0d0e21e 402
07fa94a1
JO
403There are two ways to create and load an AV. The first method creates an
404empty AV:
a0d0e21e
LW
405
406 AV* newAV();
407
54310121 408The second method both creates the AV and initially populates it with SVs:
a0d0e21e 409
c70927a6 410 AV* av_make(SSize_t num, SV **ptr);
a0d0e21e 411
5f05dabc 412The second argument points to an array containing C<num> C<SV*>'s. Once the
54310121 413AV has been created, the SVs can be destroyed, if so desired.
a0d0e21e 414
da8c5729 415Once the AV has been created, the following operations are possible on it:
a0d0e21e
LW
416
417 void av_push(AV*, SV*);
418 SV* av_pop(AV*);
419 SV* av_shift(AV*);
c70927a6 420 void av_unshift(AV*, SSize_t num);
a0d0e21e
LW
421
422These should be familiar operations, with the exception of C<av_unshift>.
423This routine adds C<num> elements at the front of the array with the C<undef>
424value. You must then use C<av_store> (described below) to assign values
425to these new elements.
426
427Here are some other functions:
428
c70927a6
FC
429 SSize_t av_top_index(AV*);
430 SV** av_fetch(AV*, SSize_t key, I32 lval);
431 SV** av_store(AV*, SSize_t key, SV* val);
a0d0e21e 432
dab460cd 433The C<av_top_index> function returns the highest index value in an array (just
5f05dabc
PP
434like $#array in Perl). If the array is empty, -1 is returned. The
435C<av_fetch> function returns the value at index C<key>, but if C<lval>
436is non-zero, then C<av_fetch> will store an undef value at that index.
04343c6d
GS
437The C<av_store> function stores the value C<val> at index C<key>, and does
438not increment the reference count of C<val>. Thus the caller is responsible
439for taking care of that, and if C<av_store> returns NULL, the caller will
440have to decrement the reference count to avoid a memory leak. Note that
441C<av_fetch> and C<av_store> both return C<SV**>'s, not C<SV*>'s as their
442return value.
d1b91892 443
da8c5729
MH
444A few more:
445
a0d0e21e 446 void av_clear(AV*);
a0d0e21e 447 void av_undef(AV*);
c70927a6 448 void av_extend(AV*, SSize_t key);
5f05dabc
PP
449
450The C<av_clear> function deletes all the elements in the AV* array, but
451does not actually delete the array itself. The C<av_undef> function will
452delete all the elements in the array plus the array itself. The
adc882cf
GS
453C<av_extend> function extends the array so that it contains at least C<key+1>
454elements. If C<key+1> is less than the currently allocated length of the array,
455then nothing is done.
a0d0e21e
LW
456
457If you know the name of an array variable, you can get a pointer to its AV
458by using the following:
459
cbfd0a87 460 AV* get_av("package::varname", 0);
a0d0e21e
LW
461
462This returns NULL if the variable does not exist.
463
5a0de581 464See L</Understanding the Magic of Tied Hashes and Arrays> for more
04343c6d
GS
465information on how to use the array access functions on tied arrays.
466
54310121 467=head2 Working with HVs
a0d0e21e
LW
468
469To create an HV, you use the following routine:
470
471 HV* newHV();
472
da8c5729 473Once the HV has been created, the following operations are possible on it:
a0d0e21e 474
08105a92
GS
475 SV** hv_store(HV*, const char* key, U32 klen, SV* val, U32 hash);
476 SV** hv_fetch(HV*, const char* key, U32 klen, I32 lval);
a0d0e21e 477
5f05dabc
PP
478The C<klen> parameter is the length of the key being passed in (Note that
479you cannot pass 0 in as a value of C<klen> to tell Perl to measure the
480length of the key). The C<val> argument contains the SV pointer to the
54310121 481scalar being stored, and C<hash> is the precomputed hash value (zero if
5f05dabc
PP
482you want C<hv_store> to calculate it for you). The C<lval> parameter
483indicates whether this fetch is actually a part of a store operation, in
484which case a new undefined value will be added to the HV with the supplied
485key and C<hv_fetch> will return as if the value had already existed.
a0d0e21e 486
5f05dabc
PP
487Remember that C<hv_store> and C<hv_fetch> return C<SV**>'s and not just
488C<SV*>. To access the scalar value, you must first dereference the return
489value. However, you should check to make sure that the return value is
490not NULL before dereferencing it.
a0d0e21e 491
da8c5729
MH
492The first of these two functions checks if a hash table entry exists, and the
493second deletes it.
a0d0e21e 494
08105a92
GS
495 bool hv_exists(HV*, const char* key, U32 klen);
496 SV* hv_delete(HV*, const char* key, U32 klen, I32 flags);
a0d0e21e 497
5f05dabc
PP
498If C<flags> does not include the C<G_DISCARD> flag then C<hv_delete> will
499create and return a mortal copy of the deleted value.
500
a0d0e21e
LW
501And more miscellaneous functions:
502
503 void hv_clear(HV*);
a0d0e21e 504 void hv_undef(HV*);
5f05dabc
PP
505
506Like their AV counterparts, C<hv_clear> deletes all the entries in the hash
507table but does not actually delete the hash table. The C<hv_undef> deletes
508both the entries and the hash table itself.
a0d0e21e 509
a9b0660e 510Perl keeps the actual data in a linked list of structures with a typedef of HE.
d1b91892
AD
511These contain the actual key and value pointers (plus extra administrative
512overhead). The key is a string pointer; the value is an C<SV*>. However,
513once you have an C<HE*>, to get the actual key and value, use the routines
514specified below.
515
a0d0e21e
LW
516 I32 hv_iterinit(HV*);
517 /* Prepares starting point to traverse hash table */
518 HE* hv_iternext(HV*);
519 /* Get the next entry, and return a pointer to a
520 structure that has both the key and value */
521 char* hv_iterkey(HE* entry, I32* retlen);
522 /* Get the key from an HE structure and also return
523 the length of the key string */
cb1a09d0 524 SV* hv_iterval(HV*, HE* entry);
d1be9408 525 /* Return an SV pointer to the value of the HE
a0d0e21e 526 structure */
cb1a09d0 527 SV* hv_iternextsv(HV*, char** key, I32* retlen);
d1b91892
AD
528 /* This convenience routine combines hv_iternext,
529 hv_iterkey, and hv_iterval. The key and retlen
530 arguments are return values for the key and its
531 length. The value is returned in the SV* argument */
a0d0e21e
LW
532
533If you know the name of a hash variable, you can get a pointer to its HV
534by using the following:
535
6673a63c 536 HV* get_hv("package::varname", 0);
a0d0e21e
LW
537
538This returns NULL if the variable does not exist.
539
a43e7901 540The hash algorithm is defined in the C<PERL_HASH> macro:
a0d0e21e 541
a43e7901 542 PERL_HASH(hash, key, klen)
ab192400 543
a43e7901
YO
544The exact implementation of this macro varies by architecture and version
545of perl, and the return value may change per invocation, so the value
546is only valid for the duration of a single perl process.
a0d0e21e 547
5a0de581 548See L</Understanding the Magic of Tied Hashes and Arrays> for more
04343c6d
GS
549information on how to use the hash access functions on tied hashes.
550
1e422769
PP
551=head2 Hash API Extensions
552
553Beginning with version 5.004, the following functions are also supported:
554
555 HE* hv_fetch_ent (HV* tb, SV* key, I32 lval, U32 hash);
556 HE* hv_store_ent (HV* tb, SV* key, SV* val, U32 hash);
c47ff5f1 557
1e422769
PP
558 bool hv_exists_ent (HV* tb, SV* key, U32 hash);
559 SV* hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
c47ff5f1 560
1e422769
PP
561 SV* hv_iterkeysv (HE* entry);
562
563Note that these functions take C<SV*> keys, which simplifies writing
564of extension code that deals with hash structures. These functions
565also allow passing of C<SV*> keys to C<tie> functions without forcing
566you to stringify the keys (unlike the previous set of functions).
567
568They also return and accept whole hash entries (C<HE*>), making their
569use more efficient (since the hash number for a particular string
4a4eefd0
GS
570doesn't have to be recomputed every time). See L<perlapi> for detailed
571descriptions.
1e422769
PP
572
573The following macros must always be used to access the contents of hash
574entries. Note that the arguments to these macros must be simple
575variables, since they may get evaluated more than once. See
4a4eefd0 576L<perlapi> for detailed descriptions of these macros.
1e422769
PP
577
578 HePV(HE* he, STRLEN len)
579 HeVAL(HE* he)
580 HeHASH(HE* he)
581 HeSVKEY(HE* he)
582 HeSVKEY_force(HE* he)
583 HeSVKEY_set(HE* he, SV* sv)
584
585These two lower level macros are defined, but must only be used when
586dealing with keys that are not C<SV*>s:
587
588 HeKEY(HE* he)
589 HeKLEN(HE* he)
590
04343c6d
GS
591Note that both C<hv_store> and C<hv_store_ent> do not increment the
592reference count of the stored C<val>, which is the caller's responsibility.
593If these functions return a NULL value, the caller will usually have to
594decrement the reference count of C<val> to avoid a memory leak.
1e422769 595
a9381218
MHM
596=head2 AVs, HVs and undefined values
597
10e2eb10
FC
598Sometimes you have to store undefined values in AVs or HVs. Although
599this may be a rare case, it can be tricky. That's because you're
a9381218
MHM
600used to using C<&PL_sv_undef> if you need an undefined SV.
601
602For example, intuition tells you that this XS code:
603
604 AV *av = newAV();
605 av_store( av, 0, &PL_sv_undef );
606
607is equivalent to this Perl code:
608
609 my @av;
610 $av[0] = undef;
611
f3c4ec28 612Unfortunately, this isn't true. In perl 5.18 and earlier, AVs use C<&PL_sv_undef> as a marker
a9381218
MHM
613for indicating that an array element has not yet been initialized.
614Thus, C<exists $av[0]> would be true for the above Perl code, but
f3c4ec28
FC
615false for the array generated by the XS code. In perl 5.20, storing
616&PL_sv_undef will create a read-only element, because the scalar
617&PL_sv_undef itself is stored, not a copy.
a9381218 618
f3c4ec28 619Similar problems can occur when storing C<&PL_sv_undef> in HVs:
a9381218
MHM
620
621 hv_store( hv, "key", 3, &PL_sv_undef, 0 );
622
623This will indeed make the value C<undef>, but if you try to modify
624the value of C<key>, you'll get the following error:
625
626 Modification of non-creatable hash value attempted
627
628In perl 5.8.0, C<&PL_sv_undef> was also used to mark placeholders
10e2eb10 629in restricted hashes. This caused such hash entries not to appear
a9381218
MHM
630when iterating over the hash or when checking for the keys
631with the C<hv_exists> function.
632
8abccac8 633You can run into similar problems when you store C<&PL_sv_yes> or
10e2eb10 634C<&PL_sv_no> into AVs or HVs. Trying to modify such elements
a9381218
MHM
635will give you the following error:
636
637 Modification of a read-only value attempted
638
639To make a long story short, you can use the special variables
8abccac8 640C<&PL_sv_undef>, C<&PL_sv_yes> and C<&PL_sv_no> with AVs and
a9381218
MHM
641HVs, but you have to make sure you know what you're doing.
642
643Generally, if you want to store an undefined value in an AV
644or HV, you should not use C<&PL_sv_undef>, but rather create a
645new undefined value using the C<newSV> function, for example:
646
647 av_store( av, 42, newSV(0) );
648 hv_store( hv, "foo", 3, newSV(0), 0 );
649
a0d0e21e
LW
650=head2 References
651
d1b91892 652References are a special type of scalar that point to other data types
a9b0660e 653(including other references).
a0d0e21e 654
07fa94a1 655To create a reference, use either of the following functions:
a0d0e21e 656
5f05dabc
PP
657 SV* newRV_inc((SV*) thing);
658 SV* newRV_noinc((SV*) thing);
a0d0e21e 659
5f05dabc 660The C<thing> argument can be any of an C<SV*>, C<AV*>, or C<HV*>. The
07fa94a1
JO
661functions are identical except that C<newRV_inc> increments the reference
662count of the C<thing>, while C<newRV_noinc> does not. For historical
663reasons, C<newRV> is a synonym for C<newRV_inc>.
664
665Once you have a reference, you can use the following macro to dereference
666the reference:
a0d0e21e
LW
667
668 SvRV(SV*)
669
670then call the appropriate routines, casting the returned C<SV*> to either an
d1b91892 671C<AV*> or C<HV*>, if required.
a0d0e21e 672
d1b91892 673To determine if an SV is a reference, you can use the following macro:
a0d0e21e
LW
674
675 SvROK(SV*)
676
07fa94a1
JO
677To discover what type of value the reference refers to, use the following
678macro and then check the return value.
d1b91892
AD
679
680 SvTYPE(SvRV(SV*))
681
682The most useful types that will be returned are:
683
a5e62da0
FC
684 < SVt_PVAV Scalar
685 SVt_PVAV Array
686 SVt_PVHV Hash
687 SVt_PVCV Code
688 SVt_PVGV Glob (possibly a file handle)
689
690See L<perlapi/svtype> for more details.
d1b91892 691
cb1a09d0
AD
692=head2 Blessed References and Class Objects
693
06f6df17 694References are also used to support object-oriented programming. In perl's
cb1a09d0
AD
695OO lexicon, an object is simply a reference that has been blessed into a
696package (or class). Once blessed, the programmer may now use the reference
697to access the various methods in the class.
698
699A reference can be blessed into a package with the following function:
700
701 SV* sv_bless(SV* sv, HV* stash);
702
06f6df17
RGS
703The C<sv> argument must be a reference value. The C<stash> argument
704specifies which class the reference will belong to. See
5a0de581 705L</Stashes and Globs> for information on converting class names into stashes.
cb1a09d0
AD
706
707/* Still under construction */
708
ddd2cc91
DM
709The following function upgrades rv to reference if not already one.
710Creates a new SV for rv to point to. If C<classname> is non-null, the SV
711is blessed into the specified class. SV is returned.
cb1a09d0 712
08105a92 713 SV* newSVrv(SV* rv, const char* classname);
cb1a09d0 714
ddd2cc91
DM
715The following three functions copy integer, unsigned integer or double
716into an SV whose reference is C<rv>. SV is blessed if C<classname> is
717non-null.
cb1a09d0 718
08105a92 719 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
e1c57cef 720 SV* sv_setref_uv(SV* rv, const char* classname, UV uv);
08105a92 721 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
cb1a09d0 722
ddd2cc91
DM
723The following function copies the pointer value (I<the address, not the
724string!>) into an SV whose reference is rv. SV is blessed if C<classname>
725is non-null.
cb1a09d0 726
ddd2cc91 727 SV* sv_setref_pv(SV* rv, const char* classname, void* pv);
cb1a09d0 728
a9b0660e 729The following function copies a string into an SV whose reference is C<rv>.
ddd2cc91
DM
730Set length to 0 to let Perl calculate the string length. SV is blessed if
731C<classname> is non-null.
cb1a09d0 732
a9b0660e
KW
733 SV* sv_setref_pvn(SV* rv, const char* classname, char* pv,
734 STRLEN length);
cb1a09d0 735
ddd2cc91
DM
736The following function tests whether the SV is blessed into the specified
737class. It does not check inheritance relationships.
9abd00ed 738
08105a92 739 int sv_isa(SV* sv, const char* name);
9abd00ed 740
ddd2cc91 741The following function tests whether the SV is a reference to a blessed object.
9abd00ed
GS
742
743 int sv_isobject(SV* sv);
744
ddd2cc91 745The following function tests whether the SV is derived from the specified
10e2eb10
FC
746class. SV can be either a reference to a blessed object or a string
747containing a class name. This is the function implementing the
ddd2cc91 748C<UNIVERSAL::isa> functionality.
9abd00ed 749
08105a92 750 bool sv_derived_from(SV* sv, const char* name);
9abd00ed 751
00aadd71 752To check if you've got an object derived from a specific class you have
9abd00ed
GS
753to write:
754
755 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
cb1a09d0 756
5f05dabc 757=head2 Creating New Variables
cb1a09d0 758
5f05dabc
PP
759To create a new Perl variable with an undef value which can be accessed from
760your Perl script, use the following routines, depending on the variable type.
cb1a09d0 761
64ace3f8 762 SV* get_sv("package::varname", GV_ADD);
cbfd0a87 763 AV* get_av("package::varname", GV_ADD);
6673a63c 764 HV* get_hv("package::varname", GV_ADD);
cb1a09d0 765
058a5f6c 766Notice the use of GV_ADD as the second parameter. The new variable can now
cb1a09d0
AD
767be set, using the routines appropriate to the data type.
768
5f05dabc 769There are additional macros whose values may be bitwise OR'ed with the
058a5f6c 770C<GV_ADD> argument to enable certain extra features. Those bits are:
cb1a09d0 771
9a68f1db
SB
772=over
773
774=item GV_ADDMULTI
775
776Marks the variable as multiply defined, thus preventing the:
777
778 Name <varname> used only once: possible typo
779
780warning.
781
9a68f1db
SB
782=item GV_ADDWARN
783
784Issues the warning:
785
786 Had to create <varname> unexpectedly
787
788if the variable did not exist before the function was called.
789
790=back
cb1a09d0 791
07fa94a1
JO
792If you do not specify a package name, the variable is created in the current
793package.
cb1a09d0 794
5f05dabc 795=head2 Reference Counts and Mortality
a0d0e21e 796
10e2eb10 797Perl uses a reference count-driven garbage collection mechanism. SVs,
54310121 798AVs, or HVs (xV for short in the following) start their life with a
55497cff 799reference count of 1. If the reference count of an xV ever drops to 0,
07fa94a1 800then it will be destroyed and its memory made available for reuse.
55497cff
PP
801
802This normally doesn't happen at the Perl level unless a variable is
5f05dabc
PP
803undef'ed or the last variable holding a reference to it is changed or
804overwritten. At the internal level, however, reference counts can be
55497cff
PP
805manipulated with the following macros:
806
807 int SvREFCNT(SV* sv);
5f05dabc 808 SV* SvREFCNT_inc(SV* sv);
55497cff
PP
809 void SvREFCNT_dec(SV* sv);
810
811However, there is one other function which manipulates the reference
07fa94a1
JO
812count of its argument. The C<newRV_inc> function, you will recall,
813creates a reference to the specified argument. As a side effect,
814it increments the argument's reference count. If this is not what
815you want, use C<newRV_noinc> instead.
816
817For example, imagine you want to return a reference from an XSUB function.
818Inside the XSUB routine, you create an SV which initially has a reference
819count of one. Then you call C<newRV_inc>, passing it the just-created SV.
5f05dabc
PP
820This returns the reference as a new SV, but the reference count of the
821SV you passed to C<newRV_inc> has been incremented to two. Now you
07fa94a1
JO
822return the reference from the XSUB routine and forget about the SV.
823But Perl hasn't! Whenever the returned reference is destroyed, the
824reference count of the original SV is decreased to one and nothing happens.
825The SV will hang around without any way to access it until Perl itself
826terminates. This is a memory leak.
5f05dabc
PP
827
828The correct procedure, then, is to use C<newRV_noinc> instead of
faed5253
JO
829C<newRV_inc>. Then, if and when the last reference is destroyed,
830the reference count of the SV will go to zero and it will be destroyed,
07fa94a1 831stopping any memory leak.
55497cff 832
5f05dabc 833There are some convenience functions available that can help with the
54310121 834destruction of xVs. These functions introduce the concept of "mortality".
07fa94a1
JO
835An xV that is mortal has had its reference count marked to be decremented,
836but not actually decremented, until "a short time later". Generally the
837term "short time later" means a single Perl statement, such as a call to
54310121 838an XSUB function. The actual determinant for when mortal xVs have their
07fa94a1
JO
839reference count decremented depends on two macros, SAVETMPS and FREETMPS.
840See L<perlcall> and L<perlxs> for more details on these macros.
55497cff
PP
841
842"Mortalization" then is at its simplest a deferred C<SvREFCNT_dec>.
843However, if you mortalize a variable twice, the reference count will
844later be decremented twice.
845
00aadd71
NIS
846"Mortal" SVs are mainly used for SVs that are placed on perl's stack.
847For example an SV which is created just to pass a number to a called sub
06f6df17 848is made mortal to have it cleaned up automatically when it's popped off
10e2eb10 849the stack. Similarly, results returned by XSUBs (which are pushed on the
06f6df17 850stack) are often made mortal.
a0d0e21e
LW
851
852To create a mortal variable, use the functions:
853
854 SV* sv_newmortal()
855 SV* sv_2mortal(SV*)
856 SV* sv_mortalcopy(SV*)
857
00aadd71 858The first call creates a mortal SV (with no value), the second converts an existing
5f05dabc
PP
859SV to a mortal SV (and thus defers a call to C<SvREFCNT_dec>), and the
860third creates a mortal copy of an existing SV.
da8c5729 861Because C<sv_newmortal> gives the new SV no value, it must normally be given one
9a68f1db 862via C<sv_setpv>, C<sv_setiv>, etc. :
00aadd71
NIS
863
864 SV *tmp = sv_newmortal();
865 sv_setiv(tmp, an_integer);
866
867As that is multiple C statements it is quite common so see this idiom instead:
868
869 SV *tmp = sv_2mortal(newSViv(an_integer));
870
871
872You should be careful about creating mortal variables. Strange things
873can happen if you make the same value mortal within multiple contexts,
10e2eb10
FC
874or if you make a variable mortal multiple
875times. Thinking of "Mortalization"
00aadd71 876as deferred C<SvREFCNT_dec> should help to minimize such problems.
da8c5729 877For example if you are passing an SV which you I<know> has a high enough REFCNT
00aadd71
NIS
878to survive its use on the stack you need not do any mortalization.
879If you are not sure then doing an C<SvREFCNT_inc> and C<sv_2mortal>, or
880making a C<sv_mortalcopy> is safer.
a0d0e21e 881
ac036724 882The mortal routines are not just for SVs; AVs and HVs can be
faed5253 883made mortal by passing their address (type-casted to C<SV*>) to the
07fa94a1 884C<sv_2mortal> or C<sv_mortalcopy> routines.
a0d0e21e 885
5f05dabc 886=head2 Stashes and Globs
a0d0e21e 887
06f6df17
RGS
888A B<stash> is a hash that contains all variables that are defined
889within a package. Each key of the stash is a symbol
aa689395
PP
890name (shared by all the different types of objects that have the same
891name), and each value in the hash table is a GV (Glob Value). This GV
892in turn contains references to the various objects of that name,
893including (but not limited to) the following:
cb1a09d0 894
a0d0e21e
LW
895 Scalar Value
896 Array Value
897 Hash Value
a3cb178b 898 I/O Handle
a0d0e21e
LW
899 Format
900 Subroutine
901
06f6df17
RGS
902There is a single stash called C<PL_defstash> that holds the items that exist
903in the C<main> package. To get at the items in other packages, append the
904string "::" to the package name. The items in the C<Foo> package are in
905the stash C<Foo::> in PL_defstash. The items in the C<Bar::Baz> package are
906in the stash C<Baz::> in C<Bar::>'s stash.
a0d0e21e 907
d1b91892 908To get the stash pointer for a particular package, use the function:
a0d0e21e 909
da51bb9b
NC
910 HV* gv_stashpv(const char* name, I32 flags)
911 HV* gv_stashsv(SV*, I32 flags)
a0d0e21e
LW
912
913The first function takes a literal string, the second uses the string stored
d1b91892 914in the SV. Remember that a stash is just a hash table, so you get back an
da51bb9b 915C<HV*>. The C<flags> flag will create a new package if it is set to GV_ADD.
a0d0e21e
LW
916
917The name that C<gv_stash*v> wants is the name of the package whose symbol table
918you want. The default package is called C<main>. If you have multiply nested
d1b91892
AD
919packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
920language itself.
a0d0e21e
LW
921
922Alternately, if you have an SV that is a blessed reference, you can find
923out the stash pointer by using:
924
925 HV* SvSTASH(SvRV(SV*));
926
927then use the following to get the package name itself:
928
929 char* HvNAME(HV* stash);
930
5f05dabc
PP
931If you need to bless or re-bless an object you can use the following
932function:
a0d0e21e
LW
933
934 SV* sv_bless(SV*, HV* stash)
935
936where the first argument, an C<SV*>, must be a reference, and the second
937argument is a stash. The returned C<SV*> can now be used in the same way
938as any other SV.
939
d1b91892
AD
940For more information on references and blessings, consult L<perlref>.
941
54310121 942=head2 Double-Typed SVs
0a753a76
PP
943
944Scalar variables normally contain only one type of value, an integer,
945double, pointer, or reference. Perl will automatically convert the
946actual scalar data from the stored type into the requested type.
947
948Some scalar variables contain more than one type of scalar data. For
949example, the variable C<$!> contains either the numeric value of C<errno>
950or its string equivalent from either C<strerror> or C<sys_errlist[]>.
951
952To force multiple data values into an SV, you must do two things: use the
953C<sv_set*v> routines to add the additional scalar type, then set a flag
954so that Perl will believe it contains more than one type of data. The
955four macros to set the flags are:
956
957 SvIOK_on
958 SvNOK_on
959 SvPOK_on
960 SvROK_on
961
962The particular macro you must use depends on which C<sv_set*v> routine
963you called first. This is because every C<sv_set*v> routine turns on
964only the bit for the particular type of data being set, and turns off
965all the rest.
966
967For example, to create a new Perl variable called "dberror" that contains
968both the numeric and descriptive string error values, you could use the
969following code:
970
971 extern int dberror;
972 extern char *dberror_list;
973
64ace3f8 974 SV* sv = get_sv("dberror", GV_ADD);
0a753a76
PP
975 sv_setiv(sv, (IV) dberror);
976 sv_setpv(sv, dberror_list[dberror]);
977 SvIOK_on(sv);
978
979If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
980macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
981
4f4531b8
FC
982=head2 Read-Only Values
983
984In Perl 5.16 and earlier, copy-on-write (see the next section) shared a
985flag bit with read-only scalars. So the only way to test whether
986C<sv_setsv>, etc., will raise a "Modification of a read-only value" error
987in those versions is:
988
989 SvREADONLY(sv) && !SvIsCOW(sv)
990
991Under Perl 5.18 and later, SvREADONLY only applies to read-only variables,
992and, under 5.20, copy-on-write scalars can also be read-only, so the above
993check is incorrect. You just want:
994
995 SvREADONLY(sv)
996
997If you need to do this check often, define your own macro like this:
998
999 #if PERL_VERSION >= 18
1000 # define SvTRULYREADONLY(sv) SvREADONLY(sv)
1001 #else
1002 # define SvTRULYREADONLY(sv) (SvREADONLY(sv) && !SvIsCOW(sv))
1003 #endif
1004
1005=head2 Copy on Write
1006
1007Perl implements a copy-on-write (COW) mechanism for scalars, in which
1008string copies are not immediately made when requested, but are deferred
1009until made necessary by one or the other scalar changing. This is mostly
1010transparent, but one must take care not to modify string buffers that are
1011shared by multiple SVs.
1012
1013You can test whether an SV is using copy-on-write with C<SvIsCOW(sv)>.
1014
1015You can force an SV to make its own copy of its string buffer by calling C<sv_force_normal(sv)> or SvPV_force_nolen(sv).
1016
1017If you want to make the SV drop its string buffer, use
1018C<sv_force_normal_flags(sv, SV_COW_DROP_PV)> or simply
1019C<sv_setsv(sv, NULL)>.
1020
1021All of these functions will croak on read-only scalars (see the previous
1022section for more on those).
1023
1024To test that your code is behaving correctly and not modifying COW buffers,
1025on systems that support L<mmap(2)> (i.e., Unix) you can configure perl with
1026C<-Accflags=-DPERL_DEBUG_READONLY_COW> and it will turn buffer violations
1027into crashes. You will find it to be marvellously slow, so you may want to
1028skip perl's own tests.
1029
0a753a76 1030=head2 Magic Variables
a0d0e21e 1031
d1b91892
AD
1032[This section still under construction. Ignore everything here. Post no
1033bills. Everything not permitted is forbidden.]
1034
d1b91892
AD
1035Any SV may be magical, that is, it has special features that a normal
1036SV does not have. These features are stored in the SV structure in a
5f05dabc 1037linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
d1b91892
AD
1038
1039 struct magic {
1040 MAGIC* mg_moremagic;
1041 MGVTBL* mg_virtual;
1042 U16 mg_private;
1043 char mg_type;
1044 U8 mg_flags;
b205eb13 1045 I32 mg_len;
d1b91892
AD
1046 SV* mg_obj;
1047 char* mg_ptr;
d1b91892
AD
1048 };
1049
1050Note this is current as of patchlevel 0, and could change at any time.
1051
1052=head2 Assigning Magic
1053
1054Perl adds magic to an SV using the sv_magic function:
1055
a9b0660e 1056 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
d1b91892
AD
1057
1058The C<sv> argument is a pointer to the SV that is to acquire a new magical
1059feature.
1060
1061If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
10e2eb10
FC
1062convert C<sv> to type C<SVt_PVMG>.
1063Perl then continues by adding new magic
645c22ef
DM
1064to the beginning of the linked list of magical features. Any prior entry
1065of the same type of magic is deleted. Note that this can be overridden,
1066and multiple instances of the same type of magic can be associated with an
1067SV.
d1b91892 1068
54310121 1069The C<name> and C<namlen> arguments are used to associate a string with
10e2eb10 1070the magic, typically the name of a variable. C<namlen> is stored in the
2d8d5d5a
SH
1071C<mg_len> field and if C<name> is non-null then either a C<savepvn> copy of
1072C<name> or C<name> itself is stored in the C<mg_ptr> field, depending on
1073whether C<namlen> is greater than zero or equal to zero respectively. As a
1074special case, if C<(name && namlen == HEf_SVKEY)> then C<name> is assumed
1075to contain an C<SV*> and is stored as-is with its REFCNT incremented.
d1b91892
AD
1076
1077The sv_magic function uses C<how> to determine which, if any, predefined
1078"Magic Virtual Table" should be assigned to the C<mg_virtual> field.
5a0de581 1079See the L</Magic Virtual Tables> section below. The C<how> argument is also
10e2eb10
FC
1080stored in the C<mg_type> field. The value of
1081C<how> should be chosen from the set of macros
1082C<PERL_MAGIC_foo> found in F<perl.h>. Note that before
645c22ef 1083these macros were added, Perl internals used to directly use character
14befaf4 1084literals, so you may occasionally come across old code or documentation
75d0f26d 1085referring to 'U' magic rather than C<PERL_MAGIC_uvar> for example.
d1b91892
AD
1086
1087The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
1088structure. If it is not the same as the C<sv> argument, the reference
1089count of the C<obj> object is incremented. If it is the same, or if
27deb0cf
YO
1090the C<how> argument is C<PERL_MAGIC_arylen>, C<PERL_MAGIC_regdatum>,
1091C<PERL_MAGIC_regdata>, or if it is a NULL pointer, then C<obj> is merely
1092stored, without the reference count being incremented.
d1b91892 1093
2d8d5d5a
SH
1094See also C<sv_magicext> in L<perlapi> for a more flexible way to add magic
1095to an SV.
1096
cb1a09d0
AD
1097There is also a function to add magic to an C<HV>:
1098
1099 void hv_magic(HV *hv, GV *gv, int how);
1100
1101This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
1102
1103To remove the magic from an SV, call the function sv_unmagic:
1104
70a53b35 1105 int sv_unmagic(SV *sv, int type);
cb1a09d0
AD
1106
1107The C<type> argument should be equal to the C<how> value when the C<SV>
1108was initially made magical.
1109
f6ee7b17 1110However, note that C<sv_unmagic> removes all magic of a certain C<type> from the
10e2eb10
FC
1111C<SV>. If you want to remove only certain
1112magic of a C<type> based on the magic
f6ee7b17
FR
1113virtual table, use C<sv_unmagicext> instead:
1114
1115 int sv_unmagicext(SV *sv, int type, MGVTBL *vtbl);
1116
d1b91892
AD
1117=head2 Magic Virtual Tables
1118
d1be9408 1119The C<mg_virtual> field in the C<MAGIC> structure is a pointer to an
d1b91892
AD
1120C<MGVTBL>, which is a structure of function pointers and stands for
1121"Magic Virtual Table" to handle the various operations that might be
1122applied to that variable.
1123
301cb7e8
DM
1124The C<MGVTBL> has five (or sometimes eight) pointers to the following
1125routine types:
d1b91892 1126
e97ca230
DM
1127 int (*svt_get) (pTHX_ SV* sv, MAGIC* mg);
1128 int (*svt_set) (pTHX_ SV* sv, MAGIC* mg);
1129 U32 (*svt_len) (pTHX_ SV* sv, MAGIC* mg);
1130 int (*svt_clear)(pTHX_ SV* sv, MAGIC* mg);
1131 int (*svt_free) (pTHX_ SV* sv, MAGIC* mg);
d1b91892 1132
e97ca230 1133 int (*svt_copy) (pTHX_ SV *sv, MAGIC* mg, SV *nsv,
a9b0660e 1134 const char *name, I32 namlen);
e97ca230
DM
1135 int (*svt_dup) (pTHX_ MAGIC *mg, CLONE_PARAMS *param);
1136 int (*svt_local)(pTHX_ SV *nsv, MAGIC *mg);
301cb7e8
DM
1137
1138
06f6df17 1139This MGVTBL structure is set at compile-time in F<perl.h> and there are
b7a0f54c
SM
1140currently 32 types. These different structures contain pointers to various
1141routines that perform additional actions depending on which function is
1142being called.
d1b91892 1143
a9b0660e
KW
1144 Function pointer Action taken
1145 ---------------- ------------
1146 svt_get Do something before the value of the SV is
1147 retrieved.
1148 svt_set Do something after the SV is assigned a value.
1149 svt_len Report on the SV's length.
1150 svt_clear Clear something the SV represents.
1151 svt_free Free any extra storage associated with the SV.
d1b91892 1152
a9b0660e
KW
1153 svt_copy copy tied variable magic to a tied element
1154 svt_dup duplicate a magic structure during thread cloning
1155 svt_local copy magic to local value during 'local'
301cb7e8 1156
d1b91892 1157For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
14befaf4 1158to an C<mg_type> of C<PERL_MAGIC_sv>) contains:
d1b91892
AD
1159
1160 { magic_get, magic_set, magic_len, 0, 0 }
1161
14befaf4
DM
1162Thus, when an SV is determined to be magical and of type C<PERL_MAGIC_sv>,
1163if a get operation is being performed, the routine C<magic_get> is
1164called. All the various routines for the various magical types begin
1165with C<magic_>. NOTE: the magic routines are not considered part of
1166the Perl API, and may not be exported by the Perl library.
d1b91892 1167
301cb7e8
DM
1168The last three slots are a recent addition, and for source code
1169compatibility they are only checked for if one of the three flags
10e2eb10
FC
1170MGf_COPY, MGf_DUP or MGf_LOCAL is set in mg_flags.
1171This means that most code can continue declaring
1172a vtable as a 5-element value. These three are
301cb7e8
DM
1173currently used exclusively by the threading code, and are highly subject
1174to change.
1175
d1b91892
AD
1176The current kinds of Magic Virtual Tables are:
1177
f1f5ddd7
FC
1178=for comment
1179This table is generated by regen/mg_vtable.pl. Any changes made here
1180will be lost.
1181
1182=for mg_vtable.pl begin
1183
a9b0660e 1184 mg_type
bd6e6c12
FC
1185 (old-style char and macro) MGVTBL Type of magic
1186 -------------------------- ------ -------------
1187 \0 PERL_MAGIC_sv vtbl_sv Special scalar variable
1188 # PERL_MAGIC_arylen vtbl_arylen Array length ($#ary)
e5e1ee61 1189 % PERL_MAGIC_rhash (none) Extra data for restricted
bd6e6c12 1190 hashes
a6d69523
TC
1191 * PERL_MAGIC_debugvar vtbl_debugvar $DB::single, signal, trace
1192 vars
bd6e6c12 1193 . PERL_MAGIC_pos vtbl_pos pos() lvalue
e5e1ee61 1194 : PERL_MAGIC_symtab (none) Extra data for symbol
bd6e6c12 1195 tables
e5e1ee61
FC
1196 < PERL_MAGIC_backref vtbl_backref For weak ref data
1197 @ PERL_MAGIC_arylen_p (none) To move arylen out of XPVAV
bd6e6c12
FC
1198 B PERL_MAGIC_bm vtbl_regexp Boyer-Moore
1199 (fast string search)
1200 c PERL_MAGIC_overload_table vtbl_ovrld Holds overload table
1201 (AMT) on stash
1202 D PERL_MAGIC_regdata vtbl_regdata Regex match position data
1203 (@+ and @- vars)
1204 d PERL_MAGIC_regdatum vtbl_regdatum Regex match position data
1205 element
1206 E PERL_MAGIC_env vtbl_env %ENV hash
1207 e PERL_MAGIC_envelem vtbl_envelem %ENV hash element
eccba044 1208 f PERL_MAGIC_fm vtbl_regexp Formline
bd6e6c12 1209 ('compiled' format)
bd6e6c12
FC
1210 g PERL_MAGIC_regex_global vtbl_mglob m//g target
1211 H PERL_MAGIC_hints vtbl_hints %^H hash
1212 h PERL_MAGIC_hintselem vtbl_hintselem %^H hash element
1213 I PERL_MAGIC_isa vtbl_isa @ISA array
1214 i PERL_MAGIC_isaelem vtbl_isaelem @ISA array element
1215 k PERL_MAGIC_nkeys vtbl_nkeys scalar(keys()) lvalue
1216 L PERL_MAGIC_dbfile (none) Debugger %_<filename
1217 l PERL_MAGIC_dbline vtbl_dbline Debugger %_<filename
1218 element
1219 N PERL_MAGIC_shared (none) Shared between threads
1220 n PERL_MAGIC_shared_scalar (none) Shared between threads
1221 o PERL_MAGIC_collxfrm vtbl_collxfrm Locale transformation
1222 P PERL_MAGIC_tied vtbl_pack Tied array or hash
1223 p PERL_MAGIC_tiedelem vtbl_packelem Tied array or hash element
1224 q PERL_MAGIC_tiedscalar vtbl_packelem Tied scalar or handle
e5e1ee61 1225 r PERL_MAGIC_qr vtbl_regexp Precompiled qr// regex
bd6e6c12
FC
1226 S PERL_MAGIC_sig (none) %SIG hash
1227 s PERL_MAGIC_sigelem vtbl_sigelem %SIG hash element
1228 t PERL_MAGIC_taint vtbl_taint Taintedness
1229 U PERL_MAGIC_uvar vtbl_uvar Available for use by
1230 extensions
1231 u PERL_MAGIC_uvar_elem (none) Reserved for use by
1232 extensions
4499db73 1233 V PERL_MAGIC_vstring (none) SV was vstring literal
bd6e6c12
FC
1234 v PERL_MAGIC_vec vtbl_vec vec() lvalue
1235 w PERL_MAGIC_utf8 vtbl_utf8 Cached UTF-8 information
1236 x PERL_MAGIC_substr vtbl_substr substr() lvalue
1237 y PERL_MAGIC_defelem vtbl_defelem Shadow "foreach" iterator
1238 variable / smart parameter
1239 vivification
baabe3fb
FC
1240 \ PERL_MAGIC_lvref vtbl_lvref Lvalue reference
1241 constructor
e5e1ee61 1242 ] PERL_MAGIC_checkcall vtbl_checkcall Inlining/mutation of call
bd6e6c12
FC
1243 to this CV
1244 ~ PERL_MAGIC_ext (none) Available for use by
1245 extensions
0cbee0a4 1246
f1f5ddd7 1247=for mg_vtable.pl end
d1b91892 1248
68dc0745 1249When an uppercase and lowercase letter both exist in the table, then the
92f0c265
JP
1250uppercase letter is typically used to represent some kind of composite type
1251(a list or a hash), and the lowercase letter is used to represent an element
10e2eb10 1252of that composite type. Some internals code makes use of this case
92f0c265 1253relationship. However, 'v' and 'V' (vec and v-string) are in no way related.
14befaf4
DM
1254
1255The C<PERL_MAGIC_ext> and C<PERL_MAGIC_uvar> magic types are defined
1256specifically for use by extensions and will not be used by perl itself.
1257Extensions can use C<PERL_MAGIC_ext> magic to 'attach' private information
1258to variables (typically objects). This is especially useful because
1259there is no way for normal perl code to corrupt this private information
1260(unlike using extra elements of a hash object).
1261
1262Similarly, C<PERL_MAGIC_uvar> magic can be used much like tie() to call a
1263C function any time a scalar's value is used or changed. The C<MAGIC>'s
bdbeb323
SM
1264C<mg_ptr> field points to a C<ufuncs> structure:
1265
1266 struct ufuncs {
a9402793
AB
1267 I32 (*uf_val)(pTHX_ IV, SV*);
1268 I32 (*uf_set)(pTHX_ IV, SV*);
bdbeb323
SM
1269 IV uf_index;
1270 };
1271
1272When the SV is read from or written to, the C<uf_val> or C<uf_set>
14befaf4
DM
1273function will be called with C<uf_index> as the first arg and a pointer to
1274the SV as the second. A simple example of how to add C<PERL_MAGIC_uvar>
1526ead6
AB
1275magic is shown below. Note that the ufuncs structure is copied by
1276sv_magic, so you can safely allocate it on the stack.
1277
1278 void
1279 Umagic(sv)
1280 SV *sv;
1281 PREINIT:
1282 struct ufuncs uf;
1283 CODE:
1284 uf.uf_val = &my_get_fn;
1285 uf.uf_set = &my_set_fn;
1286 uf.uf_index = 0;
14befaf4 1287 sv_magic(sv, 0, PERL_MAGIC_uvar, (char*)&uf, sizeof(uf));
5f05dabc 1288
1e73acc8
AS
1289Attaching C<PERL_MAGIC_uvar> to arrays is permissible but has no effect.
1290
1291For hashes there is a specialized hook that gives control over hash
1292keys (but not values). This hook calls C<PERL_MAGIC_uvar> 'get' magic
1293if the "set" function in the C<ufuncs> structure is NULL. The hook
1294is activated whenever the hash is accessed with a key specified as
1295an C<SV> through the functions C<hv_store_ent>, C<hv_fetch_ent>,
1296C<hv_delete_ent>, and C<hv_exists_ent>. Accessing the key as a string
1297through the functions without the C<..._ent> suffix circumvents the
4509d391 1298hook. See L<Hash::Util::FieldHash/GUTS> for a detailed description.
1e73acc8 1299
14befaf4
DM
1300Note that because multiple extensions may be using C<PERL_MAGIC_ext>
1301or C<PERL_MAGIC_uvar> magic, it is important for extensions to take
1302extra care to avoid conflict. Typically only using the magic on
1303objects blessed into the same class as the extension is sufficient.
2f07f21a
FR
1304For C<PERL_MAGIC_ext> magic, it is usually a good idea to define an
1305C<MGVTBL>, even if all its fields will be C<0>, so that individual
1306C<MAGIC> pointers can be identified as a particular kind of magic
10e2eb10 1307using their magic virtual table. C<mg_findext> provides an easy way
f6ee7b17 1308to do that:
2f07f21a
FR
1309
1310 STATIC MGVTBL my_vtbl = { 0, 0, 0, 0, 0, 0, 0, 0 };
1311
1312 MAGIC *mg;
f6ee7b17
FR
1313 if ((mg = mg_findext(sv, PERL_MAGIC_ext, &my_vtbl))) {
1314 /* this is really ours, not another module's PERL_MAGIC_ext */
1315 my_priv_data_t *priv = (my_priv_data_t *)mg->mg_ptr;
1316 ...
2f07f21a 1317 }
5f05dabc 1318
ef50df4b
GS
1319Also note that the C<sv_set*()> and C<sv_cat*()> functions described
1320earlier do B<not> invoke 'set' magic on their targets. This must
1321be done by the user either by calling the C<SvSETMAGIC()> macro after
1322calling these functions, or by using one of the C<sv_set*_mg()> or
1323C<sv_cat*_mg()> functions. Similarly, generic C code must call the
1324C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
1325obtained from external sources in functions that don't handle magic.
4a4eefd0 1326See L<perlapi> for a description of these functions.
189b2af5
GS
1327For example, calls to the C<sv_cat*()> functions typically need to be
1328followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
1329since their implementation handles 'get' magic.
1330
d1b91892
AD
1331=head2 Finding Magic
1332
a9b0660e
KW
1333 MAGIC *mg_find(SV *sv, int type); /* Finds the magic pointer of that
1334 * type */
f6ee7b17
FR
1335
1336This routine returns a pointer to a C<MAGIC> structure stored in the SV.
10e2eb10
FC
1337If the SV does not have that magical
1338feature, C<NULL> is returned. If the
f6ee7b17 1339SV has multiple instances of that magical feature, the first one will be
10e2eb10
FC
1340returned. C<mg_findext> can be used
1341to find a C<MAGIC> structure of an SV
da8c5729 1342based on both its magic type and its magic virtual table:
f6ee7b17
FR
1343
1344 MAGIC *mg_findext(SV *sv, int type, MGVTBL *vtbl);
d1b91892 1345
f6ee7b17
FR
1346Also, if the SV passed to C<mg_find> or C<mg_findext> is not of type
1347SVt_PVMG, Perl may core dump.
d1b91892 1348
08105a92 1349 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
d1b91892
AD
1350
1351This routine checks to see what types of magic C<sv> has. If the mg_type
68dc0745
PP
1352field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
1353the mg_type field is changed to be the lowercase letter.
a0d0e21e 1354
04343c6d
GS
1355=head2 Understanding the Magic of Tied Hashes and Arrays
1356
14befaf4
DM
1357Tied hashes and arrays are magical beasts of the C<PERL_MAGIC_tied>
1358magic type.
9edb2b46
GS
1359
1360WARNING: As of the 5.004 release, proper usage of the array and hash
1361access functions requires understanding a few caveats. Some
1362of these caveats are actually considered bugs in the API, to be fixed
10e2eb10 1363in later releases, and are bracketed with [MAYCHANGE] below. If
9edb2b46
GS
1364you find yourself actually applying such information in this section, be
1365aware that the behavior may change in the future, umm, without warning.
04343c6d 1366
1526ead6 1367The perl tie function associates a variable with an object that implements
9a68f1db 1368the various GET, SET, etc methods. To perform the equivalent of the perl
1526ead6 1369tie function from an XSUB, you must mimic this behaviour. The code below
61ad4b94 1370carries out the necessary steps -- firstly it creates a new hash, and then
1526ead6 1371creates a second hash which it blesses into the class which will implement
10e2eb10 1372the tie methods. Lastly it ties the two hashes together, and returns a
1526ead6
AB
1373reference to the new tied hash. Note that the code below does NOT call the
1374TIEHASH method in the MyTie class -
5a0de581 1375see L</Calling Perl Routines from within C Programs> for details on how
1526ead6
AB
1376to do this.
1377
1378 SV*
1379 mytie()
1380 PREINIT:
1381 HV *hash;
1382 HV *stash;
1383 SV *tie;
1384 CODE:
1385 hash = newHV();
1386 tie = newRV_noinc((SV*)newHV());
da51bb9b 1387 stash = gv_stashpv("MyTie", GV_ADD);
1526ead6 1388 sv_bless(tie, stash);
899e16d0 1389 hv_magic(hash, (GV*)tie, PERL_MAGIC_tied);
1526ead6
AB
1390 RETVAL = newRV_noinc(hash);
1391 OUTPUT:
1392 RETVAL
1393
04343c6d
GS
1394The C<av_store> function, when given a tied array argument, merely
1395copies the magic of the array onto the value to be "stored", using
1396C<mg_copy>. It may also return NULL, indicating that the value did not
9edb2b46
GS
1397actually need to be stored in the array. [MAYCHANGE] After a call to
1398C<av_store> on a tied array, the caller will usually need to call
1399C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1400TIEARRAY object. If C<av_store> did return NULL, a call to
1401C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1402leak. [/MAYCHANGE]
04343c6d
GS
1403
1404The previous paragraph is applicable verbatim to tied hash access using the
1405C<hv_store> and C<hv_store_ent> functions as well.
1406
1407C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1408C<hv_fetch_ent> actually return an undefined mortal value whose magic
1409has been initialized using C<mg_copy>. Note the value so returned does not
9edb2b46
GS
1410need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1411need to call C<mg_get()> on the returned value in order to actually invoke
1412the perl level "FETCH" method on the underlying TIE object. Similarly,
04343c6d
GS
1413you may also call C<mg_set()> on the return value after possibly assigning
1414a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
9edb2b46 1415method on the TIE object. [/MAYCHANGE]
04343c6d 1416
9edb2b46 1417[MAYCHANGE]
04343c6d
GS
1418In other words, the array or hash fetch/store functions don't really
1419fetch and store actual values in the case of tied arrays and hashes. They
1420merely call C<mg_copy> to attach magic to the values that were meant to be
1421"stored" or "fetched". Later calls to C<mg_get> and C<mg_set> actually
1422do the job of invoking the TIE methods on the underlying objects. Thus
9edb2b46 1423the magic mechanism currently implements a kind of lazy access to arrays
04343c6d
GS
1424and hashes.
1425
1426Currently (as of perl version 5.004), use of the hash and array access
1427functions requires the user to be aware of whether they are operating on
9edb2b46
GS
1428"normal" hashes and arrays, or on their tied variants. The API may be
1429changed to provide more transparent access to both tied and normal data
1430types in future versions.
1431[/MAYCHANGE]
04343c6d
GS
1432
1433You would do well to understand that the TIEARRAY and TIEHASH interfaces
1434are mere sugar to invoke some perl method calls while using the uniform hash
1435and array syntax. The use of this sugar imposes some overhead (typically
1436about two to four extra opcodes per FETCH/STORE operation, in addition to
1437the creation of all the mortal variables required to invoke the methods).
1438This overhead will be comparatively small if the TIE methods are themselves
1439substantial, but if they are only a few statements long, the overhead
1440will not be insignificant.
1441
d1c897a1
IZ
1442=head2 Localizing changes
1443
1444Perl has a very handy construction
1445
1446 {
1447 local $var = 2;
1448 ...
1449 }
1450
1451This construction is I<approximately> equivalent to
1452
1453 {
1454 my $oldvar = $var;
1455 $var = 2;
1456 ...
1457 $var = $oldvar;
1458 }
1459
1460The biggest difference is that the first construction would
1461reinstate the initial value of $var, irrespective of how control exits
10e2eb10 1462the block: C<goto>, C<return>, C<die>/C<eval>, etc. It is a little bit
d1c897a1
IZ
1463more efficient as well.
1464
1465There is a way to achieve a similar task from C via Perl API: create a
1466I<pseudo-block>, and arrange for some changes to be automatically
1467undone at the end of it, either explicit, or via a non-local exit (via
10e2eb10 1468die()). A I<block>-like construct is created by a pair of
b687b08b
TC
1469C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1470Such a construct may be created specially for some important localized
1471task, or an existing one (like boundaries of enclosing Perl
1472subroutine/block, or an existing pair for freeing TMPs) may be
10e2eb10
FC
1473used. (In the second case the overhead of additional localization must
1474be almost negligible.) Note that any XSUB is automatically enclosed in
b687b08b 1475an C<ENTER>/C<LEAVE> pair.
d1c897a1
IZ
1476
1477Inside such a I<pseudo-block> the following service is available:
1478
13a2d996 1479=over 4
d1c897a1
IZ
1480
1481=item C<SAVEINT(int i)>
1482
1483=item C<SAVEIV(IV i)>
1484
1485=item C<SAVEI32(I32 i)>
1486
1487=item C<SAVELONG(long i)>
1488
1489These macros arrange things to restore the value of integer variable
1490C<i> at the end of enclosing I<pseudo-block>.
1491
1492=item C<SAVESPTR(s)>
1493
1494=item C<SAVEPPTR(p)>
1495
1496These macros arrange things to restore the value of pointers C<s> and
10e2eb10 1497C<p>. C<s> must be a pointer of a type which survives conversion to
d1c897a1
IZ
1498C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1499and back.
1500
1501=item C<SAVEFREESV(SV *sv)>
1502
06f1e0b6 1503The refcount of C<sv> will be decremented at the end of
26d9b02f
JH
1504I<pseudo-block>. This is similar to C<sv_2mortal> in that it is also a
1505mechanism for doing a delayed C<SvREFCNT_dec>. However, while C<sv_2mortal>
1506extends the lifetime of C<sv> until the beginning of the next statement,
1507C<SAVEFREESV> extends it until the end of the enclosing scope. These
1508lifetimes can be wildly different.
1509
1510Also compare C<SAVEMORTALIZESV>.
1511
1512=item C<SAVEMORTALIZESV(SV *sv)>
1513
1514Just like C<SAVEFREESV>, but mortalizes C<sv> at the end of the current
1515scope instead of decrementing its reference count. This usually has the
1516effect of keeping C<sv> alive until the statement that called the currently
1517live scope has finished executing.
d1c897a1
IZ
1518
1519=item C<SAVEFREEOP(OP *op)>
1520
1521The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1522
1523=item C<SAVEFREEPV(p)>
1524
1525The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1526end of I<pseudo-block>.
1527
1528=item C<SAVECLEARSV(SV *sv)>
1529
1530Clears a slot in the current scratchpad which corresponds to C<sv> at
1531the end of I<pseudo-block>.
1532
1533=item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1534
10e2eb10 1535The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
d1c897a1
IZ
1536string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1537short-lived storage, the corresponding string may be reallocated like
1538this:
1539
9cde0e7f 1540 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
d1c897a1 1541
c76ac1ee 1542=item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
d1c897a1
IZ
1543
1544At the end of I<pseudo-block> the function C<f> is called with the
c76ac1ee
GS
1545only argument C<p>.
1546
1547=item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1548
1549At the end of I<pseudo-block> the function C<f> is called with the
1550implicit context argument (if any), and C<p>.
d1c897a1
IZ
1551
1552=item C<SAVESTACK_POS()>
1553
1554The current offset on the Perl internal stack (cf. C<SP>) is restored
1555at the end of I<pseudo-block>.
1556
1557=back
1558
1559The following API list contains functions, thus one needs to
1560provide pointers to the modifiable data explicitly (either C pointers,
00aadd71 1561or Perlish C<GV *>s). Where the above macros take C<int>, a similar
d1c897a1
IZ
1562function takes C<int *>.
1563
13a2d996 1564=over 4
d1c897a1
IZ
1565
1566=item C<SV* save_scalar(GV *gv)>
1567
1568Equivalent to Perl code C<local $gv>.
1569
1570=item C<AV* save_ary(GV *gv)>
1571
1572=item C<HV* save_hash(GV *gv)>
1573
1574Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1575
1576=item C<void save_item(SV *item)>
1577
1578Duplicates the current value of C<SV>, on the exit from the current
1579C<ENTER>/C<LEAVE> I<pseudo-block> will restore the value of C<SV>
10e2eb10 1580using the stored value. It doesn't handle magic. Use C<save_scalar> if
038fcae3 1581magic is affected.
d1c897a1
IZ
1582
1583=item C<void save_list(SV **sarg, I32 maxsarg)>
1584
1585A variant of C<save_item> which takes multiple arguments via an array
1586C<sarg> of C<SV*> of length C<maxsarg>.
1587
1588=item C<SV* save_svref(SV **sptr)>
1589
d1be9408 1590Similar to C<save_scalar>, but will reinstate an C<SV *>.
d1c897a1
IZ
1591
1592=item C<void save_aptr(AV **aptr)>
1593
1594=item C<void save_hptr(HV **hptr)>
1595
1596Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1597
1598=back
1599
1600The C<Alias> module implements localization of the basic types within the
1601I<caller's scope>. People who are interested in how to localize things in
1602the containing scope should take a look there too.
1603
0a753a76 1604=head1 Subroutines
a0d0e21e 1605
68dc0745 1606=head2 XSUBs and the Argument Stack
5f05dabc
PP
1607
1608The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1609An XSUB routine will have a stack that contains the arguments from the Perl
1610program, and a way to map from the Perl data structures to a C equivalent.
1611
1612The stack arguments are accessible through the C<ST(n)> macro, which returns
1613the C<n>'th stack argument. Argument 0 is the first argument passed in the
1614Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1615an C<SV*> is used.
1616
1617Most of the time, output from the C routine can be handled through use of
1618the RETVAL and OUTPUT directives. However, there are some cases where the
1619argument stack is not already long enough to handle all the return values.
1620An example is the POSIX tzname() call, which takes no arguments, but returns
1621two, the local time zone's standard and summer time abbreviations.
1622
1623To handle this situation, the PPCODE directive is used and the stack is
1624extended using the macro:
1625
924508f0 1626 EXTEND(SP, num);
5f05dabc 1627
924508f0
GS
1628where C<SP> is the macro that represents the local copy of the stack pointer,
1629and C<num> is the number of elements the stack should be extended by.
5f05dabc 1630
00aadd71 1631Now that there is room on the stack, values can be pushed on it using C<PUSHs>
10e2eb10 1632macro. The pushed values will often need to be "mortal" (See
d82b684c 1633L</Reference Counts and Mortality>):
5f05dabc 1634
00aadd71 1635 PUSHs(sv_2mortal(newSViv(an_integer)))
d82b684c
SH
1636 PUSHs(sv_2mortal(newSVuv(an_unsigned_integer)))
1637 PUSHs(sv_2mortal(newSVnv(a_double)))
00aadd71 1638 PUSHs(sv_2mortal(newSVpv("Some String",0)))
a9b0660e
KW
1639 /* Although the last example is better written as the more
1640 * efficient: */
a3179684 1641 PUSHs(newSVpvs_flags("Some String", SVs_TEMP))
5f05dabc
PP
1642
1643And now the Perl program calling C<tzname>, the two values will be assigned
1644as in:
1645
1646 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1647
1648An alternate (and possibly simpler) method to pushing values on the stack is
00aadd71 1649to use the macro:
5f05dabc 1650
5f05dabc
PP
1651 XPUSHs(SV*)
1652
da8c5729 1653This macro automatically adjusts the stack for you, if needed. Thus, you
5f05dabc 1654do not need to call C<EXTEND> to extend the stack.
00aadd71
NIS
1655
1656Despite their suggestions in earlier versions of this document the macros
d82b684c
SH
1657C<(X)PUSH[iunp]> are I<not> suited to XSUBs which return multiple results.
1658For that, either stick to the C<(X)PUSHs> macros shown above, or use the new
1659C<m(X)PUSH[iunp]> macros instead; see L</Putting a C value on Perl stack>.
5f05dabc
PP
1660
1661For more information, consult L<perlxs> and L<perlxstut>.
1662
5b36e945
FC
1663=head2 Autoloading with XSUBs
1664
1665If an AUTOLOAD routine is an XSUB, as with Perl subroutines, Perl puts the
1666fully-qualified name of the autoloaded subroutine in the $AUTOLOAD variable
1667of the XSUB's package.
1668
1669But it also puts the same information in certain fields of the XSUB itself:
1670
1671 HV *stash = CvSTASH(cv);
1672 const char *subname = SvPVX(cv);
1673 STRLEN name_length = SvCUR(cv); /* in bytes */
1674 U32 is_utf8 = SvUTF8(cv);
f703fc96 1675
5b36e945 1676C<SvPVX(cv)> contains just the sub name itself, not including the package.
d8893903
FC
1677For an AUTOLOAD routine in UNIVERSAL or one of its superclasses,
1678C<CvSTASH(cv)> returns NULL during a method call on a nonexistent package.
5b36e945
FC
1679
1680B<Note>: Setting $AUTOLOAD stopped working in 5.6.1, which did not support
1681XS AUTOLOAD subs at all. Perl 5.8.0 introduced the use of fields in the
1682XSUB itself. Perl 5.16.0 restored the setting of $AUTOLOAD. If you need
1683to support 5.8-5.14, use the XSUB's fields.
1684
5f05dabc 1685=head2 Calling Perl Routines from within C Programs
a0d0e21e
LW
1686
1687There are four routines that can be used to call a Perl subroutine from
1688within a C program. These four are:
1689
954c1994
GS
1690 I32 call_sv(SV*, I32);
1691 I32 call_pv(const char*, I32);
1692 I32 call_method(const char*, I32);
5aaab254 1693 I32 call_argv(const char*, I32, char**);
a0d0e21e 1694
954c1994 1695The routine most often used is C<call_sv>. The C<SV*> argument
d1b91892
AD
1696contains either the name of the Perl subroutine to be called, or a
1697reference to the subroutine. The second argument consists of flags
1698that control the context in which the subroutine is called, whether
1699or not the subroutine is being passed arguments, how errors should be
1700trapped, and how to treat return values.
a0d0e21e
LW
1701
1702All four routines return the number of arguments that the subroutine returned
1703on the Perl stack.
1704
9a68f1db 1705These routines used to be called C<perl_call_sv>, etc., before Perl v5.6.0,
954c1994
GS
1706but those names are now deprecated; macros of the same name are provided for
1707compatibility.
1708
1709When using any of these routines (except C<call_argv>), the programmer
d1b91892
AD
1710must manipulate the Perl stack. These include the following macros and
1711functions:
a0d0e21e
LW
1712
1713 dSP
924508f0 1714 SP
a0d0e21e
LW
1715 PUSHMARK()
1716 PUTBACK
1717 SPAGAIN
1718 ENTER
1719 SAVETMPS
1720 FREETMPS
1721 LEAVE
1722 XPUSH*()
cb1a09d0 1723 POP*()
a0d0e21e 1724
5f05dabc
PP
1725For a detailed description of calling conventions from C to Perl,
1726consult L<perlcall>.
a0d0e21e 1727
8ebc5c01 1728=head2 Putting a C value on Perl stack
ce3d39e2
IZ
1729
1730A lot of opcodes (this is an elementary operation in the internal perl
10e2eb10
FC
1731stack machine) put an SV* on the stack. However, as an optimization
1732the corresponding SV is (usually) not recreated each time. The opcodes
ce3d39e2
IZ
1733reuse specially assigned SVs (I<target>s) which are (as a corollary)
1734not constantly freed/created.
1735
0a753a76 1736Each of the targets is created only once (but see
5a0de581 1737L</Scratchpads and recursion> below), and when an opcode needs to put
ce3d39e2
IZ
1738an integer, a double, or a string on stack, it just sets the
1739corresponding parts of its I<target> and puts the I<target> on stack.
1740
1741The macro to put this target on stack is C<PUSHTARG>, and it is
1742directly used in some opcodes, as well as indirectly in zillions of
d82b684c 1743others, which use it via C<(X)PUSH[iunp]>.
ce3d39e2 1744
1bd1c0d5 1745Because the target is reused, you must be careful when pushing multiple
10e2eb10 1746values on the stack. The following code will not do what you think:
1bd1c0d5
SC
1747
1748 XPUSHi(10);
1749 XPUSHi(20);
1750
1751This translates as "set C<TARG> to 10, push a pointer to C<TARG> onto
1752the stack; set C<TARG> to 20, push a pointer to C<TARG> onto the stack".
1753At the end of the operation, the stack does not contain the values 10
1754and 20, but actually contains two pointers to C<TARG>, which we have set
d82b684c 1755to 20.
1bd1c0d5 1756
d82b684c
SH
1757If you need to push multiple different values then you should either use
1758the C<(X)PUSHs> macros, or else use the new C<m(X)PUSH[iunp]> macros,
1759none of which make use of C<TARG>. The C<(X)PUSHs> macros simply push an
1760SV* on the stack, which, as noted under L</XSUBs and the Argument Stack>,
1761will often need to be "mortal". The new C<m(X)PUSH[iunp]> macros make
1762this a little easier to achieve by creating a new mortal for you (via
1763C<(X)PUSHmortal>), pushing that onto the stack (extending it if necessary
1764in the case of the C<mXPUSH[iunp]> macros), and then setting its value.
1765Thus, instead of writing this to "fix" the example above:
1766
1767 XPUSHs(sv_2mortal(newSViv(10)))
1768 XPUSHs(sv_2mortal(newSViv(20)))
1769
1770you can simply write:
1771
1772 mXPUSHi(10)
1773 mXPUSHi(20)
1774
1775On a related note, if you do use C<(X)PUSH[iunp]>, then you're going to
1bd1c0d5 1776need a C<dTARG> in your variable declarations so that the C<*PUSH*>
d82b684c
SH
1777macros can make use of the local variable C<TARG>. See also C<dTARGET>
1778and C<dXSTARG>.
1bd1c0d5 1779
8ebc5c01 1780=head2 Scratchpads
ce3d39e2 1781
54310121 1782The question remains on when the SVs which are I<target>s for opcodes
10e2eb10 1783are created. The answer is that they are created when the current
ac036724 1784unit--a subroutine or a file (for opcodes for statements outside of
10e2eb10 1785subroutines)--is compiled. During this time a special anonymous Perl
ac036724 1786array is created, which is called a scratchpad for the current unit.
ce3d39e2 1787
54310121 1788A scratchpad keeps SVs which are lexicals for the current unit and are
d777b41a
FC
1789targets for opcodes. A previous version of this document
1790stated that one can deduce that an SV lives on a scratchpad
ce3d39e2 1791by looking on its flags: lexicals have C<SVs_PADMY> set, and
eee3e302 1792I<target>s have C<SVs_PADTMP> set. But this has never been fully true.
d777b41a
FC
1793C<SVs_PADMY> could be set on a variable that no longer resides in any pad.
1794While I<target>s do have C<SVs_PADTMP> set, it can also be set on variables
eed77337
FC
1795that have never resided in a pad, but nonetheless act like I<target>s. As
1796of perl 5.21.5, the C<SVs_PADMY> flag is no longer used and is defined as
17970. C<SvPADMY()> now returns true for anything without C<SVs_PADTMP>.
ce3d39e2 1798
10e2eb10 1799The correspondence between OPs and I<target>s is not 1-to-1. Different
54310121 1800OPs in the compile tree of the unit can use the same target, if this
ce3d39e2
IZ
1801would not conflict with the expected life of the temporary.
1802
2ae324a7 1803=head2 Scratchpads and recursion
ce3d39e2
IZ
1804
1805In fact it is not 100% true that a compiled unit contains a pointer to
10e2eb10
FC
1806the scratchpad AV. In fact it contains a pointer to an AV of
1807(initially) one element, and this element is the scratchpad AV. Why do
ce3d39e2
IZ
1808we need an extra level of indirection?
1809
10e2eb10 1810The answer is B<recursion>, and maybe B<threads>. Both
ce3d39e2 1811these can create several execution pointers going into the same
10e2eb10 1812subroutine. For the subroutine-child not write over the temporaries
ce3d39e2
IZ
1813for the subroutine-parent (lifespan of which covers the call to the
1814child), the parent and the child should have different
10e2eb10 1815scratchpads. (I<And> the lexicals should be separate anyway!)
ce3d39e2 1816
5f05dabc
PP
1817So each subroutine is born with an array of scratchpads (of length 1).
1818On each entry to the subroutine it is checked that the current
ce3d39e2
IZ
1819depth of the recursion is not more than the length of this array, and
1820if it is, new scratchpad is created and pushed into the array.
1821
1822The I<target>s on this scratchpad are C<undef>s, but they are already
1823marked with correct flags.
1824
22d36020
FC
1825=head1 Memory Allocation
1826
1827=head2 Allocation
1828
1829All memory meant to be used with the Perl API functions should be manipulated
1830using the macros described in this section. The macros provide the necessary
1831transparency between differences in the actual malloc implementation that is
1832used within perl.
1833
1834It is suggested that you enable the version of malloc that is distributed
1835with Perl. It keeps pools of various sizes of unallocated memory in
1836order to satisfy allocation requests more quickly. However, on some
1837platforms, it may cause spurious malloc or free errors.
1838
1839The following three macros are used to initially allocate memory :
1840
1841 Newx(pointer, number, type);
1842 Newxc(pointer, number, type, cast);
1843 Newxz(pointer, number, type);
1844
1845The first argument C<pointer> should be the name of a variable that will
1846point to the newly allocated memory.
1847
1848The second and third arguments C<number> and C<type> specify how many of
1849the specified type of data structure should be allocated. The argument
1850C<type> is passed to C<sizeof>. The final argument to C<Newxc>, C<cast>,
1851should be used if the C<pointer> argument is different from the C<type>
1852argument.
1853
1854Unlike the C<Newx> and C<Newxc> macros, the C<Newxz> macro calls C<memzero>
1855to zero out all the newly allocated memory.
1856
1857=head2 Reallocation
1858
1859 Renew(pointer, number, type);
1860 Renewc(pointer, number, type, cast);
1861 Safefree(pointer)
1862
1863These three macros are used to change a memory buffer size or to free a
1864piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1865match those of C<New> and C<Newc> with the exception of not needing the
1866"magic cookie" argument.
1867
1868=head2 Moving
1869
1870 Move(source, dest, number, type);
1871 Copy(source, dest, number, type);
1872 Zero(dest, number, type);
1873
1874These three macros are used to move, copy, or zero out previously allocated
1875memory. The C<source> and C<dest> arguments point to the source and
1876destination starting points. Perl will move, copy, or zero out C<number>
1877instances of the size of the C<type> data structure (using the C<sizeof>
1878function).
1879
1880=head1 PerlIO
1881
1882The most recent development releases of Perl have been experimenting with
1883removing Perl's dependency on the "normal" standard I/O suite and allowing
1884other stdio implementations to be used. This involves creating a new
1885abstraction layer that then calls whichever implementation of stdio Perl
1886was compiled with. All XSUBs should now use the functions in the PerlIO
1887abstraction layer and not make any assumptions about what kind of stdio
1888is being used.
1889
1890For a complete description of the PerlIO abstraction, consult L<perlapio>.
1891
0a753a76
PP
1892=head1 Compiled code
1893
1894=head2 Code tree
1895
1896Here we describe the internal form your code is converted to by
10e2eb10 1897Perl. Start with a simple example:
0a753a76
PP
1898
1899 $a = $b + $c;
1900
1901This is converted to a tree similar to this one:
1902
1903 assign-to
1904 / \
1905 + $a
1906 / \
1907 $b $c
1908
7b8d334a 1909(but slightly more complicated). This tree reflects the way Perl
0a753a76
PP
1910parsed your code, but has nothing to do with the execution order.
1911There is an additional "thread" going through the nodes of the tree
1912which shows the order of execution of the nodes. In our simplified
1913example above it looks like:
1914
1915 $b ---> $c ---> + ---> $a ---> assign-to
1916
1917But with the actual compile tree for C<$a = $b + $c> it is different:
1918some nodes I<optimized away>. As a corollary, though the actual tree
1919contains more nodes than our simplified example, the execution order
1920is the same as in our example.
1921
1922=head2 Examining the tree
1923
06f6df17
RGS
1924If you have your perl compiled for debugging (usually done with
1925C<-DDEBUGGING> on the C<Configure> command line), you may examine the
0a753a76
PP
1926compiled tree by specifying C<-Dx> on the Perl command line. The
1927output takes several lines per node, and for C<$b+$c> it looks like
1928this:
1929
1930 5 TYPE = add ===> 6
1931 TARG = 1
1932 FLAGS = (SCALAR,KIDS)
1933 {
1934 TYPE = null ===> (4)
1935 (was rv2sv)
1936 FLAGS = (SCALAR,KIDS)
1937 {
1938 3 TYPE = gvsv ===> 4
1939 FLAGS = (SCALAR)
1940 GV = main::b
1941 }
1942 }
1943 {
1944 TYPE = null ===> (5)
1945 (was rv2sv)
1946 FLAGS = (SCALAR,KIDS)
1947 {
1948 4 TYPE = gvsv ===> 5
1949 FLAGS = (SCALAR)
1950 GV = main::c
1951 }
1952 }
1953
1954This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
1955not optimized away (one per number in the left column). The immediate
1956children of the given node correspond to C<{}> pairs on the same level
1957of indentation, thus this listing corresponds to the tree:
1958
1959 add
1960 / \
1961 null null
1962 | |
1963 gvsv gvsv
1964
1965The execution order is indicated by C<===E<gt>> marks, thus it is C<3
19664 5 6> (node C<6> is not included into above listing), i.e.,
1967C<gvsv gvsv add whatever>.
1968
9afa14e3 1969Each of these nodes represents an op, a fundamental operation inside the
10e2eb10 1970Perl core. The code which implements each operation can be found in the
9afa14e3 1971F<pp*.c> files; the function which implements the op with type C<gvsv>
10e2eb10 1972is C<pp_gvsv>, and so on. As the tree above shows, different ops have
9afa14e3 1973different numbers of children: C<add> is a binary operator, as one would
10e2eb10 1974expect, and so has two children. To accommodate the various different
9afa14e3
SC
1975numbers of children, there are various types of op data structure, and
1976they link together in different ways.
1977
10e2eb10 1978The simplest type of op structure is C<OP>: this has no children. Unary
9afa14e3 1979operators, C<UNOP>s, have one child, and this is pointed to by the
10e2eb10
FC
1980C<op_first> field. Binary operators (C<BINOP>s) have not only an
1981C<op_first> field but also an C<op_last> field. The most complex type of
1982op is a C<LISTOP>, which has any number of children. In this case, the
9afa14e3 1983first child is pointed to by C<op_first> and the last child by
10e2eb10 1984C<op_last>. The children in between can be found by iteratively
86cd3a13 1985following the C<OpSIBLING> pointer from the first child to the last (but
29e61fd9 1986see below).
9afa14e3 1987
29e61fd9 1988There are also some other op types: a C<PMOP> holds a regular expression,
10e2eb10
FC
1989and has no children, and a C<LOOP> may or may not have children. If the
1990C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
9afa14e3
SC
1991complicate matters, if a C<UNOP> is actually a C<null> op after
1992optimization (see L</Compile pass 2: context propagation>) it will still
1993have children in accordance with its former type.
1994
29e61fd9
DM
1995Finally, there is a C<LOGOP>, or logic op. Like a C<LISTOP>, this has one
1996or more children, but it doesn't have an C<op_last> field: so you have to
86cd3a13 1997follow C<op_first> and then the C<OpSIBLING> chain itself to find the
29e61fd9
DM
1998last child. Instead it has an C<op_other> field, which is comparable to
1999the C<op_next> field described below, and represents an alternate
2000execution path. Operators like C<and>, C<or> and C<?> are C<LOGOP>s. Note
2001that in general, C<op_other> may not point to any of the direct children
2002of the C<LOGOP>.
2003
2004Starting in version 5.21.2, perls built with the experimental
2005define C<-DPERL_OP_PARENT> add an extra boolean flag for each op,
87b5a8b9 2006C<op_moresib>. When not set, this indicates that this is the last op in an
86cd3a13
DM
2007C<OpSIBLING> chain. This frees up the C<op_sibling> field on the last
2008sibling to point back to the parent op. Under this build, that field is
2009also renamed C<op_sibparent> to reflect its joint role. The macro
2010C<OpSIBLING(o)> wraps this special behaviour, and always returns NULL on
2011the last sibling. With this build the C<op_parent(o)> function can be
2012used to find the parent of any op. Thus for forward compatibility, you
2013should always use the C<OpSIBLING(o)> macro rather than accessing
2014C<op_sibling> directly.
29e61fd9 2015
06f6df17
RGS
2016Another way to examine the tree is to use a compiler back-end module, such
2017as L<B::Concise>.
2018
0a753a76
PP
2019=head2 Compile pass 1: check routines
2020
8870b5c7 2021The tree is created by the compiler while I<yacc> code feeds it
10e2eb10 2022the constructions it recognizes. Since I<yacc> works bottom-up, so does
0a753a76
PP
2023the first pass of perl compilation.
2024
2025What makes this pass interesting for perl developers is that some
2026optimization may be performed on this pass. This is optimization by
8870b5c7 2027so-called "check routines". The correspondence between node names
0a753a76
PP
2028and corresponding check routines is described in F<opcode.pl> (do not
2029forget to run C<make regen_headers> if you modify this file).
2030
2031A check routine is called when the node is fully constructed except
7b8d334a 2032for the execution-order thread. Since at this time there are no
0a753a76
PP
2033back-links to the currently constructed node, one can do most any
2034operation to the top-level node, including freeing it and/or creating
2035new nodes above/below it.
2036
2037The check routine returns the node which should be inserted into the
2038tree (if the top-level node was not modified, check routine returns
2039its argument).
2040
10e2eb10 2041By convention, check routines have names C<ck_*>. They are usually
0a753a76
PP
2042called from C<new*OP> subroutines (or C<convert>) (which in turn are
2043called from F<perly.y>).
2044
2045=head2 Compile pass 1a: constant folding
2046
2047Immediately after the check routine is called the returned node is
2048checked for being compile-time executable. If it is (the value is
2049judged to be constant) it is immediately executed, and a I<constant>
2050node with the "return value" of the corresponding subtree is
2051substituted instead. The subtree is deleted.
2052
2053If constant folding was not performed, the execution-order thread is
2054created.
2055
2056=head2 Compile pass 2: context propagation
2057
2058When a context for a part of compile tree is known, it is propagated
a3cb178b 2059down through the tree. At this time the context can have 5 values
0a753a76
PP
2060(instead of 2 for runtime context): void, boolean, scalar, list, and
2061lvalue. In contrast with the pass 1 this pass is processed from top
2062to bottom: a node's context determines the context for its children.
2063
2064Additional context-dependent optimizations are performed at this time.
2065Since at this moment the compile tree contains back-references (via
2066"thread" pointers), nodes cannot be free()d now. To allow
2067optimized-away nodes at this stage, such nodes are null()ified instead
2068of free()ing (i.e. their type is changed to OP_NULL).
2069
2070=head2 Compile pass 3: peephole optimization
2071
2072After the compile tree for a subroutine (or for an C<eval> or a file)
10e2eb10 2073is created, an additional pass over the code is performed. This pass
0a753a76 2074is neither top-down or bottom-up, but in the execution order (with
9ea12537
Z
2075additional complications for conditionals). Optimizations performed
2076at this stage are subject to the same restrictions as in the pass 2.
2077
2078Peephole optimizations are done by calling the function pointed to
2079by the global variable C<PL_peepp>. By default, C<PL_peepp> just
2080calls the function pointed to by the global variable C<PL_rpeepp>.
2081By default, that performs some basic op fixups and optimisations along
2082the execution-order op chain, and recursively calls C<PL_rpeepp> for
2083each side chain of ops (resulting from conditionals). Extensions may
2084provide additional optimisations or fixups, hooking into either the
2085per-subroutine or recursive stage, like this:
2086
2087 static peep_t prev_peepp;
2088 static void my_peep(pTHX_ OP *o)
2089 {
2090 /* custom per-subroutine optimisation goes here */
f0358462 2091 prev_peepp(aTHX_ o);
9ea12537
Z
2092 /* custom per-subroutine optimisation may also go here */
2093 }
2094 BOOT:
2095 prev_peepp = PL_peepp;
2096 PL_peepp = my_peep;
2097
2098 static peep_t prev_rpeepp;
2099 static void my_rpeep(pTHX_ OP *o)
2100 {
2101 OP *orig_o = o;
2102 for(; o; o = o->op_next) {
2103 /* custom per-op optimisation goes here */
2104 }
f0358462 2105 prev_rpeepp(aTHX_ orig_o);
9ea12537
Z
2106 }
2107 BOOT:
2108 prev_rpeepp = PL_rpeepp;
2109 PL_rpeepp = my_rpeep;
0a753a76 2110
1ba7f851
PJ
2111=head2 Pluggable runops
2112
2113The compile tree is executed in a runops function. There are two runops
1388f78e
RGS
2114functions, in F<run.c> and in F<dump.c>. C<Perl_runops_debug> is used
2115with DEBUGGING and C<Perl_runops_standard> is used otherwise. For fine
2116control over the execution of the compile tree it is possible to provide
2117your own runops function.
1ba7f851
PJ
2118
2119It's probably best to copy one of the existing runops functions and
2120change it to suit your needs. Then, in the BOOT section of your XS
2121file, add the line:
2122
2123 PL_runops = my_runops;
2124
2125This function should be as efficient as possible to keep your programs
2126running as fast as possible.
2127
fd85fad2
BM
2128=head2 Compile-time scope hooks
2129
2130As of perl 5.14 it is possible to hook into the compile-time lexical
10e2eb10 2131scope mechanism using C<Perl_blockhook_register>. This is used like
fd85fad2
BM
2132this:
2133
2134 STATIC void my_start_hook(pTHX_ int full);
2135 STATIC BHK my_hooks;
2136
2137 BOOT:
a88d97bf 2138 BhkENTRY_set(&my_hooks, bhk_start, my_start_hook);
fd85fad2
BM
2139 Perl_blockhook_register(aTHX_ &my_hooks);
2140
2141This will arrange to have C<my_start_hook> called at the start of
10e2eb10 2142compiling every lexical scope. The available hooks are:
fd85fad2
BM
2143
2144=over 4
2145
a88d97bf 2146=item C<void bhk_start(pTHX_ int full)>
fd85fad2 2147
10e2eb10 2148This is called just after starting a new lexical scope. Note that Perl
fd85fad2
BM
2149code like
2150
2151 if ($x) { ... }
2152
2153creates two scopes: the first starts at the C<(> and has C<full == 1>,
10e2eb10 2154the second starts at the C<{> and has C<full == 0>. Both end at the
f185f654 2155C<}>, so calls to C<start> and C<pre>/C<post_end> will match. Anything
fd85fad2
BM
2156pushed onto the save stack by this hook will be popped just before the
2157scope ends (between the C<pre_> and C<post_end> hooks, in fact).
2158
a88d97bf 2159=item C<void bhk_pre_end(pTHX_ OP **o)>
fd85fad2
BM
2160
2161This is called at the end of a lexical scope, just before unwinding the
10e2eb10 2162stack. I<o> is the root of the optree representing the scope; it is a
fd85fad2
BM
2163double pointer so you can replace the OP if you need to.
2164
a88d97bf 2165=item C<void bhk_post_end(pTHX_ OP **o)>
fd85fad2
BM
2166
2167This is called at the end of a lexical scope, just after unwinding the
10e2eb10 2168stack. I<o> is as above. Note that it is possible for calls to C<pre_>
fd85fad2
BM
2169and C<post_end> to nest, if there is something on the save stack that
2170calls string eval.
2171
a88d97bf 2172=item C<void bhk_eval(pTHX_ OP *const o)>
fd85fad2
BM
2173
2174This is called just before starting to compile an C<eval STRING>, C<do
10e2eb10 2175FILE>, C<require> or C<use>, after the eval has been set up. I<o> is the
fd85fad2
BM
2176OP that requested the eval, and will normally be an C<OP_ENTEREVAL>,
2177C<OP_DOFILE> or C<OP_REQUIRE>.
2178
2179=back
2180
2181Once you have your hook functions, you need a C<BHK> structure to put
10e2eb10
FC
2182them in. It's best to allocate it statically, since there is no way to
2183free it once it's registered. The function pointers should be inserted
fd85fad2 2184into this structure using the C<BhkENTRY_set> macro, which will also set
10e2eb10 2185flags indicating which entries are valid. If you do need to allocate
fd85fad2
BM
2186your C<BHK> dynamically for some reason, be sure to zero it before you
2187start.
2188
2189Once registered, there is no mechanism to switch these hooks off, so if
10e2eb10 2190that is necessary you will need to do this yourself. An entry in C<%^H>
a3e07c87
BM
2191is probably the best way, so the effect is lexically scoped; however it
2192is also possible to use the C<BhkDISABLE> and C<BhkENABLE> macros to
10e2eb10 2193temporarily switch entries on and off. You should also be aware that
a3e07c87 2194generally speaking at least one scope will have opened before your
f185f654 2195extension is loaded, so you will see some C<pre>/C<post_end> pairs that
a3e07c87 2196didn't have a matching C<start>.
fd85fad2 2197
9afa14e3
SC
2198=head1 Examining internal data structures with the C<dump> functions
2199
2200To aid debugging, the source file F<dump.c> contains a number of
2201functions which produce formatted output of internal data structures.
2202
2203The most commonly used of these functions is C<Perl_sv_dump>; it's used
10e2eb10 2204for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
9afa14e3 2205C<sv_dump> to produce debugging output from Perl-space, so users of that
00aadd71 2206module should already be familiar with its format.
9afa14e3
SC
2207
2208C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
210b36aa 2209derivatives, and produces output similar to C<perl -Dx>; in fact,
9afa14e3
SC
2210C<Perl_dump_eval> will dump the main root of the code being evaluated,
2211exactly like C<-Dx>.
2212
2213Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
2214op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
2215subroutines in a package like so: (Thankfully, these are all xsubs, so
2216there is no op tree)
2217
2218 (gdb) print Perl_dump_packsubs(PL_defstash)
2219
2220 SUB attributes::bootstrap = (xsub 0x811fedc 0)
2221
2222 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
2223
2224 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
2225
2226 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
2227
2228 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
2229
2230and C<Perl_dump_all>, which dumps all the subroutines in the stash and
2231the op tree of the main root.
2232
954c1994 2233=head1 How multiple interpreters and concurrency are supported
ee072b34 2234
ee072b34
GS
2235=head2 Background and PERL_IMPLICIT_CONTEXT
2236
2237The Perl interpreter can be regarded as a closed box: it has an API
2238for feeding it code or otherwise making it do things, but it also has
2239functions for its own use. This smells a lot like an object, and
2240there are ways for you to build Perl so that you can have multiple
acfe0abc
GS
2241interpreters, with one interpreter represented either as a C structure,
2242or inside a thread-specific structure. These structures contain all
2243the context, the state of that interpreter.
2244
10e2eb10 2245One macro controls the major Perl build flavor: MULTIPLICITY. The
7b52221d 2246MULTIPLICITY build has a C structure that packages all the interpreter
10e2eb10 2247state. With multiplicity-enabled perls, PERL_IMPLICIT_CONTEXT is also
7b52221d 2248normally defined, and enables the support for passing in a "hidden" first
10e2eb10 2249argument that represents all three data structures. MULTIPLICITY makes
1a64a5e6 2250multi-threaded perls possible (with the ithreads threading model, related
7b52221d 2251to the macro USE_ITHREADS.)
54aff467 2252
27da23d5
JH
2253Two other "encapsulation" macros are the PERL_GLOBAL_STRUCT and
2254PERL_GLOBAL_STRUCT_PRIVATE (the latter turns on the former, and the
2255former turns on MULTIPLICITY.) The PERL_GLOBAL_STRUCT causes all the
2256internal variables of Perl to be wrapped inside a single global struct,
2257struct perl_vars, accessible as (globals) &PL_Vars or PL_VarsPtr or
2258the function Perl_GetVars(). The PERL_GLOBAL_STRUCT_PRIVATE goes
2259one step further, there is still a single struct (allocated in main()
2260either from heap or from stack) but there are no global data symbols
3bf17896 2261pointing to it. In either case the global struct should be initialized
27da23d5
JH
2262as the very first thing in main() using Perl_init_global_struct() and
2263correspondingly tear it down after perl_free() using Perl_free_global_struct(),
2264please see F<miniperlmain.c> for usage details. You may also need
2265to use C<dVAR> in your coding to "declare the global variables"
2266when you are using them. dTHX does this for you automatically.
2267
9aa97215
JH
2268To see whether you have non-const data you can use a BSD (or GNU)
2269compatible C<nm>:
bc028b6b
JH
2270
2271 nm libperl.a | grep -v ' [TURtr] '
2272
9aa97215
JH
2273If this displays any C<D> or C<d> symbols (or possibly C<C> or C<c>),
2274you have non-const data. The symbols the C<grep> removed are as follows:
2275C<Tt> are I<text>, or code, the C<Rr> are I<read-only> (const) data,
2276and the C<U> is <undefined>, external symbols referred to.
2277
2278The test F<t/porting/libperl.t> does this kind of symbol sanity
2279checking on C<libperl.a>.
bc028b6b 2280
27da23d5
JH
2281For backward compatibility reasons defining just PERL_GLOBAL_STRUCT
2282doesn't actually hide all symbols inside a big global struct: some
2283PerlIO_xxx vtables are left visible. The PERL_GLOBAL_STRUCT_PRIVATE
2284then hides everything (see how the PERLIO_FUNCS_DECL is used).
2285
54aff467 2286All this obviously requires a way for the Perl internal functions to be
acfe0abc 2287either subroutines taking some kind of structure as the first
ee072b34 2288argument, or subroutines taking nothing as the first argument. To
acfe0abc 2289enable these two very different ways of building the interpreter,
ee072b34
GS
2290the Perl source (as it does in so many other situations) makes heavy
2291use of macros and subroutine naming conventions.
2292
54aff467 2293First problem: deciding which functions will be public API functions and
00aadd71 2294which will be private. All functions whose names begin C<S_> are private
954c1994
GS
2295(think "S" for "secret" or "static"). All other functions begin with
2296"Perl_", but just because a function begins with "Perl_" does not mean it is
10e2eb10
FC
2297part of the API. (See L</Internal
2298Functions>.) The easiest way to be B<sure> a
00aadd71
NIS
2299function is part of the API is to find its entry in L<perlapi>.
2300If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
2301think it should be (i.e., you need it for your extension), send mail via
a422fd2d 2302L<perlbug> explaining why you think it should be.
ee072b34
GS
2303
2304Second problem: there must be a syntax so that the same subroutine
2305declarations and calls can pass a structure as their first argument,
2306or pass nothing. To solve this, the subroutines are named and
2307declared in a particular way. Here's a typical start of a static
2308function used within the Perl guts:
2309
2310 STATIC void
2311 S_incline(pTHX_ char *s)
2312
acfe0abc 2313STATIC becomes "static" in C, and may be #define'd to nothing in some
da8c5729 2314configurations in the future.
ee072b34 2315
651a3225
GS
2316A public function (i.e. part of the internal API, but not necessarily
2317sanctioned for use in extensions) begins like this:
ee072b34
GS
2318
2319 void
2307c6d0 2320 Perl_sv_setiv(pTHX_ SV* dsv, IV num)
ee072b34 2321
0147cd53 2322C<pTHX_> is one of a number of macros (in F<perl.h>) that hide the
ee072b34
GS
2323details of the interpreter's context. THX stands for "thread", "this",
2324or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
2325The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
a7486cbb
JH
2326or 'd' for B<d>eclaration, so we have C<pTHX>, C<aTHX> and C<dTHX>, and
2327their variants.
ee072b34 2328
a7486cbb
JH
2329When Perl is built without options that set PERL_IMPLICIT_CONTEXT, there is no
2330first argument containing the interpreter's context. The trailing underscore
ee072b34
GS
2331in the pTHX_ macro indicates that the macro expansion needs a comma
2332after the context argument because other arguments follow it. If
2333PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
54aff467
GS
2334subroutine is not prototyped to take the extra argument. The form of the
2335macro without the trailing underscore is used when there are no additional
ee072b34
GS
2336explicit arguments.
2337
54aff467 2338When a core function calls another, it must pass the context. This
2307c6d0 2339is normally hidden via macros. Consider C<sv_setiv>. It expands into
ee072b34
GS
2340something like this:
2341
2307c6d0
SB
2342 #ifdef PERL_IMPLICIT_CONTEXT
2343 #define sv_setiv(a,b) Perl_sv_setiv(aTHX_ a, b)
ee072b34 2344 /* can't do this for vararg functions, see below */
2307c6d0
SB
2345 #else
2346 #define sv_setiv Perl_sv_setiv
2347 #endif
ee072b34
GS
2348
2349This works well, and means that XS authors can gleefully write:
2350
2307c6d0 2351 sv_setiv(foo, bar);
ee072b34
GS
2352
2353and still have it work under all the modes Perl could have been
2354compiled with.
2355
ee072b34
GS
2356This doesn't work so cleanly for varargs functions, though, as macros
2357imply that the number of arguments is known in advance. Instead we
2358either need to spell them out fully, passing C<aTHX_> as the first
2359argument (the Perl core tends to do this with functions like
2360Perl_warner), or use a context-free version.
2361
2362The context-free version of Perl_warner is called
2363Perl_warner_nocontext, and does not take the extra argument. Instead
2364it does dTHX; to get the context from thread-local storage. We
2365C<#define warner Perl_warner_nocontext> so that extensions get source
2366compatibility at the expense of performance. (Passing an arg is
2367cheaper than grabbing it from thread-local storage.)
2368
acfe0abc 2369You can ignore [pad]THXx when browsing the Perl headers/sources.
ee072b34
GS
2370Those are strictly for use within the core. Extensions and embedders
2371need only be aware of [pad]THX.
2372
a7486cbb
JH
2373=head2 So what happened to dTHR?
2374
2375C<dTHR> was introduced in perl 5.005 to support the older thread model.
2376The older thread model now uses the C<THX> mechanism to pass context
2377pointers around, so C<dTHR> is not useful any more. Perl 5.6.0 and
2378later still have it for backward source compatibility, but it is defined
2379to be a no-op.
2380
ee072b34
GS
2381=head2 How do I use all this in extensions?
2382
2383When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
2384any functions in the Perl API will need to pass the initial context
2385argument somehow. The kicker is that you will need to write it in
2386such a way that the extension still compiles when Perl hasn't been
2387built with PERL_IMPLICIT_CONTEXT enabled.
2388
2389There are three ways to do this. First, the easy but inefficient way,
2390which is also the default, in order to maintain source compatibility
0147cd53 2391with extensions: whenever F<XSUB.h> is #included, it redefines the aTHX
ee072b34
GS
2392and aTHX_ macros to call a function that will return the context.
2393Thus, something like:
2394
2307c6d0 2395 sv_setiv(sv, num);
ee072b34 2396
4375e838 2397in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
54aff467 2398in effect:
ee072b34 2399
2307c6d0 2400 Perl_sv_setiv(Perl_get_context(), sv, num);
ee072b34 2401
54aff467 2402or to this otherwise:
ee072b34 2403
2307c6d0 2404 Perl_sv_setiv(sv, num);
ee072b34 2405
da8c5729 2406You don't have to do anything new in your extension to get this; since
2fa86c13 2407the Perl library provides Perl_get_context(), it will all just
ee072b34
GS
2408work.
2409
2410The second, more efficient way is to use the following template for
2411your Foo.xs:
2412
c52f9dcd
JH
2413 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2414 #include "EXTERN.h"
2415 #include "perl.h"
2416 #include "XSUB.h"
ee072b34 2417
fd061412 2418 STATIC void my_private_function(int arg1, int arg2);
ee072b34 2419
fd061412 2420 STATIC void
c52f9dcd
JH
2421 my_private_function(int arg1, int arg2)
2422 {
2423 dTHX; /* fetch context */
2424 ... call many Perl API functions ...
2425 }
ee072b34
GS
2426
2427 [... etc ...]
2428
c52f9dcd 2429 MODULE = Foo PACKAGE = Foo
ee072b34 2430
c52f9dcd 2431 /* typical XSUB */
ee072b34 2432
c52f9dcd
JH
2433 void
2434 my_xsub(arg)
2435 int arg
2436 CODE:
2437 my_private_function(arg, 10);
ee072b34
GS
2438
2439Note that the only two changes from the normal way of writing an
2440extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
2441including the Perl headers, followed by a C<dTHX;> declaration at
2442the start of every function that will call the Perl API. (You'll
2443know which functions need this, because the C compiler will complain
2444that there's an undeclared identifier in those functions.) No changes
2445are needed for the XSUBs themselves, because the XS() macro is
2446correctly defined to pass in the implicit context if needed.
2447
2448The third, even more efficient way is to ape how it is done within
2449the Perl guts:
2450
2451
c52f9dcd
JH
2452 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2453 #include "EXTERN.h"
2454 #include "perl.h"
2455 #include "XSUB.h"
ee072b34
GS
2456
2457 /* pTHX_ only needed for functions that call Perl API */
fd061412 2458 STATIC void my_private_function(pTHX_ int arg1, int arg2);
ee072b34 2459
fd061412 2460 STATIC void
c52f9dcd
JH
2461 my_private_function(pTHX_ int arg1, int arg2)
2462 {
2463 /* dTHX; not needed here, because THX is an argument */
2464 ... call Perl API functions ...
2465 }
ee072b34
GS
2466
2467 [... etc ...]
2468
c52f9dcd 2469 MODULE = Foo PACKAGE = Foo
ee072b34 2470
c52f9dcd 2471 /* typical XSUB */
ee072b34 2472
c52f9dcd
JH
2473 void
2474 my_xsub(arg)
2475 int arg
2476 CODE:
2477 my_private_function(aTHX_ arg, 10);
ee072b34
GS
2478
2479This implementation never has to fetch the context using a function
2480call, since it is always passed as an extra argument. Depending on
2481your needs for simplicity or efficiency, you may mix the previous
2482two approaches freely.
2483
651a3225
GS
2484Never add a comma after C<pTHX> yourself--always use the form of the
2485macro with the underscore for functions that take explicit arguments,
2486or the form without the argument for functions with no explicit arguments.
ee072b34 2487
27da23d5
JH
2488If one is compiling Perl with the C<-DPERL_GLOBAL_STRUCT> the C<dVAR>
2489definition is needed if the Perl global variables (see F<perlvars.h>
2490or F<globvar.sym>) are accessed in the function and C<dTHX> is not
2491used (the C<dTHX> includes the C<dVAR> if necessary). One notices
2492the need for C<dVAR> only with the said compile-time define, because
2493otherwise the Perl global variables are visible as-is.
2494
a7486cbb
JH
2495=head2 Should I do anything special if I call perl from multiple threads?
2496
2497If you create interpreters in one thread and then proceed to call them in
2498another, you need to make sure perl's own Thread Local Storage (TLS) slot is
2499initialized correctly in each of those threads.
2500
2501The C<perl_alloc> and C<perl_clone> API functions will automatically set
2502the TLS slot to the interpreter they created, so that there is no need to do
2503anything special if the interpreter is always accessed in the same thread that
2504created it, and that thread did not create or call any other interpreters
2505afterwards. If that is not the case, you have to set the TLS slot of the
2506thread before calling any functions in the Perl API on that particular
2507interpreter. This is done by calling the C<PERL_SET_CONTEXT> macro in that
2508thread as the first thing you do:
2509
2510 /* do this before doing anything else with some_perl */
2511 PERL_SET_CONTEXT(some_perl);
2512
2513 ... other Perl API calls on some_perl go here ...
2514
ee072b34
GS
2515=head2 Future Plans and PERL_IMPLICIT_SYS
2516
2517Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
2518that the interpreter knows about itself and pass it around, so too are
2519there plans to allow the interpreter to bundle up everything it knows
2520about the environment it's running on. This is enabled with the
7b52221d
RGS
2521PERL_IMPLICIT_SYS macro. Currently it only works with USE_ITHREADS on
2522Windows.
ee072b34
GS
2523
2524This allows the ability to provide an extra pointer (called the "host"
2525environment) for all the system calls. This makes it possible for
2526all the system stuff to maintain their own state, broken down into
2527seven C structures. These are thin wrappers around the usual system
0147cd53 2528calls (see F<win32/perllib.c>) for the default perl executable, but for a
ee072b34
GS
2529more ambitious host (like the one that would do fork() emulation) all
2530the extra work needed to pretend that different interpreters are
2531actually different "processes", would be done here.
2532
2533The Perl engine/interpreter and the host are orthogonal entities.
2534There could be one or more interpreters in a process, and one or
2535more "hosts", with free association between them.
2536
a422fd2d
SC
2537=head1 Internal Functions
2538
2539All of Perl's internal functions which will be exposed to the outside
06f6df17 2540world are prefixed by C<Perl_> so that they will not conflict with XS
a422fd2d 2541functions or functions used in a program in which Perl is embedded.
10e2eb10 2542Similarly, all global variables begin with C<PL_>. (By convention,
06f6df17 2543static functions start with C<S_>.)
a422fd2d 2544
0972ecdf
DM
2545Inside the Perl core (C<PERL_CORE> defined), you can get at the functions
2546either with or without the C<Perl_> prefix, thanks to a bunch of defines
10e2eb10 2547that live in F<embed.h>. Note that extension code should I<not> set
0972ecdf
DM
2548C<PERL_CORE>; this exposes the full perl internals, and is likely to cause
2549breakage of the XS in each new perl release.
2550
2551The file F<embed.h> is generated automatically from
10e2eb10 2552F<embed.pl> and F<embed.fnc>. F<embed.pl> also creates the prototyping
dc9b1d22 2553header files for the internal functions, generates the documentation
10e2eb10 2554and a lot of other bits and pieces. It's important that when you add
dc9b1d22 2555a new function to the core or change an existing one, you change the
10e2eb10 2556data in the table in F<embed.fnc> as well. Here's a sample entry from
dc9b1d22 2557that table:
a422fd2d
SC
2558
2559 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
2560
10e2eb10
FC
2561The second column is the return type, the third column the name. Columns
2562after that are the arguments. The first column is a set of flags:
a422fd2d
SC
2563
2564=over 3
2565
2566=item A
2567
10e2eb10
FC
2568This function is a part of the public
2569API. All such functions should also
1aa6ea50 2570have 'd', very few do not.
a422fd2d
SC
2571
2572=item p
2573
1aa6ea50
JC
2574This function has a C<Perl_> prefix; i.e. it is defined as
2575C<Perl_av_fetch>.
a422fd2d
SC
2576
2577=item d
2578
2579This function has documentation using the C<apidoc> feature which we'll
1aa6ea50 2580look at in a second. Some functions have 'd' but not 'A'; docs are good.
a422fd2d
SC
2581
2582=back
2583
2584Other available flags are:
2585
2586=over 3
2587
2588=item s
2589
1aa6ea50
JC
2590This is a static function and is defined as C<STATIC S_whatever>, and
2591usually called within the sources as C<whatever(...)>.
a422fd2d
SC
2592
2593=item n
2594
da8c5729 2595This does not need an interpreter context, so the definition has no
1aa6ea50 2596C<pTHX>, and it follows that callers don't use C<aTHX>. (See
d3a43cd8 2597L</Background and PERL_IMPLICIT_CONTEXT>.)
a422fd2d
SC
2598
2599=item r
2600
2601This function never returns; C<croak>, C<exit> and friends.
2602
2603=item f
2604
2605This function takes a variable number of arguments, C<printf> style.
2606The argument list should end with C<...>, like this:
2607
2608 Afprd |void |croak |const char* pat|...
2609
a7486cbb 2610=item M
a422fd2d 2611
00aadd71 2612This function is part of the experimental development API, and may change
a422fd2d
SC
2613or disappear without notice.
2614
2615=item o
2616
2617This function should not have a compatibility macro to define, say,
10e2eb10 2618C<Perl_parse> to C<parse>. It must be called as C<Perl_parse>.
a422fd2d 2619
a422fd2d
SC
2620=item x
2621
2622This function isn't exported out of the Perl core.
2623
dc9b1d22
MHM
2624=item m
2625
2626This is implemented as a macro.
2627
2628=item X
2629
2630This function is explicitly exported.
2631
2632=item E
2633
2634This function is visible to extensions included in the Perl core.
2635
2636=item b
2637
2638Binary backward compatibility; this function is a macro but also has
2639a C<Perl_> implementation (which is exported).
2640
1aa6ea50
JC
2641=item others
2642
2643See the comments at the top of C<embed.fnc> for others.
2644
a422fd2d
SC
2645=back
2646
dc9b1d22
MHM
2647If you edit F<embed.pl> or F<embed.fnc>, you will need to run
2648C<make regen_headers> to force a rebuild of F<embed.h> and other
2649auto-generated files.
a422fd2d 2650
6b4667fc 2651=head2 Formatted Printing of IVs, UVs, and NVs
9dd9db0b 2652
6b4667fc
A
2653If you are printing IVs, UVs, or NVS instead of the stdio(3) style
2654formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
2655following macros for portability
9dd9db0b 2656
c52f9dcd
JH
2657 IVdf IV in decimal
2658 UVuf UV in decimal
2659 UVof UV in octal
2660 UVxf UV in hexadecimal
2661 NVef NV %e-like
2662 NVff NV %f-like
2663 NVgf NV %g-like
9dd9db0b 2664
6b4667fc
A
2665These will take care of 64-bit integers and long doubles.
2666For example:
2667
c52f9dcd 2668 printf("IV is %"IVdf"\n", iv);
6b4667fc
A
2669
2670The IVdf will expand to whatever is the correct format for the IVs.
9dd9db0b 2671
aacf4ea2
JH
2672Note that there are different "long doubles": Perl will use
2673whatever the compiler has.
2674
8908e76d
JH
2675If you are printing addresses of pointers, use UVxf combined
2676with PTR2UV(), do not use %lx or %p.
2677
e613617c 2678=head2 Formatted Printing of C<Size_t> and C<SSize_t>
5862f74e
KW
2679
2680The most general way to do this is to cast them to a UV or IV, and
2681print as in the
2682L<previous section|/Formatted Printing of IVs, UVs, and NVs>.
2683
2684But if you're using C<PerlIO_printf()>, it's less typing and visual
2685clutter to use the C<"%z"> length modifier (for I<siZe>):
2686
2687 PerlIO_printf("STRLEN is %zu\n", len);
2688
2689This modifier is not portable, so its use should be restricted to
2690C<PerlIO_printf()>.
2691
8908e76d
JH
2692=head2 Pointer-To-Integer and Integer-To-Pointer
2693
2694Because pointer size does not necessarily equal integer size,
2695use the follow macros to do it right.
2696
c52f9dcd
JH
2697 PTR2UV(pointer)
2698 PTR2IV(pointer)
2699 PTR2NV(pointer)
2700 INT2PTR(pointertotype, integer)
8908e76d
JH
2701
2702For example:
2703
c52f9dcd
JH
2704 IV iv = ...;
2705 SV *sv = INT2PTR(SV*, iv);
8908e76d
JH
2706
2707and
2708
c52f9dcd
JH
2709 AV *av = ...;
2710 UV uv = PTR2UV(av);
8908e76d 2711
0ca3a874
MHM
2712=head2 Exception Handling
2713
9b5c3821 2714There are a couple of macros to do very basic exception handling in XS
10e2eb10 2715modules. You have to define C<NO_XSLOCKS> before including F<XSUB.h> to
9b5c3821
MHM
2716be able to use these macros:
2717
2718 #define NO_XSLOCKS
2719 #include "XSUB.h"
2720
2721You can use these macros if you call code that may croak, but you need
10e2eb10 2722to do some cleanup before giving control back to Perl. For example:
0ca3a874 2723
d7f8936a 2724 dXCPT; /* set up necessary variables */
0ca3a874
MHM
2725
2726 XCPT_TRY_START {
2727 code_that_may_croak();
2728 } XCPT_TRY_END
2729
2730 XCPT_CATCH
2731 {
2732 /* do cleanup here */
2733 XCPT_RETHROW;
2734 }
2735
2736Note that you always have to rethrow an exception that has been
10e2eb10
FC
2737caught. Using these macros, it is not possible to just catch the
2738exception and ignore it. If you have to ignore the exception, you
0ca3a874
MHM
2739have to use the C<call_*> function.
2740
2741The advantage of using the above macros is that you don't have
2742to setup an extra function for C<call_*>, and that using these
2743macros is faster than using C<call_*>.
2744
a422fd2d
SC
2745=head2 Source Documentation
2746
2747There's an effort going on to document the internal functions and
61ad4b94 2748automatically produce reference manuals from them -- L<perlapi> is one
a422fd2d 2749such manual which details all the functions which are available to XS
10e2eb10 2750writers. L<perlintern> is the autogenerated manual for the functions
a422fd2d
SC
2751which are not part of the API and are supposedly for internal use only.
2752
2753Source documentation is created by putting POD comments into the C
2754source, like this:
2755
2756 /*
2757 =for apidoc sv_setiv
2758
2759 Copies an integer into the given SV. Does not handle 'set' magic. See
a95b3d6a 2760 L<perlapi/sv_setiv_mg>.
a422fd2d
SC
2761
2762 =cut
2763 */
2764
2765Please try and supply some documentation if you add functions to the
2766Perl core.
2767
0d098d33
MHM
2768=head2 Backwards compatibility
2769
10e2eb10
FC
2770The Perl API changes over time. New functions are
2771added or the interfaces of existing functions are
2772changed. The C<Devel::PPPort> module tries to
0d098d33
MHM
2773provide compatibility code for some of these changes, so XS writers don't
2774have to code it themselves when supporting multiple versions of Perl.
2775
2776C<Devel::PPPort> generates a C header file F<ppport.h> that can also
10e2eb10 2777be run as a Perl script. To generate F<ppport.h>, run:
0d098d33
MHM
2778
2779 perl -MDevel::PPPort -eDevel::PPPort::WriteFile
2780
2781Besides checking existing XS code, the script can also be used to retrieve
2782compatibility information for various API calls using the C<--api-info>
10e2eb10 2783command line switch. For example:
0d098d33
MHM
2784
2785 % perl ppport.h --api-info=sv_magicext
2786
2787For details, see C<perldoc ppport.h>.
2788
a422fd2d
SC
2789=head1 Unicode Support
2790
10e2eb10 2791Perl 5.6.0 introduced Unicode support. It's important for porters and XS
a422fd2d
SC
2792writers to understand this support and make sure that the code they
2793write does not corrupt Unicode data.
2794
2795=head2 What B<is> Unicode, anyway?
2796
10e2eb10
FC
2797In the olden, less enlightened times, we all used to use ASCII. Most of
2798us did, anyway. The big problem with ASCII is that it's American. Well,
a422fd2d 2799no, that's not actually the problem; the problem is that it's not
10e2eb10 2800particularly useful for people who don't use the Roman alphabet. What
a422fd2d 2801used to happen was that particular languages would stick their own
10e2eb10 2802alphabet in the upper range of the sequence, between 128 and 255. Of
a422fd2d
SC
2803course, we then ended up with plenty of variants that weren't quite
2804ASCII, and the whole point of it being a standard was lost.
2805
2806Worse still, if you've got a language like Chinese or
2807Japanese that has hundreds or thousands of characters, then you really
2808can't fit them into a mere 256, so they had to forget about ASCII
2809altogether, and build their own systems using pairs of numbers to refer
2810to one character.
2811
2812To fix this, some people formed Unicode, Inc. and
2813produced a new character set containing all the characters you can
10e2eb10
FC
2814possibly think of and more. There are several ways of representing these
2815characters, and the one Perl uses is called UTF-8. UTF-8 uses
2816a variable number of bytes to represent a character. You can learn more
2575c402 2817about Unicode and Perl's Unicode model in L<perlunicode>.
a422fd2d 2818
3ad86f0e
KW
2819(On EBCDIC platforms, Perl uses instead UTF-EBCDIC, which is a form of
2820UTF-8 adapted for EBCDIC platforms. Below, we just talk about UTF-8.
2821UTF-EBCDIC is like UTF-8, but the details are different. The macros
2822hide the differences from you, just remember that the particular numbers
2823and bit patterns presented below will differ in UTF-EBCDIC.)
2824
1e54db1a 2825=head2 How can I recognise a UTF-8 string?
a422fd2d 2826
10e2eb10
FC
2827You can't. This is because UTF-8 data is stored in bytes just like
2828non-UTF-8 data. The Unicode character 200, (C<0xC8> for you hex types)
a422fd2d 2829capital E with a grave accent, is represented by the two bytes
10e2eb10 2830C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
61ad4b94 2831has that byte sequence as well. So you can't tell just by looking -- this
a422fd2d
SC
2832is what makes Unicode input an interesting problem.
2833
2575c402
JW
2834In general, you either have to know what you're dealing with, or you
2835have to guess. The API function C<is_utf8_string> can help; it'll tell
61ad4b94
KW
2836you if a string contains only valid UTF-8 characters, and the chances
2837of a non-UTF-8 string looking like valid UTF-8 become very small very
2838quickly with increasing string length. On a character-by-character
2839basis, C<isUTF8_CHAR>
2575c402 2840will tell you whether the current character in a string is valid UTF-8.
a422fd2d 2841
1e54db1a 2842=head2 How does UTF-8 represent Unicode characters?
a422fd2d 2843
1e54db1a 2844As mentioned above, UTF-8 uses a variable number of bytes to store a
10e2eb10
FC
2845character. Characters with values 0...127 are stored in one
2846byte, just like good ol' ASCII. Character 128 is stored as
2847C<v194.128>; this continues up to character 191, which is
2848C<v194.191>. Now we've run out of bits (191 is binary
61ad4b94 2849C<10111111>) so we move on; character 192 is C<v195.128>. And
a422fd2d 2850so it goes on, moving to three bytes at character 2048.
6e31cdd1 2851L<perlunicode/Unicode Encodings> has pictures of how this works.
a422fd2d 2852
1e54db1a 2853Assuming you know you're dealing with a UTF-8 string, you can find out
a422fd2d
SC
2854how long the first character in it is with the C<UTF8SKIP> macro:
2855
2856 char *utf = "\305\233\340\240\201";
2857 I32 len;
2858
2859 len = UTF8SKIP(utf); /* len is 2 here */
2860 utf += len;
2861 len = UTF8SKIP(utf); /* len is 3 here */
2862
1e54db1a 2863Another way to skip over characters in a UTF-8 string is to use
a422fd2d 2864C<utf8_hop>, which takes a string and a number of characters to skip
10e2eb10 2865over. You're on your own about bounds checking, though, so don't use it
a422fd2d
SC
2866lightly.
2867
1e54db1a 2868All bytes in a multi-byte UTF-8 character will have the high bit set,
3a2263fe 2869so you can test if you need to do something special with this
61ad4b94 2870character like this (the C<UTF8_IS_INVARIANT()> is a macro that tests
9f98c7fe 2871whether the byte is encoded as a single byte even in UTF-8):
a422fd2d 2872
3a2263fe 2873 U8 *utf;
4b88fb76 2874 U8 *utf_end; /* 1 beyond buffer pointed to by utf */
3a2263fe 2875 UV uv; /* Note: a UV, not a U8, not a char */
95701e00 2876 STRLEN len; /* length of character in bytes */
a422fd2d 2877
3a2263fe 2878 if (!UTF8_IS_INVARIANT(*utf))
1e54db1a 2879 /* Must treat this as UTF-8 */
4b88fb76 2880 uv = utf8_to_uvchr_buf(utf, utf_end, &len);
a422fd2d
SC
2881 else
2882 /* OK to treat this character as a byte */
2883 uv = *utf;
2884
4b88fb76 2885You can also see in that example that we use C<utf8_to_uvchr_buf> to get the
95701e00 2886value of the character; the inverse function C<uvchr_to_utf8> is available
1e54db1a 2887for putting a UV into UTF-8:
a422fd2d 2888
61ad4b94 2889 if (!UVCHR_IS_INVARIANT(uv))
a422fd2d 2890 /* Must treat this as UTF8 */
95701e00 2891 utf8 = uvchr_to_utf8(utf8, uv);
a422fd2d
SC
2892 else
2893 /* OK to treat this character as a byte */
2894 *utf8++ = uv;
2895
2896You B<must> convert characters to UVs using the above functions if
1e54db1a 2897you're ever in a situation where you have to match UTF-8 and non-UTF-8
10e2eb10 2898characters. You may not skip over UTF-8 characters in this case. If you
1e54db1a
JH
2899do this, you'll lose the ability to match hi-bit non-UTF-8 characters;
2900for instance, if your UTF-8 string contains C<v196.172>, and you skip
2901that character, you can never match a C<chr(200)> in a non-UTF-8 string.
a422fd2d
SC
2902So don't do that!
2903
61ad4b94
KW
2904(Note that we don't have to test for invariant characters in the
2905examples above. The functions work on any well-formed UTF-8 input.
2906It's just that its faster to avoid the function overhead when it's not
2907needed.)
2908
1e54db1a 2909=head2 How does Perl store UTF-8 strings?
a422fd2d 2910
61ad4b94 2911Currently, Perl deals with UTF-8 strings and non-UTF-8 strings
10e2eb10
FC
2912slightly differently. A flag in the SV, C<SVf_UTF8>, indicates that the
2913string is internally encoded as UTF-8. Without it, the byte value is the
61ad4b94
KW
2914codepoint number and vice versa. This flag is only meaningful if the SV
2915is C<SvPOK> or immediately after stringification via C<SvPV> or a
2916similar macro. You can check and manipulate this flag with the
2575c402 2917following macros:
a422fd2d
SC
2918
2919 SvUTF8(sv)
2920 SvUTF8_on(sv)
2921 SvUTF8_off(sv)
2922
2923This flag has an important effect on Perl's treatment of the string: if
61ad4b94 2924UTF-8 data is not properly distinguished, regular expressions,
a422fd2d 2925C<length>, C<substr> and other string handling operations will have
61ad4b94 2926undesirable (wrong) results.
a422fd2d
SC
2927
2928The problem comes when you have, for instance, a string that isn't
61ad4b94 2929flagged as UTF-8, and contains a byte sequence that could be UTF-8 --
1e54db1a 2930especially when combining non-UTF-8 and UTF-8 strings.
a422fd2d 2931
61ad4b94
KW
2932Never forget that the C<SVf_UTF8> flag is separate from the PV value; you
2933need to be sure you don't accidentally knock it off while you're
10e2eb10 2934manipulating SVs. More specifically, you cannot expect to do this:
a422fd2d
SC
2935
2936 SV *sv;
2937 SV *nsv;
2938 STRLEN len;
2939 char *p;
2940
2941 p = SvPV(sv, len);
2942 frobnicate(p);
2943 nsv = newSVpvn(p, len);
2944
2945The C<char*> string does not tell you the whole story, and you can't
10e2eb10 2946copy or reconstruct an SV just by copying the string value. Check if the
c31cc9fc
FC
2947old SV has the UTF8 flag set (I<after> the C<SvPV> call), and act
2948accordingly:
a422fd2d
SC
2949
2950 p = SvPV(sv, len);
6db25795
KW
2951 is_utf8 = SvUTF8(sv);
2952 frobnicate(p, is_utf8);
a422fd2d 2953 nsv = newSVpvn(p, len);
6db25795 2954 if (is_utf8)
a422fd2d
SC
2955 SvUTF8_on(nsv);
2956
6db25795
KW
2957In the above, your C<frobnicate> function has been changed to be made
2958aware of whether or not it's dealing with UTF-8 data, so that it can
2959handle the string appropriately.
a422fd2d 2960
3a2263fe 2961Since just passing an SV to an XS function and copying the data of
2575c402 2962the SV is not enough to copy the UTF8 flags, even less right is just
61ad4b94 2963passing a S<C<char *>> to an XS function.
3a2263fe 2964
dc83bf8e 2965For full generality, use the L<C<DO_UTF8>|perlapi/DO_UTF8> macro to see if the
6db25795
KW
2966string in an SV is to be I<treated> as UTF-8. This takes into account
2967if the call to the XS function is being made from within the scope of
2968L<S<C<use bytes>>|bytes>. If so, the underlying bytes that comprise the
2969UTF-8 string are to be exposed, rather than the character they
2970represent. But this pragma should only really be used for debugging and
2971perhaps low-level testing at the byte level. Hence most XS code need
2972not concern itself with this, but various areas of the perl core do need
2973to support it.
2974
2975And this isn't the whole story. Starting in Perl v5.12, strings that
2976aren't encoded in UTF-8 may also be treated as Unicode under various
6e31cdd1 2977conditions (see L<perlunicode/ASCII Rules versus Unicode Rules>).
6db25795
KW
2978This is only really a problem for characters whose ordinals are between
2979128 and 255, and their behavior varies under ASCII versus Unicode rules
2980in ways that your code cares about (see L<perlunicode/The "Unicode Bug">).
2981There is no published API for dealing with this, as it is subject to
2982change, but you can look at the code for C<pp_lc> in F<pp.c> for an
2983example as to how it's currently done.
2984
1e54db1a 2985=head2 How do I convert a string to UTF-8?
a422fd2d 2986
2575c402 2987If you're mixing UTF-8 and non-UTF-8 strings, it is necessary to upgrade
61ad4b94 2988the non-UTF-8 strings to UTF-8. If you've got an SV, the easiest way to do
2575c402 2989this is:
a422fd2d
SC
2990
2991 sv_utf8_upgrade(sv);
2992
2993However, you must not do this, for example:
2994
2995 if (!SvUTF8(left))
2996 sv_utf8_upgrade(left);
2997
2998If you do this in a binary operator, you will actually change one of the
b1866b2d 2999strings that came into the operator, and, while it shouldn't be noticeable
2575c402 3000by the end user, it can cause problems in deficient code.
a422fd2d 3001
1e54db1a 3002Instead, C<bytes_to_utf8> will give you a UTF-8-encoded B<copy> of its
10e2eb10
FC
3003string argument. This is useful for having the data available for
3004comparisons and so on, without harming the original SV. There's also
a422fd2d
SC
3005C<utf8_to_bytes> to go the other way, but naturally, this will fail if
3006the string contains any characters above 255 that can't be represented
3007in a single byte.
3008
6db25795
KW
3009=head2 How do I compare strings?
3010
3011L<perlapi/sv_cmp> and L<perlapi/sv_cmp_flags> do a lexigraphic
3012comparison of two SV's, and handle UTF-8ness properly. Note, however,
3013that Unicode specifies a much fancier mechanism for collation, available
3014via the L<Unicode::Collate> module.
3015
3016To just compare two strings for equality/non-equality, you can just use
3017L<C<memEQ()>|perlapi/memEQ> and L<C<memNE()>|perlapi/memEQ> as usual,
3018except the strings must be both UTF-8 or not UTF-8 encoded.
3019
3020To compare two strings case-insensitively, use
3021L<C<foldEQ_utf8()>|perlapi/foldEQ_utf8> (the strings don't have to have
3022the same UTF-8ness).
3023
a422fd2d
SC
3024=head2 Is there anything else I need to know?
3025
10e2eb10 3026Not really. Just remember these things:
a422fd2d
SC
3027
3028=over 3
3029
3030=item *
3031
6db25795
KW
3032There's no way to tell if a S<C<char *>> or S<C<U8 *>> string is UTF-8
3033or not. But you can tell if an SV is to be treated as UTF-8 by calling
3034C<DO_UTF8> on it, after stringifying it with C<SvPV> or a similar
3035macro. And, you can tell if SV is actually UTF-8 (even if it is not to
3036be treated as such) by looking at its C<SvUTF8> flag (again after
3037stringifying it). Don't forget to set the flag if something should be
3038UTF-8.
3039Treat the flag as part of the PV, even though it's not -- if you pass on
3040the PV to somewhere, pass on the flag too.
a422fd2d
SC
3041
3042=item *
3043
4b88fb76 3044If a string is UTF-8, B<always> use C<utf8_to_uvchr_buf> to get at the value,
3a2263fe 3045unless C<UTF8_IS_INVARIANT(*s)> in which case you can use C<*s>.
a422fd2d
SC
3046
3047=item *
3048
61ad4b94
KW
3049When writing a character UV to a UTF-8 string, B<always> use
3050C<uvchr_to_utf8>, unless C<UVCHR_IS_INVARIANT(uv))> in which case
3a2263fe 3051you can use C<*s = uv>.
a422fd2d
SC
3052
3053=item *
3054
10e2eb10
FC
3055Mixing UTF-8 and non-UTF-8 strings is
3056tricky. Use C<bytes_to_utf8> to get
2bbc8d55 3057a new string which is UTF-8 encoded, and then combine them.
a422fd2d
SC
3058
3059=back
3060
53e06cf0
SC
3061=head1 Custom Operators
3062
2a0fd0f1 3063Custom operator support is an experimental feature that allows you to
10e2eb10 3064define your own ops. This is primarily to allow the building of
53e06cf0
SC
3065interpreters for other languages in the Perl core, but it also allows
3066optimizations through the creation of "macro-ops" (ops which perform the
3067functions of multiple ops which are usually executed together, such as
1aa6ea50 3068C<gvsv, gvsv, add>.)
53e06cf0 3069
10e2eb10 3070This feature is implemented as a new op type, C<OP_CUSTOM>. The Perl
53e06cf0 3071core does not "know" anything special about this op type, and so it will
10e2eb10 3072not be involved in any optimizations. This also means that you can
61ad4b94
KW
3073define your custom ops to be any op structure -- unary, binary, list and
3074so on -- you like.
53e06cf0 3075
10e2eb10
FC
3076It's important to know what custom operators won't do for you. They
3077won't let you add new syntax to Perl, directly. They won't even let you
3078add new keywords, directly. In fact, they won't change the way Perl
3079compiles a program at all. You have to do those changes yourself, after
3080Perl has compiled the program. You do this either by manipulating the op
53e06cf0
SC
3081tree using a C<CHECK> block and the C<B::Generate> module, or by adding
3082a custom peephole optimizer with the C<optimize> module.
3083
3084When you do this, you replace ordinary Perl ops with custom ops by
407f86e1 3085creating ops with the type C<OP_CUSTOM> and the C<op_ppaddr> of your own
10e2eb10
FC
3086PP function. This should be defined in XS code, and should look like
3087the PP ops in C<pp_*.c>. You are responsible for ensuring that your op
53e06cf0
SC
3088takes the appropriate number of values from the stack, and you are
3089responsible for adding stack marks if necessary.
3090
3091You should also "register" your op with the Perl interpreter so that it
10e2eb10 3092can produce sensible error and warning messages. Since it is possible to
53e06cf0 3093have multiple custom ops within the one "logical" op type C<OP_CUSTOM>,
9733086d 3094Perl uses the value of C<< o->op_ppaddr >> to determine which custom op
10e2eb10 3095it is dealing with. You should create an C<XOP> structure for each
9733086d
BM
3096ppaddr you use, set the properties of the custom op with
3097C<XopENTRY_set>, and register the structure against the ppaddr using
10e2eb10 3098C<Perl_custom_op_register>. A trivial example might look like:
9733086d
BM
3099
3100 static XOP my_xop;
3101 static OP *my_pp(pTHX);
3102
3103 BOOT:
3104 XopENTRY_set(&my_xop, xop_name, "myxop");
3105 XopENTRY_set(&my_xop, xop_desc, "Useless custom op");
3106 Perl_custom_op_register(aTHX_ my_pp, &my_xop);
3107
3108The available fields in the structure are:
3109
3110=over 4
3111
3112=item xop_name
3113
10e2eb10 3114A short name for your op. This will be included in some error messages,
9733086d
BM
3115and will also be returned as C<< $op->name >> by the L<B|B> module, so
3116it will appear in the output of module like L<B::Concise|B::Concise>.
3117
3118=item xop_desc
3119
3120A short description of the function of the op.
3121
3122=item xop_class
3123
10e2eb10 3124Which of the various C<*OP> structures this op uses. This should be one of
9733086d
BM
3125the C<OA_*> constants from F<op.h>, namely
3126
3127=over 4
3128
3129=item OA_BASEOP
3130
3131=item OA_UNOP
3132
3133=item OA_BINOP
3134
3135=item OA_LOGOP
3136
3137=item OA_LISTOP
3138
3139=item OA_PMOP
3140
3141=item OA_SVOP
3142
3143=item OA_PADOP
3144
3145=item OA_PVOP_OR_SVOP
3146
10e2eb10 3147This should be interpreted as 'C<PVOP>' only. The C<_OR_SVOP> is because
9733086d
BM
3148the only core C<PVOP>, C<OP_TRANS>, can sometimes be a C<SVOP> instead.
3149
3150=item OA_LOOP
3151
3152=item OA_COP
3153
3154=back
3155
3156The other C<OA_*> constants should not be used.
3157
3158=item xop_peep
3159
3160This member is of type C<Perl_cpeep_t>, which expands to C<void
10e2eb10 3161(*Perl_cpeep_t)(aTHX_ OP *o, OP *oldop)>. If it is set, this function
9733086d 3162will be called from C<Perl_rpeep> when ops of this type are encountered
10e2eb10 3163by the peephole optimizer. I<o> is the OP that needs optimizing;
9733086d
BM
3164I<oldop> is the previous OP optimized, whose C<op_next> points to I<o>.
3165
3166=back
53e06cf0 3167
e7d4c058 3168C<B::Generate> directly supports the creation of custom ops by name.
53e06cf0 3169
bf5e9371
DM
3170
3171=head1 Dynamic Scope and the Context Stack
3172
3173B<Note:> this section describes a non-public internal API that is subject
3174to change without notice.
3175
3176=head2 Introduction to the context stack
3177
3178In Perl, dynamic scoping refers to the runtime nesting of things like
3179subroutine calls, evals etc, as well as the entering and exiting of block
3180scopes. For example, the restoring of a C<local>ised variable is
3181determined by the dynamic scope.
3182
3183Perl tracks the dynamic scope by a data structure called the context
3184stack, which is an array of C<PERL_CONTEXT> structures, and which is
3185itself a big union for all the types of context. Whenever a new scope is
3186entered (such as a block, a C<for> loop, or a subroutine call), a new
3187context entry is pushed onto the stack. Similarly when leaving a block or
3188returning from a subroutine call etc. a context is popped. Since the
3189context stack represents the current dynamic scope, it can be searched.
3190For example, C<next LABEL> searches back through the stack looking for a
3191loop context that matches the label; C<return> pops contexts until it
3192finds a sub or eval context or similar; C<caller> examines sub contexts on
3193the stack.
3194
3195Each context entry is labelled with a context type, C<cx_type>. Typical
3196context types are C<CXt_SUB>, C<CXt_EVAL> etc., as well as C<CXt_BLOCK>
3197and C<CXt_NULL> which represent a basic scope (as pushed by C<pp_enter>)
3198and a sort block. The type determines which part of the context union are
3199valid.
3200
3201The main division in the context struct is between a substitution scope
3202(C<CXt_SUBST>) and block scopes, which are everything else. The former is
d7c7f8cb 3203just used while executing C<s///e>, and won't be discussed further
bf5e9371
DM
3204here.
3205
3206All the block scope types share a common base, which corresponds to
3207C<CXt_BLOCK>. This stores the old values of various scope-related
3208variables like C<PL_curpm>, as well as information about the current
3209scope, such as C<gimme>. On scope exit, the old variables are restored.
3210
3211Particular block scope types store extra per-type information. For
3212example, C<CXt_SUB> stores the currently executing CV, while the various
3213for loop types might hold the original loop variable SV. On scope exit,
3214the per-type data is processed; for example the CV has its reference count
3215decremented, and the original loop variable is restored.
3216
3217The macro C<cxstack> returns the base of the current context stack, while
3218C<cxstack_ix> is the index of the current frame within that stack.
3219
3220In fact, the context stack is actually part of a stack-of-stacks system;
3221whenever something unusual is done such as calling a C<DESTROY> or tie
3222handler, a new stack is pushed, then popped at the end.
3223
3224Note that the API described here changed considerably in perl 5.24; prior
3225to that, big macros like C<PUSHBLOCK> and C<POPSUB> were used; in 5.24
3226they were replaced by the inline static functions described below. In
3227addition, the ordering and detail of how these macros/function work
3228changed in many ways, often subtly. In particular they didn't handle
3229saving the savestack and temps stack positions, and required additional
3230C<ENTER>, C<SAVETMPS> and C<LEAVE> compared to the new functions. The
3231old-style macros will not be described further.
3232
3233
3234=head2 Pushing contexts
3235
3236For pushing a new context, the two basic functions are
3237C<cx = cx_pushblock()>, which pushes a new basic context block and returns
3238its address, and a family of similar functions with names like
3239C<cx_pushsub(cx)> which populate the additional type-dependent fields in
3240the C<cx> struct. Note that C<CXt_NULL> and C<CXt_BLOCK> don't have their
3241own push functions, as they don't store any data beyond that pushed by
3242C<cx_pushblock>.
3243
3244The fields of the context struct and the arguments to the C<cx_*>
3245functions are subject to change between perl releases, representing
3246whatever is convenient or efficient for that release.
3247
3248A typical context stack pushing can be found in C<pp_entersub>; the
3249following shows a simplified and stripped-down example of a non-XS call,
3250along with comments showing roughly what each function does.
3251
61f554bd
KW
3252 dMARK;
3253 U8 gimme = GIMME_V;
3254 bool hasargs = cBOOL(PL_op->op_flags & OPf_STACKED);
3255 OP *retop = PL_op->op_next;
3256 I32 old_ss_ix = PL_savestack_ix;
3257 CV *cv = ....;
3258
3259 /* ... make mortal copies of stack args which are PADTMPs here ... */
3260
3261 /* ... do any additional savestack pushes here ... */
3262
3263 /* Now push a new context entry of type 'CXt_SUB'; initially just
3264 * doing the actions common to all block types: */
3265
3266 cx = cx_pushblock(CXt_SUB, gimme, MARK, old_ss_ix);
3267
3268 /* this does (approximately):
3269 CXINC; /* cxstack_ix++ (grow if necessary) */
3270 cx = CX_CUR(); /* and get the address of new frame */
3271 cx->cx_type = CXt_SUB;
3272 cx->blk_gimme = gimme;
3273 cx->blk_oldsp = MARK - PL_stack_base;
3274 cx->blk_oldsaveix = old_ss_ix;
3275 cx->blk_oldcop = PL_curcop;
3276 cx->blk_oldmarksp = PL_markstack_ptr - PL_markstack;
3277 cx->blk_oldscopesp = PL_scopestack_ix;
3278 cx->blk_oldpm = PL_curpm;
3279 cx->blk_old_tmpsfloor = PL_tmps_floor;
3280
3281 PL_tmps_floor = PL_tmps_ix;
3282 */
bf5e9371
DM
3283
3284
61f554bd
KW
3285 /* then update the new context frame with subroutine-specific info,
3286 * such as the CV about to be executed: */
bf5e9371 3287
61f554bd 3288 cx_pushsub(cx, cv, retop, hasargs);
bf5e9371 3289
61f554bd
KW
3290 /* this does (approximately):
3291 cx->blk_sub.cv = cv;
3292 cx->blk_sub.olddepth = CvDEPTH(cv);
3293 cx->blk_sub.prevcomppad = PL_comppad;
3294 cx->cx_type |= (hasargs) ? CXp_HASARGS : 0;
3295 cx->blk_sub.retop = retop;
3296 SvREFCNT_inc_simple_void_NN(cv);
3297 */
bf5e9371
DM
3298
3299Note that C<cx_pushblock()> sets two new floors: for the args stack (to
3300C<MARK>) and the temps stack (to C<PL_tmps_ix>). While executing at this
3301scope level, every C<nextstate> (amongst others) will reset the args and
3302tmps stack levels to these floors. Note that since C<cx_pushblock> uses
3303the current value of C<PL_tmps_ix> rather than it being passed as an arg,
3304this dictates at what point C<cx_pushblock> should be called. In
3305particular, any new mortals which should be freed only on scope exit
3306(rather than at the next C<nextstate>) should be created first.
3307
3308Most callers of C<cx_pushblock> simply set the new args stack floor to the
3309top of the previous stack frame, but for C<CXt_LOOP_LIST> it stores the
3310items being iterated over on the stack, and so sets C<blk_oldsp> to the
3311top of these items instead. Note that, contrary to its name, C<blk_oldsp>
3312doesn't always represent the value to restore C<PL_stack_sp> to on scope
3313exit.
3314
3315Note the early capture of C<PL_savestack_ix> to C<old_ss_ix>, which is
3316later passed as an arg to C<cx_pushblock>. In the case of C<pp_entersub>,
3317this is because, although most values needing saving are stored in fields
3318of the context struct, an extra value needs saving only when the debugger
3319is running, and it doesn't make sense to bloat the struct for this rare
3320case. So instead it is saved on the savestack. Since this value gets
3321calculated and saved before the context is pushed, it is necessary to pass
3322the old value of C<PL_savestack_ix> to C<cx_pushblock>, to ensure that the
3323saved value gets freed during scope exit. For most users of
3324C<cx_pushblock>, where nothing needs pushing on the save stack,
3325C<PL_savestack_ix> is just passed directly as an arg to C<cx_pushblock>.
3326
3327Note that where possible, values should be saved in the context struct
3328rather than on the save stack; it's much faster that way.
3329
3330Normally C<cx_pushblock> should be immediately followed by the appropriate
3331C<cx_pushfoo>, with nothing between them; this is because if code
3332in-between could die (e.g. a warning upgraded to fatal), then the context
3333stack unwinding code in C<dounwind> would see (in the example above) a
3334C<CXt_SUB> context frame, but without all the subroutine-specific fields
3335set, and crashes would soon ensue.
3336
3337Where the two must be separate, initially set the type to C<CXt_NULL> or
3338C<CXt_BLOCK>, and later change it to C<CXt_foo> when doing the
3339C<cx_pushfoo>. This is exactly what C<pp_enteriter> does, once it's
3340determined which type of loop it's pushing.
3341
3342=head2 Popping contexts
3343
3344Contexts are popped using C<cx_popsub()> etc. and C<cx_popblock()>. Note
3345however, that unlike C<cx_pushblock>, neither of these functions actually
3346decrement the current context stack index; this is done separately using
3347C<CX_POP()>.
3348
3349There are two main ways that contexts are popped. During normal execution
3350as scopes are exited, functions like C<pp_leave>, C<pp_leaveloop> and
3351C<pp_leavesub> process and pop just one context using C<cx_popfoo> and
3352C<cx_popblock>. On the other hand, things like C<pp_return> and C<next>
3353may have to pop back several scopes until a sub or loop context is found,
3354and exceptions (such as C<die>) need to pop back contexts until an eval
3355context is found. Both of these are accomplished by C<dounwind()>, which
3356is capable of processing and popping all contexts above the target one.
3357
3358Here is a typical example of context popping, as found in C<pp_leavesub>
3359(simplified slightly):
3360
61f554bd
KW
3361 U8 gimme;
3362 PERL_CONTEXT *cx;
3363 SV **oldsp;
3364 OP *retop;
bf5e9371 3365
61f554bd 3366 cx = CX_CUR();
bf5e9371 3367
61f554bd
KW
3368 gimme = cx->blk_gimme;
3369 oldsp = PL_stack_base + cx->blk_oldsp; /* last arg of previous frame */
bf5e9371 3370
61f554bd
KW
3371 if (gimme == G_VOID)
3372 PL_stack_sp = oldsp;
3373 else
3374 leave_adjust_stacks(oldsp, oldsp, gimme, 0);
bf5e9371 3375
61f554bd
KW
3376 CX_LEAVE_SCOPE(cx);
3377 cx_popsub(cx);
3378 cx_popblock(cx);
3379 retop = cx->blk_sub.retop;
3380 CX_POP(cx);
bf5e9371 3381
61f554bd 3382 return retop;
bf5e9371
DM
3383
3384The steps above are in a very specific order, designed to be the reverse
3385order of when the context was pushed. The first thing to do is to copy
3386and/or protect any any return arguments and free any temps in the current
3387scope. Scope exits like an rvalue sub normally return a mortal copy of
3388their return args (as opposed to lvalue subs). It is important to make
3389this copy before the save stack is popped or variables are restored, or
3390bad things like the following can happen:
3391
3392 sub f { my $x =...; $x } # $x freed before we get to copy it
3393 sub f { /(...)/; $1 } # PL_curpm restored before $1 copied
3394
3395Although we wish to free any temps at the same time, we have to be careful
3396not to free any temps which are keeping return args alive; nor to free the
3397temps we have just created while mortal copying return args. Fortunately,
3398C<leave_adjust_stacks()> is capable of making mortal copies of return args,
3399shifting args down the stack, and only processing those entries on the
3400temps stack that are safe to do so.
3401
3402In void context no args are returned, so it's more efficient to skip
3403calling C<leave_adjust_stacks()>. Also in void context, a C<nextstate> op
3404is likely to be imminently called which will do a C<FREETMPS>, so there's
3405no need to do that either.
3406
3407The next step is to pop savestack entries: C<CX_LEAVE_SCOPE(cx)> is just
3408defined as C<<LEAVE_SCOPE(cx->blk_oldsaveix)>>. Note that during the
3409popping, it's possible for perl to call destructors, call C<STORE> to undo
3410localisations of tied vars, and so on. Any of these can die or call
3411C<exit()>. In this case, C<dounwind()> will be called, and the current
3412context stack frame will be re-processed. Thus it is vital that all steps
3413in popping a context are done in such a way to support reentrancy. The
3414other alternative, of decrementing C<cxstack_ix> I<before> processing the
3415frame, would lead to leaks and the like if something died halfway through,
3416or overwriting of the current frame.
3417
3418C<CX_LEAVE_SCOPE> itself is safely re-entrant: if only half the savestack
3419items have been popped before dying and getting trapped by eval, then the
3420C<CX_LEAVE_SCOPE>s in C<dounwind> or C<pp_leaveeval> will continue where
3421the first one left off.
3422
3423The next step is the type-specific context processing; in this case
3424C<cx_popsub>. In part, this looks like:
3425
3426 cv = cx->blk_sub.cv;
3427 CvDEPTH(cv) = cx->blk_sub.olddepth;
3428 cx->blk_sub.cv = NULL;
3429 SvREFCNT_dec(cv);
3430
3431where its processing the just-executed CV. Note that before it decrements
3432the CV's reference count, it nulls the C<blk_sub.cv>. This means that if
3433it re-enters, the CV won't be freed twice. It also means that you can't
3434rely on such type-specific fields having useful values after the return
3435from C<cx_popfoo>.
3436
3437Next, C<cx_popblock> restores all the various interpreter vars to their
3438previous values or previous high water marks; it expands to:
3439
3440 PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
3441 PL_scopestack_ix = cx->blk_oldscopesp;
3442 PL_curpm = cx->blk_oldpm;
3443 PL_curcop = cx->blk_oldcop;
3444 PL_tmps_floor = cx->blk_old_tmpsfloor;
3445
3446Note that it I<doesn't> restore C<PL_stack_sp>; as mentioned earlier,
3447which value to restore it to depends on the context type (specifically
3448C<for (list) {}>), and what args (if any) it returns; and that will
3449already have been sorted out earlier by C<leave_adjust_stacks()>.
3450
3451Finally, the context stack pointer is actually decremented by C<CX_POP(cx)>.
3452After this point, it's possible that that the current context frame could
3453be overwritten by other contexts being pushed. Although things like ties
3454and C<DESTROY> are supposed to work within a new context stack, it's best
3455not to assume this. Indeed on debugging builds, C<CX_POP(cx)> deliberately
3456sets C<cx> to null to detect code that is still relying on the field
3457values in that context frame. Note in the C<pp_leavesub()> example above,
3458we grab C<blk_sub.retop> I<before> calling C<CX_POP>.
3459
3460=head2 Redoing contexts
3461
3462Finally, there is C<cx_topblock(cx)>, which acts like a super-C<nextstate>
3463as regards to resetting various vars to their base values. It is used in
3464places like C<pp_next>, C<pp_redo> and C<pp_goto> where rather than
3465exiting a scope, we want to re-initialise the scope. As well as resetting
3466C<PL_stack_sp> like C<nextstate>, it also resets C<PL_markstack_ptr>,
3467C<PL_scopestack_ix> and C<PL_curpm>. Note that it doesn't do a
3468C<FREETMPS>.
3469
3470
954c1994 3471=head1 AUTHORS
e89caa19 3472
954c1994 3473Until May 1997, this document was maintained by Jeff Okamoto
9b5bb84f
SB
3474E<lt>okamoto@corp.hp.comE<gt>. It is now maintained as part of Perl
3475itself by the Perl 5 Porters E<lt>perl5-porters@perl.orgE<gt>.
cb1a09d0 3476
954c1994
GS
3477With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
3478Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
3479Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
3480Stephen McCamant, and Gurusamy Sarathy.
cb1a09d0 3481
954c1994 3482=head1 SEE ALSO