This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
perlguts: Document save_item
[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 Array
685 SVt_PVHV Hash
686 SVt_PVCV Code
687 SVt_PVGV Glob (possibly a file handle)
688
2d0e7d1f
DM
689Any numerical value returned which is less than SVt_PVAV will be a scalar
690of some form.
691
a5e62da0 692See L<perlapi/svtype> for more details.
d1b91892 693
cb1a09d0
AD
694=head2 Blessed References and Class Objects
695
06f6df17 696References are also used to support object-oriented programming. In perl's
cb1a09d0
AD
697OO lexicon, an object is simply a reference that has been blessed into a
698package (or class). Once blessed, the programmer may now use the reference
699to access the various methods in the class.
700
701A reference can be blessed into a package with the following function:
702
703 SV* sv_bless(SV* sv, HV* stash);
704
06f6df17
RGS
705The C<sv> argument must be a reference value. The C<stash> argument
706specifies which class the reference will belong to. See
5a0de581 707L</Stashes and Globs> for information on converting class names into stashes.
cb1a09d0
AD
708
709/* Still under construction */
710
ddd2cc91
DM
711The following function upgrades rv to reference if not already one.
712Creates a new SV for rv to point to. If C<classname> is non-null, the SV
713is blessed into the specified class. SV is returned.
cb1a09d0 714
08105a92 715 SV* newSVrv(SV* rv, const char* classname);
cb1a09d0 716
ddd2cc91
DM
717The following three functions copy integer, unsigned integer or double
718into an SV whose reference is C<rv>. SV is blessed if C<classname> is
719non-null.
cb1a09d0 720
08105a92 721 SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
e1c57cef 722 SV* sv_setref_uv(SV* rv, const char* classname, UV uv);
08105a92 723 SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
cb1a09d0 724
ddd2cc91
DM
725The following function copies the pointer value (I<the address, not the
726string!>) into an SV whose reference is rv. SV is blessed if C<classname>
727is non-null.
cb1a09d0 728
ddd2cc91 729 SV* sv_setref_pv(SV* rv, const char* classname, void* pv);
cb1a09d0 730
a9b0660e 731The following function copies a string into an SV whose reference is C<rv>.
ddd2cc91
DM
732Set length to 0 to let Perl calculate the string length. SV is blessed if
733C<classname> is non-null.
cb1a09d0 734
a9b0660e
KW
735 SV* sv_setref_pvn(SV* rv, const char* classname, char* pv,
736 STRLEN length);
cb1a09d0 737
ddd2cc91
DM
738The following function tests whether the SV is blessed into the specified
739class. It does not check inheritance relationships.
9abd00ed 740
08105a92 741 int sv_isa(SV* sv, const char* name);
9abd00ed 742
ddd2cc91 743The following function tests whether the SV is a reference to a blessed object.
9abd00ed
GS
744
745 int sv_isobject(SV* sv);
746
ddd2cc91 747The following function tests whether the SV is derived from the specified
10e2eb10
FC
748class. SV can be either a reference to a blessed object or a string
749containing a class name. This is the function implementing the
ddd2cc91 750C<UNIVERSAL::isa> functionality.
9abd00ed 751
08105a92 752 bool sv_derived_from(SV* sv, const char* name);
9abd00ed 753
00aadd71 754To check if you've got an object derived from a specific class you have
9abd00ed
GS
755to write:
756
757 if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
cb1a09d0 758
5f05dabc 759=head2 Creating New Variables
cb1a09d0 760
5f05dabc
PP
761To create a new Perl variable with an undef value which can be accessed from
762your Perl script, use the following routines, depending on the variable type.
cb1a09d0 763
64ace3f8 764 SV* get_sv("package::varname", GV_ADD);
cbfd0a87 765 AV* get_av("package::varname", GV_ADD);
6673a63c 766 HV* get_hv("package::varname", GV_ADD);
cb1a09d0 767
058a5f6c 768Notice the use of GV_ADD as the second parameter. The new variable can now
cb1a09d0
AD
769be set, using the routines appropriate to the data type.
770
5f05dabc 771There are additional macros whose values may be bitwise OR'ed with the
058a5f6c 772C<GV_ADD> argument to enable certain extra features. Those bits are:
cb1a09d0 773
9a68f1db
SB
774=over
775
776=item GV_ADDMULTI
777
778Marks the variable as multiply defined, thus preventing the:
779
780 Name <varname> used only once: possible typo
781
782warning.
783
9a68f1db
SB
784=item GV_ADDWARN
785
786Issues the warning:
787
788 Had to create <varname> unexpectedly
789
790if the variable did not exist before the function was called.
791
792=back
cb1a09d0 793
07fa94a1
JO
794If you do not specify a package name, the variable is created in the current
795package.
cb1a09d0 796
5f05dabc 797=head2 Reference Counts and Mortality
a0d0e21e 798
10e2eb10 799Perl uses a reference count-driven garbage collection mechanism. SVs,
54310121 800AVs, or HVs (xV for short in the following) start their life with a
55497cff 801reference count of 1. If the reference count of an xV ever drops to 0,
07fa94a1 802then it will be destroyed and its memory made available for reuse.
3d2ba989
Z
803At the most basic internal level, reference counts can be manipulated
804with the following macros:
55497cff
PP
805
806 int SvREFCNT(SV* sv);
5f05dabc 807 SV* SvREFCNT_inc(SV* sv);
55497cff
PP
808 void SvREFCNT_dec(SV* sv);
809
3d2ba989
Z
810(There are also suffixed versions of the increment and decrement macros,
811for situations where the full generality of these basic macros can be
812exchanged for some performance.)
813
814However, the way a programmer should think about references is not so
815much in terms of the bare reference count, but in terms of I<ownership>
816of references. A reference to an xV can be owned by any of a variety
817of entities: another xV, the Perl interpreter, an XS data structure,
818a piece of running code, or a dynamic scope. An xV generally does not
819know what entities own the references to it; it only knows how many
820references there are, which is the reference count.
821
822To correctly maintain reference counts, it is essential to keep track
823of what references the XS code is manipulating. The programmer should
824always know where a reference has come from and who owns it, and be
825aware of any creation or destruction of references, and any transfers
826of ownership. Because ownership isn't represented explicitly in the xV
827data structures, only the reference count need be actually maintained
828by the code, and that means that this understanding of ownership is not
829actually evident in the code. For example, transferring ownership of a
830reference from one owner to another doesn't change the reference count
831at all, so may be achieved with no actual code. (The transferring code
832doesn't touch the referenced object, but does need to ensure that the
833former owner knows that it no longer owns the reference, and that the
834new owner knows that it now does.)
835
836An xV that is visible at the Perl level should not become unreferenced
837and thus be destroyed. Normally, an object will only become unreferenced
838when it is no longer visible, often by the same means that makes it
839invisible. For example, a Perl reference value (RV) owns a reference to
840its referent, so if the RV is overwritten that reference gets destroyed,
841and the no-longer-reachable referent may be destroyed as a result.
842
843Many functions have some kind of reference manipulation as
844part of their purpose. Sometimes this is documented in terms
845of ownership of references, and sometimes it is (less helpfully)
846documented in terms of changes to reference counts. For example, the
847L<newRV_inc()|perlapi/newRV_inc> function is documented to create a new RV
848(with reference count 1) and increment the reference count of the referent
849that was supplied by the caller. This is best understood as creating
850a new reference to the referent, which is owned by the created RV,
851and returning to the caller ownership of the sole reference to the RV.
852The L<newRV_noinc()|perlapi/newRV_noinc> function instead does not
853increment the reference count of the referent, but the RV nevertheless
854ends up owning a reference to the referent. It is therefore implied
855that the caller of C<newRV_noinc()> is relinquishing a reference to the
856referent, making this conceptually a more complicated operation even
857though it does less to the data structures.
858
859For example, imagine you want to return a reference from an XSUB
860function. Inside the XSUB routine, you create an SV which initially
861has just a single reference, owned by the XSUB routine. This reference
862needs to be disposed of before the routine is complete, otherwise it
863will leak, preventing the SV from ever being destroyed. So to create
864an RV referencing the SV, it is most convenient to pass the SV to
865C<newRV_noinc()>, which consumes that reference. Now the XSUB routine
866no longer owns a reference to the SV, but does own a reference to the RV,
867which in turn owns a reference to the SV. The ownership of the reference
868to the RV is then transferred by the process of returning the RV from
869the XSUB.
55497cff 870
5f05dabc 871There are some convenience functions available that can help with the
54310121 872destruction of xVs. These functions introduce the concept of "mortality".
3d2ba989
Z
873Much documentation speaks of an xV itself being mortal, but this is
874misleading. It is really I<a reference to> an xV that is mortal, and it
875is possible for there to be more than one mortal reference to a single xV.
876For a reference to be mortal means that it is owned by the temps stack,
877one of perl's many internal stacks, which will destroy that reference
878"a short time later". Usually the "short time later" is the end of
879the current Perl statement. However, it gets more complicated around
880dynamic scopes: there can be multiple sets of mortal references hanging
881around at the same time, with different death dates. Internally, the
882actual determinant for when mortal xV references are destroyed depends
883on two macros, SAVETMPS and FREETMPS. See L<perlcall> and L<perlxs>
884for more details on these macros.
885
886Mortal references are mainly used for xVs that are placed on perl's
887main stack. The stack is problematic for reference tracking, because it
888contains a lot of xV references, but doesn't own those references: they
889are not counted. Currently, there are many bugs resulting from xVs being
890destroyed while referenced by the stack, because the stack's uncounted
891references aren't enough to keep the xVs alive. So when putting an
892(uncounted) reference on the stack, it is vitally important to ensure that
893there will be a counted reference to the same xV that will last at least
894as long as the uncounted reference. But it's also important that that
895counted reference be cleaned up at an appropriate time, and not unduly
896prolong the xV's life. For there to be a mortal reference is often the
897best way to satisfy this requirement, especially if the xV was created
898especially to be put on the stack and would otherwise be unreferenced.
899
900To create a mortal reference, use the functions:
a0d0e21e
LW
901
902 SV* sv_newmortal()
a0d0e21e 903 SV* sv_mortalcopy(SV*)
3d2ba989 904 SV* sv_2mortal(SV*)
a0d0e21e 905
3d2ba989
Z
906C<sv_newmortal()> creates an SV (with the undefined value) whose sole
907reference is mortal. C<sv_mortalcopy()> creates an xV whose value is a
908copy of a supplied xV and whose sole reference is mortal. C<sv_2mortal()>
909mortalises an existing xV reference: it transfers ownership of a reference
910from the caller to the temps stack. Because C<sv_newmortal> gives the new
911SV no value, it must normally be given one via C<sv_setpv>, C<sv_setiv>,
912etc. :
00aadd71
NIS
913
914 SV *tmp = sv_newmortal();
915 sv_setiv(tmp, an_integer);
916
917As that is multiple C statements it is quite common so see this idiom instead:
918
919 SV *tmp = sv_2mortal(newSViv(an_integer));
920
ac036724 921The mortal routines are not just for SVs; AVs and HVs can be
faed5253 922made mortal by passing their address (type-casted to C<SV*>) to the
07fa94a1 923C<sv_2mortal> or C<sv_mortalcopy> routines.
a0d0e21e 924
5f05dabc 925=head2 Stashes and Globs
a0d0e21e 926
06f6df17
RGS
927A B<stash> is a hash that contains all variables that are defined
928within a package. Each key of the stash is a symbol
aa689395
PP
929name (shared by all the different types of objects that have the same
930name), and each value in the hash table is a GV (Glob Value). This GV
931in turn contains references to the various objects of that name,
932including (but not limited to) the following:
cb1a09d0 933
a0d0e21e
LW
934 Scalar Value
935 Array Value
936 Hash Value
a3cb178b 937 I/O Handle
a0d0e21e
LW
938 Format
939 Subroutine
940
06f6df17
RGS
941There is a single stash called C<PL_defstash> that holds the items that exist
942in the C<main> package. To get at the items in other packages, append the
943string "::" to the package name. The items in the C<Foo> package are in
944the stash C<Foo::> in PL_defstash. The items in the C<Bar::Baz> package are
945in the stash C<Baz::> in C<Bar::>'s stash.
a0d0e21e 946
d1b91892 947To get the stash pointer for a particular package, use the function:
a0d0e21e 948
da51bb9b
NC
949 HV* gv_stashpv(const char* name, I32 flags)
950 HV* gv_stashsv(SV*, I32 flags)
a0d0e21e
LW
951
952The first function takes a literal string, the second uses the string stored
d1b91892 953in the SV. Remember that a stash is just a hash table, so you get back an
da51bb9b 954C<HV*>. The C<flags> flag will create a new package if it is set to GV_ADD.
a0d0e21e
LW
955
956The name that C<gv_stash*v> wants is the name of the package whose symbol table
957you want. The default package is called C<main>. If you have multiply nested
d1b91892
AD
958packages, pass their names to C<gv_stash*v>, separated by C<::> as in the Perl
959language itself.
a0d0e21e
LW
960
961Alternately, if you have an SV that is a blessed reference, you can find
962out the stash pointer by using:
963
964 HV* SvSTASH(SvRV(SV*));
965
966then use the following to get the package name itself:
967
968 char* HvNAME(HV* stash);
969
5f05dabc
PP
970If you need to bless or re-bless an object you can use the following
971function:
a0d0e21e
LW
972
973 SV* sv_bless(SV*, HV* stash)
974
975where the first argument, an C<SV*>, must be a reference, and the second
976argument is a stash. The returned C<SV*> can now be used in the same way
977as any other SV.
978
d1b91892
AD
979For more information on references and blessings, consult L<perlref>.
980
54310121 981=head2 Double-Typed SVs
0a753a76
PP
982
983Scalar variables normally contain only one type of value, an integer,
984double, pointer, or reference. Perl will automatically convert the
985actual scalar data from the stored type into the requested type.
986
987Some scalar variables contain more than one type of scalar data. For
988example, the variable C<$!> contains either the numeric value of C<errno>
989or its string equivalent from either C<strerror> or C<sys_errlist[]>.
990
991To force multiple data values into an SV, you must do two things: use the
992C<sv_set*v> routines to add the additional scalar type, then set a flag
993so that Perl will believe it contains more than one type of data. The
994four macros to set the flags are:
995
996 SvIOK_on
997 SvNOK_on
998 SvPOK_on
999 SvROK_on
1000
1001The particular macro you must use depends on which C<sv_set*v> routine
1002you called first. This is because every C<sv_set*v> routine turns on
1003only the bit for the particular type of data being set, and turns off
1004all the rest.
1005
1006For example, to create a new Perl variable called "dberror" that contains
1007both the numeric and descriptive string error values, you could use the
1008following code:
1009
1010 extern int dberror;
1011 extern char *dberror_list;
1012
64ace3f8 1013 SV* sv = get_sv("dberror", GV_ADD);
0a753a76
PP
1014 sv_setiv(sv, (IV) dberror);
1015 sv_setpv(sv, dberror_list[dberror]);
1016 SvIOK_on(sv);
1017
1018If the order of C<sv_setiv> and C<sv_setpv> had been reversed, then the
1019macro C<SvPOK_on> would need to be called instead of C<SvIOK_on>.
1020
4f4531b8
FC
1021=head2 Read-Only Values
1022
1023In Perl 5.16 and earlier, copy-on-write (see the next section) shared a
1024flag bit with read-only scalars. So the only way to test whether
1025C<sv_setsv>, etc., will raise a "Modification of a read-only value" error
1026in those versions is:
1027
1028 SvREADONLY(sv) && !SvIsCOW(sv)
1029
1030Under Perl 5.18 and later, SvREADONLY only applies to read-only variables,
1031and, under 5.20, copy-on-write scalars can also be read-only, so the above
1032check is incorrect. You just want:
1033
1034 SvREADONLY(sv)
1035
1036If you need to do this check often, define your own macro like this:
1037
1038 #if PERL_VERSION >= 18
1039 # define SvTRULYREADONLY(sv) SvREADONLY(sv)
1040 #else
1041 # define SvTRULYREADONLY(sv) (SvREADONLY(sv) && !SvIsCOW(sv))
1042 #endif
1043
1044=head2 Copy on Write
1045
1046Perl implements a copy-on-write (COW) mechanism for scalars, in which
1047string copies are not immediately made when requested, but are deferred
1048until made necessary by one or the other scalar changing. This is mostly
1049transparent, but one must take care not to modify string buffers that are
1050shared by multiple SVs.
1051
1052You can test whether an SV is using copy-on-write with C<SvIsCOW(sv)>.
1053
1054You 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).
1055
1056If you want to make the SV drop its string buffer, use
1057C<sv_force_normal_flags(sv, SV_COW_DROP_PV)> or simply
1058C<sv_setsv(sv, NULL)>.
1059
1060All of these functions will croak on read-only scalars (see the previous
1061section for more on those).
1062
1063To test that your code is behaving correctly and not modifying COW buffers,
1064on systems that support L<mmap(2)> (i.e., Unix) you can configure perl with
1065C<-Accflags=-DPERL_DEBUG_READONLY_COW> and it will turn buffer violations
1066into crashes. You will find it to be marvellously slow, so you may want to
1067skip perl's own tests.
1068
0a753a76 1069=head2 Magic Variables
a0d0e21e 1070
d1b91892
AD
1071[This section still under construction. Ignore everything here. Post no
1072bills. Everything not permitted is forbidden.]
1073
d1b91892
AD
1074Any SV may be magical, that is, it has special features that a normal
1075SV does not have. These features are stored in the SV structure in a
5f05dabc 1076linked list of C<struct magic>'s, typedef'ed to C<MAGIC>.
d1b91892
AD
1077
1078 struct magic {
1079 MAGIC* mg_moremagic;
1080 MGVTBL* mg_virtual;
1081 U16 mg_private;
1082 char mg_type;
1083 U8 mg_flags;
b205eb13 1084 I32 mg_len;
d1b91892
AD
1085 SV* mg_obj;
1086 char* mg_ptr;
d1b91892
AD
1087 };
1088
1089Note this is current as of patchlevel 0, and could change at any time.
1090
1091=head2 Assigning Magic
1092
1093Perl adds magic to an SV using the sv_magic function:
1094
a9b0660e 1095 void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
d1b91892
AD
1096
1097The C<sv> argument is a pointer to the SV that is to acquire a new magical
1098feature.
1099
1100If C<sv> is not already magical, Perl uses the C<SvUPGRADE> macro to
10e2eb10
FC
1101convert C<sv> to type C<SVt_PVMG>.
1102Perl then continues by adding new magic
645c22ef
DM
1103to the beginning of the linked list of magical features. Any prior entry
1104of the same type of magic is deleted. Note that this can be overridden,
1105and multiple instances of the same type of magic can be associated with an
1106SV.
d1b91892 1107
54310121 1108The C<name> and C<namlen> arguments are used to associate a string with
10e2eb10 1109the magic, typically the name of a variable. C<namlen> is stored in the
2d8d5d5a
SH
1110C<mg_len> field and if C<name> is non-null then either a C<savepvn> copy of
1111C<name> or C<name> itself is stored in the C<mg_ptr> field, depending on
1112whether C<namlen> is greater than zero or equal to zero respectively. As a
1113special case, if C<(name && namlen == HEf_SVKEY)> then C<name> is assumed
1114to contain an C<SV*> and is stored as-is with its REFCNT incremented.
d1b91892
AD
1115
1116The sv_magic function uses C<how> to determine which, if any, predefined
1117"Magic Virtual Table" should be assigned to the C<mg_virtual> field.
5a0de581 1118See the L</Magic Virtual Tables> section below. The C<how> argument is also
10e2eb10
FC
1119stored in the C<mg_type> field. The value of
1120C<how> should be chosen from the set of macros
1121C<PERL_MAGIC_foo> found in F<perl.h>. Note that before
645c22ef 1122these macros were added, Perl internals used to directly use character
14befaf4 1123literals, so you may occasionally come across old code or documentation
75d0f26d 1124referring to 'U' magic rather than C<PERL_MAGIC_uvar> for example.
d1b91892
AD
1125
1126The C<obj> argument is stored in the C<mg_obj> field of the C<MAGIC>
1127structure. If it is not the same as the C<sv> argument, the reference
1128count of the C<obj> object is incremented. If it is the same, or if
27deb0cf
YO
1129the C<how> argument is C<PERL_MAGIC_arylen>, C<PERL_MAGIC_regdatum>,
1130C<PERL_MAGIC_regdata>, or if it is a NULL pointer, then C<obj> is merely
1131stored, without the reference count being incremented.
d1b91892 1132
2d8d5d5a
SH
1133See also C<sv_magicext> in L<perlapi> for a more flexible way to add magic
1134to an SV.
1135
cb1a09d0
AD
1136There is also a function to add magic to an C<HV>:
1137
1138 void hv_magic(HV *hv, GV *gv, int how);
1139
1140This simply calls C<sv_magic> and coerces the C<gv> argument into an C<SV>.
1141
1142To remove the magic from an SV, call the function sv_unmagic:
1143
70a53b35 1144 int sv_unmagic(SV *sv, int type);
cb1a09d0
AD
1145
1146The C<type> argument should be equal to the C<how> value when the C<SV>
1147was initially made magical.
1148
f6ee7b17 1149However, note that C<sv_unmagic> removes all magic of a certain C<type> from the
10e2eb10
FC
1150C<SV>. If you want to remove only certain
1151magic of a C<type> based on the magic
f6ee7b17
FR
1152virtual table, use C<sv_unmagicext> instead:
1153
1154 int sv_unmagicext(SV *sv, int type, MGVTBL *vtbl);
1155
d1b91892
AD
1156=head2 Magic Virtual Tables
1157
d1be9408 1158The C<mg_virtual> field in the C<MAGIC> structure is a pointer to an
d1b91892
AD
1159C<MGVTBL>, which is a structure of function pointers and stands for
1160"Magic Virtual Table" to handle the various operations that might be
1161applied to that variable.
1162
301cb7e8
DM
1163The C<MGVTBL> has five (or sometimes eight) pointers to the following
1164routine types:
d1b91892 1165
e97ca230
DM
1166 int (*svt_get) (pTHX_ SV* sv, MAGIC* mg);
1167 int (*svt_set) (pTHX_ SV* sv, MAGIC* mg);
1168 U32 (*svt_len) (pTHX_ SV* sv, MAGIC* mg);
1169 int (*svt_clear)(pTHX_ SV* sv, MAGIC* mg);
1170 int (*svt_free) (pTHX_ SV* sv, MAGIC* mg);
d1b91892 1171
e97ca230 1172 int (*svt_copy) (pTHX_ SV *sv, MAGIC* mg, SV *nsv,
a9b0660e 1173 const char *name, I32 namlen);
e97ca230
DM
1174 int (*svt_dup) (pTHX_ MAGIC *mg, CLONE_PARAMS *param);
1175 int (*svt_local)(pTHX_ SV *nsv, MAGIC *mg);
301cb7e8
DM
1176
1177
06f6df17 1178This MGVTBL structure is set at compile-time in F<perl.h> and there are
b7a0f54c
SM
1179currently 32 types. These different structures contain pointers to various
1180routines that perform additional actions depending on which function is
1181being called.
d1b91892 1182
a9b0660e
KW
1183 Function pointer Action taken
1184 ---------------- ------------
1185 svt_get Do something before the value of the SV is
1186 retrieved.
1187 svt_set Do something after the SV is assigned a value.
1188 svt_len Report on the SV's length.
1189 svt_clear Clear something the SV represents.
1190 svt_free Free any extra storage associated with the SV.
d1b91892 1191
a9b0660e
KW
1192 svt_copy copy tied variable magic to a tied element
1193 svt_dup duplicate a magic structure during thread cloning
1194 svt_local copy magic to local value during 'local'
301cb7e8 1195
d1b91892 1196For instance, the MGVTBL structure called C<vtbl_sv> (which corresponds
14befaf4 1197to an C<mg_type> of C<PERL_MAGIC_sv>) contains:
d1b91892
AD
1198
1199 { magic_get, magic_set, magic_len, 0, 0 }
1200
14befaf4
DM
1201Thus, when an SV is determined to be magical and of type C<PERL_MAGIC_sv>,
1202if a get operation is being performed, the routine C<magic_get> is
1203called. All the various routines for the various magical types begin
1204with C<magic_>. NOTE: the magic routines are not considered part of
1205the Perl API, and may not be exported by the Perl library.
d1b91892 1206
301cb7e8
DM
1207The last three slots are a recent addition, and for source code
1208compatibility they are only checked for if one of the three flags
10e2eb10
FC
1209MGf_COPY, MGf_DUP or MGf_LOCAL is set in mg_flags.
1210This means that most code can continue declaring
1211a vtable as a 5-element value. These three are
301cb7e8
DM
1212currently used exclusively by the threading code, and are highly subject
1213to change.
1214
d1b91892
AD
1215The current kinds of Magic Virtual Tables are:
1216
f1f5ddd7
FC
1217=for comment
1218This table is generated by regen/mg_vtable.pl. Any changes made here
1219will be lost.
1220
1221=for mg_vtable.pl begin
1222
a9b0660e 1223 mg_type
bd6e6c12
FC
1224 (old-style char and macro) MGVTBL Type of magic
1225 -------------------------- ------ -------------
1226 \0 PERL_MAGIC_sv vtbl_sv Special scalar variable
1227 # PERL_MAGIC_arylen vtbl_arylen Array length ($#ary)
e5e1ee61 1228 % PERL_MAGIC_rhash (none) Extra data for restricted
bd6e6c12 1229 hashes
a6d69523
TC
1230 * PERL_MAGIC_debugvar vtbl_debugvar $DB::single, signal, trace
1231 vars
bd6e6c12 1232 . PERL_MAGIC_pos vtbl_pos pos() lvalue
e5e1ee61 1233 : PERL_MAGIC_symtab (none) Extra data for symbol
bd6e6c12 1234 tables
e5e1ee61
FC
1235 < PERL_MAGIC_backref vtbl_backref For weak ref data
1236 @ PERL_MAGIC_arylen_p (none) To move arylen out of XPVAV
bd6e6c12
FC
1237 B PERL_MAGIC_bm vtbl_regexp Boyer-Moore
1238 (fast string search)
1239 c PERL_MAGIC_overload_table vtbl_ovrld Holds overload table
1240 (AMT) on stash
1241 D PERL_MAGIC_regdata vtbl_regdata Regex match position data
1242 (@+ and @- vars)
1243 d PERL_MAGIC_regdatum vtbl_regdatum Regex match position data
1244 element
1245 E PERL_MAGIC_env vtbl_env %ENV hash
1246 e PERL_MAGIC_envelem vtbl_envelem %ENV hash element
eccba044 1247 f PERL_MAGIC_fm vtbl_regexp Formline
bd6e6c12 1248 ('compiled' format)
bd6e6c12
FC
1249 g PERL_MAGIC_regex_global vtbl_mglob m//g target
1250 H PERL_MAGIC_hints vtbl_hints %^H hash
1251 h PERL_MAGIC_hintselem vtbl_hintselem %^H hash element
1252 I PERL_MAGIC_isa vtbl_isa @ISA array
1253 i PERL_MAGIC_isaelem vtbl_isaelem @ISA array element
1254 k PERL_MAGIC_nkeys vtbl_nkeys scalar(keys()) lvalue
1255 L PERL_MAGIC_dbfile (none) Debugger %_<filename
1256 l PERL_MAGIC_dbline vtbl_dbline Debugger %_<filename
1257 element
1258 N PERL_MAGIC_shared (none) Shared between threads
1259 n PERL_MAGIC_shared_scalar (none) Shared between threads
1260 o PERL_MAGIC_collxfrm vtbl_collxfrm Locale transformation
1261 P PERL_MAGIC_tied vtbl_pack Tied array or hash
1262 p PERL_MAGIC_tiedelem vtbl_packelem Tied array or hash element
1263 q PERL_MAGIC_tiedscalar vtbl_packelem Tied scalar or handle
e5e1ee61 1264 r PERL_MAGIC_qr vtbl_regexp Precompiled qr// regex
bd6e6c12
FC
1265 S PERL_MAGIC_sig (none) %SIG hash
1266 s PERL_MAGIC_sigelem vtbl_sigelem %SIG hash element
1267 t PERL_MAGIC_taint vtbl_taint Taintedness
1268 U PERL_MAGIC_uvar vtbl_uvar Available for use by
1269 extensions
1270 u PERL_MAGIC_uvar_elem (none) Reserved for use by
1271 extensions
4499db73 1272 V PERL_MAGIC_vstring (none) SV was vstring literal
bd6e6c12
FC
1273 v PERL_MAGIC_vec vtbl_vec vec() lvalue
1274 w PERL_MAGIC_utf8 vtbl_utf8 Cached UTF-8 information
1275 x PERL_MAGIC_substr vtbl_substr substr() lvalue
1f1dcfb5
FC
1276 Y PERL_MAGIC_nonelem vtbl_nonelem Array element that does not
1277 exist
bd6e6c12
FC
1278 y PERL_MAGIC_defelem vtbl_defelem Shadow "foreach" iterator
1279 variable / smart parameter
1280 vivification
baabe3fb
FC
1281 \ PERL_MAGIC_lvref vtbl_lvref Lvalue reference
1282 constructor
e5e1ee61 1283 ] PERL_MAGIC_checkcall vtbl_checkcall Inlining/mutation of call
bd6e6c12
FC
1284 to this CV
1285 ~ PERL_MAGIC_ext (none) Available for use by
1286 extensions
0cbee0a4 1287
f2f5335a
KW
1288
1289=for apidoc Amnh||PERL_MAGIC_sv
1290=for apidoc Amnh||PERL_MAGIC_arylen
1291=for apidoc Amnh||PERL_MAGIC_rhash
1292=for apidoc Amnh||PERL_MAGIC_debugvar
1293=for apidoc Amnh||PERL_MAGIC_pos
1294=for apidoc Amnh||PERL_MAGIC_symtab
1295=for apidoc Amnh||PERL_MAGIC_backref
1296=for apidoc Amnh||PERL_MAGIC_arylen_p
1297=for apidoc Amnh||PERL_MAGIC_bm
1298=for apidoc Amnh||PERL_MAGIC_overload_table
1299=for apidoc Amnh||PERL_MAGIC_regdata
1300=for apidoc Amnh||PERL_MAGIC_regdatum
1301=for apidoc Amnh||PERL_MAGIC_env
1302=for apidoc Amnh||PERL_MAGIC_envelem
1303=for apidoc Amnh||PERL_MAGIC_fm
1304=for apidoc Amnh||PERL_MAGIC_regex_global
1305=for apidoc Amnh||PERL_MAGIC_hints
1306=for apidoc Amnh||PERL_MAGIC_hintselem
1307=for apidoc Amnh||PERL_MAGIC_isa
1308=for apidoc Amnh||PERL_MAGIC_isaelem
1309=for apidoc Amnh||PERL_MAGIC_nkeys
1310=for apidoc Amnh||PERL_MAGIC_dbfile
1311=for apidoc Amnh||PERL_MAGIC_dbline
1312=for apidoc Amnh||PERL_MAGIC_shared
1313=for apidoc Amnh||PERL_MAGIC_shared_scalar
1314=for apidoc Amnh||PERL_MAGIC_collxfrm
1315=for apidoc Amnh||PERL_MAGIC_tied
1316=for apidoc Amnh||PERL_MAGIC_tiedelem
1317=for apidoc Amnh||PERL_MAGIC_tiedscalar
1318=for apidoc Amnh||PERL_MAGIC_qr
1319=for apidoc Amnh||PERL_MAGIC_sig
1320=for apidoc Amnh||PERL_MAGIC_sigelem
1321=for apidoc Amnh||PERL_MAGIC_taint
1322=for apidoc Amnh||PERL_MAGIC_uvar
1323=for apidoc Amnh||PERL_MAGIC_uvar_elem
1324=for apidoc Amnh||PERL_MAGIC_vstring
1325=for apidoc Amnh||PERL_MAGIC_vec
1326=for apidoc Amnh||PERL_MAGIC_utf8
1327=for apidoc Amnh||PERL_MAGIC_substr
1328=for apidoc Amnh||PERL_MAGIC_nonelem
1329=for apidoc Amnh||PERL_MAGIC_defelem
1330=for apidoc Amnh||PERL_MAGIC_lvref
1331=for apidoc Amnh||PERL_MAGIC_checkcall
1332=for apidoc Amnh||PERL_MAGIC_ext
1333
f1f5ddd7 1334=for mg_vtable.pl end
d1b91892 1335
68dc0745 1336When an uppercase and lowercase letter both exist in the table, then the
92f0c265
JP
1337uppercase letter is typically used to represent some kind of composite type
1338(a list or a hash), and the lowercase letter is used to represent an element
10e2eb10 1339of that composite type. Some internals code makes use of this case
92f0c265 1340relationship. However, 'v' and 'V' (vec and v-string) are in no way related.
14befaf4
DM
1341
1342The C<PERL_MAGIC_ext> and C<PERL_MAGIC_uvar> magic types are defined
1343specifically for use by extensions and will not be used by perl itself.
1344Extensions can use C<PERL_MAGIC_ext> magic to 'attach' private information
1345to variables (typically objects). This is especially useful because
1346there is no way for normal perl code to corrupt this private information
1347(unlike using extra elements of a hash object).
1348
1349Similarly, C<PERL_MAGIC_uvar> magic can be used much like tie() to call a
1350C function any time a scalar's value is used or changed. The C<MAGIC>'s
bdbeb323
SM
1351C<mg_ptr> field points to a C<ufuncs> structure:
1352
1353 struct ufuncs {
a9402793
AB
1354 I32 (*uf_val)(pTHX_ IV, SV*);
1355 I32 (*uf_set)(pTHX_ IV, SV*);
bdbeb323
SM
1356 IV uf_index;
1357 };
1358
1359When the SV is read from or written to, the C<uf_val> or C<uf_set>
14befaf4
DM
1360function will be called with C<uf_index> as the first arg and a pointer to
1361the SV as the second. A simple example of how to add C<PERL_MAGIC_uvar>
1526ead6
AB
1362magic is shown below. Note that the ufuncs structure is copied by
1363sv_magic, so you can safely allocate it on the stack.
1364
1365 void
1366 Umagic(sv)
1367 SV *sv;
1368 PREINIT:
1369 struct ufuncs uf;
1370 CODE:
1371 uf.uf_val = &my_get_fn;
1372 uf.uf_set = &my_set_fn;
1373 uf.uf_index = 0;
14befaf4 1374 sv_magic(sv, 0, PERL_MAGIC_uvar, (char*)&uf, sizeof(uf));
5f05dabc 1375
1e73acc8
AS
1376Attaching C<PERL_MAGIC_uvar> to arrays is permissible but has no effect.
1377
1378For hashes there is a specialized hook that gives control over hash
1379keys (but not values). This hook calls C<PERL_MAGIC_uvar> 'get' magic
1380if the "set" function in the C<ufuncs> structure is NULL. The hook
1381is activated whenever the hash is accessed with a key specified as
1382an C<SV> through the functions C<hv_store_ent>, C<hv_fetch_ent>,
1383C<hv_delete_ent>, and C<hv_exists_ent>. Accessing the key as a string
1384through the functions without the C<..._ent> suffix circumvents the
4509d391 1385hook. See L<Hash::Util::FieldHash/GUTS> for a detailed description.
1e73acc8 1386
14befaf4
DM
1387Note that because multiple extensions may be using C<PERL_MAGIC_ext>
1388or C<PERL_MAGIC_uvar> magic, it is important for extensions to take
1389extra care to avoid conflict. Typically only using the magic on
1390objects blessed into the same class as the extension is sufficient.
2f07f21a
FR
1391For C<PERL_MAGIC_ext> magic, it is usually a good idea to define an
1392C<MGVTBL>, even if all its fields will be C<0>, so that individual
1393C<MAGIC> pointers can be identified as a particular kind of magic
10e2eb10 1394using their magic virtual table. C<mg_findext> provides an easy way
f6ee7b17 1395to do that:
2f07f21a
FR
1396
1397 STATIC MGVTBL my_vtbl = { 0, 0, 0, 0, 0, 0, 0, 0 };
1398
1399 MAGIC *mg;
f6ee7b17
FR
1400 if ((mg = mg_findext(sv, PERL_MAGIC_ext, &my_vtbl))) {
1401 /* this is really ours, not another module's PERL_MAGIC_ext */
1402 my_priv_data_t *priv = (my_priv_data_t *)mg->mg_ptr;
1403 ...
2f07f21a 1404 }
5f05dabc 1405
ef50df4b
GS
1406Also note that the C<sv_set*()> and C<sv_cat*()> functions described
1407earlier do B<not> invoke 'set' magic on their targets. This must
1408be done by the user either by calling the C<SvSETMAGIC()> macro after
1409calling these functions, or by using one of the C<sv_set*_mg()> or
1410C<sv_cat*_mg()> functions. Similarly, generic C code must call the
1411C<SvGETMAGIC()> macro to invoke any 'get' magic if they use an SV
1412obtained from external sources in functions that don't handle magic.
4a4eefd0 1413See L<perlapi> for a description of these functions.
189b2af5
GS
1414For example, calls to the C<sv_cat*()> functions typically need to be
1415followed by C<SvSETMAGIC()>, but they don't need a prior C<SvGETMAGIC()>
1416since their implementation handles 'get' magic.
1417
d1b91892
AD
1418=head2 Finding Magic
1419
a9b0660e
KW
1420 MAGIC *mg_find(SV *sv, int type); /* Finds the magic pointer of that
1421 * type */
f6ee7b17
FR
1422
1423This routine returns a pointer to a C<MAGIC> structure stored in the SV.
10e2eb10
FC
1424If the SV does not have that magical
1425feature, C<NULL> is returned. If the
f6ee7b17 1426SV has multiple instances of that magical feature, the first one will be
10e2eb10
FC
1427returned. C<mg_findext> can be used
1428to find a C<MAGIC> structure of an SV
da8c5729 1429based on both its magic type and its magic virtual table:
f6ee7b17
FR
1430
1431 MAGIC *mg_findext(SV *sv, int type, MGVTBL *vtbl);
d1b91892 1432
f6ee7b17
FR
1433Also, if the SV passed to C<mg_find> or C<mg_findext> is not of type
1434SVt_PVMG, Perl may core dump.
d1b91892 1435
08105a92 1436 int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
d1b91892
AD
1437
1438This routine checks to see what types of magic C<sv> has. If the mg_type
68dc0745
PP
1439field is an uppercase letter, then the mg_obj is copied to C<nsv>, but
1440the mg_type field is changed to be the lowercase letter.
a0d0e21e 1441
04343c6d
GS
1442=head2 Understanding the Magic of Tied Hashes and Arrays
1443
14befaf4
DM
1444Tied hashes and arrays are magical beasts of the C<PERL_MAGIC_tied>
1445magic type.
9edb2b46
GS
1446
1447WARNING: As of the 5.004 release, proper usage of the array and hash
1448access functions requires understanding a few caveats. Some
1449of these caveats are actually considered bugs in the API, to be fixed
10e2eb10 1450in later releases, and are bracketed with [MAYCHANGE] below. If
9edb2b46
GS
1451you find yourself actually applying such information in this section, be
1452aware that the behavior may change in the future, umm, without warning.
04343c6d 1453
1526ead6 1454The perl tie function associates a variable with an object that implements
9a68f1db 1455the various GET, SET, etc methods. To perform the equivalent of the perl
1526ead6 1456tie function from an XSUB, you must mimic this behaviour. The code below
61ad4b94 1457carries out the necessary steps -- firstly it creates a new hash, and then
1526ead6 1458creates a second hash which it blesses into the class which will implement
10e2eb10 1459the tie methods. Lastly it ties the two hashes together, and returns a
1526ead6
AB
1460reference to the new tied hash. Note that the code below does NOT call the
1461TIEHASH method in the MyTie class -
5a0de581 1462see L</Calling Perl Routines from within C Programs> for details on how
1526ead6
AB
1463to do this.
1464
1465 SV*
1466 mytie()
1467 PREINIT:
1468 HV *hash;
1469 HV *stash;
1470 SV *tie;
1471 CODE:
1472 hash = newHV();
1473 tie = newRV_noinc((SV*)newHV());
da51bb9b 1474 stash = gv_stashpv("MyTie", GV_ADD);
1526ead6 1475 sv_bless(tie, stash);
899e16d0 1476 hv_magic(hash, (GV*)tie, PERL_MAGIC_tied);
1526ead6
AB
1477 RETVAL = newRV_noinc(hash);
1478 OUTPUT:
1479 RETVAL
1480
04343c6d
GS
1481The C<av_store> function, when given a tied array argument, merely
1482copies the magic of the array onto the value to be "stored", using
1483C<mg_copy>. It may also return NULL, indicating that the value did not
9edb2b46
GS
1484actually need to be stored in the array. [MAYCHANGE] After a call to
1485C<av_store> on a tied array, the caller will usually need to call
1486C<mg_set(val)> to actually invoke the perl level "STORE" method on the
1487TIEARRAY object. If C<av_store> did return NULL, a call to
1488C<SvREFCNT_dec(val)> will also be usually necessary to avoid a memory
1489leak. [/MAYCHANGE]
04343c6d
GS
1490
1491The previous paragraph is applicable verbatim to tied hash access using the
1492C<hv_store> and C<hv_store_ent> functions as well.
1493
1494C<av_fetch> and the corresponding hash functions C<hv_fetch> and
1495C<hv_fetch_ent> actually return an undefined mortal value whose magic
1496has been initialized using C<mg_copy>. Note the value so returned does not
9edb2b46
GS
1497need to be deallocated, as it is already mortal. [MAYCHANGE] But you will
1498need to call C<mg_get()> on the returned value in order to actually invoke
1499the perl level "FETCH" method on the underlying TIE object. Similarly,
04343c6d
GS
1500you may also call C<mg_set()> on the return value after possibly assigning
1501a suitable value to it using C<sv_setsv>, which will invoke the "STORE"
9edb2b46 1502method on the TIE object. [/MAYCHANGE]
04343c6d 1503
9edb2b46 1504[MAYCHANGE]
04343c6d
GS
1505In other words, the array or hash fetch/store functions don't really
1506fetch and store actual values in the case of tied arrays and hashes. They
1507merely call C<mg_copy> to attach magic to the values that were meant to be
1508"stored" or "fetched". Later calls to C<mg_get> and C<mg_set> actually
1509do the job of invoking the TIE methods on the underlying objects. Thus
9edb2b46 1510the magic mechanism currently implements a kind of lazy access to arrays
04343c6d
GS
1511and hashes.
1512
1513Currently (as of perl version 5.004), use of the hash and array access
1514functions requires the user to be aware of whether they are operating on
9edb2b46
GS
1515"normal" hashes and arrays, or on their tied variants. The API may be
1516changed to provide more transparent access to both tied and normal data
1517types in future versions.
1518[/MAYCHANGE]
04343c6d
GS
1519
1520You would do well to understand that the TIEARRAY and TIEHASH interfaces
1521are mere sugar to invoke some perl method calls while using the uniform hash
1522and array syntax. The use of this sugar imposes some overhead (typically
1523about two to four extra opcodes per FETCH/STORE operation, in addition to
1524the creation of all the mortal variables required to invoke the methods).
1525This overhead will be comparatively small if the TIE methods are themselves
1526substantial, but if they are only a few statements long, the overhead
1527will not be insignificant.
1528
d1c897a1
IZ
1529=head2 Localizing changes
1530
1531Perl has a very handy construction
1532
1533 {
1534 local $var = 2;
1535 ...
1536 }
1537
1538This construction is I<approximately> equivalent to
1539
1540 {
1541 my $oldvar = $var;
1542 $var = 2;
1543 ...
1544 $var = $oldvar;
1545 }
1546
1547The biggest difference is that the first construction would
1548reinstate the initial value of $var, irrespective of how control exits
10e2eb10 1549the block: C<goto>, C<return>, C<die>/C<eval>, etc. It is a little bit
d1c897a1
IZ
1550more efficient as well.
1551
1552There is a way to achieve a similar task from C via Perl API: create a
1553I<pseudo-block>, and arrange for some changes to be automatically
1554undone at the end of it, either explicit, or via a non-local exit (via
10e2eb10 1555die()). A I<block>-like construct is created by a pair of
b687b08b
TC
1556C<ENTER>/C<LEAVE> macros (see L<perlcall/"Returning a Scalar">).
1557Such a construct may be created specially for some important localized
1558task, or an existing one (like boundaries of enclosing Perl
1559subroutine/block, or an existing pair for freeing TMPs) may be
10e2eb10
FC
1560used. (In the second case the overhead of additional localization must
1561be almost negligible.) Note that any XSUB is automatically enclosed in
b687b08b 1562an C<ENTER>/C<LEAVE> pair.
d1c897a1
IZ
1563
1564Inside such a I<pseudo-block> the following service is available:
1565
13a2d996 1566=over 4
d1c897a1
IZ
1567
1568=item C<SAVEINT(int i)>
1569
1570=item C<SAVEIV(IV i)>
1571
1572=item C<SAVEI32(I32 i)>
1573
1574=item C<SAVELONG(long i)>
1575
1576These macros arrange things to restore the value of integer variable
1577C<i> at the end of enclosing I<pseudo-block>.
1578
1579=item C<SAVESPTR(s)>
1580
1581=item C<SAVEPPTR(p)>
1582
1583These macros arrange things to restore the value of pointers C<s> and
10e2eb10 1584C<p>. C<s> must be a pointer of a type which survives conversion to
d1c897a1
IZ
1585C<SV*> and back, C<p> should be able to survive conversion to C<char*>
1586and back.
1587
1588=item C<SAVEFREESV(SV *sv)>
1589
06f1e0b6 1590The refcount of C<sv> will be decremented at the end of
26d9b02f
JH
1591I<pseudo-block>. This is similar to C<sv_2mortal> in that it is also a
1592mechanism for doing a delayed C<SvREFCNT_dec>. However, while C<sv_2mortal>
1593extends the lifetime of C<sv> until the beginning of the next statement,
1594C<SAVEFREESV> extends it until the end of the enclosing scope. These
1595lifetimes can be wildly different.
1596
1597Also compare C<SAVEMORTALIZESV>.
1598
1599=item C<SAVEMORTALIZESV(SV *sv)>
1600
1601Just like C<SAVEFREESV>, but mortalizes C<sv> at the end of the current
1602scope instead of decrementing its reference count. This usually has the
1603effect of keeping C<sv> alive until the statement that called the currently
1604live scope has finished executing.
d1c897a1
IZ
1605
1606=item C<SAVEFREEOP(OP *op)>
1607
1608The C<OP *> is op_free()ed at the end of I<pseudo-block>.
1609
1610=item C<SAVEFREEPV(p)>
1611
1612The chunk of memory which is pointed to by C<p> is Safefree()ed at the
1613end of I<pseudo-block>.
1614
1615=item C<SAVECLEARSV(SV *sv)>
1616
1617Clears a slot in the current scratchpad which corresponds to C<sv> at
1618the end of I<pseudo-block>.
1619
1620=item C<SAVEDELETE(HV *hv, char *key, I32 length)>
1621
10e2eb10 1622The key C<key> of C<hv> is deleted at the end of I<pseudo-block>. The
d1c897a1
IZ
1623string pointed to by C<key> is Safefree()ed. If one has a I<key> in
1624short-lived storage, the corresponding string may be reallocated like
1625this:
1626
9cde0e7f 1627 SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
d1c897a1 1628
c76ac1ee 1629=item C<SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)>
d1c897a1
IZ
1630
1631At the end of I<pseudo-block> the function C<f> is called with the
c76ac1ee
GS
1632only argument C<p>.
1633
1634=item C<SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)>
1635
1636At the end of I<pseudo-block> the function C<f> is called with the
1637implicit context argument (if any), and C<p>.
d1c897a1
IZ
1638
1639=item C<SAVESTACK_POS()>
1640
1641The current offset on the Perl internal stack (cf. C<SP>) is restored
1642at the end of I<pseudo-block>.
1643
1644=back
1645
1646The following API list contains functions, thus one needs to
1647provide pointers to the modifiable data explicitly (either C pointers,
00aadd71 1648or Perlish C<GV *>s). Where the above macros take C<int>, a similar
d1c897a1
IZ
1649function takes C<int *>.
1650
13a2d996 1651=over 4
d1c897a1
IZ
1652
1653=item C<SV* save_scalar(GV *gv)>
1654
1655Equivalent to Perl code C<local $gv>.
1656
1657=item C<AV* save_ary(GV *gv)>
1658
1659=item C<HV* save_hash(GV *gv)>
1660
1661Similar to C<save_scalar>, but localize C<@gv> and C<%gv>.
1662
1663=item C<void save_item(SV *item)>
1664
53dedf6f
KW
1665=for apidoc save_item
1666
1667Duplicates the current value of C<SV>. On the exit from the current
1668C<ENTER>/C<LEAVE> I<pseudo-block> the value of C<SV> will be restored
10e2eb10 1669using the stored value. It doesn't handle magic. Use C<save_scalar> if
038fcae3 1670magic is affected.
d1c897a1
IZ
1671
1672=item C<void save_list(SV **sarg, I32 maxsarg)>
1673
1674A variant of C<save_item> which takes multiple arguments via an array
1675C<sarg> of C<SV*> of length C<maxsarg>.
1676
1677=item C<SV* save_svref(SV **sptr)>
1678
d1be9408 1679Similar to C<save_scalar>, but will reinstate an C<SV *>.
d1c897a1
IZ
1680
1681=item C<void save_aptr(AV **aptr)>
1682
1683=item C<void save_hptr(HV **hptr)>
1684
1685Similar to C<save_svref>, but localize C<AV *> and C<HV *>.
1686
1687=back
1688
1689The C<Alias> module implements localization of the basic types within the
1690I<caller's scope>. People who are interested in how to localize things in
1691the containing scope should take a look there too.
1692
0a753a76 1693=head1 Subroutines
a0d0e21e 1694
68dc0745 1695=head2 XSUBs and the Argument Stack
5f05dabc
PP
1696
1697The XSUB mechanism is a simple way for Perl programs to access C subroutines.
1698An XSUB routine will have a stack that contains the arguments from the Perl
1699program, and a way to map from the Perl data structures to a C equivalent.
1700
1701The stack arguments are accessible through the C<ST(n)> macro, which returns
1702the C<n>'th stack argument. Argument 0 is the first argument passed in the
1703Perl subroutine call. These arguments are C<SV*>, and can be used anywhere
1704an C<SV*> is used.
1705
1706Most of the time, output from the C routine can be handled through use of
1707the RETVAL and OUTPUT directives. However, there are some cases where the
1708argument stack is not already long enough to handle all the return values.
1709An example is the POSIX tzname() call, which takes no arguments, but returns
1710two, the local time zone's standard and summer time abbreviations.
1711
1712To handle this situation, the PPCODE directive is used and the stack is
1713extended using the macro:
1714
924508f0 1715 EXTEND(SP, num);
5f05dabc 1716
924508f0
GS
1717where C<SP> is the macro that represents the local copy of the stack pointer,
1718and C<num> is the number of elements the stack should be extended by.
5f05dabc 1719
00aadd71 1720Now that there is room on the stack, values can be pushed on it using C<PUSHs>
10e2eb10 1721macro. The pushed values will often need to be "mortal" (See
d82b684c 1722L</Reference Counts and Mortality>):
5f05dabc 1723
00aadd71 1724 PUSHs(sv_2mortal(newSViv(an_integer)))
d82b684c
SH
1725 PUSHs(sv_2mortal(newSVuv(an_unsigned_integer)))
1726 PUSHs(sv_2mortal(newSVnv(a_double)))
00aadd71 1727 PUSHs(sv_2mortal(newSVpv("Some String",0)))
a9b0660e
KW
1728 /* Although the last example is better written as the more
1729 * efficient: */
a3179684 1730 PUSHs(newSVpvs_flags("Some String", SVs_TEMP))
5f05dabc
PP
1731
1732And now the Perl program calling C<tzname>, the two values will be assigned
1733as in:
1734
1735 ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1736
1737An alternate (and possibly simpler) method to pushing values on the stack is
00aadd71 1738to use the macro:
5f05dabc 1739
5f05dabc
PP
1740 XPUSHs(SV*)
1741
da8c5729 1742This macro automatically adjusts the stack for you, if needed. Thus, you
5f05dabc 1743do not need to call C<EXTEND> to extend the stack.
00aadd71
NIS
1744
1745Despite their suggestions in earlier versions of this document the macros
d82b684c
SH
1746C<(X)PUSH[iunp]> are I<not> suited to XSUBs which return multiple results.
1747For that, either stick to the C<(X)PUSHs> macros shown above, or use the new
1748C<m(X)PUSH[iunp]> macros instead; see L</Putting a C value on Perl stack>.
5f05dabc
PP
1749
1750For more information, consult L<perlxs> and L<perlxstut>.
1751
5b36e945
FC
1752=head2 Autoloading with XSUBs
1753
1754If an AUTOLOAD routine is an XSUB, as with Perl subroutines, Perl puts the
1755fully-qualified name of the autoloaded subroutine in the $AUTOLOAD variable
1756of the XSUB's package.
1757
1758But it also puts the same information in certain fields of the XSUB itself:
1759
1760 HV *stash = CvSTASH(cv);
1761 const char *subname = SvPVX(cv);
1762 STRLEN name_length = SvCUR(cv); /* in bytes */
1763 U32 is_utf8 = SvUTF8(cv);
f703fc96 1764
5b36e945 1765C<SvPVX(cv)> contains just the sub name itself, not including the package.
d8893903
FC
1766For an AUTOLOAD routine in UNIVERSAL or one of its superclasses,
1767C<CvSTASH(cv)> returns NULL during a method call on a nonexistent package.
5b36e945
FC
1768
1769B<Note>: Setting $AUTOLOAD stopped working in 5.6.1, which did not support
1770XS AUTOLOAD subs at all. Perl 5.8.0 introduced the use of fields in the
1771XSUB itself. Perl 5.16.0 restored the setting of $AUTOLOAD. If you need
1772to support 5.8-5.14, use the XSUB's fields.
1773
5f05dabc 1774=head2 Calling Perl Routines from within C Programs
a0d0e21e
LW
1775
1776There are four routines that can be used to call a Perl subroutine from
1777within a C program. These four are:
1778
954c1994
GS
1779 I32 call_sv(SV*, I32);
1780 I32 call_pv(const char*, I32);
1781 I32 call_method(const char*, I32);
5aaab254 1782 I32 call_argv(const char*, I32, char**);
a0d0e21e 1783
954c1994 1784The routine most often used is C<call_sv>. The C<SV*> argument
d1b91892
AD
1785contains either the name of the Perl subroutine to be called, or a
1786reference to the subroutine. The second argument consists of flags
1787that control the context in which the subroutine is called, whether
1788or not the subroutine is being passed arguments, how errors should be
1789trapped, and how to treat return values.
a0d0e21e
LW
1790
1791All four routines return the number of arguments that the subroutine returned
1792on the Perl stack.
1793
9a68f1db 1794These routines used to be called C<perl_call_sv>, etc., before Perl v5.6.0,
954c1994
GS
1795but those names are now deprecated; macros of the same name are provided for
1796compatibility.
1797
1798When using any of these routines (except C<call_argv>), the programmer
d1b91892
AD
1799must manipulate the Perl stack. These include the following macros and
1800functions:
a0d0e21e
LW
1801
1802 dSP
924508f0 1803 SP
a0d0e21e
LW
1804 PUSHMARK()
1805 PUTBACK
1806 SPAGAIN
1807 ENTER
1808 SAVETMPS
1809 FREETMPS
1810 LEAVE
1811 XPUSH*()
cb1a09d0 1812 POP*()
a0d0e21e 1813
5f05dabc
PP
1814For a detailed description of calling conventions from C to Perl,
1815consult L<perlcall>.
a0d0e21e 1816
8ebc5c01 1817=head2 Putting a C value on Perl stack
ce3d39e2
IZ
1818
1819A lot of opcodes (this is an elementary operation in the internal perl
10e2eb10
FC
1820stack machine) put an SV* on the stack. However, as an optimization
1821the corresponding SV is (usually) not recreated each time. The opcodes
ce3d39e2
IZ
1822reuse specially assigned SVs (I<target>s) which are (as a corollary)
1823not constantly freed/created.
1824
0a753a76 1825Each of the targets is created only once (but see
5a0de581 1826L</Scratchpads and recursion> below), and when an opcode needs to put
ce3d39e2
IZ
1827an integer, a double, or a string on stack, it just sets the
1828corresponding parts of its I<target> and puts the I<target> on stack.
1829
1830The macro to put this target on stack is C<PUSHTARG>, and it is
1831directly used in some opcodes, as well as indirectly in zillions of
d82b684c 1832others, which use it via C<(X)PUSH[iunp]>.
ce3d39e2 1833
1bd1c0d5 1834Because the target is reused, you must be careful when pushing multiple
10e2eb10 1835values on the stack. The following code will not do what you think:
1bd1c0d5
SC
1836
1837 XPUSHi(10);
1838 XPUSHi(20);
1839
1840This translates as "set C<TARG> to 10, push a pointer to C<TARG> onto
1841the stack; set C<TARG> to 20, push a pointer to C<TARG> onto the stack".
1842At the end of the operation, the stack does not contain the values 10
1843and 20, but actually contains two pointers to C<TARG>, which we have set
d82b684c 1844to 20.
1bd1c0d5 1845
d82b684c
SH
1846If you need to push multiple different values then you should either use
1847the C<(X)PUSHs> macros, or else use the new C<m(X)PUSH[iunp]> macros,
1848none of which make use of C<TARG>. The C<(X)PUSHs> macros simply push an
1849SV* on the stack, which, as noted under L</XSUBs and the Argument Stack>,
1850will often need to be "mortal". The new C<m(X)PUSH[iunp]> macros make
1851this a little easier to achieve by creating a new mortal for you (via
1852C<(X)PUSHmortal>), pushing that onto the stack (extending it if necessary
1853in the case of the C<mXPUSH[iunp]> macros), and then setting its value.
1854Thus, instead of writing this to "fix" the example above:
1855
1856 XPUSHs(sv_2mortal(newSViv(10)))
1857 XPUSHs(sv_2mortal(newSViv(20)))
1858
1859you can simply write:
1860
1861 mXPUSHi(10)
1862 mXPUSHi(20)
1863
1864On a related note, if you do use C<(X)PUSH[iunp]>, then you're going to
1bd1c0d5 1865need a C<dTARG> in your variable declarations so that the C<*PUSH*>
d82b684c
SH
1866macros can make use of the local variable C<TARG>. See also C<dTARGET>
1867and C<dXSTARG>.
1bd1c0d5 1868
8ebc5c01 1869=head2 Scratchpads
ce3d39e2 1870
54310121 1871The question remains on when the SVs which are I<target>s for opcodes
10e2eb10 1872are created. The answer is that they are created when the current
ac036724 1873unit--a subroutine or a file (for opcodes for statements outside of
10e2eb10 1874subroutines)--is compiled. During this time a special anonymous Perl
ac036724 1875array is created, which is called a scratchpad for the current unit.
ce3d39e2 1876
54310121 1877A scratchpad keeps SVs which are lexicals for the current unit and are
d777b41a
FC
1878targets for opcodes. A previous version of this document
1879stated that one can deduce that an SV lives on a scratchpad
ce3d39e2 1880by looking on its flags: lexicals have C<SVs_PADMY> set, and
eee3e302 1881I<target>s have C<SVs_PADTMP> set. But this has never been fully true.
d777b41a
FC
1882C<SVs_PADMY> could be set on a variable that no longer resides in any pad.
1883While I<target>s do have C<SVs_PADTMP> set, it can also be set on variables
eed77337
FC
1884that have never resided in a pad, but nonetheless act like I<target>s. As
1885of perl 5.21.5, the C<SVs_PADMY> flag is no longer used and is defined as
18860. C<SvPADMY()> now returns true for anything without C<SVs_PADTMP>.
ce3d39e2 1887
10e2eb10 1888The correspondence between OPs and I<target>s is not 1-to-1. Different
54310121 1889OPs in the compile tree of the unit can use the same target, if this
ce3d39e2
IZ
1890would not conflict with the expected life of the temporary.
1891
2ae324a7 1892=head2 Scratchpads and recursion
ce3d39e2
IZ
1893
1894In fact it is not 100% true that a compiled unit contains a pointer to
10e2eb10
FC
1895the scratchpad AV. In fact it contains a pointer to an AV of
1896(initially) one element, and this element is the scratchpad AV. Why do
ce3d39e2
IZ
1897we need an extra level of indirection?
1898
10e2eb10 1899The answer is B<recursion>, and maybe B<threads>. Both
ce3d39e2 1900these can create several execution pointers going into the same
10e2eb10 1901subroutine. For the subroutine-child not write over the temporaries
ce3d39e2
IZ
1902for the subroutine-parent (lifespan of which covers the call to the
1903child), the parent and the child should have different
10e2eb10 1904scratchpads. (I<And> the lexicals should be separate anyway!)
ce3d39e2 1905
5f05dabc
PP
1906So each subroutine is born with an array of scratchpads (of length 1).
1907On each entry to the subroutine it is checked that the current
ce3d39e2
IZ
1908depth of the recursion is not more than the length of this array, and
1909if it is, new scratchpad is created and pushed into the array.
1910
1911The I<target>s on this scratchpad are C<undef>s, but they are already
1912marked with correct flags.
1913
22d36020
FC
1914=head1 Memory Allocation
1915
1916=head2 Allocation
1917
1918All memory meant to be used with the Perl API functions should be manipulated
1919using the macros described in this section. The macros provide the necessary
1920transparency between differences in the actual malloc implementation that is
1921used within perl.
1922
1923It is suggested that you enable the version of malloc that is distributed
1924with Perl. It keeps pools of various sizes of unallocated memory in
1925order to satisfy allocation requests more quickly. However, on some
1926platforms, it may cause spurious malloc or free errors.
1927
1928The following three macros are used to initially allocate memory :
1929
1930 Newx(pointer, number, type);
1931 Newxc(pointer, number, type, cast);
1932 Newxz(pointer, number, type);
1933
1934The first argument C<pointer> should be the name of a variable that will
1935point to the newly allocated memory.
1936
1937The second and third arguments C<number> and C<type> specify how many of
1938the specified type of data structure should be allocated. The argument
1939C<type> is passed to C<sizeof>. The final argument to C<Newxc>, C<cast>,
1940should be used if the C<pointer> argument is different from the C<type>
1941argument.
1942
1943Unlike the C<Newx> and C<Newxc> macros, the C<Newxz> macro calls C<memzero>
1944to zero out all the newly allocated memory.
1945
1946=head2 Reallocation
1947
1948 Renew(pointer, number, type);
1949 Renewc(pointer, number, type, cast);
1950 Safefree(pointer)
1951
1952These three macros are used to change a memory buffer size or to free a
1953piece of memory no longer needed. The arguments to C<Renew> and C<Renewc>
1954match those of C<New> and C<Newc> with the exception of not needing the
1955"magic cookie" argument.
1956
1957=head2 Moving
1958
1959 Move(source, dest, number, type);
1960 Copy(source, dest, number, type);
1961 Zero(dest, number, type);
1962
1963These three macros are used to move, copy, or zero out previously allocated
1964memory. The C<source> and C<dest> arguments point to the source and
1965destination starting points. Perl will move, copy, or zero out C<number>
1966instances of the size of the C<type> data structure (using the C<sizeof>
1967function).
1968
1969=head1 PerlIO
1970
1971The most recent development releases of Perl have been experimenting with
1972removing Perl's dependency on the "normal" standard I/O suite and allowing
1973other stdio implementations to be used. This involves creating a new
1974abstraction layer that then calls whichever implementation of stdio Perl
1975was compiled with. All XSUBs should now use the functions in the PerlIO
1976abstraction layer and not make any assumptions about what kind of stdio
1977is being used.
1978
1979For a complete description of the PerlIO abstraction, consult L<perlapio>.
1980
0a753a76
PP
1981=head1 Compiled code
1982
1983=head2 Code tree
1984
1985Here we describe the internal form your code is converted to by
10e2eb10 1986Perl. Start with a simple example:
0a753a76
PP
1987
1988 $a = $b + $c;
1989
1990This is converted to a tree similar to this one:
1991
1992 assign-to
1993 / \
1994 + $a
1995 / \
1996 $b $c
1997
7b8d334a 1998(but slightly more complicated). This tree reflects the way Perl
0a753a76
PP
1999parsed your code, but has nothing to do with the execution order.
2000There is an additional "thread" going through the nodes of the tree
2001which shows the order of execution of the nodes. In our simplified
2002example above it looks like:
2003
2004 $b ---> $c ---> + ---> $a ---> assign-to
2005
2006But with the actual compile tree for C<$a = $b + $c> it is different:
2007some nodes I<optimized away>. As a corollary, though the actual tree
2008contains more nodes than our simplified example, the execution order
2009is the same as in our example.
2010
2011=head2 Examining the tree
2012
06f6df17
RGS
2013If you have your perl compiled for debugging (usually done with
2014C<-DDEBUGGING> on the C<Configure> command line), you may examine the
0a753a76
PP
2015compiled tree by specifying C<-Dx> on the Perl command line. The
2016output takes several lines per node, and for C<$b+$c> it looks like
2017this:
2018
2019 5 TYPE = add ===> 6
2020 TARG = 1
2021 FLAGS = (SCALAR,KIDS)
2022 {
2023 TYPE = null ===> (4)
2024 (was rv2sv)
2025 FLAGS = (SCALAR,KIDS)
2026 {
2027 3 TYPE = gvsv ===> 4
2028 FLAGS = (SCALAR)
2029 GV = main::b
2030 }
2031 }
2032 {
2033 TYPE = null ===> (5)
2034 (was rv2sv)
2035 FLAGS = (SCALAR,KIDS)
2036 {
2037 4 TYPE = gvsv ===> 5
2038 FLAGS = (SCALAR)
2039 GV = main::c
2040 }
2041 }
2042
2043This tree has 5 nodes (one per C<TYPE> specifier), only 3 of them are
2044not optimized away (one per number in the left column). The immediate
2045children of the given node correspond to C<{}> pairs on the same level
2046of indentation, thus this listing corresponds to the tree:
2047
2048 add
2049 / \
2050 null null
2051 | |
2052 gvsv gvsv
2053
2054The execution order is indicated by C<===E<gt>> marks, thus it is C<3
20554 5 6> (node C<6> is not included into above listing), i.e.,
2056C<gvsv gvsv add whatever>.
2057
9afa14e3 2058Each of these nodes represents an op, a fundamental operation inside the
10e2eb10 2059Perl core. The code which implements each operation can be found in the
9afa14e3 2060F<pp*.c> files; the function which implements the op with type C<gvsv>
10e2eb10 2061is C<pp_gvsv>, and so on. As the tree above shows, different ops have
9afa14e3 2062different numbers of children: C<add> is a binary operator, as one would
10e2eb10 2063expect, and so has two children. To accommodate the various different
9afa14e3
SC
2064numbers of children, there are various types of op data structure, and
2065they link together in different ways.
2066
10e2eb10 2067The simplest type of op structure is C<OP>: this has no children. Unary
9afa14e3 2068operators, C<UNOP>s, have one child, and this is pointed to by the
10e2eb10
FC
2069C<op_first> field. Binary operators (C<BINOP>s) have not only an
2070C<op_first> field but also an C<op_last> field. The most complex type of
2071op is a C<LISTOP>, which has any number of children. In this case, the
9afa14e3 2072first child is pointed to by C<op_first> and the last child by
10e2eb10 2073C<op_last>. The children in between can be found by iteratively
86cd3a13 2074following the C<OpSIBLING> pointer from the first child to the last (but
29e61fd9 2075see below).
9afa14e3 2076
29e61fd9 2077There are also some other op types: a C<PMOP> holds a regular expression,
10e2eb10
FC
2078and has no children, and a C<LOOP> may or may not have children. If the
2079C<op_children> field is non-zero, it behaves like a C<LISTOP>. To
9afa14e3
SC
2080complicate matters, if a C<UNOP> is actually a C<null> op after
2081optimization (see L</Compile pass 2: context propagation>) it will still
2082have children in accordance with its former type.
2083
29e61fd9
DM
2084Finally, there is a C<LOGOP>, or logic op. Like a C<LISTOP>, this has one
2085or more children, but it doesn't have an C<op_last> field: so you have to
86cd3a13 2086follow C<op_first> and then the C<OpSIBLING> chain itself to find the
29e61fd9
DM
2087last child. Instead it has an C<op_other> field, which is comparable to
2088the C<op_next> field described below, and represents an alternate
2089execution path. Operators like C<and>, C<or> and C<?> are C<LOGOP>s. Note
2090that in general, C<op_other> may not point to any of the direct children
2091of the C<LOGOP>.
2092
2093Starting in version 5.21.2, perls built with the experimental
2094define C<-DPERL_OP_PARENT> add an extra boolean flag for each op,
87b5a8b9 2095C<op_moresib>. When not set, this indicates that this is the last op in an
86cd3a13
DM
2096C<OpSIBLING> chain. This frees up the C<op_sibling> field on the last
2097sibling to point back to the parent op. Under this build, that field is
2098also renamed C<op_sibparent> to reflect its joint role. The macro
2099C<OpSIBLING(o)> wraps this special behaviour, and always returns NULL on
2100the last sibling. With this build the C<op_parent(o)> function can be
2101used to find the parent of any op. Thus for forward compatibility, you
2102should always use the C<OpSIBLING(o)> macro rather than accessing
2103C<op_sibling> directly.
29e61fd9 2104
06f6df17
RGS
2105Another way to examine the tree is to use a compiler back-end module, such
2106as L<B::Concise>.
2107
0a753a76
PP
2108=head2 Compile pass 1: check routines
2109
8870b5c7 2110The tree is created by the compiler while I<yacc> code feeds it
10e2eb10 2111the constructions it recognizes. Since I<yacc> works bottom-up, so does
0a753a76
PP
2112the first pass of perl compilation.
2113
2114What makes this pass interesting for perl developers is that some
2115optimization may be performed on this pass. This is optimization by
8870b5c7 2116so-called "check routines". The correspondence between node names
0a753a76
PP
2117and corresponding check routines is described in F<opcode.pl> (do not
2118forget to run C<make regen_headers> if you modify this file).
2119
2120A check routine is called when the node is fully constructed except
7b8d334a 2121for the execution-order thread. Since at this time there are no
0a753a76
PP
2122back-links to the currently constructed node, one can do most any
2123operation to the top-level node, including freeing it and/or creating
2124new nodes above/below it.
2125
2126The check routine returns the node which should be inserted into the
2127tree (if the top-level node was not modified, check routine returns
2128its argument).
2129
10e2eb10 2130By convention, check routines have names C<ck_*>. They are usually
0a753a76
PP
2131called from C<new*OP> subroutines (or C<convert>) (which in turn are
2132called from F<perly.y>).
2133
2134=head2 Compile pass 1a: constant folding
2135
2136Immediately after the check routine is called the returned node is
2137checked for being compile-time executable. If it is (the value is
2138judged to be constant) it is immediately executed, and a I<constant>
2139node with the "return value" of the corresponding subtree is
2140substituted instead. The subtree is deleted.
2141
2142If constant folding was not performed, the execution-order thread is
2143created.
2144
2145=head2 Compile pass 2: context propagation
2146
2147When a context for a part of compile tree is known, it is propagated
a3cb178b 2148down through the tree. At this time the context can have 5 values
0a753a76
PP
2149(instead of 2 for runtime context): void, boolean, scalar, list, and
2150lvalue. In contrast with the pass 1 this pass is processed from top
2151to bottom: a node's context determines the context for its children.
2152
2153Additional context-dependent optimizations are performed at this time.
2154Since at this moment the compile tree contains back-references (via
2155"thread" pointers), nodes cannot be free()d now. To allow
2156optimized-away nodes at this stage, such nodes are null()ified instead
2157of free()ing (i.e. their type is changed to OP_NULL).
2158
2159=head2 Compile pass 3: peephole optimization
2160
2161After the compile tree for a subroutine (or for an C<eval> or a file)
10e2eb10 2162is created, an additional pass over the code is performed. This pass
0a753a76 2163is neither top-down or bottom-up, but in the execution order (with
9ea12537
Z
2164additional complications for conditionals). Optimizations performed
2165at this stage are subject to the same restrictions as in the pass 2.
2166
2167Peephole optimizations are done by calling the function pointed to
2168by the global variable C<PL_peepp>. By default, C<PL_peepp> just
2169calls the function pointed to by the global variable C<PL_rpeepp>.
2170By default, that performs some basic op fixups and optimisations along
2171the execution-order op chain, and recursively calls C<PL_rpeepp> for
2172each side chain of ops (resulting from conditionals). Extensions may
2173provide additional optimisations or fixups, hooking into either the
2174per-subroutine or recursive stage, like this:
2175
2176 static peep_t prev_peepp;
2177 static void my_peep(pTHX_ OP *o)
2178 {
2179 /* custom per-subroutine optimisation goes here */
f0358462 2180 prev_peepp(aTHX_ o);
9ea12537
Z
2181 /* custom per-subroutine optimisation may also go here */
2182 }
2183 BOOT:
2184 prev_peepp = PL_peepp;
2185 PL_peepp = my_peep;
2186
2187 static peep_t prev_rpeepp;
2188 static void my_rpeep(pTHX_ OP *o)
2189 {
2190 OP *orig_o = o;
2191 for(; o; o = o->op_next) {
2192 /* custom per-op optimisation goes here */
2193 }
f0358462 2194 prev_rpeepp(aTHX_ orig_o);
9ea12537
Z
2195 }
2196 BOOT:
2197 prev_rpeepp = PL_rpeepp;
2198 PL_rpeepp = my_rpeep;
0a753a76 2199
1ba7f851
PJ
2200=head2 Pluggable runops
2201
2202The compile tree is executed in a runops function. There are two runops
1388f78e
RGS
2203functions, in F<run.c> and in F<dump.c>. C<Perl_runops_debug> is used
2204with DEBUGGING and C<Perl_runops_standard> is used otherwise. For fine
2205control over the execution of the compile tree it is possible to provide
2206your own runops function.
1ba7f851
PJ
2207
2208It's probably best to copy one of the existing runops functions and
2209change it to suit your needs. Then, in the BOOT section of your XS
2210file, add the line:
2211
2212 PL_runops = my_runops;
2213
2214This function should be as efficient as possible to keep your programs
2215running as fast as possible.
2216
fd85fad2
BM
2217=head2 Compile-time scope hooks
2218
2219As of perl 5.14 it is possible to hook into the compile-time lexical
10e2eb10 2220scope mechanism using C<Perl_blockhook_register>. This is used like
fd85fad2
BM
2221this:
2222
2223 STATIC void my_start_hook(pTHX_ int full);
2224 STATIC BHK my_hooks;
2225
2226 BOOT:
a88d97bf 2227 BhkENTRY_set(&my_hooks, bhk_start, my_start_hook);
fd85fad2
BM
2228 Perl_blockhook_register(aTHX_ &my_hooks);
2229
2230This will arrange to have C<my_start_hook> called at the start of
10e2eb10 2231compiling every lexical scope. The available hooks are:
fd85fad2
BM
2232
2233=over 4
2234
a88d97bf 2235=item C<void bhk_start(pTHX_ int full)>
fd85fad2 2236
10e2eb10 2237This is called just after starting a new lexical scope. Note that Perl
fd85fad2
BM
2238code like
2239
2240 if ($x) { ... }
2241
2242creates two scopes: the first starts at the C<(> and has C<full == 1>,
10e2eb10 2243the second starts at the C<{> and has C<full == 0>. Both end at the
f185f654 2244C<}>, so calls to C<start> and C<pre>/C<post_end> will match. Anything
fd85fad2
BM
2245pushed onto the save stack by this hook will be popped just before the
2246scope ends (between the C<pre_> and C<post_end> hooks, in fact).
2247
a88d97bf 2248=item C<void bhk_pre_end(pTHX_ OP **o)>
fd85fad2
BM
2249
2250This is called at the end of a lexical scope, just before unwinding the
10e2eb10 2251stack. I<o> is the root of the optree representing the scope; it is a
fd85fad2
BM
2252double pointer so you can replace the OP if you need to.
2253
a88d97bf 2254=item C<void bhk_post_end(pTHX_ OP **o)>
fd85fad2
BM
2255
2256This is called at the end of a lexical scope, just after unwinding the
10e2eb10 2257stack. I<o> is as above. Note that it is possible for calls to C<pre_>
fd85fad2
BM
2258and C<post_end> to nest, if there is something on the save stack that
2259calls string eval.
2260
a88d97bf 2261=item C<void bhk_eval(pTHX_ OP *const o)>
fd85fad2
BM
2262
2263This is called just before starting to compile an C<eval STRING>, C<do
10e2eb10 2264FILE>, C<require> or C<use>, after the eval has been set up. I<o> is the
fd85fad2
BM
2265OP that requested the eval, and will normally be an C<OP_ENTEREVAL>,
2266C<OP_DOFILE> or C<OP_REQUIRE>.
2267
2268=back
2269
2270Once you have your hook functions, you need a C<BHK> structure to put
10e2eb10
FC
2271them in. It's best to allocate it statically, since there is no way to
2272free it once it's registered. The function pointers should be inserted
fd85fad2 2273into this structure using the C<BhkENTRY_set> macro, which will also set
10e2eb10 2274flags indicating which entries are valid. If you do need to allocate
fd85fad2
BM
2275your C<BHK> dynamically for some reason, be sure to zero it before you
2276start.
2277
2278Once registered, there is no mechanism to switch these hooks off, so if
10e2eb10 2279that is necessary you will need to do this yourself. An entry in C<%^H>
a3e07c87
BM
2280is probably the best way, so the effect is lexically scoped; however it
2281is also possible to use the C<BhkDISABLE> and C<BhkENABLE> macros to
10e2eb10 2282temporarily switch entries on and off. You should also be aware that
a3e07c87 2283generally speaking at least one scope will have opened before your
f185f654 2284extension is loaded, so you will see some C<pre>/C<post_end> pairs that
a3e07c87 2285didn't have a matching C<start>.
fd85fad2 2286
9afa14e3
SC
2287=head1 Examining internal data structures with the C<dump> functions
2288
2289To aid debugging, the source file F<dump.c> contains a number of
2290functions which produce formatted output of internal data structures.
2291
2292The most commonly used of these functions is C<Perl_sv_dump>; it's used
10e2eb10 2293for dumping SVs, AVs, HVs, and CVs. The C<Devel::Peek> module calls
9afa14e3 2294C<sv_dump> to produce debugging output from Perl-space, so users of that
00aadd71 2295module should already be familiar with its format.
9afa14e3
SC
2296
2297C<Perl_op_dump> can be used to dump an C<OP> structure or any of its
210b36aa 2298derivatives, and produces output similar to C<perl -Dx>; in fact,
9afa14e3
SC
2299C<Perl_dump_eval> will dump the main root of the code being evaluated,
2300exactly like C<-Dx>.
2301
2302Other useful functions are C<Perl_dump_sub>, which turns a C<GV> into an
2303op tree, C<Perl_dump_packsubs> which calls C<Perl_dump_sub> on all the
2304subroutines in a package like so: (Thankfully, these are all xsubs, so
2305there is no op tree)
2306
2307 (gdb) print Perl_dump_packsubs(PL_defstash)
2308
2309 SUB attributes::bootstrap = (xsub 0x811fedc 0)
2310
2311 SUB UNIVERSAL::can = (xsub 0x811f50c 0)
2312
2313 SUB UNIVERSAL::isa = (xsub 0x811f304 0)
2314
2315 SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
2316
2317 SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
2318
2319and C<Perl_dump_all>, which dumps all the subroutines in the stash and
2320the op tree of the main root.
2321
954c1994 2322=head1 How multiple interpreters and concurrency are supported
ee072b34 2323
ee072b34
GS
2324=head2 Background and PERL_IMPLICIT_CONTEXT
2325
2326The Perl interpreter can be regarded as a closed box: it has an API
2327for feeding it code or otherwise making it do things, but it also has
2328functions for its own use. This smells a lot like an object, and
2329there are ways for you to build Perl so that you can have multiple
acfe0abc
GS
2330interpreters, with one interpreter represented either as a C structure,
2331or inside a thread-specific structure. These structures contain all
2332the context, the state of that interpreter.
2333
10e2eb10 2334One macro controls the major Perl build flavor: MULTIPLICITY. The
7b52221d 2335MULTIPLICITY build has a C structure that packages all the interpreter
10e2eb10 2336state. With multiplicity-enabled perls, PERL_IMPLICIT_CONTEXT is also
7b52221d 2337normally defined, and enables the support for passing in a "hidden" first
10e2eb10 2338argument that represents all three data structures. MULTIPLICITY makes
1a64a5e6 2339multi-threaded perls possible (with the ithreads threading model, related
7b52221d 2340to the macro USE_ITHREADS.)
54aff467 2341
27da23d5
JH
2342Two other "encapsulation" macros are the PERL_GLOBAL_STRUCT and
2343PERL_GLOBAL_STRUCT_PRIVATE (the latter turns on the former, and the
2344former turns on MULTIPLICITY.) The PERL_GLOBAL_STRUCT causes all the
2345internal variables of Perl to be wrapped inside a single global struct,
2346struct perl_vars, accessible as (globals) &PL_Vars or PL_VarsPtr or
2347the function Perl_GetVars(). The PERL_GLOBAL_STRUCT_PRIVATE goes
2348one step further, there is still a single struct (allocated in main()
2349either from heap or from stack) but there are no global data symbols
3bf17896 2350pointing to it. In either case the global struct should be initialized
27da23d5
JH
2351as the very first thing in main() using Perl_init_global_struct() and
2352correspondingly tear it down after perl_free() using Perl_free_global_struct(),
2353please see F<miniperlmain.c> for usage details. You may also need
2354to use C<dVAR> in your coding to "declare the global variables"
2355when you are using them. dTHX does this for you automatically.
2356
9aa97215
JH
2357To see whether you have non-const data you can use a BSD (or GNU)
2358compatible C<nm>:
bc028b6b
JH
2359
2360 nm libperl.a | grep -v ' [TURtr] '
2361
9aa97215
JH
2362If this displays any C<D> or C<d> symbols (or possibly C<C> or C<c>),
2363you have non-const data. The symbols the C<grep> removed are as follows:
2364C<Tt> are I<text>, or code, the C<Rr> are I<read-only> (const) data,
2365and the C<U> is <undefined>, external symbols referred to.
2366
2367The test F<t/porting/libperl.t> does this kind of symbol sanity
2368checking on C<libperl.a>.
bc028b6b 2369
27da23d5
JH
2370For backward compatibility reasons defining just PERL_GLOBAL_STRUCT
2371doesn't actually hide all symbols inside a big global struct: some
2372PerlIO_xxx vtables are left visible. The PERL_GLOBAL_STRUCT_PRIVATE
2373then hides everything (see how the PERLIO_FUNCS_DECL is used).
2374
54aff467 2375All this obviously requires a way for the Perl internal functions to be
acfe0abc 2376either subroutines taking some kind of structure as the first
ee072b34 2377argument, or subroutines taking nothing as the first argument. To
acfe0abc 2378enable these two very different ways of building the interpreter,
ee072b34
GS
2379the Perl source (as it does in so many other situations) makes heavy
2380use of macros and subroutine naming conventions.
2381
54aff467 2382First problem: deciding which functions will be public API functions and
00aadd71 2383which will be private. All functions whose names begin C<S_> are private
954c1994
GS
2384(think "S" for "secret" or "static"). All other functions begin with
2385"Perl_", but just because a function begins with "Perl_" does not mean it is
10e2eb10
FC
2386part of the API. (See L</Internal
2387Functions>.) The easiest way to be B<sure> a
00aadd71
NIS
2388function is part of the API is to find its entry in L<perlapi>.
2389If it exists in L<perlapi>, it's part of the API. If it doesn't, and you
2390think it should be (i.e., you need it for your extension), send mail via
a422fd2d 2391L<perlbug> explaining why you think it should be.
ee072b34
GS
2392
2393Second problem: there must be a syntax so that the same subroutine
2394declarations and calls can pass a structure as their first argument,
2395or pass nothing. To solve this, the subroutines are named and
2396declared in a particular way. Here's a typical start of a static
2397function used within the Perl guts:
2398
2399 STATIC void
2400 S_incline(pTHX_ char *s)
2401
acfe0abc 2402STATIC becomes "static" in C, and may be #define'd to nothing in some
da8c5729 2403configurations in the future.
ee072b34 2404
651a3225
GS
2405A public function (i.e. part of the internal API, but not necessarily
2406sanctioned for use in extensions) begins like this:
ee072b34
GS
2407
2408 void
2307c6d0 2409 Perl_sv_setiv(pTHX_ SV* dsv, IV num)
ee072b34 2410
0147cd53 2411C<pTHX_> is one of a number of macros (in F<perl.h>) that hide the
ee072b34
GS
2412details of the interpreter's context. THX stands for "thread", "this",
2413or "thingy", as the case may be. (And no, George Lucas is not involved. :-)
2414The first character could be 'p' for a B<p>rototype, 'a' for B<a>rgument,
a7486cbb
JH
2415or 'd' for B<d>eclaration, so we have C<pTHX>, C<aTHX> and C<dTHX>, and
2416their variants.
ee072b34 2417
a7486cbb
JH
2418When Perl is built without options that set PERL_IMPLICIT_CONTEXT, there is no
2419first argument containing the interpreter's context. The trailing underscore
ee072b34
GS
2420in the pTHX_ macro indicates that the macro expansion needs a comma
2421after the context argument because other arguments follow it. If
2422PERL_IMPLICIT_CONTEXT is not defined, pTHX_ will be ignored, and the
54aff467
GS
2423subroutine is not prototyped to take the extra argument. The form of the
2424macro without the trailing underscore is used when there are no additional
ee072b34
GS
2425explicit arguments.
2426
54aff467 2427When a core function calls another, it must pass the context. This
2307c6d0 2428is normally hidden via macros. Consider C<sv_setiv>. It expands into
ee072b34
GS
2429something like this:
2430
2307c6d0
SB
2431 #ifdef PERL_IMPLICIT_CONTEXT
2432 #define sv_setiv(a,b) Perl_sv_setiv(aTHX_ a, b)
ee072b34 2433 /* can't do this for vararg functions, see below */
2307c6d0
SB
2434 #else
2435 #define sv_setiv Perl_sv_setiv
2436 #endif
ee072b34
GS
2437
2438This works well, and means that XS authors can gleefully write:
2439
2307c6d0 2440 sv_setiv(foo, bar);
ee072b34
GS
2441
2442and still have it work under all the modes Perl could have been
2443compiled with.
2444
ee072b34
GS
2445This doesn't work so cleanly for varargs functions, though, as macros
2446imply that the number of arguments is known in advance. Instead we
2447either need to spell them out fully, passing C<aTHX_> as the first
2448argument (the Perl core tends to do this with functions like
2449Perl_warner), or use a context-free version.
2450
2451The context-free version of Perl_warner is called
2452Perl_warner_nocontext, and does not take the extra argument. Instead
2453it does dTHX; to get the context from thread-local storage. We
2454C<#define warner Perl_warner_nocontext> so that extensions get source
2455compatibility at the expense of performance. (Passing an arg is
2456cheaper than grabbing it from thread-local storage.)
2457
acfe0abc 2458You can ignore [pad]THXx when browsing the Perl headers/sources.
ee072b34
GS
2459Those are strictly for use within the core. Extensions and embedders
2460need only be aware of [pad]THX.
2461
a7486cbb
JH
2462=head2 So what happened to dTHR?
2463
2464C<dTHR> was introduced in perl 5.005 to support the older thread model.
2465The older thread model now uses the C<THX> mechanism to pass context
2466pointers around, so C<dTHR> is not useful any more. Perl 5.6.0 and
2467later still have it for backward source compatibility, but it is defined
2468to be a no-op.
2469
ee072b34
GS
2470=head2 How do I use all this in extensions?
2471
2472When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call
2473any functions in the Perl API will need to pass the initial context
2474argument somehow. The kicker is that you will need to write it in
2475such a way that the extension still compiles when Perl hasn't been
2476built with PERL_IMPLICIT_CONTEXT enabled.
2477
2478There are three ways to do this. First, the easy but inefficient way,
2479which is also the default, in order to maintain source compatibility
0147cd53 2480with extensions: whenever F<XSUB.h> is #included, it redefines the aTHX
ee072b34
GS
2481and aTHX_ macros to call a function that will return the context.
2482Thus, something like:
2483
2307c6d0 2484 sv_setiv(sv, num);
ee072b34 2485
4375e838 2486in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
54aff467 2487in effect:
ee072b34 2488
2307c6d0 2489 Perl_sv_setiv(Perl_get_context(), sv, num);
ee072b34 2490
54aff467 2491or to this otherwise:
ee072b34 2492
2307c6d0 2493 Perl_sv_setiv(sv, num);
ee072b34 2494
da8c5729 2495You don't have to do anything new in your extension to get this; since
2fa86c13 2496the Perl library provides Perl_get_context(), it will all just
ee072b34
GS
2497work.
2498
2499The second, more efficient way is to use the following template for
2500your Foo.xs:
2501
c52f9dcd
JH
2502 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2503 #include "EXTERN.h"
2504 #include "perl.h"
2505 #include "XSUB.h"
ee072b34 2506
fd061412 2507 STATIC void my_private_function(int arg1, int arg2);
ee072b34 2508
fd061412 2509 STATIC void
c52f9dcd
JH
2510 my_private_function(int arg1, int arg2)
2511 {
2512 dTHX; /* fetch context */
2513 ... call many Perl API functions ...
2514 }
ee072b34
GS
2515
2516 [... etc ...]
2517
c52f9dcd 2518 MODULE = Foo PACKAGE = Foo
ee072b34 2519
c52f9dcd 2520 /* typical XSUB */
ee072b34 2521
c52f9dcd
JH
2522 void
2523 my_xsub(arg)
2524 int arg
2525 CODE:
2526 my_private_function(arg, 10);
ee072b34
GS
2527
2528Note that the only two changes from the normal way of writing an
2529extension is the addition of a C<#define PERL_NO_GET_CONTEXT> before
2530including the Perl headers, followed by a C<dTHX;> declaration at
2531the start of every function that will call the Perl API. (You'll
2532know which functions need this, because the C compiler will complain
2533that there's an undeclared identifier in those functions.) No changes
2534are needed for the XSUBs themselves, because the XS() macro is
2535correctly defined to pass in the implicit context if needed.
2536
2537The third, even more efficient way is to ape how it is done within
2538the Perl guts:
2539
2540
c52f9dcd
JH
2541 #define PERL_NO_GET_CONTEXT /* we want efficiency */
2542 #include "EXTERN.h"
2543 #include "perl.h"
2544 #include "XSUB.h"
ee072b34
GS
2545
2546 /* pTHX_ only needed for functions that call Perl API */
fd061412 2547 STATIC void my_private_function(pTHX_ int arg1, int arg2);
ee072b34 2548
fd061412 2549 STATIC void
c52f9dcd
JH
2550 my_private_function(pTHX_ int arg1, int arg2)
2551 {
2552 /* dTHX; not needed here, because THX is an argument */
2553 ... call Perl API functions ...
2554 }
ee072b34
GS
2555
2556 [... etc ...]
2557
c52f9dcd 2558 MODULE = Foo PACKAGE = Foo
ee072b34 2559
c52f9dcd 2560 /* typical XSUB */
ee072b34 2561
c52f9dcd
JH
2562 void
2563 my_xsub(arg)
2564 int arg
2565 CODE:
2566 my_private_function(aTHX_ arg, 10);
ee072b34
GS
2567
2568This implementation never has to fetch the context using a function
2569call, since it is always passed as an extra argument. Depending on
2570your needs for simplicity or efficiency, you may mix the previous
2571two approaches freely.
2572
651a3225
GS
2573Never add a comma after C<pTHX> yourself--always use the form of the
2574macro with the underscore for functions that take explicit arguments,
2575or the form without the argument for functions with no explicit arguments.
ee072b34 2576
27da23d5
JH
2577If one is compiling Perl with the C<-DPERL_GLOBAL_STRUCT> the C<dVAR>
2578definition is needed if the Perl global variables (see F<perlvars.h>
2579or F<globvar.sym>) are accessed in the function and C<dTHX> is not
2580used (the C<dTHX> includes the C<dVAR> if necessary). One notices
2581the need for C<dVAR> only with the said compile-time define, because
2582otherwise the Perl global variables are visible as-is.
2583
a7486cbb
JH
2584=head2 Should I do anything special if I call perl from multiple threads?
2585
2586If you create interpreters in one thread and then proceed to call them in
2587another, you need to make sure perl's own Thread Local Storage (TLS) slot is
2588initialized correctly in each of those threads.
2589
2590The C<perl_alloc> and C<perl_clone> API functions will automatically set
2591the TLS slot to the interpreter they created, so that there is no need to do
2592anything special if the interpreter is always accessed in the same thread that
2593created it, and that thread did not create or call any other interpreters
2594afterwards. If that is not the case, you have to set the TLS slot of the
2595thread before calling any functions in the Perl API on that particular
2596interpreter. This is done by calling the C<PERL_SET_CONTEXT> macro in that
2597thread as the first thing you do:
2598
2599 /* do this before doing anything else with some_perl */
2600 PERL_SET_CONTEXT(some_perl);
2601
2602 ... other Perl API calls on some_perl go here ...
2603
ee072b34
GS
2604=head2 Future Plans and PERL_IMPLICIT_SYS
2605
2606Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
2607that the interpreter knows about itself and pass it around, so too are
2608there plans to allow the interpreter to bundle up everything it knows
2609about the environment it's running on. This is enabled with the
7b52221d
RGS
2610PERL_IMPLICIT_SYS macro. Currently it only works with USE_ITHREADS on
2611Windows.
ee072b34
GS
2612
2613This allows the ability to provide an extra pointer (called the "host"
2614environment) for all the system calls. This makes it possible for
2615all the system stuff to maintain their own state, broken down into
2616seven C structures. These are thin wrappers around the usual system
0147cd53 2617calls (see F<win32/perllib.c>) for the default perl executable, but for a
ee072b34
GS
2618more ambitious host (like the one that would do fork() emulation) all
2619the extra work needed to pretend that different interpreters are
2620actually different "processes", would be done here.
2621
2622The Perl engine/interpreter and the host are orthogonal entities.
2623There could be one or more interpreters in a process, and one or
2624more "hosts", with free association between them.
2625
a422fd2d
SC
2626=head1 Internal Functions
2627
2628All of Perl's internal functions which will be exposed to the outside
06f6df17 2629world are prefixed by C<Perl_> so that they will not conflict with XS
a422fd2d 2630functions or functions used in a program in which Perl is embedded.
10e2eb10 2631Similarly, all global variables begin with C<PL_>. (By convention,
06f6df17 2632static functions start with C<S_>.)
a422fd2d 2633
0972ecdf
DM
2634Inside the Perl core (C<PERL_CORE> defined), you can get at the functions
2635either with or without the C<Perl_> prefix, thanks to a bunch of defines
10e2eb10 2636that live in F<embed.h>. Note that extension code should I<not> set
0972ecdf
DM
2637C<PERL_CORE>; this exposes the full perl internals, and is likely to cause
2638breakage of the XS in each new perl release.
2639
2640The file F<embed.h> is generated automatically from
10e2eb10 2641F<embed.pl> and F<embed.fnc>. F<embed.pl> also creates the prototyping
dc9b1d22 2642header files for the internal functions, generates the documentation
10e2eb10 2643and a lot of other bits and pieces. It's important that when you add
dc9b1d22 2644a new function to the core or change an existing one, you change the
10e2eb10 2645data in the table in F<embed.fnc> as well. Here's a sample entry from
dc9b1d22 2646that table:
a422fd2d
SC
2647
2648 Apd |SV** |av_fetch |AV* ar|I32 key|I32 lval
2649
790ba721
KW
2650The first column is a set of flags, the second column the return type,
2651the third column the name. Columns after that are the arguments.
2652The flags are documented at the top of F<embed.fnc>.
a422fd2d 2653
dc9b1d22
MHM
2654If you edit F<embed.pl> or F<embed.fnc>, you will need to run
2655C<make regen_headers> to force a rebuild of F<embed.h> and other
2656auto-generated files.
a422fd2d 2657
6b4667fc 2658=head2 Formatted Printing of IVs, UVs, and NVs
9dd9db0b 2659
6b4667fc
A
2660If you are printing IVs, UVs, or NVS instead of the stdio(3) style
2661formatting codes like C<%d>, C<%ld>, C<%f>, you should use the
2662following macros for portability
9dd9db0b 2663
c52f9dcd
JH
2664 IVdf IV in decimal
2665 UVuf UV in decimal
2666 UVof UV in octal
2667 UVxf UV in hexadecimal
2668 NVef NV %e-like
2669 NVff NV %f-like
2670 NVgf NV %g-like
9dd9db0b 2671
6b4667fc
A
2672These will take care of 64-bit integers and long doubles.
2673For example:
2674
c52f9dcd 2675 printf("IV is %"IVdf"\n", iv);
6b4667fc
A
2676
2677The IVdf will expand to whatever is the correct format for the IVs.
9dd9db0b 2678
aacf4ea2
JH
2679Note that there are different "long doubles": Perl will use
2680whatever the compiler has.
2681
8908e76d
JH
2682If you are printing addresses of pointers, use UVxf combined
2683with PTR2UV(), do not use %lx or %p.
2684
e613617c 2685=head2 Formatted Printing of C<Size_t> and C<SSize_t>
5862f74e
KW
2686
2687The most general way to do this is to cast them to a UV or IV, and
2688print as in the
2689L<previous section|/Formatted Printing of IVs, UVs, and NVs>.
2690
2691But if you're using C<PerlIO_printf()>, it's less typing and visual
2692clutter to use the C<"%z"> length modifier (for I<siZe>):
2693
2694 PerlIO_printf("STRLEN is %zu\n", len);
2695
2696This modifier is not portable, so its use should be restricted to
2697C<PerlIO_printf()>.
2698
8908e76d
JH
2699=head2 Pointer-To-Integer and Integer-To-Pointer
2700
2701Because pointer size does not necessarily equal integer size,
2702use the follow macros to do it right.
2703
c52f9dcd
JH
2704 PTR2UV(pointer)
2705 PTR2IV(pointer)
2706 PTR2NV(pointer)
2707 INT2PTR(pointertotype, integer)
8908e76d
JH
2708
2709For example:
2710
c52f9dcd
JH
2711 IV iv = ...;
2712 SV *sv = INT2PTR(SV*, iv);
8908e76d
JH
2713
2714and
2715
c52f9dcd
JH
2716 AV *av = ...;
2717 UV uv = PTR2UV(av);
8908e76d 2718
0ca3a874
MHM
2719=head2 Exception Handling
2720
9b5c3821 2721There are a couple of macros to do very basic exception handling in XS
10e2eb10 2722modules. You have to define C<NO_XSLOCKS> before including F<XSUB.h> to
9b5c3821
MHM
2723be able to use these macros:
2724
2725 #define NO_XSLOCKS
2726 #include "XSUB.h"
2727
2728You can use these macros if you call code that may croak, but you need
10e2eb10 2729to do some cleanup before giving control back to Perl. For example:
0ca3a874 2730
d7f8936a 2731 dXCPT; /* set up necessary variables */
0ca3a874
MHM
2732
2733 XCPT_TRY_START {
2734 code_that_may_croak();
2735 } XCPT_TRY_END
2736
2737 XCPT_CATCH
2738 {
2739 /* do cleanup here */
2740 XCPT_RETHROW;
2741 }
2742
2743Note that you always have to rethrow an exception that has been
10e2eb10
FC
2744caught. Using these macros, it is not possible to just catch the
2745exception and ignore it. If you have to ignore the exception, you
0ca3a874
MHM
2746have to use the C<call_*> function.
2747
2748The advantage of using the above macros is that you don't have
2749to setup an extra function for C<call_*>, and that using these
2750macros is faster than using C<call_*>.
2751
a422fd2d
SC
2752=head2 Source Documentation
2753
2754There's an effort going on to document the internal functions and
61ad4b94 2755automatically produce reference manuals from them -- L<perlapi> is one
a422fd2d 2756such manual which details all the functions which are available to XS
10e2eb10 2757writers. L<perlintern> is the autogenerated manual for the functions
a422fd2d
SC
2758which are not part of the API and are supposedly for internal use only.
2759
2760Source documentation is created by putting POD comments into the C
2761source, like this:
2762
2763 /*
2764 =for apidoc sv_setiv
2765
2766 Copies an integer into the given SV. Does not handle 'set' magic. See
a95b3d6a 2767 L<perlapi/sv_setiv_mg>.
a422fd2d
SC
2768
2769 =cut
2770 */
2771
2772Please try and supply some documentation if you add functions to the
2773Perl core.
2774
0d098d33
MHM
2775=head2 Backwards compatibility
2776
10e2eb10
FC
2777The Perl API changes over time. New functions are
2778added or the interfaces of existing functions are
2779changed. The C<Devel::PPPort> module tries to
0d098d33
MHM
2780provide compatibility code for some of these changes, so XS writers don't
2781have to code it themselves when supporting multiple versions of Perl.
2782
2783C<Devel::PPPort> generates a C header file F<ppport.h> that can also
10e2eb10 2784be run as a Perl script. To generate F<ppport.h>, run:
0d098d33
MHM
2785
2786 perl -MDevel::PPPort -eDevel::PPPort::WriteFile
2787
2788Besides checking existing XS code, the script can also be used to retrieve
2789compatibility information for various API calls using the C<--api-info>
10e2eb10 2790command line switch. For example:
0d098d33
MHM
2791
2792 % perl ppport.h --api-info=sv_magicext
2793
2794For details, see C<perldoc ppport.h>.
2795
a422fd2d
SC
2796=head1 Unicode Support
2797
10e2eb10 2798Perl 5.6.0 introduced Unicode support. It's important for porters and XS
a422fd2d
SC
2799writers to understand this support and make sure that the code they
2800write does not corrupt Unicode data.
2801
2802=head2 What B<is> Unicode, anyway?
2803
10e2eb10
FC
2804In the olden, less enlightened times, we all used to use ASCII. Most of
2805us did, anyway. The big problem with ASCII is that it's American. Well,
a422fd2d 2806no, that's not actually the problem; the problem is that it's not
10e2eb10 2807particularly useful for people who don't use the Roman alphabet. What
a422fd2d 2808used to happen was that particular languages would stick their own
10e2eb10 2809alphabet in the upper range of the sequence, between 128 and 255. Of
a422fd2d
SC
2810course, we then ended up with plenty of variants that weren't quite
2811ASCII, and the whole point of it being a standard was lost.
2812
2813Worse still, if you've got a language like Chinese or
2814Japanese that has hundreds or thousands of characters, then you really
2815can't fit them into a mere 256, so they had to forget about ASCII
2816altogether, and build their own systems using pairs of numbers to refer
2817to one character.
2818
2819To fix this, some people formed Unicode, Inc. and
2820produced a new character set containing all the characters you can
10e2eb10
FC
2821possibly think of and more. There are several ways of representing these
2822characters, and the one Perl uses is called UTF-8. UTF-8 uses
2823a variable number of bytes to represent a character. You can learn more
2575c402 2824about Unicode and Perl's Unicode model in L<perlunicode>.
a422fd2d 2825
3ad86f0e
KW
2826(On EBCDIC platforms, Perl uses instead UTF-EBCDIC, which is a form of
2827UTF-8 adapted for EBCDIC platforms. Below, we just talk about UTF-8.
2828UTF-EBCDIC is like UTF-8, but the details are different. The macros
2829hide the differences from you, just remember that the particular numbers
2830and bit patterns presented below will differ in UTF-EBCDIC.)
2831
1e54db1a 2832=head2 How can I recognise a UTF-8 string?
a422fd2d 2833
10e2eb10
FC
2834You can't. This is because UTF-8 data is stored in bytes just like
2835non-UTF-8 data. The Unicode character 200, (C<0xC8> for you hex types)
a422fd2d 2836capital E with a grave accent, is represented by the two bytes
10e2eb10 2837C<v196.172>. Unfortunately, the non-Unicode string C<chr(196).chr(172)>
61ad4b94 2838has that byte sequence as well. So you can't tell just by looking -- this
a422fd2d
SC
2839is what makes Unicode input an interesting problem.
2840
2575c402
JW
2841In general, you either have to know what you're dealing with, or you
2842have to guess. The API function C<is_utf8_string> can help; it'll tell
61ad4b94
KW
2843you if a string contains only valid UTF-8 characters, and the chances
2844of a non-UTF-8 string looking like valid UTF-8 become very small very
2845quickly with increasing string length. On a character-by-character
2846basis, C<isUTF8_CHAR>
2575c402 2847will tell you whether the current character in a string is valid UTF-8.
a422fd2d 2848
1e54db1a 2849=head2 How does UTF-8 represent Unicode characters?
a422fd2d 2850
1e54db1a 2851As mentioned above, UTF-8 uses a variable number of bytes to store a
10e2eb10
FC
2852character. Characters with values 0...127 are stored in one
2853byte, just like good ol' ASCII. Character 128 is stored as
2854C<v194.128>; this continues up to character 191, which is
2855C<v194.191>. Now we've run out of bits (191 is binary
61ad4b94 2856C<10111111>) so we move on; character 192 is C<v195.128>. And
a422fd2d 2857so it goes on, moving to three bytes at character 2048.
6e31cdd1 2858L<perlunicode/Unicode Encodings> has pictures of how this works.
a422fd2d 2859
1e54db1a 2860Assuming you know you're dealing with a UTF-8 string, you can find out
a422fd2d
SC
2861how long the first character in it is with the C<UTF8SKIP> macro:
2862
2863 char *utf = "\305\233\340\240\201";
2864 I32 len;
2865
2866 len = UTF8SKIP(utf); /* len is 2 here */
2867 utf += len;
2868 len = UTF8SKIP(utf); /* len is 3 here */
2869
1e54db1a 2870Another way to skip over characters in a UTF-8 string is to use
a422fd2d 2871C<utf8_hop>, which takes a string and a number of characters to skip
10e2eb10 2872over. You're on your own about bounds checking, though, so don't use it
a422fd2d
SC
2873lightly.
2874
1e54db1a 2875All bytes in a multi-byte UTF-8 character will have the high bit set,
3a2263fe 2876so you can test if you need to do something special with this
61ad4b94 2877character like this (the C<UTF8_IS_INVARIANT()> is a macro that tests
9f98c7fe 2878whether the byte is encoded as a single byte even in UTF-8):
a422fd2d 2879
32128a7f
KW
2880 U8 *utf; /* Initialize this to point to the beginning of the
2881 sequence to convert */
2882 U8 *utf_end; /* Initialize this to 1 beyond the end of the sequence
2883 pointed to by 'utf' */
2884 UV uv; /* Returned code point; note: a UV, not a U8, not a
2885 char */
2886 STRLEN len; /* Returned length of character in bytes */
a422fd2d 2887
3a2263fe 2888 if (!UTF8_IS_INVARIANT(*utf))
1e54db1a 2889 /* Must treat this as UTF-8 */
4b88fb76 2890 uv = utf8_to_uvchr_buf(utf, utf_end, &len);
a422fd2d
SC
2891 else
2892 /* OK to treat this character as a byte */
2893 uv = *utf;
2894
4b88fb76 2895You can also see in that example that we use C<utf8_to_uvchr_buf> to get the
95701e00 2896value of the character; the inverse function C<uvchr_to_utf8> is available
1e54db1a 2897for putting a UV into UTF-8:
a422fd2d 2898
61ad4b94 2899 if (!UVCHR_IS_INVARIANT(uv))
a422fd2d 2900 /* Must treat this as UTF8 */
95701e00 2901 utf8 = uvchr_to_utf8(utf8, uv);
a422fd2d
SC
2902 else
2903 /* OK to treat this character as a byte */
2904 *utf8++ = uv;
2905
2906You B<must> convert characters to UVs using the above functions if
1e54db1a 2907you're ever in a situation where you have to match UTF-8 and non-UTF-8
10e2eb10 2908characters. You may not skip over UTF-8 characters in this case. If you
1e54db1a
JH
2909do this, you'll lose the ability to match hi-bit non-UTF-8 characters;
2910for instance, if your UTF-8 string contains C<v196.172>, and you skip
2911that character, you can never match a C<chr(200)> in a non-UTF-8 string.
a422fd2d
SC
2912So don't do that!
2913
61ad4b94
KW
2914(Note that we don't have to test for invariant characters in the
2915examples above. The functions work on any well-formed UTF-8 input.
2916It's just that its faster to avoid the function overhead when it's not
2917needed.)
2918
1e54db1a 2919=head2 How does Perl store UTF-8 strings?
a422fd2d 2920
61ad4b94 2921Currently, Perl deals with UTF-8 strings and non-UTF-8 strings
10e2eb10
FC
2922slightly differently. A flag in the SV, C<SVf_UTF8>, indicates that the
2923string is internally encoded as UTF-8. Without it, the byte value is the
61ad4b94
KW
2924codepoint number and vice versa. This flag is only meaningful if the SV
2925is C<SvPOK> or immediately after stringification via C<SvPV> or a
2926similar macro. You can check and manipulate this flag with the
2575c402 2927following macros:
a422fd2d
SC
2928
2929 SvUTF8(sv)
2930 SvUTF8_on(sv)
2931 SvUTF8_off(sv)
2932
2933This flag has an important effect on Perl's treatment of the string: if
61ad4b94 2934UTF-8 data is not properly distinguished, regular expressions,
a422fd2d 2935C<length>, C<substr> and other string handling operations will have
61ad4b94 2936undesirable (wrong) results.
a422fd2d
SC
2937
2938The problem comes when you have, for instance, a string that isn't
61ad4b94 2939flagged as UTF-8, and contains a byte sequence that could be UTF-8 --
1e54db1a 2940especially when combining non-UTF-8 and UTF-8 strings.
a422fd2d 2941
61ad4b94
KW
2942Never forget that the C<SVf_UTF8> flag is separate from the PV value; you
2943need to be sure you don't accidentally knock it off while you're
10e2eb10 2944manipulating SVs. More specifically, you cannot expect to do this:
a422fd2d
SC
2945
2946 SV *sv;
2947 SV *nsv;
2948 STRLEN len;
2949 char *p;
2950
2951 p = SvPV(sv, len);
2952 frobnicate(p);
2953 nsv = newSVpvn(p, len);
2954
2955The C<char*> string does not tell you the whole story, and you can't
10e2eb10 2956copy or reconstruct an SV just by copying the string value. Check if the
c31cc9fc
FC
2957old SV has the UTF8 flag set (I<after> the C<SvPV> call), and act
2958accordingly:
a422fd2d
SC
2959
2960 p = SvPV(sv, len);
6db25795
KW
2961 is_utf8 = SvUTF8(sv);
2962 frobnicate(p, is_utf8);
a422fd2d 2963 nsv = newSVpvn(p, len);
6db25795 2964 if (is_utf8)
a422fd2d
SC
2965 SvUTF8_on(nsv);
2966
6db25795
KW
2967In the above, your C<frobnicate> function has been changed to be made
2968aware of whether or not it's dealing with UTF-8 data, so that it can
2969handle the string appropriately.
a422fd2d 2970
3a2263fe 2971Since just passing an SV to an XS function and copying the data of
2575c402 2972the SV is not enough to copy the UTF8 flags, even less right is just
61ad4b94 2973passing a S<C<char *>> to an XS function.
3a2263fe 2974
dc83bf8e 2975For full generality, use the L<C<DO_UTF8>|perlapi/DO_UTF8> macro to see if the
6db25795
KW
2976string in an SV is to be I<treated> as UTF-8. This takes into account
2977if the call to the XS function is being made from within the scope of
2978L<S<C<use bytes>>|bytes>. If so, the underlying bytes that comprise the
2979UTF-8 string are to be exposed, rather than the character they
2980represent. But this pragma should only really be used for debugging and
2981perhaps low-level testing at the byte level. Hence most XS code need
2982not concern itself with this, but various areas of the perl core do need
2983to support it.
2984
2985And this isn't the whole story. Starting in Perl v5.12, strings that
2986aren't encoded in UTF-8 may also be treated as Unicode under various
6e31cdd1 2987conditions (see L<perlunicode/ASCII Rules versus Unicode Rules>).
6db25795
KW
2988This is only really a problem for characters whose ordinals are between
2989128 and 255, and their behavior varies under ASCII versus Unicode rules
2990in ways that your code cares about (see L<perlunicode/The "Unicode Bug">).
2991There is no published API for dealing with this, as it is subject to
2992change, but you can look at the code for C<pp_lc> in F<pp.c> for an
2993example as to how it's currently done.
2994
1e54db1a 2995=head2 How do I convert a string to UTF-8?
a422fd2d 2996
2575c402 2997If you're mixing UTF-8 and non-UTF-8 strings, it is necessary to upgrade
61ad4b94 2998the non-UTF-8 strings to UTF-8. If you've got an SV, the easiest way to do
2575c402 2999this is:
a422fd2d
SC
3000
3001 sv_utf8_upgrade(sv);
3002
3003However, you must not do this, for example:
3004
3005 if (!SvUTF8(left))
3006 sv_utf8_upgrade(left);
3007
3008If you do this in a binary operator, you will actually change one of the
b1866b2d 3009strings that came into the operator, and, while it shouldn't be noticeable
2575c402 3010by the end user, it can cause problems in deficient code.
a422fd2d 3011
1e54db1a 3012Instead, C<bytes_to_utf8> will give you a UTF-8-encoded B<copy> of its
10e2eb10
FC
3013string argument. This is useful for having the data available for
3014comparisons and so on, without harming the original SV. There's also
a422fd2d
SC
3015C<utf8_to_bytes> to go the other way, but naturally, this will fail if
3016the string contains any characters above 255 that can't be represented
3017in a single byte.
3018
6db25795
KW
3019=head2 How do I compare strings?
3020
3021L<perlapi/sv_cmp> and L<perlapi/sv_cmp_flags> do a lexigraphic
3022comparison of two SV's, and handle UTF-8ness properly. Note, however,
3023that Unicode specifies a much fancier mechanism for collation, available
3024via the L<Unicode::Collate> module.
3025
3026To just compare two strings for equality/non-equality, you can just use
3027L<C<memEQ()>|perlapi/memEQ> and L<C<memNE()>|perlapi/memEQ> as usual,
3028except the strings must be both UTF-8 or not UTF-8 encoded.
3029
3030To compare two strings case-insensitively, use
3031L<C<foldEQ_utf8()>|perlapi/foldEQ_utf8> (the strings don't have to have
3032the same UTF-8ness).
3033
a422fd2d
SC
3034=head2 Is there anything else I need to know?
3035
10e2eb10 3036Not really. Just remember these things:
a422fd2d
SC
3037
3038=over 3
3039
3040=item *
3041
6db25795
KW
3042There's no way to tell if a S<C<char *>> or S<C<U8 *>> string is UTF-8
3043or not. But you can tell if an SV is to be treated as UTF-8 by calling
3044C<DO_UTF8> on it, after stringifying it with C<SvPV> or a similar
3045macro. And, you can tell if SV is actually UTF-8 (even if it is not to
3046be treated as such) by looking at its C<SvUTF8> flag (again after
3047stringifying it). Don't forget to set the flag if something should be
3048UTF-8.
3049Treat the flag as part of the PV, even though it's not -- if you pass on
3050the PV to somewhere, pass on the flag too.
a422fd2d
SC
3051
3052=item *
3053
4b88fb76 3054If a string is UTF-8, B<always> use C<utf8_to_uvchr_buf> to get at the value,
3a2263fe 3055unless C<UTF8_IS_INVARIANT(*s)> in which case you can use C<*s>.
a422fd2d
SC
3056
3057=item *
3058
61ad4b94
KW
3059When writing a character UV to a UTF-8 string, B<always> use
3060C<uvchr_to_utf8>, unless C<UVCHR_IS_INVARIANT(uv))> in which case
3a2263fe 3061you can use C<*s = uv>.
a422fd2d
SC
3062
3063=item *
3064
10e2eb10
FC
3065Mixing UTF-8 and non-UTF-8 strings is
3066tricky. Use C<bytes_to_utf8> to get
2bbc8d55 3067a new string which is UTF-8 encoded, and then combine them.
a422fd2d
SC
3068
3069=back
3070
53e06cf0
SC
3071=head1 Custom Operators
3072
2a0fd0f1 3073Custom operator support is an experimental feature that allows you to
10e2eb10 3074define your own ops. This is primarily to allow the building of
53e06cf0
SC
3075interpreters for other languages in the Perl core, but it also allows
3076optimizations through the creation of "macro-ops" (ops which perform the
3077functions of multiple ops which are usually executed together, such as
1aa6ea50 3078C<gvsv, gvsv, add>.)
53e06cf0 3079
10e2eb10 3080This feature is implemented as a new op type, C<OP_CUSTOM>. The Perl
53e06cf0 3081core does not "know" anything special about this op type, and so it will
10e2eb10 3082not be involved in any optimizations. This also means that you can
61ad4b94
KW
3083define your custom ops to be any op structure -- unary, binary, list and
3084so on -- you like.
53e06cf0 3085
10e2eb10
FC
3086It's important to know what custom operators won't do for you. They
3087won't let you add new syntax to Perl, directly. They won't even let you
3088add new keywords, directly. In fact, they won't change the way Perl
3089compiles a program at all. You have to do those changes yourself, after
3090Perl has compiled the program. You do this either by manipulating the op
53e06cf0
SC
3091tree using a C<CHECK> block and the C<B::Generate> module, or by adding
3092a custom peephole optimizer with the C<optimize> module.
3093
3094When you do this, you replace ordinary Perl ops with custom ops by
407f86e1 3095creating ops with the type C<OP_CUSTOM> and the C<op_ppaddr> of your own
10e2eb10
FC
3096PP function. This should be defined in XS code, and should look like
3097the PP ops in C<pp_*.c>. You are responsible for ensuring that your op
53e06cf0
SC
3098takes the appropriate number of values from the stack, and you are
3099responsible for adding stack marks if necessary.
3100
3101You should also "register" your op with the Perl interpreter so that it
10e2eb10 3102can produce sensible error and warning messages. Since it is possible to
53e06cf0 3103have multiple custom ops within the one "logical" op type C<OP_CUSTOM>,
9733086d 3104Perl uses the value of C<< o->op_ppaddr >> to determine which custom op
10e2eb10 3105it is dealing with. You should create an C<XOP> structure for each
9733086d
BM
3106ppaddr you use, set the properties of the custom op with
3107C<XopENTRY_set>, and register the structure against the ppaddr using
10e2eb10 3108C<Perl_custom_op_register>. A trivial example might look like:
9733086d
BM
3109
3110 static XOP my_xop;
3111 static OP *my_pp(pTHX);
3112
3113 BOOT:
3114 XopENTRY_set(&my_xop, xop_name, "myxop");
3115 XopENTRY_set(&my_xop, xop_desc, "Useless custom op");
3116 Perl_custom_op_register(aTHX_ my_pp, &my_xop);
3117
3118The available fields in the structure are:
3119
3120=over 4
3121
3122=item xop_name
3123
10e2eb10 3124A short name for your op. This will be included in some error messages,
9733086d
BM
3125and will also be returned as C<< $op->name >> by the L<B|B> module, so
3126it will appear in the output of module like L<B::Concise|B::Concise>.
3127
3128=item xop_desc
3129
3130A short description of the function of the op.
3131
3132=item xop_class
3133
10e2eb10 3134Which of the various C<*OP> structures this op uses. This should be one of
9733086d
BM
3135the C<OA_*> constants from F<op.h>, namely
3136
3137=over 4
3138
3139=item OA_BASEOP
3140
3141=item OA_UNOP
3142
3143=item OA_BINOP
3144
3145=item OA_LOGOP
3146
3147=item OA_LISTOP
3148
3149=item OA_PMOP
3150
3151=item OA_SVOP
3152
3153=item OA_PADOP
3154
3155=item OA_PVOP_OR_SVOP
3156
10e2eb10 3157This should be interpreted as 'C<PVOP>' only. The C<_OR_SVOP> is because
9733086d
BM
3158the only core C<PVOP>, C<OP_TRANS>, can sometimes be a C<SVOP> instead.
3159
3160=item OA_LOOP
3161
3162=item OA_COP
3163
3164=back
3165
3166The other C<OA_*> constants should not be used.
3167
3168=item xop_peep
3169
3170This member is of type C<Perl_cpeep_t>, which expands to C<void
10e2eb10 3171(*Perl_cpeep_t)(aTHX_ OP *o, OP *oldop)>. If it is set, this function
9733086d 3172will be called from C<Perl_rpeep> when ops of this type are encountered
10e2eb10 3173by the peephole optimizer. I<o> is the OP that needs optimizing;
9733086d
BM
3174I<oldop> is the previous OP optimized, whose C<op_next> points to I<o>.
3175
3176=back
53e06cf0 3177
e7d4c058 3178C<B::Generate> directly supports the creation of custom ops by name.
53e06cf0 3179
bf5e9371
DM
3180
3181=head1 Dynamic Scope and the Context Stack
3182
3183B<Note:> this section describes a non-public internal API that is subject
3184to change without notice.
3185
3186=head2 Introduction to the context stack
3187
3188In Perl, dynamic scoping refers to the runtime nesting of things like
3189subroutine calls, evals etc, as well as the entering and exiting of block
3190scopes. For example, the restoring of a C<local>ised variable is
3191determined by the dynamic scope.
3192
3193Perl tracks the dynamic scope by a data structure called the context
3194stack, which is an array of C<PERL_CONTEXT> structures, and which is
3195itself a big union for all the types of context. Whenever a new scope is
3196entered (such as a block, a C<for> loop, or a subroutine call), a new
3197context entry is pushed onto the stack. Similarly when leaving a block or
3198returning from a subroutine call etc. a context is popped. Since the
3199context stack represents the current dynamic scope, it can be searched.
3200For example, C<next LABEL> searches back through the stack looking for a
3201loop context that matches the label; C<return> pops contexts until it
3202finds a sub or eval context or similar; C<caller> examines sub contexts on
3203the stack.
3204
3205Each context entry is labelled with a context type, C<cx_type>. Typical
3206context types are C<CXt_SUB>, C<CXt_EVAL> etc., as well as C<CXt_BLOCK>
3207and C<CXt_NULL> which represent a basic scope (as pushed by C<pp_enter>)
3208and a sort block. The type determines which part of the context union are
3209valid.
3210
3211The main division in the context struct is between a substitution scope
3212(C<CXt_SUBST>) and block scopes, which are everything else. The former is
d7c7f8cb 3213just used while executing C<s///e>, and won't be discussed further
bf5e9371
DM
3214here.
3215
3216All the block scope types share a common base, which corresponds to
3217C<CXt_BLOCK>. This stores the old values of various scope-related
3218variables like C<PL_curpm>, as well as information about the current
3219scope, such as C<gimme>. On scope exit, the old variables are restored.
3220
3221Particular block scope types store extra per-type information. For
3222example, C<CXt_SUB> stores the currently executing CV, while the various
3223for loop types might hold the original loop variable SV. On scope exit,
3224the per-type data is processed; for example the CV has its reference count
3225decremented, and the original loop variable is restored.
3226
3227The macro C<cxstack> returns the base of the current context stack, while
3228C<cxstack_ix> is the index of the current frame within that stack.
3229
3230In fact, the context stack is actually part of a stack-of-stacks system;
3231whenever something unusual is done such as calling a C<DESTROY> or tie
3232handler, a new stack is pushed, then popped at the end.
3233
3234Note that the API described here changed considerably in perl 5.24; prior
3235to that, big macros like C<PUSHBLOCK> and C<POPSUB> were used; in 5.24
3236they were replaced by the inline static functions described below. In
3237addition, the ordering and detail of how these macros/function work
3238changed in many ways, often subtly. In particular they didn't handle
3239saving the savestack and temps stack positions, and required additional
3240C<ENTER>, C<SAVETMPS> and C<LEAVE> compared to the new functions. The
3241old-style macros will not be described further.
3242
3243
3244=head2 Pushing contexts
3245
3246For pushing a new context, the two basic functions are
3247C<cx = cx_pushblock()>, which pushes a new basic context block and returns
3248its address, and a family of similar functions with names like
3249C<cx_pushsub(cx)> which populate the additional type-dependent fields in
3250the C<cx> struct. Note that C<CXt_NULL> and C<CXt_BLOCK> don't have their
3251own push functions, as they don't store any data beyond that pushed by
3252C<cx_pushblock>.
3253
3254The fields of the context struct and the arguments to the C<cx_*>
3255functions are subject to change between perl releases, representing
3256whatever is convenient or efficient for that release.
3257
3258A typical context stack pushing can be found in C<pp_entersub>; the
3259following shows a simplified and stripped-down example of a non-XS call,
3260along with comments showing roughly what each function does.
3261
61f554bd
KW
3262 dMARK;
3263 U8 gimme = GIMME_V;
3264 bool hasargs = cBOOL(PL_op->op_flags & OPf_STACKED);
3265 OP *retop = PL_op->op_next;
3266 I32 old_ss_ix = PL_savestack_ix;
3267 CV *cv = ....;
3268
3269 /* ... make mortal copies of stack args which are PADTMPs here ... */
3270
3271 /* ... do any additional savestack pushes here ... */
3272
3273 /* Now push a new context entry of type 'CXt_SUB'; initially just
3274 * doing the actions common to all block types: */
3275
3276 cx = cx_pushblock(CXt_SUB, gimme, MARK, old_ss_ix);
3277
3278 /* this does (approximately):
3279 CXINC; /* cxstack_ix++ (grow if necessary) */
3280 cx = CX_CUR(); /* and get the address of new frame */
3281 cx->cx_type = CXt_SUB;
3282 cx->blk_gimme = gimme;
3283 cx->blk_oldsp = MARK - PL_stack_base;
3284 cx->blk_oldsaveix = old_ss_ix;
3285 cx->blk_oldcop = PL_curcop;
3286 cx->blk_oldmarksp = PL_markstack_ptr - PL_markstack;
3287 cx->blk_oldscopesp = PL_scopestack_ix;
3288 cx->blk_oldpm = PL_curpm;
3289 cx->blk_old_tmpsfloor = PL_tmps_floor;
3290
3291 PL_tmps_floor = PL_tmps_ix;
3292 */
bf5e9371
DM
3293
3294
61f554bd
KW
3295 /* then update the new context frame with subroutine-specific info,
3296 * such as the CV about to be executed: */
bf5e9371 3297
61f554bd 3298 cx_pushsub(cx, cv, retop, hasargs);
bf5e9371 3299
61f554bd
KW
3300 /* this does (approximately):
3301 cx->blk_sub.cv = cv;
3302 cx->blk_sub.olddepth = CvDEPTH(cv);
3303 cx->blk_sub.prevcomppad = PL_comppad;
3304 cx->cx_type |= (hasargs) ? CXp_HASARGS : 0;
3305 cx->blk_sub.retop = retop;
3306 SvREFCNT_inc_simple_void_NN(cv);
3307 */
bf5e9371
DM
3308
3309Note that C<cx_pushblock()> sets two new floors: for the args stack (to
3310C<MARK>) and the temps stack (to C<PL_tmps_ix>). While executing at this
3311scope level, every C<nextstate> (amongst others) will reset the args and
3312tmps stack levels to these floors. Note that since C<cx_pushblock> uses
3313the current value of C<PL_tmps_ix> rather than it being passed as an arg,
3314this dictates at what point C<cx_pushblock> should be called. In
3315particular, any new mortals which should be freed only on scope exit
3316(rather than at the next C<nextstate>) should be created first.
3317
3318Most callers of C<cx_pushblock> simply set the new args stack floor to the
3319top of the previous stack frame, but for C<CXt_LOOP_LIST> it stores the
3320items being iterated over on the stack, and so sets C<blk_oldsp> to the
3321top of these items instead. Note that, contrary to its name, C<blk_oldsp>
3322doesn't always represent the value to restore C<PL_stack_sp> to on scope
3323exit.
3324
3325Note the early capture of C<PL_savestack_ix> to C<old_ss_ix>, which is
3326later passed as an arg to C<cx_pushblock>. In the case of C<pp_entersub>,
3327this is because, although most values needing saving are stored in fields
3328of the context struct, an extra value needs saving only when the debugger
3329is running, and it doesn't make sense to bloat the struct for this rare
3330case. So instead it is saved on the savestack. Since this value gets
3331calculated and saved before the context is pushed, it is necessary to pass
3332the old value of C<PL_savestack_ix> to C<cx_pushblock>, to ensure that the
3333saved value gets freed during scope exit. For most users of
3334C<cx_pushblock>, where nothing needs pushing on the save stack,
3335C<PL_savestack_ix> is just passed directly as an arg to C<cx_pushblock>.
3336
3337Note that where possible, values should be saved in the context struct
3338rather than on the save stack; it's much faster that way.
3339
3340Normally C<cx_pushblock> should be immediately followed by the appropriate
3341C<cx_pushfoo>, with nothing between them; this is because if code
3342in-between could die (e.g. a warning upgraded to fatal), then the context
3343stack unwinding code in C<dounwind> would see (in the example above) a
3344C<CXt_SUB> context frame, but without all the subroutine-specific fields
3345set, and crashes would soon ensue.
3346
3347Where the two must be separate, initially set the type to C<CXt_NULL> or
3348C<CXt_BLOCK>, and later change it to C<CXt_foo> when doing the
3349C<cx_pushfoo>. This is exactly what C<pp_enteriter> does, once it's
3350determined which type of loop it's pushing.
3351
3352=head2 Popping contexts
3353
3354Contexts are popped using C<cx_popsub()> etc. and C<cx_popblock()>. Note
3355however, that unlike C<cx_pushblock>, neither of these functions actually
3356decrement the current context stack index; this is done separately using
3357C<CX_POP()>.
3358
3359There are two main ways that contexts are popped. During normal execution
3360as scopes are exited, functions like C<pp_leave>, C<pp_leaveloop> and
3361C<pp_leavesub> process and pop just one context using C<cx_popfoo> and
3362C<cx_popblock>. On the other hand, things like C<pp_return> and C<next>
3363may have to pop back several scopes until a sub or loop context is found,
3364and exceptions (such as C<die>) need to pop back contexts until an eval
3365context is found. Both of these are accomplished by C<dounwind()>, which
3366is capable of processing and popping all contexts above the target one.
3367
3368Here is a typical example of context popping, as found in C<pp_leavesub>
3369(simplified slightly):
3370
61f554bd
KW
3371 U8 gimme;
3372 PERL_CONTEXT *cx;
3373 SV **oldsp;
3374 OP *retop;
bf5e9371 3375
61f554bd 3376 cx = CX_CUR();
bf5e9371 3377
61f554bd
KW
3378 gimme = cx->blk_gimme;
3379 oldsp = PL_stack_base + cx->blk_oldsp; /* last arg of previous frame */
bf5e9371 3380
61f554bd
KW
3381 if (gimme == G_VOID)
3382 PL_stack_sp = oldsp;
3383 else
3384 leave_adjust_stacks(oldsp, oldsp, gimme, 0);
bf5e9371 3385
61f554bd
KW
3386 CX_LEAVE_SCOPE(cx);
3387 cx_popsub(cx);
3388 cx_popblock(cx);
3389 retop = cx->blk_sub.retop;
3390 CX_POP(cx);
bf5e9371 3391
61f554bd 3392 return retop;
bf5e9371
DM
3393
3394The steps above are in a very specific order, designed to be the reverse
3395order of when the context was pushed. The first thing to do is to copy
3396and/or protect any any return arguments and free any temps in the current
3397scope. Scope exits like an rvalue sub normally return a mortal copy of
3398their return args (as opposed to lvalue subs). It is important to make
3399this copy before the save stack is popped or variables are restored, or
3400bad things like the following can happen:
3401
3402 sub f { my $x =...; $x } # $x freed before we get to copy it
3403 sub f { /(...)/; $1 } # PL_curpm restored before $1 copied
3404
3405Although we wish to free any temps at the same time, we have to be careful
3406not to free any temps which are keeping return args alive; nor to free the
3407temps we have just created while mortal copying return args. Fortunately,
3408C<leave_adjust_stacks()> is capable of making mortal copies of return args,
3409shifting args down the stack, and only processing those entries on the
3410temps stack that are safe to do so.
3411
3412In void context no args are returned, so it's more efficient to skip
3413calling C<leave_adjust_stacks()>. Also in void context, a C<nextstate> op
3414is likely to be imminently called which will do a C<FREETMPS>, so there's
3415no need to do that either.
3416
3417The next step is to pop savestack entries: C<CX_LEAVE_SCOPE(cx)> is just
4b93f2ab 3418defined as C<< LEAVE_SCOPE(cx->blk_oldsaveix) >>. Note that during the
bf5e9371
DM
3419popping, it's possible for perl to call destructors, call C<STORE> to undo
3420localisations of tied vars, and so on. Any of these can die or call
3421C<exit()>. In this case, C<dounwind()> will be called, and the current
3422context stack frame will be re-processed. Thus it is vital that all steps
3423in popping a context are done in such a way to support reentrancy. The
3424other alternative, of decrementing C<cxstack_ix> I<before> processing the
3425frame, would lead to leaks and the like if something died halfway through,
3426or overwriting of the current frame.
3427
3428C<CX_LEAVE_SCOPE> itself is safely re-entrant: if only half the savestack
3429items have been popped before dying and getting trapped by eval, then the
3430C<CX_LEAVE_SCOPE>s in C<dounwind> or C<pp_leaveeval> will continue where
3431the first one left off.
3432
3433The next step is the type-specific context processing; in this case
3434C<cx_popsub>. In part, this looks like:
3435
3436 cv = cx->blk_sub.cv;
3437 CvDEPTH(cv) = cx->blk_sub.olddepth;
3438 cx->blk_sub.cv = NULL;
3439 SvREFCNT_dec(cv);
3440
3441where its processing the just-executed CV. Note that before it decrements
3442the CV's reference count, it nulls the C<blk_sub.cv>. This means that if
3443it re-enters, the CV won't be freed twice. It also means that you can't
3444rely on such type-specific fields having useful values after the return
3445from C<cx_popfoo>.
3446
3447Next, C<cx_popblock> restores all the various interpreter vars to their
3448previous values or previous high water marks; it expands to:
3449
3450 PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
3451 PL_scopestack_ix = cx->blk_oldscopesp;
3452 PL_curpm = cx->blk_oldpm;
3453 PL_curcop = cx->blk_oldcop;
3454 PL_tmps_floor = cx->blk_old_tmpsfloor;
3455
3456Note that it I<doesn't> restore C<PL_stack_sp>; as mentioned earlier,
3457which value to restore it to depends on the context type (specifically
3458C<for (list) {}>), and what args (if any) it returns; and that will
3459already have been sorted out earlier by C<leave_adjust_stacks()>.
3460
3461Finally, the context stack pointer is actually decremented by C<CX_POP(cx)>.
3462After this point, it's possible that that the current context frame could
3463be overwritten by other contexts being pushed. Although things like ties
3464and C<DESTROY> are supposed to work within a new context stack, it's best
3465not to assume this. Indeed on debugging builds, C<CX_POP(cx)> deliberately
3466sets C<cx> to null to detect code that is still relying on the field
3467values in that context frame. Note in the C<pp_leavesub()> example above,
3468we grab C<blk_sub.retop> I<before> calling C<CX_POP>.
3469
3470=head2 Redoing contexts
3471
3472Finally, there is C<cx_topblock(cx)>, which acts like a super-C<nextstate>
3473as regards to resetting various vars to their base values. It is used in
3474places like C<pp_next>, C<pp_redo> and C<pp_goto> where rather than
3475exiting a scope, we want to re-initialise the scope. As well as resetting
3476C<PL_stack_sp> like C<nextstate>, it also resets C<PL_markstack_ptr>,
3477C<PL_scopestack_ix> and C<PL_curpm>. Note that it doesn't do a
3478C<FREETMPS>.
3479
3480
091ff1b6
AC
3481=head1 Slab-based operator allocation
3482
3483B<Note:> this section describes a non-public internal API that is subject
3484to change without notice.
3485
3486Perl's internal error-handling mechanisms implement C<die> (and its internal
3487equivalents) using longjmp. If this occurs during lexing, parsing or
3488compilation, we must ensure that any ops allocated as part of the compilation
3489process are freed. (Older Perl versions did not adequately handle this
3490situation: when failing a parse, they would leak ops that were stored in
3491C C<auto> variables and not linked anywhere else.)
3492
3493To handle this situation, Perl uses I<op slabs> that are attached to the
3494currently-compiling CV. A slab is a chunk of allocated memory. New ops are
3495allocated as regions of the slab. If the slab fills up, a new one is created
3496(and linked from the previous one). When an error occurs and the CV is freed,
3497any ops remaining are freed.
3498
3499Each op is preceded by two pointers: one points to the next op in the slab, and
3500the other points to the slab that owns it. The next-op pointer is needed so
3501that Perl can iterate over a slab and free all its ops. (Op structures are of
3502different sizes, so the slab's ops can't merely be treated as a dense array.)
3503The slab pointer is needed for accessing a reference count on the slab: when