1PERLGUTS(1)            Perl Programmers Reference Guide            PERLGUTS(1)
2
3
4

NAME

6       perlguts - Introduction to the Perl API
7

DESCRIPTION

9       This document attempts to describe how to use the Perl API, as well as
10       to provide some info on the basic workings of the Perl core.  It is far
11       from complete and probably contains many errors.  Please refer any
12       questions or comments to the author below.
13

Variables

15   Datatypes
16       Perl 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
22       Each typedef has specific routines that manipulate the various data
23       types.
24
25   What is an "IV"?
26       Perl uses a special typedef IV which is a simple signed integer type
27       that is guaranteed to be large enough to hold a pointer (as well as an
28       integer).  Additionally, there is the UV, which is simply an unsigned
29       IV.
30
31       Perl also uses two special typedefs, I32 and I16, which will always be
32       at least 32-bits and 16-bits long, respectively.  (Again, there are U32
33       and U16, as well.)  They will usually be exactly 32 and 16 bits long,
34       but on Crays they will both be 64 bits.
35
36   Working with SVs
37       An SV can be created and loaded with one command.  There are five types
38       of values that can be loaded: an integer value (IV), an unsigned
39       integer value (UV), a double (NV), a string (PV), and another scalar
40       (SV).  ("PV" stands for "Pointer Value".  You might think that it is
41       misnamed because it is described as pointing only to strings.  However,
42       it is possible to have it point to other things.  For example, it could
43       point to an array of UVs.  But, using it for non-strings requires care,
44       as the underlying assumption of much of the internals is that PVs are
45       just for strings.  Often, for example, a trailing "NUL" is tacked on
46       automatically.  The non-string use is documented only in this
47       paragraph.)
48
49       The seven routines are:
50
51           SV*  newSViv(IV);
52           SV*  newSVuv(UV);
53           SV*  newSVnv(double);
54           SV*  newSVpv(const char*, STRLEN);
55           SV*  newSVpvn(const char*, STRLEN);
56           SV*  newSVpvf(const char*, ...);
57           SV*  newSVsv(SV*);
58
59       "STRLEN" is an integer type ("Size_t", usually defined as "size_t" in
60       config.h) guaranteed to be large enough to represent the size of any
61       string that perl can handle.
62
63       In the unlikely case of a SV requiring more complex initialization, you
64       can create an empty SV with newSV(len).  If "len" is 0 an empty SV of
65       type NULL is returned, else an SV of type PV is returned with len + 1
66       (for the "NUL") bytes of storage allocated, accessible via SvPVX.  In
67       both cases the SV has the undef value.
68
69           SV *sv = newSV(0);   /* no storage allocated  */
70           SV *sv = newSV(10);  /* 10 (+1) bytes of uninitialised storage
71                                 * allocated */
72
73       To change the value of an already-existing SV, there are eight
74       routines:
75
76           void  sv_setiv(SV*, IV);
77           void  sv_setuv(SV*, UV);
78           void  sv_setnv(SV*, double);
79           void  sv_setpv(SV*, const char*);
80           void  sv_setpvn(SV*, const char*, STRLEN)
81           void  sv_setpvf(SV*, const char*, ...);
82           void  sv_vsetpvfn(SV*, const char*, STRLEN, va_list *,
83                                               SV **, Size_t, bool *);
84           void  sv_setsv(SV*, SV*);
85
86       Notice that you can choose to specify the length of the string to be
87       assigned by using "sv_setpvn", "newSVpvn", or "newSVpv", or you may
88       allow Perl to calculate the length by using "sv_setpv" or by specifying
89       0 as the second argument to "newSVpv".  Be warned, though, that Perl
90       will determine the string's length by using "strlen", which depends on
91       the string terminating with a "NUL" character, and not otherwise
92       containing NULs.
93
94       The arguments of "sv_setpvf" are processed like "sprintf", and the
95       formatted output becomes the value.
96
97       "sv_vsetpvfn" is an analogue of "vsprintf", but it allows you to
98       specify either a pointer to a variable argument list or the address and
99       length of an array of SVs.  The last argument points to a boolean; on
100       return, if that boolean is true, then locale-specific information has
101       been used to format the string, and the string's contents are therefore
102       untrustworthy (see perlsec).  This pointer may be NULL if that
103       information is not important.  Note that this function requires you to
104       specify the length of the format.
105
106       The "sv_set*()" functions are not generic enough to operate on values
107       that have "magic".  See "Magic Virtual Tables" later in this document.
108
109       All SVs that contain strings should be terminated with a "NUL"
110       character.  If it is not "NUL"-terminated there is a risk of core dumps
111       and corruptions from code which passes the string to C functions or
112       system calls which expect a "NUL"-terminated string.  Perl's own
113       functions typically add a trailing "NUL" for this reason.
114       Nevertheless, you should be very careful when you pass a string stored
115       in an SV to a C function or system call.
116
117       To access the actual value that an SV points to, you can use the
118       macros:
119
120           SvIV(SV*)
121           SvUV(SV*)
122           SvNV(SV*)
123           SvPV(SV*, STRLEN len)
124           SvPV_nolen(SV*)
125
126       which will automatically coerce the actual scalar type into an IV, UV,
127       double, or string.
128
129       In the "SvPV" macro, the length of the string returned is placed into
130       the variable "len" (this is a macro, so you do not use &len).  If you
131       do not care what the length of the data is, use the "SvPV_nolen" macro.
132       Historically the "SvPV" macro with the global variable "PL_na" has been
133       used in this case.  But that can be quite inefficient because "PL_na"
134       must be accessed in thread-local storage in threaded Perl.  In any
135       case, remember that Perl allows arbitrary strings of data that may both
136       contain NULs and might not be terminated by a "NUL".
137
138       Also remember that C doesn't allow you to safely say "foo(SvPV(s, len),
139       len);".  It might work with your compiler, but it won't work for
140       everyone.  Break this sort of statement up into separate assignments:
141
142           SV *s;
143           STRLEN len;
144           char *ptr;
145           ptr = SvPV(s, len);
146           foo(ptr, len);
147
148       If you want to know if the scalar value is TRUE, you can use:
149
150           SvTRUE(SV*)
151
152       Although Perl will automatically grow strings for you, if you need to
153       force Perl to allocate more memory for your SV, you can use the macro
154
155           SvGROW(SV*, STRLEN newlen)
156
157       which will determine if more memory needs to be allocated.  If so, it
158       will call the function "sv_grow".  Note that "SvGROW" can only
159       increase, not decrease, the allocated memory of an SV and that it does
160       not automatically add space for the trailing "NUL" byte (perl's own
161       string functions typically do "SvGROW(sv, len + 1)").
162
163       If you want to write to an existing SV's buffer and set its value to a
164       string, use SvPV_force() or one of its variants to force the SV to be a
165       PV.  This will remove any of various types of non-stringness from the
166       SV while preserving the content of the SV in the PV.  This can be used,
167       for example, to append data from an API function to a buffer without
168       extra copying:
169
170           (void)SvPVbyte_force(sv, len);
171           s = SvGROW(sv, len + needlen + 1);
172           /* something that modifies up to needlen bytes at s+len, but
173              modifies newlen bytes
174                eg. newlen = read(fd, s + len, needlen);
175              ignoring errors for these examples
176            */
177           s[len + newlen] = '\0';
178           SvCUR_set(sv, len + newlen);
179           SvUTF8_off(sv);
180           SvSETMAGIC(sv);
181
182       If you already have the data in memory or if you want to keep your code
183       simple, you can use one of the sv_cat*() variants, such as sv_catpvn().
184       If you want to insert anywhere in the string you can use sv_insert() or
185       sv_insert_flags().
186
187       If you don't need the existing content of the SV, you can avoid some
188       copying with:
189
190           SvPVCLEAR(sv);
191           s = SvGROW(sv, needlen + 1);
192           /* something that modifies up to needlen bytes at s, but modifies
193              newlen bytes
194                eg. newlen = read(fd, s. needlen);
195            */
196           s[newlen] = '\0';
197           SvCUR_set(sv, newlen);
198           SvPOK_only(sv); /* also clears SVf_UTF8 */
199           SvSETMAGIC(sv);
200
201       Again, if you already have the data in memory or want to avoid the
202       complexity of the above, you can use sv_setpvn().
203
204       If you have a buffer allocated with Newx() and want to set that as the
205       SV's value, you can use sv_usepvn_flags().  That has some requirements
206       if you want to avoid perl re-allocating the buffer to fit the trailing
207       NUL:
208
209          Newx(buf, somesize+1, char);
210          /* ... fill in buf ... */
211          buf[somesize] = '\0';
212          sv_usepvn_flags(sv, buf, somesize, SV_SMAGIC | SV_HAS_TRAILING_NUL);
213          /* buf now belongs to perl, don't release it */
214
215       If you have an SV and want to know what kind of data Perl thinks is
216       stored in it, you can use the following macros to check the type of SV
217       you have.
218
219           SvIOK(SV*)
220           SvNOK(SV*)
221           SvPOK(SV*)
222
223       You can get and set the current length of the string stored in an SV
224       with the following macros:
225
226           SvCUR(SV*)
227           SvCUR_set(SV*, I32 val)
228
229       You can also get a pointer to the end of the string stored in the SV
230       with the macro:
231
232           SvEND(SV*)
233
234       But note that these last three macros are valid only if "SvPOK()" is
235       true.
236
237       If you want to append something to the end of string stored in an
238       "SV*", you can use the following functions:
239
240           void  sv_catpv(SV*, const char*);
241           void  sv_catpvn(SV*, const char*, STRLEN);
242           void  sv_catpvf(SV*, const char*, ...);
243           void  sv_vcatpvfn(SV*, const char*, STRLEN, va_list *, SV **,
244                                                                    I32, bool);
245           void  sv_catsv(SV*, SV*);
246
247       The first function calculates the length of the string to be appended
248       by using "strlen".  In the second, you specify the length of the string
249       yourself.  The third function processes its arguments like "sprintf"
250       and appends the formatted output.  The fourth function works like
251       "vsprintf".  You can specify the address and length of an array of SVs
252       instead of the va_list argument.  The fifth function extends the string
253       stored in the first SV with the string stored in the second SV.  It
254       also forces the second SV to be interpreted as a string.
255
256       The "sv_cat*()" functions are not generic enough to operate on values
257       that have "magic".  See "Magic Virtual Tables" later in this document.
258
259       If you know the name of a scalar variable, you can get a pointer to its
260       SV by using the following:
261
262           SV*  get_sv("package::varname", 0);
263
264       This returns NULL if the variable does not exist.
265
266       If you want to know if this variable (or any other SV) is actually
267       "defined", you can call:
268
269           SvOK(SV*)
270
271       The scalar "undef" value is stored in an SV instance called
272       "PL_sv_undef".
273
274       Its address can be used whenever an "SV*" is needed.  Make sure that
275       you don't try to compare a random sv with &PL_sv_undef.  For example
276       when interfacing Perl code, it'll work correctly for:
277
278         foo(undef);
279
280       But won't work when called as:
281
282         $x = undef;
283         foo($x);
284
285       So to repeat always use SvOK() to check whether an sv is defined.
286
287       Also you have to be careful when using &PL_sv_undef as a value in AVs
288       or HVs (see "AVs, HVs and undefined values").
289
290       There are also the two values "PL_sv_yes" and "PL_sv_no", which contain
291       boolean TRUE and FALSE values, respectively.  Like "PL_sv_undef", their
292       addresses can be used whenever an "SV*" is needed.
293
294       Do not be fooled into thinking that "(SV *) 0" is the same as
295       &PL_sv_undef.  Take this code:
296
297           SV* sv = (SV*) 0;
298           if (I-am-to-return-a-real-value) {
299                   sv = sv_2mortal(newSViv(42));
300           }
301           sv_setsv(ST(0), sv);
302
303       This code tries to return a new SV (which contains the value 42) if it
304       should return a real value, or undef otherwise.  Instead it has
305       returned a NULL pointer which, somewhere down the line, will cause a
306       segmentation violation, bus error, or just weird results.  Change the
307       zero to &PL_sv_undef in the first line and all will be well.
308
309       To free an SV that you've created, call "SvREFCNT_dec(SV*)".  Normally
310       this call is not necessary (see "Reference Counts and Mortality").
311
312   Offsets
313       Perl provides the function "sv_chop" to efficiently remove characters
314       from the beginning of a string; you give it an SV and a pointer to
315       somewhere inside the PV, and it discards everything before the pointer.
316       The efficiency comes by means of a little hack: instead of actually
317       removing the characters, "sv_chop" sets the flag "OOK" (offset OK) to
318       signal to other functions that the offset hack is in effect, and it
319       moves the PV pointer (called "SvPVX") forward by the number of bytes
320       chopped off, and adjusts "SvCUR" and "SvLEN" accordingly.  (A portion
321       of the space between the old and new PV pointers is used to store the
322       count of chopped bytes.)
323
324       Hence, at this point, the start of the buffer that we allocated lives
325       at "SvPVX(sv) - SvIV(sv)" in memory and the PV pointer is pointing into
326       the middle of this allocated storage.
327
328       This is best demonstrated by example.  Normally copy-on-write will
329       prevent the substitution from operator from using this hack, but if you
330       can craft a string for which copy-on-write is not possible, you can see
331       it in play.  In the current implementation, the final byte of a string
332       buffer is used as a copy-on-write reference count.  If the buffer is
333       not big enough, then copy-on-write is skipped.  First have a look at an
334       empty string:
335
336         % ./perl -Ilib -MDevel::Peek -le '$a=""; $a .= ""; Dump $a'
337         SV = PV(0x7ffb7c008a70) at 0x7ffb7c030390
338           REFCNT = 1
339           FLAGS = (POK,pPOK)
340           PV = 0x7ffb7bc05b50 ""\0
341           CUR = 0
342           LEN = 10
343
344       Notice here the LEN is 10.  (It may differ on your platform.)  Extend
345       the length of the string to one less than 10, and do a substitution:
346
347        % ./perl -Ilib -MDevel::Peek -le '$a=""; $a.="123456789"; $a=~s/.//; \
348                                                                   Dump($a)'
349        SV = PV(0x7ffa04008a70) at 0x7ffa04030390
350          REFCNT = 1
351          FLAGS = (POK,OOK,pPOK)
352          OFFSET = 1
353          PV = 0x7ffa03c05b61 ( "\1" . ) "23456789"\0
354          CUR = 8
355          LEN = 9
356
357       Here the number of bytes chopped off (1) is shown next as the OFFSET.
358       The portion of the string between the "real" and the "fake" beginnings
359       is shown in parentheses, and the values of "SvCUR" and "SvLEN" reflect
360       the fake beginning, not the real one.  (The first character of the
361       string buffer happens to have changed to "\1" here, not "1", because
362       the current implementation stores the offset count in the string
363       buffer.  This is subject to change.)
364
365       Something similar to the offset hack is performed on AVs to enable
366       efficient shifting and splicing off the beginning of the array; while
367       "AvARRAY" points to the first element in the array that is visible from
368       Perl, "AvALLOC" points to the real start of the C array.  These are
369       usually the same, but a "shift" operation can be carried out by
370       increasing "AvARRAY" by one and decreasing "AvFILL" and "AvMAX".
371       Again, the location of the real start of the C array only comes into
372       play when freeing the array.  See "av_shift" in av.c.
373
374   What's Really Stored in an SV?
375       Recall that the usual method of determining the type of scalar you have
376       is to use "Sv*OK" macros.  Because a scalar can be both a number and a
377       string, usually these macros will always return TRUE and calling the
378       "Sv*V" macros will do the appropriate conversion of string to
379       integer/double or integer/double to string.
380
381       If you really need to know if you have an integer, double, or string
382       pointer in an SV, you can use the following three macros instead:
383
384           SvIOKp(SV*)
385           SvNOKp(SV*)
386           SvPOKp(SV*)
387
388       These will tell you if you truly have an integer, double, or string
389       pointer stored in your SV.  The "p" stands for private.
390
391       There are various ways in which the private and public flags may
392       differ.  For example, in perl 5.16 and earlier a tied SV may have a
393       valid underlying value in the IV slot (so SvIOKp is true), but the data
394       should be accessed via the FETCH routine rather than directly, so SvIOK
395       is false.  (In perl 5.18 onwards, tied scalars use the flags the same
396       way as untied scalars.)  Another is when numeric conversion has
397       occurred and precision has been lost: only the private flag is set on
398       'lossy' values.  So when an NV is converted to an IV with loss, SvIOKp,
399       SvNOKp and SvNOK will be set, while SvIOK wont be.
400
401       In general, though, it's best to use the "Sv*V" macros.
402
403   Working with AVs
404       There are two ways to create and load an AV.  The first method creates
405       an empty AV:
406
407           AV*  newAV();
408
409       The second method both creates the AV and initially populates it with
410       SVs:
411
412           AV*  av_make(SSize_t num, SV **ptr);
413
414       The second argument points to an array containing "num" "SV*"'s.  Once
415       the AV has been created, the SVs can be destroyed, if so desired.
416
417       Once the AV has been created, the following operations are possible on
418       it:
419
420           void  av_push(AV*, SV*);
421           SV*   av_pop(AV*);
422           SV*   av_shift(AV*);
423           void  av_unshift(AV*, SSize_t num);
424
425       These should be familiar operations, with the exception of
426       "av_unshift".  This routine adds "num" elements at the front of the
427       array with the "undef" value.  You must then use "av_store" (described
428       below) to assign values to these new elements.
429
430       Here are some other functions:
431
432           SSize_t av_top_index(AV*);
433           SV**    av_fetch(AV*, SSize_t key, I32 lval);
434           SV**    av_store(AV*, SSize_t key, SV* val);
435
436       The "av_top_index" function returns the highest index value in an array
437       (just like $#array in Perl).  If the array is empty, -1 is returned.
438       The "av_fetch" function returns the value at index "key", but if "lval"
439       is non-zero, then "av_fetch" will store an undef value at that index.
440       The "av_store" function stores the value "val" at index "key", and does
441       not increment the reference count of "val".  Thus the caller is
442       responsible for taking care of that, and if "av_store" returns NULL,
443       the caller will have to decrement the reference count to avoid a memory
444       leak.  Note that "av_fetch" and "av_store" both return "SV**"'s, not
445       "SV*"'s as their return value.
446
447       A few more:
448
449           void  av_clear(AV*);
450           void  av_undef(AV*);
451           void  av_extend(AV*, SSize_t key);
452
453       The "av_clear" function deletes all the elements in the AV* array, but
454       does not actually delete the array itself.  The "av_undef" function
455       will delete all the elements in the array plus the array itself.  The
456       "av_extend" function extends the array so that it contains at least
457       "key+1" elements.  If "key+1" is less than the currently allocated
458       length of the array, then nothing is done.
459
460       If you know the name of an array variable, you can get a pointer to its
461       AV by using the following:
462
463           AV*  get_av("package::varname", 0);
464
465       This returns NULL if the variable does not exist.
466
467       See "Understanding the Magic of Tied Hashes and Arrays" for more
468       information on how to use the array access functions on tied arrays.
469
470   Working with HVs
471       To create an HV, you use the following routine:
472
473           HV*  newHV();
474
475       Once the HV has been created, the following operations are possible on
476       it:
477
478           SV**  hv_store(HV*, const char* key, U32 klen, SV* val, U32 hash);
479           SV**  hv_fetch(HV*, const char* key, U32 klen, I32 lval);
480
481       The "klen" parameter is the length of the key being passed in (Note
482       that you cannot pass 0 in as a value of "klen" to tell Perl to measure
483       the length of the key).  The "val" argument contains the SV pointer to
484       the scalar being stored, and "hash" is the precomputed hash value (zero
485       if you want "hv_store" to calculate it for you).  The "lval" parameter
486       indicates whether this fetch is actually a part of a store operation,
487       in which case a new undefined value will be added to the HV with the
488       supplied key and "hv_fetch" will return as if the value had already
489       existed.
490
491       Remember that "hv_store" and "hv_fetch" return "SV**"'s and not just
492       "SV*".  To access the scalar value, you must first dereference the
493       return value.  However, you should check to make sure that the return
494       value is not NULL before dereferencing it.
495
496       The first of these two functions checks if a hash table entry exists,
497       and the second deletes it.
498
499           bool  hv_exists(HV*, const char* key, U32 klen);
500           SV*   hv_delete(HV*, const char* key, U32 klen, I32 flags);
501
502       If "flags" does not include the "G_DISCARD" flag then "hv_delete" will
503       create and return a mortal copy of the deleted value.
504
505       And more miscellaneous functions:
506
507           void   hv_clear(HV*);
508           void   hv_undef(HV*);
509
510       Like their AV counterparts, "hv_clear" deletes all the entries in the
511       hash table but does not actually delete the hash table.  The "hv_undef"
512       deletes both the entries and the hash table itself.
513
514       Perl keeps the actual data in a linked list of structures with a
515       typedef of HE.  These contain the actual key and value pointers (plus
516       extra administrative overhead).  The key is a string pointer; the value
517       is an "SV*".  However, once you have an "HE*", to get the actual key
518       and value, use the routines specified below.
519
520           I32    hv_iterinit(HV*);
521                   /* Prepares starting point to traverse hash table */
522           HE*    hv_iternext(HV*);
523                   /* Get the next entry, and return a pointer to a
524                      structure that has both the key and value */
525           char*  hv_iterkey(HE* entry, I32* retlen);
526                   /* Get the key from an HE structure and also return
527                      the length of the key string */
528           SV*    hv_iterval(HV*, HE* entry);
529                   /* Return an SV pointer to the value of the HE
530                      structure */
531           SV*    hv_iternextsv(HV*, char** key, I32* retlen);
532                   /* This convenience routine combines hv_iternext,
533                      hv_iterkey, and hv_iterval.  The key and retlen
534                      arguments are return values for the key and its
535                      length.  The value is returned in the SV* argument */
536
537       If you know the name of a hash variable, you can get a pointer to its
538       HV by using the following:
539
540           HV*  get_hv("package::varname", 0);
541
542       This returns NULL if the variable does not exist.
543
544       The hash algorithm is defined in the "PERL_HASH" macro:
545
546           PERL_HASH(hash, key, klen)
547
548       The exact implementation of this macro varies by architecture and
549       version of perl, and the return value may change per invocation, so the
550       value is only valid for the duration of a single perl process.
551
552       See "Understanding the Magic of Tied Hashes and Arrays" for more
553       information on how to use the hash access functions on tied hashes.
554
555   Hash API Extensions
556       Beginning with version 5.004, the following functions are also
557       supported:
558
559           HE*     hv_fetch_ent  (HV* tb, SV* key, I32 lval, U32 hash);
560           HE*     hv_store_ent  (HV* tb, SV* key, SV* val, U32 hash);
561
562           bool    hv_exists_ent (HV* tb, SV* key, U32 hash);
563           SV*     hv_delete_ent (HV* tb, SV* key, I32 flags, U32 hash);
564
565           SV*     hv_iterkeysv  (HE* entry);
566
567       Note that these functions take "SV*" keys, which simplifies writing of
568       extension code that deals with hash structures.  These functions also
569       allow passing of "SV*" keys to "tie" functions without forcing you to
570       stringify the keys (unlike the previous set of functions).
571
572       They also return and accept whole hash entries ("HE*"), making their
573       use more efficient (since the hash number for a particular string
574       doesn't have to be recomputed every time).  See perlapi for detailed
575       descriptions.
576
577       The following macros must always be used to access the contents of hash
578       entries.  Note that the arguments to these macros must be simple
579       variables, since they may get evaluated more than once.  See perlapi
580       for detailed descriptions of these macros.
581
582           HePV(HE* he, STRLEN len)
583           HeVAL(HE* he)
584           HeHASH(HE* he)
585           HeSVKEY(HE* he)
586           HeSVKEY_force(HE* he)
587           HeSVKEY_set(HE* he, SV* sv)
588
589       These two lower level macros are defined, but must only be used when
590       dealing with keys that are not "SV*"s:
591
592           HeKEY(HE* he)
593           HeKLEN(HE* he)
594
595       Note that both "hv_store" and "hv_store_ent" do not increment the
596       reference count of the stored "val", which is the caller's
597       responsibility.  If these functions return a NULL value, the caller
598       will usually have to decrement the reference count of "val" to avoid a
599       memory leak.
600
601   AVs, HVs and undefined values
602       Sometimes you have to store undefined values in AVs or HVs.  Although
603       this may be a rare case, it can be tricky.  That's because you're used
604       to using &PL_sv_undef if you need an undefined SV.
605
606       For example, intuition tells you that this XS code:
607
608           AV *av = newAV();
609           av_store( av, 0, &PL_sv_undef );
610
611       is equivalent to this Perl code:
612
613           my @av;
614           $av[0] = undef;
615
616       Unfortunately, this isn't true.  In perl 5.18 and earlier, AVs use
617       &PL_sv_undef as a marker for indicating that an array element has not
618       yet been initialized.  Thus, "exists $av[0]" would be true for the
619       above Perl code, but false for the array generated by the XS code.  In
620       perl 5.20, storing &PL_sv_undef will create a read-only element,
621       because the scalar &PL_sv_undef itself is stored, not a copy.
622
623       Similar problems can occur when storing &PL_sv_undef in HVs:
624
625           hv_store( hv, "key", 3, &PL_sv_undef, 0 );
626
627       This will indeed make the value "undef", but if you try to modify the
628       value of "key", you'll get the following error:
629
630           Modification of non-creatable hash value attempted
631
632       In perl 5.8.0, &PL_sv_undef was also used to mark placeholders in
633       restricted hashes.  This caused such hash entries not to appear when
634       iterating over the hash or when checking for the keys with the
635       "hv_exists" function.
636
637       You can run into similar problems when you store &PL_sv_yes or
638       &PL_sv_no into AVs or HVs.  Trying to modify such elements will give
639       you the following error:
640
641           Modification of a read-only value attempted
642
643       To make a long story short, you can use the special variables
644       &PL_sv_undef, &PL_sv_yes and &PL_sv_no with AVs and HVs, but you have
645       to make sure you know what you're doing.
646
647       Generally, if you want to store an undefined value in an AV or HV, you
648       should not use &PL_sv_undef, but rather create a new undefined value
649       using the "newSV" function, for example:
650
651           av_store( av, 42, newSV(0) );
652           hv_store( hv, "foo", 3, newSV(0), 0 );
653
654   References
655       References are a special type of scalar that point to other data types
656       (including other references).
657
658       To create a reference, use either of the following functions:
659
660           SV* newRV_inc((SV*) thing);
661           SV* newRV_noinc((SV*) thing);
662
663       The "thing" argument can be any of an "SV*", "AV*", or "HV*".  The
664       functions are identical except that "newRV_inc" increments the
665       reference count of the "thing", while "newRV_noinc" does not.  For
666       historical reasons, "newRV" is a synonym for "newRV_inc".
667
668       Once you have a reference, you can use the following macro to
669       dereference the reference:
670
671           SvRV(SV*)
672
673       then call the appropriate routines, casting the returned "SV*" to
674       either an "AV*" or "HV*", if required.
675
676       To determine if an SV is a reference, you can use the following macro:
677
678           SvROK(SV*)
679
680       To discover what type of value the reference refers to, use the
681       following macro and then check the return value.
682
683           SvTYPE(SvRV(SV*))
684
685       The most useful types that will be returned are:
686
687           SVt_PVAV    Array
688           SVt_PVHV    Hash
689           SVt_PVCV    Code
690           SVt_PVGV    Glob (possibly a file handle)
691
692       Any numerical value returned which is less than SVt_PVAV will be a
693       scalar of some form.
694
695       See "svtype" in perlapi for more details.
696
697   Blessed References and Class Objects
698       References are also used to support object-oriented programming.  In
699       perl's OO lexicon, an object is simply a reference that has been
700       blessed into a package (or class).  Once blessed, the programmer may
701       now use the reference to access the various methods in the class.
702
703       A reference can be blessed into a package with the following function:
704
705           SV* sv_bless(SV* sv, HV* stash);
706
707       The "sv" argument must be a reference value.  The "stash" argument
708       specifies which class the reference will belong to.  See "Stashes and
709       Globs" for information on converting class names into stashes.
710
711       /* Still under construction */
712
713       The following function upgrades rv to reference if not already one.
714       Creates a new SV for rv to point to.  If "classname" is non-null, the
715       SV is blessed into the specified class.  SV is returned.
716
717               SV* newSVrv(SV* rv, const char* classname);
718
719       The following three functions copy integer, unsigned integer or double
720       into an SV whose reference is "rv".  SV is blessed if "classname" is
721       non-null.
722
723               SV* sv_setref_iv(SV* rv, const char* classname, IV iv);
724               SV* sv_setref_uv(SV* rv, const char* classname, UV uv);
725               SV* sv_setref_nv(SV* rv, const char* classname, NV iv);
726
727       The following function copies the pointer value (the address, not the
728       string!) into an SV whose reference is rv.  SV is blessed if
729       "classname" is non-null.
730
731               SV* sv_setref_pv(SV* rv, const char* classname, void* pv);
732
733       The following function copies a string into an SV whose reference is
734       "rv".  Set length to 0 to let Perl calculate the string length.  SV is
735       blessed if "classname" is non-null.
736
737           SV* sv_setref_pvn(SV* rv, const char* classname, char* pv,
738                                                                STRLEN length);
739
740       The following function tests whether the SV is blessed into the
741       specified class.  It does not check inheritance relationships.
742
743               int  sv_isa(SV* sv, const char* name);
744
745       The following function tests whether the SV is a reference to a blessed
746       object.
747
748               int  sv_isobject(SV* sv);
749
750       The following function tests whether the SV is derived from the
751       specified class.  SV can be either a reference to a blessed object or a
752       string containing a class name.  This is the function implementing the
753       "UNIVERSAL::isa" functionality.
754
755               bool sv_derived_from(SV* sv, const char* name);
756
757       To check if you've got an object derived from a specific class you have
758       to write:
759
760               if (sv_isobject(sv) && sv_derived_from(sv, class)) { ... }
761
762   Creating New Variables
763       To create a new Perl variable with an undef value which can be accessed
764       from your Perl script, use the following routines, depending on the
765       variable type.
766
767           SV*  get_sv("package::varname", GV_ADD);
768           AV*  get_av("package::varname", GV_ADD);
769           HV*  get_hv("package::varname", GV_ADD);
770
771       Notice the use of GV_ADD as the second parameter.  The new variable can
772       now be set, using the routines appropriate to the data type.
773
774       There are additional macros whose values may be bitwise OR'ed with the
775       "GV_ADD" argument to enable certain extra features.  Those bits are:
776
777       GV_ADDMULTI
778           Marks the variable as multiply defined, thus preventing the:
779
780             Name <varname> used only once: possible typo
781
782           warning.
783
784       GV_ADDWARN
785           Issues the warning:
786
787             Had to create <varname> unexpectedly
788
789           if the variable did not exist before the function was called.
790
791       If you do not specify a package name, the variable is created in the
792       current package.
793
794   Reference Counts and Mortality
795       Perl uses a reference count-driven garbage collection mechanism.  SVs,
796       AVs, or HVs (xV for short in the following) start their life with a
797       reference count of 1.  If the reference count of an xV ever drops to 0,
798       then it will be destroyed and its memory made available for reuse.  At
799       the most basic internal level, reference counts can be manipulated with
800       the following macros:
801
802           int SvREFCNT(SV* sv);
803           SV* SvREFCNT_inc(SV* sv);
804           void SvREFCNT_dec(SV* sv);
805
806       (There are also suffixed versions of the increment and decrement
807       macros, for situations where the full generality of these basic macros
808       can be exchanged for some performance.)
809
810       However, the way a programmer should think about references is not so
811       much in terms of the bare reference count, but in terms of ownership of
812       references.  A reference to an xV can be owned by any of a variety of
813       entities: another xV, the Perl interpreter, an XS data structure, a
814       piece of running code, or a dynamic scope.  An xV generally does not
815       know what entities own the references to it; it only knows how many
816       references there are, which is the reference count.
817
818       To correctly maintain reference counts, it is essential to keep track
819       of what references the XS code is manipulating.  The programmer should
820       always know where a reference has come from and who owns it, and be
821       aware of any creation or destruction of references, and any transfers
822       of ownership.  Because ownership isn't represented explicitly in the xV
823       data structures, only the reference count need be actually maintained
824       by the code, and that means that this understanding of ownership is not
825       actually evident in the code.  For example, transferring ownership of a
826       reference from one owner to another doesn't change the reference count
827       at all, so may be achieved with no actual code.  (The transferring code
828       doesn't touch the referenced object, but does need to ensure that the
829       former owner knows that it no longer owns the reference, and that the
830       new owner knows that it now does.)
831
832       An xV that is visible at the Perl level should not become unreferenced
833       and thus be destroyed.  Normally, an object will only become
834       unreferenced when it is no longer visible, often by the same means that
835       makes it invisible.  For example, a Perl reference value (RV) owns a
836       reference to its referent, so if the RV is overwritten that reference
837       gets destroyed, and the no-longer-reachable referent may be destroyed
838       as a result.
839
840       Many functions have some kind of reference manipulation as part of
841       their purpose.  Sometimes this is documented in terms of ownership of
842       references, and sometimes it is (less helpfully) documented in terms of
843       changes to reference counts.  For example, the newRV_inc() function is
844       documented to create a new RV (with reference count 1) and increment
845       the reference count of the referent that was supplied by the caller.
846       This is best understood as creating a new reference to the referent,
847       which is owned by the created RV, and returning to the caller ownership
848       of the sole reference to the RV.  The newRV_noinc() function instead
849       does not increment the reference count of the referent, but the RV
850       nevertheless ends up owning a reference to the referent.  It is
851       therefore implied that the caller of "newRV_noinc()" is relinquishing a
852       reference to the referent, making this conceptually a more complicated
853       operation even though it does less to the data structures.
854
855       For example, imagine you want to return a reference from an XSUB
856       function.  Inside the XSUB routine, you create an SV which initially
857       has just a single reference, owned by the XSUB routine.  This reference
858       needs to be disposed of before the routine is complete, otherwise it
859       will leak, preventing the SV from ever being destroyed.  So to create
860       an RV referencing the SV, it is most convenient to pass the SV to
861       "newRV_noinc()", which consumes that reference.  Now the XSUB routine
862       no longer owns a reference to the SV, but does own a reference to the
863       RV, which in turn owns a reference to the SV.  The ownership of the
864       reference to the RV is then transferred by the process of returning the
865       RV from the XSUB.
866
867       There are some convenience functions available that can help with the
868       destruction of xVs.  These functions introduce the concept of
869       "mortality".  Much documentation speaks of an xV itself being mortal,
870       but this is misleading.  It is really a reference to an xV that is
871       mortal, and it is possible for there to be more than one mortal
872       reference to a single xV.  For a reference to be mortal means that it
873       is owned by the temps stack, one of perl's many internal stacks, which
874       will destroy that reference "a short time later".  Usually the "short
875       time later" is the end of the current Perl statement.  However, it gets
876       more complicated around dynamic scopes: there can be multiple sets of
877       mortal references hanging around at the same time, with different death
878       dates.  Internally, the actual determinant for when mortal xV
879       references are destroyed depends on two macros, SAVETMPS and FREETMPS.
880       See perlcall and perlxs and "Temporaries Stack" below for more details
881       on these macros.
882
883       Mortal references are mainly used for xVs that are placed on perl's
884       main stack.  The stack is problematic for reference tracking, because
885       it contains a lot of xV references, but doesn't own those references:
886       they are not counted.  Currently, there are many bugs resulting from
887       xVs being destroyed while referenced by the stack, because the stack's
888       uncounted references aren't enough to keep the xVs alive.  So when
889       putting an (uncounted) reference on the stack, it is vitally important
890       to ensure that there will be a counted reference to the same xV that
891       will last at least as long as the uncounted reference.  But it's also
892       important that that counted reference be cleaned up at an appropriate
893       time, and not unduly prolong the xV's life.  For there to be a mortal
894       reference is often the best way to satisfy this requirement, especially
895       if the xV was created especially to be put on the stack and would
896       otherwise be unreferenced.
897
898       To create a mortal reference, use the functions:
899
900           SV*  sv_newmortal()
901           SV*  sv_mortalcopy(SV*)
902           SV*  sv_2mortal(SV*)
903
904       "sv_newmortal()" creates an SV (with the undefined value) whose sole
905       reference is mortal.  "sv_mortalcopy()" creates an xV whose value is a
906       copy of a supplied xV and whose sole reference is mortal.
907       "sv_2mortal()" mortalises an existing xV reference: it transfers
908       ownership of a reference from the caller to the temps stack.  Because
909       "sv_newmortal" gives the new SV no value, it must normally be given one
910       via "sv_setpv", "sv_setiv", etc. :
911
912           SV *tmp = sv_newmortal();
913           sv_setiv(tmp, an_integer);
914
915       As that is multiple C statements it is quite common so see this idiom
916       instead:
917
918           SV *tmp = sv_2mortal(newSViv(an_integer));
919
920       The mortal routines are not just for SVs; AVs and HVs can be made
921       mortal by passing their address (type-casted to "SV*") to the
922       "sv_2mortal" or "sv_mortalcopy" routines.
923
924   Stashes and Globs
925       A stash is a hash that contains all variables that are defined within a
926       package.  Each key of the stash is a symbol name (shared by all the
927       different types of objects that have the same name), and each value in
928       the hash table is a GV (Glob Value).  This GV in turn contains
929       references to the various objects of that name, including (but not
930       limited to) the following:
931
932           Scalar Value
933           Array Value
934           Hash Value
935           I/O Handle
936           Format
937           Subroutine
938
939       There is a single stash called "PL_defstash" that holds the items that
940       exist in the "main" package.  To get at the items in other packages,
941       append the string "::" to the package name.  The items in the "Foo"
942       package are in the stash "Foo::" in PL_defstash.  The items in the
943       "Bar::Baz" package are in the stash "Baz::" in "Bar::"'s stash.
944
945       To get the stash pointer for a particular package, use the function:
946
947           HV*  gv_stashpv(const char* name, I32 flags)
948           HV*  gv_stashsv(SV*, I32 flags)
949
950       The first function takes a literal string, the second uses the string
951       stored in the SV.  Remember that a stash is just a hash table, so you
952       get back an "HV*".  The "flags" flag will create a new package if it is
953       set to GV_ADD.
954
955       The name that "gv_stash*v" wants is the name of the package whose
956       symbol table you want.  The default package is called "main".  If you
957       have multiply nested packages, pass their names to "gv_stash*v",
958       separated by "::" as in the Perl language itself.
959
960       Alternately, if you have an SV that is a blessed reference, you can
961       find out the stash pointer by using:
962
963           HV*  SvSTASH(SvRV(SV*));
964
965       then use the following to get the package name itself:
966
967           char*  HvNAME(HV* stash);
968
969       If you need to bless or re-bless an object you can use the following
970       function:
971
972           SV*  sv_bless(SV*, HV* stash)
973
974       where the first argument, an "SV*", must be a reference, and the second
975       argument is a stash.  The returned "SV*" can now be used in the same
976       way as any other SV.
977
978       For more information on references and blessings, consult perlref.
979
980   Double-Typed SVs
981       Scalar variables normally contain only one type of value, an integer,
982       double, pointer, or reference.  Perl will automatically convert the
983       actual scalar data from the stored type into the requested type.
984
985       Some scalar variables contain more than one type of scalar data.  For
986       example, the variable $! contains either the numeric value of "errno"
987       or its string equivalent from either "strerror" or "sys_errlist[]".
988
989       To force multiple data values into an SV, you must do two things: use
990       the "sv_set*v" routines to add the additional scalar type, then set a
991       flag so that Perl will believe it contains more than one type of data.
992       The four macros to set the flags are:
993
994               SvIOK_on
995               SvNOK_on
996               SvPOK_on
997               SvROK_on
998
999       The particular macro you must use depends on which "sv_set*v" routine
1000       you called first.  This is because every "sv_set*v" routine turns on
1001       only the bit for the particular type of data being set, and turns off
1002       all the rest.
1003
1004       For example, to create a new Perl variable called "dberror" that
1005       contains both the numeric and descriptive string error values, you
1006       could use the following code:
1007
1008           extern int  dberror;
1009           extern char *dberror_list;
1010
1011           SV* sv = get_sv("dberror", GV_ADD);
1012           sv_setiv(sv, (IV) dberror);
1013           sv_setpv(sv, dberror_list[dberror]);
1014           SvIOK_on(sv);
1015
1016       If the order of "sv_setiv" and "sv_setpv" had been reversed, then the
1017       macro "SvPOK_on" would need to be called instead of "SvIOK_on".
1018
1019   Read-Only Values
1020       In Perl 5.16 and earlier, copy-on-write (see the next section) shared a
1021       flag bit with read-only scalars.  So the only way to test whether
1022       "sv_setsv", etc., will raise a "Modification of a read-only value"
1023       error in those versions is:
1024
1025           SvREADONLY(sv) && !SvIsCOW(sv)
1026
1027       Under Perl 5.18 and later, SvREADONLY only applies to read-only
1028       variables, and, under 5.20, copy-on-write scalars can also be read-
1029       only, so the above check is incorrect.  You just want:
1030
1031           SvREADONLY(sv)
1032
1033       If you need to do this check often, define your own macro like this:
1034
1035           #if PERL_VERSION >= 18
1036           # define SvTRULYREADONLY(sv) SvREADONLY(sv)
1037           #else
1038           # define SvTRULYREADONLY(sv) (SvREADONLY(sv) && !SvIsCOW(sv))
1039           #endif
1040
1041   Copy on Write
1042       Perl implements a copy-on-write (COW) mechanism for scalars, in which
1043       string copies are not immediately made when requested, but are deferred
1044       until made necessary by one or the other scalar changing.  This is
1045       mostly transparent, but one must take care not to modify string buffers
1046       that are shared by multiple SVs.
1047
1048       You can test whether an SV is using copy-on-write with "SvIsCOW(sv)".
1049
1050       You can force an SV to make its own copy of its string buffer by
1051       calling "sv_force_normal(sv)" or SvPV_force_nolen(sv).
1052
1053       If you want to make the SV drop its string buffer, use
1054       "sv_force_normal_flags(sv, SV_COW_DROP_PV)" or simply "sv_setsv(sv,
1055       NULL)".
1056
1057       All of these functions will croak on read-only scalars (see the
1058       previous section for more on those).
1059
1060       To test that your code is behaving correctly and not modifying COW
1061       buffers, on systems that support mmap(2) (i.e., Unix) you can configure
1062       perl with "-Accflags=-DPERL_DEBUG_READONLY_COW" and it will turn buffer
1063       violations into crashes.  You will find it to be marvellously slow, so
1064       you may want to skip perl's own tests.
1065
1066   Magic Variables
1067       [This section still under construction.  Ignore everything here.  Post
1068       no bills.  Everything not permitted is forbidden.]
1069
1070       Any SV may be magical, that is, it has special features that a normal
1071       SV does not have.  These features are stored in the SV structure in a
1072       linked list of "struct magic"'s, typedef'ed to "MAGIC".
1073
1074           struct magic {
1075               MAGIC*      mg_moremagic;
1076               MGVTBL*     mg_virtual;
1077               U16         mg_private;
1078               char        mg_type;
1079               U8          mg_flags;
1080               I32         mg_len;
1081               SV*         mg_obj;
1082               char*       mg_ptr;
1083           };
1084
1085       Note this is current as of patchlevel 0, and could change at any time.
1086
1087   Assigning Magic
1088       Perl adds magic to an SV using the sv_magic function:
1089
1090         void sv_magic(SV* sv, SV* obj, int how, const char* name, I32 namlen);
1091
1092       The "sv" argument is a pointer to the SV that is to acquire a new
1093       magical feature.
1094
1095       If "sv" is not already magical, Perl uses the "SvUPGRADE" macro to
1096       convert "sv" to type "SVt_PVMG".  Perl then continues by adding new
1097       magic to the beginning of the linked list of magical features.  Any
1098       prior entry of the same type of magic is deleted.  Note that this can
1099       be overridden, and multiple instances of the same type of magic can be
1100       associated with an SV.
1101
1102       The "name" and "namlen" arguments are used to associate a string with
1103       the magic, typically the name of a variable.  "namlen" is stored in the
1104       "mg_len" field and if "name" is non-null then either a "savepvn" copy
1105       of "name" or "name" itself is stored in the "mg_ptr" field, depending
1106       on whether "namlen" is greater than zero or equal to zero respectively.
1107       As a special case, if "(name && namlen == HEf_SVKEY)" then "name" is
1108       assumed to contain an "SV*" and is stored as-is with its REFCNT
1109       incremented.
1110
1111       The sv_magic function uses "how" to determine which, if any, predefined
1112       "Magic Virtual Table" should be assigned to the "mg_virtual" field.
1113       See the "Magic Virtual Tables" section below.  The "how" argument is
1114       also stored in the "mg_type" field.  The value of "how" should be
1115       chosen from the set of macros "PERL_MAGIC_foo" found in perl.h.  Note
1116       that before these macros were added, Perl internals used to directly
1117       use character literals, so you may occasionally come across old code or
1118       documentation referring to 'U' magic rather than "PERL_MAGIC_uvar" for
1119       example.
1120
1121       The "obj" argument is stored in the "mg_obj" field of the "MAGIC"
1122       structure.  If it is not the same as the "sv" argument, the reference
1123       count of the "obj" object is incremented.  If it is the same, or if the
1124       "how" argument is "PERL_MAGIC_arylen", "PERL_MAGIC_regdatum",
1125       "PERL_MAGIC_regdata", or if it is a NULL pointer, then "obj" is merely
1126       stored, without the reference count being incremented.
1127
1128       See also "sv_magicext" in perlapi for a more flexible way to add magic
1129       to an SV.
1130
1131       There is also a function to add magic to an "HV":
1132
1133           void hv_magic(HV *hv, GV *gv, int how);
1134
1135       This simply calls "sv_magic" and coerces the "gv" argument into an
1136       "SV".
1137
1138       To remove the magic from an SV, call the function sv_unmagic:
1139
1140           int sv_unmagic(SV *sv, int type);
1141
1142       The "type" argument should be equal to the "how" value when the "SV"
1143       was initially made magical.
1144
1145       However, note that "sv_unmagic" removes all magic of a certain "type"
1146       from the "SV".  If you want to remove only certain magic of a "type"
1147       based on the magic virtual table, use "sv_unmagicext" instead:
1148
1149           int sv_unmagicext(SV *sv, int type, MGVTBL *vtbl);
1150
1151   Magic Virtual Tables
1152       The "mg_virtual" field in the "MAGIC" structure is a pointer to an
1153       "MGVTBL", which is a structure of function pointers and stands for
1154       "Magic Virtual Table" to handle the various operations that might be
1155       applied to that variable.
1156
1157       The "MGVTBL" has five (or sometimes eight) pointers to the following
1158       routine types:
1159
1160           int  (*svt_get)  (pTHX_ SV* sv, MAGIC* mg);
1161           int  (*svt_set)  (pTHX_ SV* sv, MAGIC* mg);
1162           U32  (*svt_len)  (pTHX_ SV* sv, MAGIC* mg);
1163           int  (*svt_clear)(pTHX_ SV* sv, MAGIC* mg);
1164           int  (*svt_free) (pTHX_ SV* sv, MAGIC* mg);
1165
1166           int  (*svt_copy) (pTHX_ SV *sv, MAGIC* mg, SV *nsv,
1167                                                 const char *name, I32 namlen);
1168           int  (*svt_dup)  (pTHX_ MAGIC *mg, CLONE_PARAMS *param);
1169           int  (*svt_local)(pTHX_ SV *nsv, MAGIC *mg);
1170
1171       This MGVTBL structure is set at compile-time in perl.h and there are
1172       currently 32 types.  These different structures contain pointers to
1173       various routines that perform additional actions depending on which
1174       function is being called.
1175
1176          Function pointer    Action taken
1177          ----------------    ------------
1178          svt_get             Do something before the value of the SV is
1179                              retrieved.
1180          svt_set             Do something after the SV is assigned a value.
1181          svt_len             Report on the SV's length.
1182          svt_clear           Clear something the SV represents.
1183          svt_free            Free any extra storage associated with the SV.
1184
1185          svt_copy            copy tied variable magic to a tied element
1186          svt_dup             duplicate a magic structure during thread cloning
1187          svt_local           copy magic to local value during 'local'
1188
1189       For instance, the MGVTBL structure called "vtbl_sv" (which corresponds
1190       to an "mg_type" of "PERL_MAGIC_sv") contains:
1191
1192           { magic_get, magic_set, magic_len, 0, 0 }
1193
1194       Thus, when an SV is determined to be magical and of type
1195       "PERL_MAGIC_sv", if a get operation is being performed, the routine
1196       "magic_get" is called.  All the various routines for the various
1197       magical types begin with "magic_".  NOTE: the magic routines are not
1198       considered part of the Perl API, and may not be exported by the Perl
1199       library.
1200
1201       The last three slots are a recent addition, and for source code
1202       compatibility they are only checked for if one of the three flags
1203       MGf_COPY, MGf_DUP or MGf_LOCAL is set in mg_flags.  This means that
1204       most code can continue declaring a vtable as a 5-element value.  These
1205       three are currently used exclusively by the threading code, and are
1206       highly subject to change.
1207
1208       The current kinds of Magic Virtual Tables are:
1209
1210        mg_type
1211        (old-style char and macro)   MGVTBL         Type of magic
1212        --------------------------   ------         -------------
1213        \0 PERL_MAGIC_sv             vtbl_sv        Special scalar variable
1214        #  PERL_MAGIC_arylen         vtbl_arylen    Array length ($#ary)
1215        %  PERL_MAGIC_rhash          (none)         Extra data for restricted
1216                                                    hashes
1217        *  PERL_MAGIC_debugvar       vtbl_debugvar  $DB::single, signal, trace
1218                                                    vars
1219        .  PERL_MAGIC_pos            vtbl_pos       pos() lvalue
1220        :  PERL_MAGIC_symtab         (none)         Extra data for symbol
1221                                                    tables
1222        <  PERL_MAGIC_backref        vtbl_backref   For weak ref data
1223        @  PERL_MAGIC_arylen_p       (none)         To move arylen out of XPVAV
1224        B  PERL_MAGIC_bm             vtbl_regexp    Boyer-Moore
1225                                                    (fast string search)
1226        c  PERL_MAGIC_overload_table vtbl_ovrld     Holds overload table
1227                                                    (AMT) on stash
1228        D  PERL_MAGIC_regdata        vtbl_regdata   Regex match position data
1229                                                    (@+ and @- vars)
1230        d  PERL_MAGIC_regdatum       vtbl_regdatum  Regex match position data
1231                                                    element
1232        E  PERL_MAGIC_env            vtbl_env       %ENV hash
1233        e  PERL_MAGIC_envelem        vtbl_envelem   %ENV hash element
1234        f  PERL_MAGIC_fm             vtbl_regexp    Formline
1235                                                    ('compiled' format)
1236        g  PERL_MAGIC_regex_global   vtbl_mglob     m//g target
1237        H  PERL_MAGIC_hints          vtbl_hints     %^H hash
1238        h  PERL_MAGIC_hintselem      vtbl_hintselem %^H hash element
1239        I  PERL_MAGIC_isa            vtbl_isa       @ISA array
1240        i  PERL_MAGIC_isaelem        vtbl_isaelem   @ISA array element
1241        k  PERL_MAGIC_nkeys          vtbl_nkeys     scalar(keys()) lvalue
1242        L  PERL_MAGIC_dbfile         (none)         Debugger %_<filename
1243        l  PERL_MAGIC_dbline         vtbl_dbline    Debugger %_<filename
1244                                                    element
1245        N  PERL_MAGIC_shared         (none)         Shared between threads
1246        n  PERL_MAGIC_shared_scalar  (none)         Shared between threads
1247        o  PERL_MAGIC_collxfrm       vtbl_collxfrm  Locale transformation
1248        P  PERL_MAGIC_tied           vtbl_pack      Tied array or hash
1249        p  PERL_MAGIC_tiedelem       vtbl_packelem  Tied array or hash element
1250        q  PERL_MAGIC_tiedscalar     vtbl_packelem  Tied scalar or handle
1251        r  PERL_MAGIC_qr             vtbl_regexp    Precompiled qr// regex
1252        S  PERL_MAGIC_sig            (none)         %SIG hash
1253        s  PERL_MAGIC_sigelem        vtbl_sigelem   %SIG hash element
1254        t  PERL_MAGIC_taint          vtbl_taint     Taintedness
1255        U  PERL_MAGIC_uvar           vtbl_uvar      Available for use by
1256                                                    extensions
1257        u  PERL_MAGIC_uvar_elem      (none)         Reserved for use by
1258                                                    extensions
1259        V  PERL_MAGIC_vstring        (none)         SV was vstring literal
1260        v  PERL_MAGIC_vec            vtbl_vec       vec() lvalue
1261        w  PERL_MAGIC_utf8           vtbl_utf8      Cached UTF-8 information
1262        x  PERL_MAGIC_substr         vtbl_substr    substr() lvalue
1263        Y  PERL_MAGIC_nonelem        vtbl_nonelem   Array element that does not
1264                                                    exist
1265        y  PERL_MAGIC_defelem        vtbl_defelem   Shadow "foreach" iterator
1266                                                    variable / smart parameter
1267                                                    vivification
1268        \  PERL_MAGIC_lvref          vtbl_lvref     Lvalue reference
1269                                                    constructor
1270        ]  PERL_MAGIC_checkcall      vtbl_checkcall Inlining/mutation of call
1271                                                    to this CV
1272        ~  PERL_MAGIC_ext            (none)         Available for use by
1273                                                    extensions
1274
1275       When an uppercase and lowercase letter both exist in the table, then
1276       the uppercase letter is typically used to represent some kind of
1277       composite type (a list or a hash), and the lowercase letter is used to
1278       represent an element of that composite type.  Some internals code makes
1279       use of this case relationship.  However, 'v' and 'V' (vec and v-string)
1280       are in no way related.
1281
1282       The "PERL_MAGIC_ext" and "PERL_MAGIC_uvar" magic types are defined
1283       specifically for use by extensions and will not be used by perl itself.
1284       Extensions can use "PERL_MAGIC_ext" magic to 'attach' private
1285       information to variables (typically objects).  This is especially
1286       useful because there is no way for normal perl code to corrupt this
1287       private information (unlike using extra elements of a hash object).
1288
1289       Similarly, "PERL_MAGIC_uvar" magic can be used much like tie() to call
1290       a C function any time a scalar's value is used or changed.  The
1291       "MAGIC"'s "mg_ptr" field points to a "ufuncs" structure:
1292
1293           struct ufuncs {
1294               I32 (*uf_val)(pTHX_ IV, SV*);
1295               I32 (*uf_set)(pTHX_ IV, SV*);
1296               IV uf_index;
1297           };
1298
1299       When the SV is read from or written to, the "uf_val" or "uf_set"
1300       function will be called with "uf_index" as the first arg and a pointer
1301       to the SV as the second.  A simple example of how to add
1302       "PERL_MAGIC_uvar" magic is shown below.  Note that the ufuncs structure
1303       is copied by sv_magic, so you can safely allocate it on the stack.
1304
1305           void
1306           Umagic(sv)
1307               SV *sv;
1308           PREINIT:
1309               struct ufuncs uf;
1310           CODE:
1311               uf.uf_val   = &my_get_fn;
1312               uf.uf_set   = &my_set_fn;
1313               uf.uf_index = 0;
1314               sv_magic(sv, 0, PERL_MAGIC_uvar, (char*)&uf, sizeof(uf));
1315
1316       Attaching "PERL_MAGIC_uvar" to arrays is permissible but has no effect.
1317
1318       For hashes there is a specialized hook that gives control over hash
1319       keys (but not values).  This hook calls "PERL_MAGIC_uvar" 'get' magic
1320       if the "set" function in the "ufuncs" structure is NULL.  The hook is
1321       activated whenever the hash is accessed with a key specified as an "SV"
1322       through the functions "hv_store_ent", "hv_fetch_ent", "hv_delete_ent",
1323       and "hv_exists_ent".  Accessing the key as a string through the
1324       functions without the "..._ent" suffix circumvents the hook.  See
1325       "GUTS" in Hash::Util::FieldHash for a detailed description.
1326
1327       Note that because multiple extensions may be using "PERL_MAGIC_ext" or
1328       "PERL_MAGIC_uvar" magic, it is important for extensions to take extra
1329       care to avoid conflict.  Typically only using the magic on objects
1330       blessed into the same class as the extension is sufficient.  For
1331       "PERL_MAGIC_ext" magic, it is usually a good idea to define an
1332       "MGVTBL", even if all its fields will be 0, so that individual "MAGIC"
1333       pointers can be identified as a particular kind of magic using their
1334       magic virtual table.  "mg_findext" provides an easy way to do that:
1335
1336           STATIC MGVTBL my_vtbl = { 0, 0, 0, 0, 0, 0, 0, 0 };
1337
1338           MAGIC *mg;
1339           if ((mg = mg_findext(sv, PERL_MAGIC_ext, &my_vtbl))) {
1340               /* this is really ours, not another module's PERL_MAGIC_ext */
1341               my_priv_data_t *priv = (my_priv_data_t *)mg->mg_ptr;
1342               ...
1343           }
1344
1345       Also note that the "sv_set*()" and "sv_cat*()" functions described
1346       earlier do not invoke 'set' magic on their targets.  This must be done
1347       by the user either by calling the "SvSETMAGIC()" macro after calling
1348       these functions, or by using one of the "sv_set*_mg()" or
1349       "sv_cat*_mg()" functions.  Similarly, generic C code must call the
1350       "SvGETMAGIC()" macro to invoke any 'get' magic if they use an SV
1351       obtained from external sources in functions that don't handle magic.
1352       See perlapi for a description of these functions.  For example, calls
1353       to the "sv_cat*()" functions typically need to be followed by
1354       "SvSETMAGIC()", but they don't need a prior "SvGETMAGIC()" since their
1355       implementation handles 'get' magic.
1356
1357   Finding Magic
1358           MAGIC *mg_find(SV *sv, int type); /* Finds the magic pointer of that
1359                                              * type */
1360
1361       This routine returns a pointer to a "MAGIC" structure stored in the SV.
1362       If the SV does not have that magical feature, "NULL" is returned.  If
1363       the SV has multiple instances of that magical feature, the first one
1364       will be returned.  "mg_findext" can be used to find a "MAGIC" structure
1365       of an SV based on both its magic type and its magic virtual table:
1366
1367           MAGIC *mg_findext(SV *sv, int type, MGVTBL *vtbl);
1368
1369       Also, if the SV passed to "mg_find" or "mg_findext" is not of type
1370       SVt_PVMG, Perl may core dump.
1371
1372           int mg_copy(SV* sv, SV* nsv, const char* key, STRLEN klen);
1373
1374       This routine checks to see what types of magic "sv" has.  If the
1375       mg_type field is an uppercase letter, then the mg_obj is copied to
1376       "nsv", but the mg_type field is changed to be the lowercase letter.
1377
1378   Understanding the Magic of Tied Hashes and Arrays
1379       Tied hashes and arrays are magical beasts of the "PERL_MAGIC_tied"
1380       magic type.
1381
1382       WARNING: As of the 5.004 release, proper usage of the array and hash
1383       access functions requires understanding a few caveats.  Some of these
1384       caveats are actually considered bugs in the API, to be fixed in later
1385       releases, and are bracketed with [MAYCHANGE] below.  If you find
1386       yourself actually applying such information in this section, be aware
1387       that the behavior may change in the future, umm, without warning.
1388
1389       The perl tie function associates a variable with an object that
1390       implements the various GET, SET, etc methods.  To perform the
1391       equivalent of the perl tie function from an XSUB, you must mimic this
1392       behaviour.  The code below carries out the necessary steps -- firstly
1393       it creates a new hash, and then creates a second hash which it blesses
1394       into the class which will implement the tie methods.  Lastly it ties
1395       the two hashes together, and returns a reference to the new tied hash.
1396       Note that the code below does NOT call the TIEHASH method in the MyTie
1397       class - see "Calling Perl Routines from within C Programs" for details
1398       on how to do this.
1399
1400           SV*
1401           mytie()
1402           PREINIT:
1403               HV *hash;
1404               HV *stash;
1405               SV *tie;
1406           CODE:
1407               hash = newHV();
1408               tie = newRV_noinc((SV*)newHV());
1409               stash = gv_stashpv("MyTie", GV_ADD);
1410               sv_bless(tie, stash);
1411               hv_magic(hash, (GV*)tie, PERL_MAGIC_tied);
1412               RETVAL = newRV_noinc(hash);
1413           OUTPUT:
1414               RETVAL
1415
1416       The "av_store" function, when given a tied array argument, merely
1417       copies the magic of the array onto the value to be "stored", using
1418       "mg_copy".  It may also return NULL, indicating that the value did not
1419       actually need to be stored in the array.  [MAYCHANGE] After a call to
1420       "av_store" on a tied array, the caller will usually need to call
1421       "mg_set(val)" to actually invoke the perl level "STORE" method on the
1422       TIEARRAY object.  If "av_store" did return NULL, a call to
1423       "SvREFCNT_dec(val)" will also be usually necessary to avoid a memory
1424       leak. [/MAYCHANGE]
1425
1426       The previous paragraph is applicable verbatim to tied hash access using
1427       the "hv_store" and "hv_store_ent" functions as well.
1428
1429       "av_fetch" and the corresponding hash functions "hv_fetch" and
1430       "hv_fetch_ent" actually return an undefined mortal value whose magic
1431       has been initialized using "mg_copy".  Note the value so returned does
1432       not need to be deallocated, as it is already mortal.  [MAYCHANGE] But
1433       you will need to call "mg_get()" on the returned value in order to
1434       actually invoke the perl level "FETCH" method on the underlying TIE
1435       object.  Similarly, you may also call "mg_set()" on the return value
1436       after possibly assigning a suitable value to it using "sv_setsv",
1437       which will invoke the "STORE" method on the TIE object. [/MAYCHANGE]
1438
1439       [MAYCHANGE] In other words, the array or hash fetch/store functions
1440       don't really fetch and store actual values in the case of tied arrays
1441       and hashes.  They merely call "mg_copy" to attach magic to the values
1442       that were meant to be "stored" or "fetched".  Later calls to "mg_get"
1443       and "mg_set" actually do the job of invoking the TIE methods on the
1444       underlying objects.  Thus the magic mechanism currently implements a
1445       kind of lazy access to arrays and hashes.
1446
1447       Currently (as of perl version 5.004), use of the hash and array access
1448       functions requires the user to be aware of whether they are operating
1449       on "normal" hashes and arrays, or on their tied variants.  The API may
1450       be changed to provide more transparent access to both tied and normal
1451       data types in future versions.  [/MAYCHANGE]
1452
1453       You would do well to understand that the TIEARRAY and TIEHASH
1454       interfaces are mere sugar to invoke some perl method calls while using
1455       the uniform hash and array syntax.  The use of this sugar imposes some
1456       overhead (typically about two to four extra opcodes per FETCH/STORE
1457       operation, in addition to the creation of all the mortal variables
1458       required to invoke the methods).  This overhead will be comparatively
1459       small if the TIE methods are themselves substantial, but if they are
1460       only a few statements long, the overhead will not be insignificant.
1461
1462   Localizing changes
1463       Perl has a very handy construction
1464
1465         {
1466           local $var = 2;
1467           ...
1468         }
1469
1470       This construction is approximately equivalent to
1471
1472         {
1473           my $oldvar = $var;
1474           $var = 2;
1475           ...
1476           $var = $oldvar;
1477         }
1478
1479       The biggest difference is that the first construction would reinstate
1480       the initial value of $var, irrespective of how control exits the block:
1481       "goto", "return", "die"/"eval", etc.  It is a little bit more efficient
1482       as well.
1483
1484       There is a way to achieve a similar task from C via Perl API: create a
1485       pseudo-block, and arrange for some changes to be automatically undone
1486       at the end of it, either explicit, or via a non-local exit (via die()).
1487       A block-like construct is created by a pair of "ENTER"/"LEAVE" macros
1488       (see "Returning a Scalar" in perlcall).  Such a construct may be
1489       created specially for some important localized task, or an existing one
1490       (like boundaries of enclosing Perl subroutine/block, or an existing
1491       pair for freeing TMPs) may be used.  (In the second case the overhead
1492       of additional localization must be almost negligible.)  Note that any
1493       XSUB is automatically enclosed in an "ENTER"/"LEAVE" pair.
1494
1495       Inside such a pseudo-block the following service is available:
1496
1497       "SAVEINT(int i)"
1498       "SAVEIV(IV i)"
1499       "SAVEI32(I32 i)"
1500       "SAVELONG(long i)"
1501           These macros arrange things to restore the value of integer
1502           variable "i" at the end of enclosing pseudo-block.
1503
1504       SAVESPTR(s)
1505       SAVEPPTR(p)
1506           These macros arrange things to restore the value of pointers "s"
1507           and "p".  "s" must be a pointer of a type which survives conversion
1508           to "SV*" and back, "p" should be able to survive conversion to
1509           "char*" and back.
1510
1511       "SAVEFREESV(SV *sv)"
1512           The refcount of "sv" will be decremented at the end of pseudo-
1513           block.  This is similar to "sv_2mortal" in that it is also a
1514           mechanism for doing a delayed "SvREFCNT_dec".  However, while
1515           "sv_2mortal" extends the lifetime of "sv" until the beginning of
1516           the next statement, "SAVEFREESV" extends it until the end of the
1517           enclosing scope.  These lifetimes can be wildly different.
1518
1519           Also compare "SAVEMORTALIZESV".
1520
1521       "SAVEMORTALIZESV(SV *sv)"
1522           Just like "SAVEFREESV", but mortalizes "sv" at the end of the
1523           current scope instead of decrementing its reference count.  This
1524           usually has the effect of keeping "sv" alive until the statement
1525           that called the currently live scope has finished executing.
1526
1527       "SAVEFREEOP(OP *op)"
1528           The "OP *" is op_free()ed at the end of pseudo-block.
1529
1530       SAVEFREEPV(p)
1531           The chunk of memory which is pointed to by "p" is Safefree()ed at
1532           the end of pseudo-block.
1533
1534       "SAVECLEARSV(SV *sv)"
1535           Clears a slot in the current scratchpad which corresponds to "sv"
1536           at the end of pseudo-block.
1537
1538       "SAVEDELETE(HV *hv, char *key, I32 length)"
1539           The key "key" of "hv" is deleted at the end of pseudo-block.  The
1540           string pointed to by "key" is Safefree()ed.  If one has a key in
1541           short-lived storage, the corresponding string may be reallocated
1542           like this:
1543
1544             SAVEDELETE(PL_defstash, savepv(tmpbuf), strlen(tmpbuf));
1545
1546       "SAVEDESTRUCTOR(DESTRUCTORFUNC_NOCONTEXT_t f, void *p)"
1547           At the end of pseudo-block the function "f" is called with the only
1548           argument "p".
1549
1550       "SAVEDESTRUCTOR_X(DESTRUCTORFUNC_t f, void *p)"
1551           At the end of pseudo-block the function "f" is called with the
1552           implicit context argument (if any), and "p".
1553
1554       "SAVESTACK_POS()"
1555           The current offset on the Perl internal stack (cf. "SP") is
1556           restored at the end of pseudo-block.
1557
1558       The following API list contains functions, thus one needs to provide
1559       pointers to the modifiable data explicitly (either C pointers, or
1560       Perlish "GV *"s).  Where the above macros take "int", a similar
1561       function takes "int *".
1562
1563       "SV* save_scalar(GV *gv)"
1564           Equivalent to Perl code "local $gv".
1565
1566       "AV* save_ary(GV *gv)"
1567       "HV* save_hash(GV *gv)"
1568           Similar to "save_scalar", but localize @gv and %gv.
1569
1570       "void save_item(SV *item)"
1571           Duplicates the current value of "SV". On the exit from the current
1572           "ENTER"/"LEAVE" pseudo-block the value of "SV" will be restored
1573           using the stored value.  It doesn't handle magic.  Use
1574           "save_scalar" if magic is affected.
1575
1576       "void save_list(SV **sarg, I32 maxsarg)"
1577           A variant of "save_item" which takes multiple arguments via an
1578           array "sarg" of "SV*" of length "maxsarg".
1579
1580       "SV* save_svref(SV **sptr)"
1581           Similar to "save_scalar", but will reinstate an "SV *".
1582
1583       "void save_aptr(AV **aptr)"
1584       "void save_hptr(HV **hptr)"
1585           Similar to "save_svref", but localize "AV *" and "HV *".
1586
1587       The "Alias" module implements localization of the basic types within
1588       the caller's scope.  People who are interested in how to localize
1589       things in the containing scope should take a look there too.
1590

Subroutines

1592   XSUBs and the Argument Stack
1593       The XSUB mechanism is a simple way for Perl programs to access C
1594       subroutines.  An XSUB routine will have a stack that contains the
1595       arguments from the Perl program, and a way to map from the Perl data
1596       structures to a C equivalent.
1597
1598       The stack arguments are accessible through the ST(n) macro, which
1599       returns the "n"'th stack argument.  Argument 0 is the first argument
1600       passed in the Perl subroutine call.  These arguments are "SV*", and can
1601       be used anywhere an "SV*" is used.
1602
1603       Most of the time, output from the C routine can be handled through use
1604       of the RETVAL and OUTPUT directives.  However, there are some cases
1605       where the argument stack is not already long enough to handle all the
1606       return values.  An example is the POSIX tzname() call, which takes no
1607       arguments, but returns two, the local time zone's standard and summer
1608       time abbreviations.
1609
1610       To handle this situation, the PPCODE directive is used and the stack is
1611       extended using the macro:
1612
1613           EXTEND(SP, num);
1614
1615       where "SP" is the macro that represents the local copy of the stack
1616       pointer, and "num" is the number of elements the stack should be
1617       extended by.
1618
1619       Now that there is room on the stack, values can be pushed on it using
1620       "PUSHs" macro.  The pushed values will often need to be "mortal" (See
1621       "Reference Counts and Mortality"):
1622
1623           PUSHs(sv_2mortal(newSViv(an_integer)))
1624           PUSHs(sv_2mortal(newSVuv(an_unsigned_integer)))
1625           PUSHs(sv_2mortal(newSVnv(a_double)))
1626           PUSHs(sv_2mortal(newSVpv("Some String",0)))
1627           /* Although the last example is better written as the more
1628            * efficient: */
1629           PUSHs(newSVpvs_flags("Some String", SVs_TEMP))
1630
1631       And now the Perl program calling "tzname", the two values will be
1632       assigned as in:
1633
1634           ($standard_abbrev, $summer_abbrev) = POSIX::tzname;
1635
1636       An alternate (and possibly simpler) method to pushing values on the
1637       stack is to use the macro:
1638
1639           XPUSHs(SV*)
1640
1641       This macro automatically adjusts the stack for you, if needed.  Thus,
1642       you do not need to call "EXTEND" to extend the stack.
1643
1644       Despite their suggestions in earlier versions of this document the
1645       macros "(X)PUSH[iunp]" are not suited to XSUBs which return multiple
1646       results.  For that, either stick to the "(X)PUSHs" macros shown above,
1647       or use the new "m(X)PUSH[iunp]" macros instead; see "Putting a C value
1648       on Perl stack".
1649
1650       For more information, consult perlxs and perlxstut.
1651
1652   Autoloading with XSUBs
1653       If an AUTOLOAD routine is an XSUB, as with Perl subroutines, Perl puts
1654       the fully-qualified name of the autoloaded subroutine in the $AUTOLOAD
1655       variable of the XSUB's package.
1656
1657       But it also puts the same information in certain fields of the XSUB
1658       itself:
1659
1660           HV *stash           = CvSTASH(cv);
1661           const char *subname = SvPVX(cv);
1662           STRLEN name_length  = SvCUR(cv); /* in bytes */
1663           U32 is_utf8         = SvUTF8(cv);
1664
1665       "SvPVX(cv)" contains just the sub name itself, not including the
1666       package.  For an AUTOLOAD routine in UNIVERSAL or one of its
1667       superclasses, "CvSTASH(cv)" returns NULL during a method call on a
1668       nonexistent package.
1669
1670       Note: Setting $AUTOLOAD stopped working in 5.6.1, which did not support
1671       XS AUTOLOAD subs at all.  Perl 5.8.0 introduced the use of fields in
1672       the XSUB itself.  Perl 5.16.0 restored the setting of $AUTOLOAD.  If
1673       you need to support 5.8-5.14, use the XSUB's fields.
1674
1675   Calling Perl Routines from within C Programs
1676       There are four routines that can be used to call a Perl subroutine from
1677       within a C program.  These four are:
1678
1679           I32  call_sv(SV*, I32);
1680           I32  call_pv(const char*, I32);
1681           I32  call_method(const char*, I32);
1682           I32  call_argv(const char*, I32, char**);
1683
1684       The routine most often used is "call_sv".  The "SV*" argument contains
1685       either the name of the Perl subroutine to be called, or a reference to
1686       the subroutine.  The second argument consists of flags that control the
1687       context in which the subroutine is called, whether or not the
1688       subroutine is being passed arguments, how errors should be trapped, and
1689       how to treat return values.
1690
1691       All four routines return the number of arguments that the subroutine
1692       returned on the Perl stack.
1693
1694       These routines used to be called "perl_call_sv", etc., before Perl
1695       v5.6.0, but those names are now deprecated; macros of the same name are
1696       provided for compatibility.
1697
1698       When using any of these routines (except "call_argv"), the programmer
1699       must manipulate the Perl stack.  These include the following macros and
1700       functions:
1701
1702           dSP
1703           SP
1704           PUSHMARK()
1705           PUTBACK
1706           SPAGAIN
1707           ENTER
1708           SAVETMPS
1709           FREETMPS
1710           LEAVE
1711           XPUSH*()
1712           POP*()
1713
1714       For a detailed description of calling conventions from C to Perl,
1715       consult perlcall.
1716
1717   Putting a C value on Perl stack
1718       A lot of opcodes (this is an elementary operation in the internal perl
1719       stack machine) put an SV* on the stack.  However, as an optimization
1720       the corresponding SV is (usually) not recreated each time.  The opcodes
1721       reuse specially assigned SVs (targets) which are (as a corollary) not
1722       constantly freed/created.
1723
1724       Each of the targets is created only once (but see "Scratchpads and
1725       recursion" below), and when an opcode needs to put an integer, a
1726       double, or a string on stack, it just sets the corresponding parts of
1727       its target and puts the target on stack.
1728
1729       The macro to put this target on stack is "PUSHTARG", and it is directly
1730       used in some opcodes, as well as indirectly in zillions of others,
1731       which use it via "(X)PUSH[iunp]".
1732
1733       Because the target is reused, you must be careful when pushing multiple
1734       values on the stack.  The following code will not do what you think:
1735
1736           XPUSHi(10);
1737           XPUSHi(20);
1738
1739       This translates as "set "TARG" to 10, push a pointer to "TARG" onto the
1740       stack; set "TARG" to 20, push a pointer to "TARG" onto the stack".  At
1741       the end of the operation, the stack does not contain the values 10 and
1742       20, but actually contains two pointers to "TARG", which we have set to
1743       20.
1744
1745       If you need to push multiple different values then you should either
1746       use the "(X)PUSHs" macros, or else use the new "m(X)PUSH[iunp]" macros,
1747       none of which make use of "TARG".  The "(X)PUSHs" macros simply push an
1748       SV* on the stack, which, as noted under "XSUBs and the Argument Stack",
1749       will often need to be "mortal".  The new "m(X)PUSH[iunp]" macros make
1750       this a little easier to achieve by creating a new mortal for you (via
1751       "(X)PUSHmortal"), pushing that onto the stack (extending it if
1752       necessary in the case of the "mXPUSH[iunp]" macros), and then setting
1753       its value.  Thus, instead of writing this to "fix" the example above:
1754
1755           XPUSHs(sv_2mortal(newSViv(10)))
1756           XPUSHs(sv_2mortal(newSViv(20)))
1757
1758       you can simply write:
1759
1760           mXPUSHi(10)
1761           mXPUSHi(20)
1762
1763       On a related note, if you do use "(X)PUSH[iunp]", then you're going to
1764       need a "dTARG" in your variable declarations so that the "*PUSH*"
1765       macros can make use of the local variable "TARG".  See also "dTARGET"
1766       and "dXSTARG".
1767
1768   Scratchpads
1769       The question remains on when the SVs which are targets for opcodes are
1770       created.  The answer is that they are created when the current unit--a
1771       subroutine or a file (for opcodes for statements outside of
1772       subroutines)--is compiled.  During this time a special anonymous Perl
1773       array is created, which is called a scratchpad for the current unit.
1774
1775       A scratchpad keeps SVs which are lexicals for the current unit and are
1776       targets for opcodes.  A previous version of this document stated that
1777       one can deduce that an SV lives on a scratchpad by looking on its
1778       flags: lexicals have "SVs_PADMY" set, and targets have "SVs_PADTMP"
1779       set.  But this has never been fully true.  "SVs_PADMY" could be set on
1780       a variable that no longer resides in any pad.  While targets do have
1781       "SVs_PADTMP" set, it can also be set on variables that have never
1782       resided in a pad, but nonetheless act like targets.  As of perl 5.21.5,
1783       the "SVs_PADMY" flag is no longer used and is defined as 0.
1784       "SvPADMY()" now returns true for anything without "SVs_PADTMP".
1785
1786       The correspondence between OPs and targets is not 1-to-1.  Different
1787       OPs in the compile tree of the unit can use the same target, if this
1788       would not conflict with the expected life of the temporary.
1789
1790   Scratchpads and recursion
1791       In fact it is not 100% true that a compiled unit contains a pointer to
1792       the scratchpad AV.  In fact it contains a pointer to an AV of
1793       (initially) one element, and this element is the scratchpad AV.  Why do
1794       we need an extra level of indirection?
1795
1796       The answer is recursion, and maybe threads.  Both these can create
1797       several execution pointers going into the same subroutine.  For the
1798       subroutine-child not write over the temporaries for the subroutine-
1799       parent (lifespan of which covers the call to the child), the parent and
1800       the child should have different scratchpads.  (And the lexicals should
1801       be separate anyway!)
1802
1803       So each subroutine is born with an array of scratchpads (of length 1).
1804       On each entry to the subroutine it is checked that the current depth of
1805       the recursion is not more than the length of this array, and if it is,
1806       new scratchpad is created and pushed into the array.
1807
1808       The targets on this scratchpad are "undef"s, but they are already
1809       marked with correct flags.
1810

Memory Allocation

1812   Allocation
1813       All memory meant to be used with the Perl API functions should be
1814       manipulated using the macros described in this section.  The macros
1815       provide the necessary transparency between differences in the actual
1816       malloc implementation that is used within perl.
1817
1818       The following three macros are used to initially allocate memory :
1819
1820           Newx(pointer, number, type);
1821           Newxc(pointer, number, type, cast);
1822           Newxz(pointer, number, type);
1823
1824       The first argument "pointer" should be the name of a variable that will
1825       point to the newly allocated memory.
1826
1827       The second and third arguments "number" and "type" specify how many of
1828       the specified type of data structure should be allocated.  The argument
1829       "type" is passed to "sizeof".  The final argument to "Newxc", "cast",
1830       should be used if the "pointer" argument is different from the "type"
1831       argument.
1832
1833       Unlike the "Newx" and "Newxc" macros, the "Newxz" macro calls "memzero"
1834       to zero out all the newly allocated memory.
1835
1836   Reallocation
1837           Renew(pointer, number, type);
1838           Renewc(pointer, number, type, cast);
1839           Safefree(pointer)
1840
1841       These three macros are used to change a memory buffer size or to free a
1842       piece of memory no longer needed.  The arguments to "Renew" and
1843       "Renewc" match those of "New" and "Newc" with the exception of not
1844       needing the "magic cookie" argument.
1845
1846   Moving
1847           Move(source, dest, number, type);
1848           Copy(source, dest, number, type);
1849           Zero(dest, number, type);
1850
1851       These three macros are used to move, copy, or zero out previously
1852       allocated memory.  The "source" and "dest" arguments point to the
1853       source and destination starting points.  Perl will move, copy, or zero
1854       out "number" instances of the size of the "type" data structure (using
1855       the "sizeof" function).
1856

PerlIO

1858       The most recent development releases of Perl have been experimenting
1859       with removing Perl's dependency on the "normal" standard I/O suite and
1860       allowing other stdio implementations to be used.  This involves
1861       creating a new abstraction layer that then calls whichever
1862       implementation of stdio Perl was compiled with.  All XSUBs should now
1863       use the functions in the PerlIO abstraction layer and not make any
1864       assumptions about what kind of stdio is being used.
1865
1866       For a complete description of the PerlIO abstraction, consult perlapio.
1867

Compiled code

1869   Code tree
1870       Here we describe the internal form your code is converted to by Perl.
1871       Start with a simple example:
1872
1873         $a = $b + $c;
1874
1875       This is converted to a tree similar to this one:
1876
1877                    assign-to
1878                  /           \
1879                 +             $a
1880               /   \
1881             $b     $c
1882
1883       (but slightly more complicated).  This tree reflects the way Perl
1884       parsed your code, but has nothing to do with the execution order.
1885       There is an additional "thread" going through the nodes of the tree
1886       which shows the order of execution of the nodes.  In our simplified
1887       example above it looks like:
1888
1889            $b ---> $c ---> + ---> $a ---> assign-to
1890
1891       But with the actual compile tree for "$a = $b + $c" it is different:
1892       some nodes optimized away.  As a corollary, though the actual tree
1893       contains more nodes than our simplified example, the execution order is
1894       the same as in our example.
1895
1896   Examining the tree
1897       If you have your perl compiled for debugging (usually done with
1898       "-DDEBUGGING" on the "Configure" command line), you may examine the
1899       compiled tree by specifying "-Dx" on the Perl command line.  The output
1900       takes several lines per node, and for "$b+$c" it looks like this:
1901
1902           5           TYPE = add  ===> 6
1903                       TARG = 1
1904                       FLAGS = (SCALAR,KIDS)
1905                       {
1906                           TYPE = null  ===> (4)
1907                             (was rv2sv)
1908                           FLAGS = (SCALAR,KIDS)
1909                           {
1910           3                   TYPE = gvsv  ===> 4
1911                               FLAGS = (SCALAR)
1912                               GV = main::b
1913                           }
1914                       }
1915                       {
1916                           TYPE = null  ===> (5)
1917                             (was rv2sv)
1918                           FLAGS = (SCALAR,KIDS)
1919                           {
1920           4                   TYPE = gvsv  ===> 5
1921                               FLAGS = (SCALAR)
1922                               GV = main::c
1923                           }
1924                       }
1925
1926       This tree has 5 nodes (one per "TYPE" specifier), only 3 of them are
1927       not optimized away (one per number in the left column).  The immediate
1928       children of the given node correspond to "{}" pairs on the same level
1929       of indentation, thus this listing corresponds to the tree:
1930
1931                          add
1932                        /     \
1933                      null    null
1934                       |       |
1935                      gvsv    gvsv
1936
1937       The execution order is indicated by "===>" marks, thus it is "3 4 5 6"
1938       (node 6 is not included into above listing), i.e., "gvsv gvsv add
1939       whatever".
1940
1941       Each of these nodes represents an op, a fundamental operation inside
1942       the Perl core.  The code which implements each operation can be found
1943       in the pp*.c files; the function which implements the op with type
1944       "gvsv" is "pp_gvsv", and so on.  As the tree above shows, different ops
1945       have different numbers of children: "add" is a binary operator, as one
1946       would expect, and so has two children.  To accommodate the various
1947       different numbers of children, there are various types of op data
1948       structure, and they link together in different ways.
1949
1950       The simplest type of op structure is "OP": this has no children.  Unary
1951       operators, "UNOP"s, have one child, and this is pointed to by the
1952       "op_first" field.  Binary operators ("BINOP"s) have not only an
1953       "op_first" field but also an "op_last" field.  The most complex type of
1954       op is a "LISTOP", which has any number of children.  In this case, the
1955       first child is pointed to by "op_first" and the last child by
1956       "op_last".  The children in between can be found by iteratively
1957       following the "OpSIBLING" pointer from the first child to the last (but
1958       see below).
1959
1960       There are also some other op types: a "PMOP" holds a regular
1961       expression, and has no children, and a "LOOP" may or may not have
1962       children.  If the "op_children" field is non-zero, it behaves like a
1963       "LISTOP".  To complicate matters, if a "UNOP" is actually a "null" op
1964       after optimization (see "Compile pass 2: context propagation") it will
1965       still have children in accordance with its former type.
1966
1967       Finally, there is a "LOGOP", or logic op. Like a "LISTOP", this has one
1968       or more children, but it doesn't have an "op_last" field: so you have
1969       to follow "op_first" and then the "OpSIBLING" chain itself to find the
1970       last child. Instead it has an "op_other" field, which is comparable to
1971       the "op_next" field described below, and represents an alternate
1972       execution path. Operators like "and", "or" and "?" are "LOGOP"s. Note
1973       that in general, "op_other" may not point to any of the direct children
1974       of the "LOGOP".
1975
1976       Starting in version 5.21.2, perls built with the experimental define
1977       "-DPERL_OP_PARENT" add an extra boolean flag for each op, "op_moresib".
1978       When not set, this indicates that this is the last op in an "OpSIBLING"
1979       chain. This frees up the "op_sibling" field on the last sibling to
1980       point back to the parent op. Under this build, that field is also
1981       renamed "op_sibparent" to reflect its joint role. The macro
1982       OpSIBLING(o) wraps this special behaviour, and always returns NULL on
1983       the last sibling.  With this build the op_parent(o) function can be
1984       used to find the parent of any op. Thus for forward compatibility, you
1985       should always use the OpSIBLING(o) macro rather than accessing
1986       "op_sibling" directly.
1987
1988       Another way to examine the tree is to use a compiler back-end module,
1989       such as B::Concise.
1990
1991   Compile pass 1: check routines
1992       The tree is created by the compiler while yacc code feeds it the
1993       constructions it recognizes.  Since yacc works bottom-up, so does the
1994       first pass of perl compilation.
1995
1996       What makes this pass interesting for perl developers is that some
1997       optimization may be performed on this pass.  This is optimization by
1998       so-called "check routines".  The correspondence between node names and
1999       corresponding check routines is described in opcode.pl (do not forget
2000       to run "make regen_headers" if you modify this file).
2001
2002       A check routine is called when the node is fully constructed except for
2003       the execution-order thread.  Since at this time there are no back-links
2004       to the currently constructed node, one can do most any operation to the
2005       top-level node, including freeing it and/or creating new nodes
2006       above/below it.
2007
2008       The check routine returns the node which should be inserted into the
2009       tree (if the top-level node was not modified, check routine returns its
2010       argument).
2011
2012       By convention, check routines have names "ck_*".  They are usually
2013       called from "new*OP" subroutines (or "convert") (which in turn are
2014       called from perly.y).
2015
2016   Compile pass 1a: constant folding
2017       Immediately after the check routine is called the returned node is
2018       checked for being compile-time executable.  If it is (the value is
2019       judged to be constant) it is immediately executed, and a constant node
2020       with the "return value" of the corresponding subtree is substituted
2021       instead.  The subtree is deleted.
2022
2023       If constant folding was not performed, the execution-order thread is
2024       created.
2025
2026   Compile pass 2: context propagation
2027       When a context for a part of compile tree is known, it is propagated
2028       down through the tree.  At this time the context can have 5 values
2029       (instead of 2 for runtime context): void, boolean, scalar, list, and
2030       lvalue.  In contrast with the pass 1 this pass is processed from top to
2031       bottom: a node's context determines the context for its children.
2032
2033       Additional context-dependent optimizations are performed at this time.
2034       Since at this moment the compile tree contains back-references (via
2035       "thread" pointers), nodes cannot be free()d now.  To allow optimized-
2036       away nodes at this stage, such nodes are null()ified instead of
2037       free()ing (i.e. their type is changed to OP_NULL).
2038
2039   Compile pass 3: peephole optimization
2040       After the compile tree for a subroutine (or for an "eval" or a file) is
2041       created, an additional pass over the code is performed.  This pass is
2042       neither top-down or bottom-up, but in the execution order (with
2043       additional complications for conditionals).  Optimizations performed at
2044       this stage are subject to the same restrictions as in the pass 2.
2045
2046       Peephole optimizations are done by calling the function pointed to by
2047       the global variable "PL_peepp".  By default, "PL_peepp" just calls the
2048       function pointed to by the global variable "PL_rpeepp".  By default,
2049       that performs some basic op fixups and optimisations along the
2050       execution-order op chain, and recursively calls "PL_rpeepp" for each
2051       side chain of ops (resulting from conditionals).  Extensions may
2052       provide additional optimisations or fixups, hooking into either the
2053       per-subroutine or recursive stage, like this:
2054
2055           static peep_t prev_peepp;
2056           static void my_peep(pTHX_ OP *o)
2057           {
2058               /* custom per-subroutine optimisation goes here */
2059               prev_peepp(aTHX_ o);
2060               /* custom per-subroutine optimisation may also go here */
2061           }
2062           BOOT:
2063               prev_peepp = PL_peepp;
2064               PL_peepp = my_peep;
2065
2066           static peep_t prev_rpeepp;
2067           static void my_rpeep(pTHX_ OP *o)
2068           {
2069               OP *orig_o = o;
2070               for(; o; o = o->op_next) {
2071                   /* custom per-op optimisation goes here */
2072               }
2073               prev_rpeepp(aTHX_ orig_o);
2074           }
2075           BOOT:
2076               prev_rpeepp = PL_rpeepp;
2077               PL_rpeepp = my_rpeep;
2078
2079   Pluggable runops
2080       The compile tree is executed in a runops function.  There are two
2081       runops functions, in run.c and in dump.c.  "Perl_runops_debug" is used
2082       with DEBUGGING and "Perl_runops_standard" is used otherwise.  For fine
2083       control over the execution of the compile tree it is possible to
2084       provide your own runops function.
2085
2086       It's probably best to copy one of the existing runops functions and
2087       change it to suit your needs.  Then, in the BOOT section of your XS
2088       file, add the line:
2089
2090         PL_runops = my_runops;
2091
2092       This function should be as efficient as possible to keep your programs
2093       running as fast as possible.
2094
2095   Compile-time scope hooks
2096       As of perl 5.14 it is possible to hook into the compile-time lexical
2097       scope mechanism using "Perl_blockhook_register".  This is used like
2098       this:
2099
2100           STATIC void my_start_hook(pTHX_ int full);
2101           STATIC BHK my_hooks;
2102
2103           BOOT:
2104               BhkENTRY_set(&my_hooks, bhk_start, my_start_hook);
2105               Perl_blockhook_register(aTHX_ &my_hooks);
2106
2107       This will arrange to have "my_start_hook" called at the start of
2108       compiling every lexical scope.  The available hooks are:
2109
2110       "void bhk_start(pTHX_ int full)"
2111           This is called just after starting a new lexical scope.  Note that
2112           Perl code like
2113
2114               if ($x) { ... }
2115
2116           creates two scopes: the first starts at the "(" and has "full ==
2117           1", the second starts at the "{" and has "full == 0".  Both end at
2118           the "}", so calls to "start" and "pre"/"post_end" will match.
2119           Anything pushed onto the save stack by this hook will be popped
2120           just before the scope ends (between the "pre_" and "post_end"
2121           hooks, in fact).
2122
2123       "void bhk_pre_end(pTHX_ OP **o)"
2124           This is called at the end of a lexical scope, just before unwinding
2125           the stack.  o is the root of the optree representing the scope; it
2126           is a double pointer so you can replace the OP if you need to.
2127
2128       "void bhk_post_end(pTHX_ OP **o)"
2129           This is called at the end of a lexical scope, just after unwinding
2130           the stack.  o is as above.  Note that it is possible for calls to
2131           "pre_" and "post_end" to nest, if there is something on the save
2132           stack that calls string eval.
2133
2134       "void bhk_eval(pTHX_ OP *const o)"
2135           This is called just before starting to compile an "eval STRING",
2136           "do FILE", "require" or "use", after the eval has been set up.  o
2137           is the OP that requested the eval, and will normally be an
2138           "OP_ENTEREVAL", "OP_DOFILE" or "OP_REQUIRE".
2139
2140       Once you have your hook functions, you need a "BHK" structure to put
2141       them in.  It's best to allocate it statically, since there is no way to
2142       free it once it's registered.  The function pointers should be inserted
2143       into this structure using the "BhkENTRY_set" macro, which will also set
2144       flags indicating which entries are valid.  If you do need to allocate
2145       your "BHK" dynamically for some reason, be sure to zero it before you
2146       start.
2147
2148       Once registered, there is no mechanism to switch these hooks off, so if
2149       that is necessary you will need to do this yourself.  An entry in "%^H"
2150       is probably the best way, so the effect is lexically scoped; however it
2151       is also possible to use the "BhkDISABLE" and "BhkENABLE" macros to
2152       temporarily switch entries on and off.  You should also be aware that
2153       generally speaking at least one scope will have opened before your
2154       extension is loaded, so you will see some "pre"/"post_end" pairs that
2155       didn't have a matching "start".
2156

Examining internal data structures with the "dump" functions

2158       To aid debugging, the source file dump.c contains a number of functions
2159       which produce formatted output of internal data structures.
2160
2161       The most commonly used of these functions is "Perl_sv_dump"; it's used
2162       for dumping SVs, AVs, HVs, and CVs.  The "Devel::Peek" module calls
2163       "sv_dump" to produce debugging output from Perl-space, so users of that
2164       module should already be familiar with its format.
2165
2166       "Perl_op_dump" can be used to dump an "OP" structure or any of its
2167       derivatives, and produces output similar to "perl -Dx"; in fact,
2168       "Perl_dump_eval" will dump the main root of the code being evaluated,
2169       exactly like "-Dx".
2170
2171       Other useful functions are "Perl_dump_sub", which turns a "GV" into an
2172       op tree, "Perl_dump_packsubs" which calls "Perl_dump_sub" on all the
2173       subroutines in a package like so: (Thankfully, these are all xsubs, so
2174       there is no op tree)
2175
2176           (gdb) print Perl_dump_packsubs(PL_defstash)
2177
2178           SUB attributes::bootstrap = (xsub 0x811fedc 0)
2179
2180           SUB UNIVERSAL::can = (xsub 0x811f50c 0)
2181
2182           SUB UNIVERSAL::isa = (xsub 0x811f304 0)
2183
2184           SUB UNIVERSAL::VERSION = (xsub 0x811f7ac 0)
2185
2186           SUB DynaLoader::boot_DynaLoader = (xsub 0x805b188 0)
2187
2188       and "Perl_dump_all", which dumps all the subroutines in the stash and
2189       the op tree of the main root.
2190

How multiple interpreters and concurrency are supported

2192   Background and PERL_IMPLICIT_CONTEXT
2193       The Perl interpreter can be regarded as a closed box: it has an API for
2194       feeding it code or otherwise making it do things, but it also has
2195       functions for its own use.  This smells a lot like an object, and there
2196       are ways for you to build Perl so that you can have multiple
2197       interpreters, with one interpreter represented either as a C structure,
2198       or inside a thread-specific structure.  These structures contain all
2199       the context, the state of that interpreter.
2200
2201       One macro controls the major Perl build flavor: MULTIPLICITY.  The
2202       MULTIPLICITY build has a C structure that packages all the interpreter
2203       state.  With multiplicity-enabled perls, PERL_IMPLICIT_CONTEXT is also
2204       normally defined, and enables the support for passing in a "hidden"
2205       first argument that represents all three data structures.  MULTIPLICITY
2206       makes multi-threaded perls possible (with the ithreads threading model,
2207       related to the macro USE_ITHREADS.)
2208
2209       Two other "encapsulation" macros are the PERL_GLOBAL_STRUCT and
2210       PERL_GLOBAL_STRUCT_PRIVATE (the latter turns on the former, and the
2211       former turns on MULTIPLICITY.)  The PERL_GLOBAL_STRUCT causes all the
2212       internal variables of Perl to be wrapped inside a single global struct,
2213       struct perl_vars, accessible as (globals) &PL_Vars or PL_VarsPtr or the
2214       function  Perl_GetVars().  The PERL_GLOBAL_STRUCT_PRIVATE goes one step
2215       further, there is still a single struct (allocated in main() either
2216       from heap or from stack) but there are no global data symbols pointing
2217       to it.  In either case the global struct should be initialized as the
2218       very first thing in main() using Perl_init_global_struct() and
2219       correspondingly tear it down after perl_free() using
2220       Perl_free_global_struct(), please see miniperlmain.c for usage details.
2221       You may also need to use "dVAR" in your coding to "declare the global
2222       variables" when you are using them.  dTHX does this for you
2223       automatically.
2224
2225       To see whether you have non-const data you can use a BSD (or GNU)
2226       compatible "nm":
2227
2228         nm libperl.a | grep -v ' [TURtr] '
2229
2230       If this displays any "D" or "d" symbols (or possibly "C" or "c"), you
2231       have non-const data.  The symbols the "grep" removed are as follows:
2232       "Tt" are text, or code, the "Rr" are read-only (const) data, and the
2233       "U" is <undefined>, external symbols referred to.
2234
2235       The test t/porting/libperl.t does this kind of symbol sanity checking
2236       on "libperl.a".
2237
2238       For backward compatibility reasons defining just PERL_GLOBAL_STRUCT
2239       doesn't actually hide all symbols inside a big global struct: some
2240       PerlIO_xxx vtables are left visible.  The PERL_GLOBAL_STRUCT_PRIVATE
2241       then hides everything (see how the PERLIO_FUNCS_DECL is used).
2242
2243       All this obviously requires a way for the Perl internal functions to be
2244       either subroutines taking some kind of structure as the first argument,
2245       or subroutines taking nothing as the first argument.  To enable these
2246       two very different ways of building the interpreter, the Perl source
2247       (as it does in so many other situations) makes heavy use of macros and
2248       subroutine naming conventions.
2249
2250       First problem: deciding which functions will be public API functions
2251       and which will be private.  All functions whose names begin "S_" are
2252       private (think "S" for "secret" or "static").  All other functions
2253       begin with "Perl_", but just because a function begins with "Perl_"
2254       does not mean it is part of the API.  (See "Internal Functions".)  The
2255       easiest way to be sure a function is part of the API is to find its
2256       entry in perlapi.  If it exists in perlapi, it's part of the API.  If
2257       it doesn't, and you think it should be (i.e., you need it for your
2258       extension), submit an issue at <https://github.com/Perl/perl5/issues>
2259       explaining why you think it should be.
2260
2261       Second problem: there must be a syntax so that the same subroutine
2262       declarations and calls can pass a structure as their first argument, or
2263       pass nothing.  To solve this, the subroutines are named and declared in
2264       a particular way.  Here's a typical start of a static function used
2265       within the Perl guts:
2266
2267         STATIC void
2268         S_incline(pTHX_ char *s)
2269
2270       STATIC becomes "static" in C, and may be #define'd to nothing in some
2271       configurations in the future.
2272
2273       A public function (i.e. part of the internal API, but not necessarily
2274       sanctioned for use in extensions) begins like this:
2275
2276         void
2277         Perl_sv_setiv(pTHX_ SV* dsv, IV num)
2278
2279       "pTHX_" is one of a number of macros (in perl.h) that hide the details
2280       of the interpreter's context.  THX stands for "thread", "this", or
2281       "thingy", as the case may be.  (And no, George Lucas is not involved.
2282       :-) The first character could be 'p' for a prototype, 'a' for argument,
2283       or 'd' for declaration, so we have "pTHX", "aTHX" and "dTHX", and their
2284       variants.
2285
2286       When Perl is built without options that set PERL_IMPLICIT_CONTEXT,
2287       there is no first argument containing the interpreter's context.  The
2288       trailing underscore in the pTHX_ macro indicates that the macro
2289       expansion needs a comma after the context argument because other
2290       arguments follow it.  If PERL_IMPLICIT_CONTEXT is not defined, pTHX_
2291       will be ignored, and the subroutine is not prototyped to take the extra
2292       argument.  The form of the macro without the trailing underscore is
2293       used when there are no additional explicit arguments.
2294
2295       When a core function calls another, it must pass the context.  This is
2296       normally hidden via macros.  Consider "sv_setiv".  It expands into
2297       something like this:
2298
2299           #ifdef PERL_IMPLICIT_CONTEXT
2300             #define sv_setiv(a,b)      Perl_sv_setiv(aTHX_ a, b)
2301             /* can't do this for vararg functions, see below */
2302           #else
2303             #define sv_setiv           Perl_sv_setiv
2304           #endif
2305
2306       This works well, and means that XS authors can gleefully write:
2307
2308           sv_setiv(foo, bar);
2309
2310       and still have it work under all the modes Perl could have been
2311       compiled with.
2312
2313       This doesn't work so cleanly for varargs functions, though, as macros
2314       imply that the number of arguments is known in advance.  Instead we
2315       either need to spell them out fully, passing "aTHX_" as the first
2316       argument (the Perl core tends to do this with functions like
2317       Perl_warner), or use a context-free version.
2318
2319       The context-free version of Perl_warner is called
2320       Perl_warner_nocontext, and does not take the extra argument.  Instead
2321       it does "dTHX;" to get the context from thread-local storage.  We
2322       "#define warner Perl_warner_nocontext" so that extensions get source
2323       compatibility at the expense of performance.  (Passing an arg is
2324       cheaper than grabbing it from thread-local storage.)
2325
2326       You can ignore [pad]THXx when browsing the Perl headers/sources.  Those
2327       are strictly for use within the core.  Extensions and embedders need
2328       only be aware of [pad]THX.
2329
2330   So what happened to dTHR?
2331       "dTHR" was introduced in perl 5.005 to support the older thread model.
2332       The older thread model now uses the "THX" mechanism to pass context
2333       pointers around, so "dTHR" is not useful any more.  Perl 5.6.0 and
2334       later still have it for backward source compatibility, but it is
2335       defined to be a no-op.
2336
2337   How do I use all this in extensions?
2338       When Perl is built with PERL_IMPLICIT_CONTEXT, extensions that call any
2339       functions in the Perl API will need to pass the initial context
2340       argument somehow.  The kicker is that you will need to write it in such
2341       a way that the extension still compiles when Perl hasn't been built
2342       with PERL_IMPLICIT_CONTEXT enabled.
2343
2344       There are three ways to do this.  First, the easy but inefficient way,
2345       which is also the default, in order to maintain source compatibility
2346       with extensions: whenever XSUB.h is #included, it redefines the aTHX
2347       and aTHX_ macros to call a function that will return the context.
2348       Thus, something like:
2349
2350               sv_setiv(sv, num);
2351
2352       in your extension will translate to this when PERL_IMPLICIT_CONTEXT is
2353       in effect:
2354
2355               Perl_sv_setiv(Perl_get_context(), sv, num);
2356
2357       or to this otherwise:
2358
2359               Perl_sv_setiv(sv, num);
2360
2361       You don't have to do anything new in your extension to get this; since
2362       the Perl library provides Perl_get_context(), it will all just work.
2363
2364       The second, more efficient way is to use the following template for
2365       your Foo.xs:
2366
2367               #define PERL_NO_GET_CONTEXT     /* we want efficiency */
2368               #include "EXTERN.h"
2369               #include "perl.h"
2370               #include "XSUB.h"
2371
2372               STATIC void my_private_function(int arg1, int arg2);
2373
2374               STATIC void
2375               my_private_function(int arg1, int arg2)
2376               {
2377                   dTHX;       /* fetch context */
2378                   ... call many Perl API functions ...
2379               }
2380
2381               [... etc ...]
2382
2383               MODULE = Foo            PACKAGE = Foo
2384
2385               /* typical XSUB */
2386
2387               void
2388               my_xsub(arg)
2389                       int arg
2390                   CODE:
2391                       my_private_function(arg, 10);
2392
2393       Note that the only two changes from the normal way of writing an
2394       extension is the addition of a "#define PERL_NO_GET_CONTEXT" before
2395       including the Perl headers, followed by a "dTHX;" declaration at the
2396       start of every function that will call the Perl API.  (You'll know
2397       which functions need this, because the C compiler will complain that
2398       there's an undeclared identifier in those functions.)  No changes are
2399       needed for the XSUBs themselves, because the XS() macro is correctly
2400       defined to pass in the implicit context if needed.
2401
2402       The third, even more efficient way is to ape how it is done within the
2403       Perl guts:
2404
2405               #define PERL_NO_GET_CONTEXT     /* we want efficiency */
2406               #include "EXTERN.h"
2407               #include "perl.h"
2408               #include "XSUB.h"
2409
2410               /* pTHX_ only needed for functions that call Perl API */
2411               STATIC void my_private_function(pTHX_ int arg1, int arg2);
2412
2413               STATIC void
2414               my_private_function(pTHX_ int arg1, int arg2)
2415               {
2416                   /* dTHX; not needed here, because THX is an argument */
2417                   ... call Perl API functions ...
2418               }
2419
2420               [... etc ...]
2421
2422               MODULE = Foo            PACKAGE = Foo
2423
2424               /* typical XSUB */
2425
2426               void
2427               my_xsub(arg)
2428                       int arg
2429                   CODE:
2430                       my_private_function(aTHX_ arg, 10);
2431
2432       This implementation never has to fetch the context using a function
2433       call, since it is always passed as an extra argument.  Depending on
2434       your needs for simplicity or efficiency, you may mix the previous two
2435       approaches freely.
2436
2437       Never add a comma after "pTHX" yourself--always use the form of the
2438       macro with the underscore for functions that take explicit arguments,
2439       or the form without the argument for functions with no explicit
2440       arguments.
2441
2442       If one is compiling Perl with the "-DPERL_GLOBAL_STRUCT" the "dVAR"
2443       definition is needed if the Perl global variables (see perlvars.h or
2444       globvar.sym) are accessed in the function and "dTHX" is not used (the
2445       "dTHX" includes the "dVAR" if necessary).  One notices the need for
2446       "dVAR" only with the said compile-time define, because otherwise the
2447       Perl global variables are visible as-is.
2448
2449   Should I do anything special if I call perl from multiple threads?
2450       If you create interpreters in one thread and then proceed to call them
2451       in another, you need to make sure perl's own Thread Local Storage (TLS)
2452       slot is initialized correctly in each of those threads.
2453
2454       The "perl_alloc" and "perl_clone" API functions will automatically set
2455       the TLS slot to the interpreter they created, so that there is no need
2456       to do anything special if the interpreter is always accessed in the
2457       same thread that created it, and that thread did not create or call any
2458       other interpreters afterwards.  If that is not the case, you have to
2459       set the TLS slot of the thread before calling any functions in the Perl
2460       API on that particular interpreter.  This is done by calling the
2461       "PERL_SET_CONTEXT" macro in that thread as the first thing you do:
2462
2463               /* do this before doing anything else with some_perl */
2464               PERL_SET_CONTEXT(some_perl);
2465
2466               ... other Perl API calls on some_perl go here ...
2467
2468   Future Plans and PERL_IMPLICIT_SYS
2469       Just as PERL_IMPLICIT_CONTEXT provides a way to bundle up everything
2470       that the interpreter knows about itself and pass it around, so too are
2471       there plans to allow the interpreter to bundle up everything it knows
2472       about the environment it's running on.  This is enabled with the
2473       PERL_IMPLICIT_SYS macro.  Currently it only works with USE_ITHREADS on
2474       Windows.
2475
2476       This allows the ability to provide an extra pointer (called the "host"
2477       environment) for all the system calls.  This makes it possible for all
2478       the system stuff to maintain their own state, broken down into seven C
2479       structures.  These are thin wrappers around the usual system calls (see
2480       win32/perllib.c) for the default perl executable, but for a more
2481       ambitious host (like the one that would do fork() emulation) all the
2482       extra work needed to pretend that different interpreters are actually
2483       different "processes", would be done here.
2484
2485       The Perl engine/interpreter and the host are orthogonal entities.
2486       There could be one or more interpreters in a process, and one or more
2487       "hosts", with free association between them.
2488

Internal Functions

2490       All of Perl's internal functions which will be exposed to the outside
2491       world are prefixed by "Perl_" so that they will not conflict with XS
2492       functions or functions used in a program in which Perl is embedded.
2493       Similarly, all global variables begin with "PL_".  (By convention,
2494       static functions start with "S_".)
2495
2496       Inside the Perl core ("PERL_CORE" defined), you can get at the
2497       functions either with or without the "Perl_" prefix, thanks to a bunch
2498       of defines that live in embed.h.  Note that extension code should not
2499       set "PERL_CORE"; this exposes the full perl internals, and is likely to
2500       cause breakage of the XS in each new perl release.
2501
2502       The file embed.h is generated automatically from embed.pl and
2503       embed.fnc.  embed.pl also creates the prototyping header files for the
2504       internal functions, generates the documentation and a lot of other bits
2505       and pieces.  It's important that when you add a new function to the
2506       core or change an existing one, you change the data in the table in
2507       embed.fnc as well.  Here's a sample entry from that table:
2508
2509           Apd |SV**   |av_fetch   |AV* ar|I32 key|I32 lval
2510
2511       The first column is a set of flags, the second column the return type,
2512       the third column the name.  Columns after that are the arguments.  The
2513       flags are documented at the top of embed.fnc.
2514
2515       If you edit embed.pl or embed.fnc, you will need to run "make
2516       regen_headers" to force a rebuild of embed.h and other auto-generated
2517       files.
2518
2519   Formatted Printing of IVs, UVs, and NVs
2520       If you are printing IVs, UVs, or NVS instead of the stdio(3) style
2521       formatting codes like %d, %ld, %f, you should use the following macros
2522       for portability
2523
2524               IVdf            IV in decimal
2525               UVuf            UV in decimal
2526               UVof            UV in octal
2527               UVxf            UV in hexadecimal
2528               NVef            NV %e-like
2529               NVff            NV %f-like
2530               NVgf            NV %g-like
2531
2532       These will take care of 64-bit integers and long doubles.  For example:
2533
2534               printf("IV is %" IVdf "\n", iv);
2535
2536       The "IVdf" will expand to whatever is the correct format for the IVs.
2537       Note that the spaces are required around the format in case the code is
2538       compiled with C++, to maintain compliance with its standard.
2539
2540       Note that there are different "long doubles": Perl will use whatever
2541       the compiler has.
2542
2543       If you are printing addresses of pointers, use %p or UVxf combined with
2544       PTR2UV().
2545
2546   Formatted Printing of SVs
2547       The contents of SVs may be printed using the "SVf" format, like so:
2548
2549        Perl_croak(aTHX_ "This croaked because: %" SVf "\n", SvfARG(err_msg))
2550
2551       where "err_msg" is an SV.
2552
2553       Not all scalar types are printable.  Simple values certainly are: one
2554       of IV, UV, NV, or PV.  Also, if the SV is a reference to some value,
2555       either it will be dereferenced and the value printed, or information
2556       about the type of that value and its address are displayed.  The
2557       results of printing any other type of SV are undefined and likely to
2558       lead to an interpreter crash.  NVs are printed using a %g-ish format.
2559
2560       Note that the spaces are required around the "SVf" in case the code is
2561       compiled with C++, to maintain compliance with its standard.
2562
2563       Note that any filehandle being printed to under UTF-8 must be expecting
2564       UTF-8 in order to get good results and avoid Wide-character warnings.
2565       One way to do this for typical filehandles is to invoke perl with the
2566       "-C"> parameter.  (See "-C [number/list]" in perlrun.
2567
2568       You can use this to concatenate two scalars:
2569
2570        SV *var1 = get_sv("var1", GV_ADD);
2571        SV *var2 = get_sv("var2", GV_ADD);
2572        SV *var3 = newSVpvf("var1=%" SVf " and var2=%" SVf,
2573                            SVfARG(var1), SVfARG(var2));
2574
2575   Formatted Printing of Strings
2576       If you just want the bytes printed in a 7bit NUL-terminated string, you
2577       can just use %s (assuming they are all really only 7bit).  But if there
2578       is a possibility the value will be encoded as UTF-8 or contains bytes
2579       above 0x7F (and therefore 8bit), you should instead use the "UTF8f"
2580       format.  And as its parameter, use the "UTF8fARG()" macro:
2581
2582        chr * msg;
2583
2584        /* U+2018: \xE2\x80\x98 LEFT SINGLE QUOTATION MARK
2585           U+2019: \xE2\x80\x99 RIGHT SINGLE QUOTATION MARK */
2586        if (can_utf8)
2587          msg = "\xE2\x80\x98Uses fancy quotes\xE2\x80\x99";
2588        else
2589          msg = "'Uses simple quotes'";
2590
2591        Perl_croak(aTHX_ "The message is: %" UTF8f "\n",
2592                         UTF8fARG(can_utf8, strlen(msg), msg));
2593
2594       The first parameter to "UTF8fARG" is a boolean: 1 if the string is in
2595       UTF-8; 0 if string is in native byte encoding (Latin1).  The second
2596       parameter is the number of bytes in the string to print.  And the third
2597       and final parameter is a pointer to the first byte in the string.
2598
2599       Note that any filehandle being printed to under UTF-8 must be expecting
2600       UTF-8 in order to get good results and avoid Wide-character warnings.
2601       One way to do this for typical filehandles is to invoke perl with the
2602       "-C"> parameter.  (See "-C [number/list]" in perlrun.
2603
2604   Formatted Printing of "Size_t" and "SSize_t"
2605       The most general way to do this is to cast them to a UV or IV, and
2606       print as in the previous section.
2607
2608       But if you're using "PerlIO_printf()", it's less typing and visual
2609       clutter to use the %z length modifier (for siZe):
2610
2611               PerlIO_printf("STRLEN is %zu\n", len);
2612
2613       This modifier is not portable, so its use should be restricted to
2614       "PerlIO_printf()".
2615
2616   Pointer-To-Integer and Integer-To-Pointer
2617       Because pointer size does not necessarily equal integer size, use the
2618       follow macros to do it right.
2619
2620               PTR2UV(pointer)
2621               PTR2IV(pointer)
2622               PTR2NV(pointer)
2623               INT2PTR(pointertotype, integer)
2624
2625       For example:
2626
2627               IV  iv = ...;
2628               SV *sv = INT2PTR(SV*, iv);
2629
2630       and
2631
2632               AV *av = ...;
2633               UV  uv = PTR2UV(av);
2634
2635   Exception Handling
2636       There are a couple of macros to do very basic exception handling in XS
2637       modules.  You have to define "NO_XSLOCKS" before including XSUB.h to be
2638       able to use these macros:
2639
2640               #define NO_XSLOCKS
2641               #include "XSUB.h"
2642
2643       You can use these macros if you call code that may croak, but you need
2644       to do some cleanup before giving control back to Perl.  For example:
2645
2646               dXCPT;    /* set up necessary variables */
2647
2648               XCPT_TRY_START {
2649                 code_that_may_croak();
2650               } XCPT_TRY_END
2651
2652               XCPT_CATCH
2653               {
2654                 /* do cleanup here */
2655                 XCPT_RETHROW;
2656               }
2657
2658       Note that you always have to rethrow an exception that has been caught.
2659       Using these macros, it is not possible to just catch the exception and
2660       ignore it.  If you have to ignore the exception, you have to use the
2661       "call_*" function.
2662
2663       The advantage of using the above macros is that you don't have to setup
2664       an extra function for "call_*", and that using these macros is faster
2665       than using "call_*".
2666
2667   Source Documentation
2668       There's an effort going on to document the internal functions and
2669       automatically produce reference manuals from them -- perlapi is one
2670       such manual which details all the functions which are available to XS
2671       writers.  perlintern is the autogenerated manual for the functions
2672       which are not part of the API and are supposedly for internal use only.
2673
2674       Source documentation is created by putting POD comments into the C
2675       source, like this:
2676
2677        /*
2678        =for apidoc sv_setiv
2679
2680        Copies an integer into the given SV.  Does not handle 'set' magic.  See
2681        L<perlapi/sv_setiv_mg>.
2682
2683        =cut
2684        */
2685
2686       Please try and supply some documentation if you add functions to the
2687       Perl core.
2688
2689   Backwards compatibility
2690       The Perl API changes over time.  New functions are added or the
2691       interfaces of existing functions are changed.  The "Devel::PPPort"
2692       module tries to provide compatibility code for some of these changes,
2693       so XS writers don't have to code it themselves when supporting multiple
2694       versions of Perl.
2695
2696       "Devel::PPPort" generates a C header file ppport.h that can also be run
2697       as a Perl script.  To generate ppport.h, run:
2698
2699           perl -MDevel::PPPort -eDevel::PPPort::WriteFile
2700
2701       Besides checking existing XS code, the script can also be used to
2702       retrieve compatibility information for various API calls using the
2703       "--api-info" command line switch.  For example:
2704
2705         % perl ppport.h --api-info=sv_magicext
2706
2707       For details, see "perldoc ppport.h".
2708

Unicode Support

2710       Perl 5.6.0 introduced Unicode support.  It's important for porters and
2711       XS writers to understand this support and make sure that the code they
2712       write does not corrupt Unicode data.
2713
2714   What is Unicode, anyway?
2715       In the olden, less enlightened times, we all used to use ASCII.  Most
2716       of us did, anyway.  The big problem with ASCII is that it's American.
2717       Well, no, that's not actually the problem; the problem is that it's not
2718       particularly useful for people who don't use the Roman alphabet.  What
2719       used to happen was that particular languages would stick their own
2720       alphabet in the upper range of the sequence, between 128 and 255.  Of
2721       course, we then ended up with plenty of variants that weren't quite
2722       ASCII, and the whole point of it being a standard was lost.
2723
2724       Worse still, if you've got a language like Chinese or Japanese that has
2725       hundreds or thousands of characters, then you really can't fit them
2726       into a mere 256, so they had to forget about ASCII altogether, and
2727       build their own systems using pairs of numbers to refer to one
2728       character.
2729
2730       To fix this, some people formed Unicode, Inc. and produced a new
2731       character set containing all the characters you can possibly think of
2732       and more.  There are several ways of representing these characters, and
2733       the one Perl uses is called UTF-8.  UTF-8 uses a variable number of
2734       bytes to represent a character.  You can learn more about Unicode and
2735       Perl's Unicode model in perlunicode.
2736
2737       (On EBCDIC platforms, Perl uses instead UTF-EBCDIC, which is a form of
2738       UTF-8 adapted for EBCDIC platforms.  Below, we just talk about UTF-8.
2739       UTF-EBCDIC is like UTF-8, but the details are different.  The macros
2740       hide the differences from you, just remember that the particular
2741       numbers and bit patterns presented below will differ in UTF-EBCDIC.)
2742
2743   How can I recognise a UTF-8 string?
2744       You can't.  This is because UTF-8 data is stored in bytes just like
2745       non-UTF-8 data.  The Unicode character 200, (0xC8 for you hex types)
2746       capital E with a grave accent, is represented by the two bytes
2747       "v196.172".  Unfortunately, the non-Unicode string "chr(196).chr(172)"
2748       has that byte sequence as well.  So you can't tell just by looking --
2749       this is what makes Unicode input an interesting problem.
2750
2751       In general, you either have to know what you're dealing with, or you
2752       have to guess.  The API function "is_utf8_string" can help; it'll tell
2753       you if a string contains only valid UTF-8 characters, and the chances
2754       of a non-UTF-8 string looking like valid UTF-8 become very small very
2755       quickly with increasing string length.  On a character-by-character
2756       basis, "isUTF8_CHAR" will tell you whether the current character in a
2757       string is valid UTF-8.
2758
2759   How does UTF-8 represent Unicode characters?
2760       As mentioned above, UTF-8 uses a variable number of bytes to store a
2761       character.  Characters with values 0...127 are stored in one byte, just
2762       like good ol' ASCII.  Character 128 is stored as "v194.128"; this
2763       continues up to character 191, which is "v194.191".  Now we've run out
2764       of bits (191 is binary 10111111) so we move on; character 192 is
2765       "v195.128".  And so it goes on, moving to three bytes at character
2766       2048.  "Unicode Encodings" in perlunicode has pictures of how this
2767       works.
2768
2769       Assuming you know you're dealing with a UTF-8 string, you can find out
2770       how long the first character in it is with the "UTF8SKIP" macro:
2771
2772           char *utf = "\305\233\340\240\201";
2773           I32 len;
2774
2775           len = UTF8SKIP(utf); /* len is 2 here */
2776           utf += len;
2777           len = UTF8SKIP(utf); /* len is 3 here */
2778
2779       Another way to skip over characters in a UTF-8 string is to use
2780       "utf8_hop", which takes a string and a number of characters to skip
2781       over.  You're on your own about bounds checking, though, so don't use
2782       it lightly.
2783
2784       All bytes in a multi-byte UTF-8 character will have the high bit set,
2785       so you can test if you need to do something special with this character
2786       like this (the "UTF8_IS_INVARIANT()" is a macro that tests whether the
2787       byte is encoded as a single byte even in UTF-8):
2788
2789           U8 *utf;     /* Initialize this to point to the beginning of the
2790                           sequence to convert */
2791           U8 *utf_end; /* Initialize this to 1 beyond the end of the sequence
2792                           pointed to by 'utf' */
2793           UV uv;       /* Returned code point; note: a UV, not a U8, not a
2794                           char */
2795           STRLEN len; /* Returned length of character in bytes */
2796
2797           if (!UTF8_IS_INVARIANT(*utf))
2798               /* Must treat this as UTF-8 */
2799               uv = utf8_to_uvchr_buf(utf, utf_end, &len);
2800           else
2801               /* OK to treat this character as a byte */
2802               uv = *utf;
2803
2804       You can also see in that example that we use "utf8_to_uvchr_buf" to get
2805       the value of the character; the inverse function "uvchr_to_utf8" is
2806       available for putting a UV into UTF-8:
2807
2808           if (!UVCHR_IS_INVARIANT(uv))
2809               /* Must treat this as UTF8 */
2810               utf8 = uvchr_to_utf8(utf8, uv);
2811           else
2812               /* OK to treat this character as a byte */
2813               *utf8++ = uv;
2814
2815       You must convert characters to UVs using the above functions if you're
2816       ever in a situation where you have to match UTF-8 and non-UTF-8
2817       characters.  You may not skip over UTF-8 characters in this case.  If
2818       you do this, you'll lose the ability to match hi-bit non-UTF-8
2819       characters; for instance, if your UTF-8 string contains "v196.172", and
2820       you skip that character, you can never match a "chr(200)" in a
2821       non-UTF-8 string.  So don't do that!
2822
2823       (Note that we don't have to test for invariant characters in the
2824       examples above.  The functions work on any well-formed UTF-8 input.
2825       It's just that its faster to avoid the function overhead when it's not
2826       needed.)
2827
2828   How does Perl store UTF-8 strings?
2829       Currently, Perl deals with UTF-8 strings and non-UTF-8 strings slightly
2830       differently.  A flag in the SV, "SVf_UTF8", indicates that the string
2831       is internally encoded as UTF-8.  Without it, the byte value is the
2832       codepoint number and vice versa.  This flag is only meaningful if the
2833       SV is "SvPOK" or immediately after stringification via "SvPV" or a
2834       similar macro.  You can check and manipulate this flag with the
2835       following macros:
2836
2837           SvUTF8(sv)
2838           SvUTF8_on(sv)
2839           SvUTF8_off(sv)
2840
2841       This flag has an important effect on Perl's treatment of the string: if
2842       UTF-8 data is not properly distinguished, regular expressions,
2843       "length", "substr" and other string handling operations will have
2844       undesirable (wrong) results.
2845
2846       The problem comes when you have, for instance, a string that isn't
2847       flagged as UTF-8, and contains a byte sequence that could be UTF-8 --
2848       especially when combining non-UTF-8 and UTF-8 strings.
2849
2850       Never forget that the "SVf_UTF8" flag is separate from the PV value;
2851       you need to be sure you don't accidentally knock it off while you're
2852       manipulating SVs.  More specifically, you cannot expect to do this:
2853
2854           SV *sv;
2855           SV *nsv;
2856           STRLEN len;
2857           char *p;
2858
2859           p = SvPV(sv, len);
2860           frobnicate(p);
2861           nsv = newSVpvn(p, len);
2862
2863       The "char*" string does not tell you the whole story, and you can't
2864       copy or reconstruct an SV just by copying the string value.  Check if
2865       the old SV has the UTF8 flag set (after the "SvPV" call), and act
2866       accordingly:
2867
2868           p = SvPV(sv, len);
2869           is_utf8 = SvUTF8(sv);
2870           frobnicate(p, is_utf8);
2871           nsv = newSVpvn(p, len);
2872           if (is_utf8)
2873               SvUTF8_on(nsv);
2874
2875       In the above, your "frobnicate" function has been changed to be made
2876       aware of whether or not it's dealing with UTF-8 data, so that it can
2877       handle the string appropriately.
2878
2879       Since just passing an SV to an XS function and copying the data of the
2880       SV is not enough to copy the UTF8 flags, even less right is just
2881       passing a "char *" to an XS function.
2882
2883       For full generality, use the "DO_UTF8" macro to see if the string in an
2884       SV is to be treated as UTF-8.  This takes into account if the call to
2885       the XS function is being made from within the scope of "use bytes".  If
2886       so, the underlying bytes that comprise the UTF-8 string are to be
2887       exposed, rather than the character they represent.  But this pragma
2888       should only really be used for debugging and perhaps low-level testing
2889       at the byte level.  Hence most XS code need not concern itself with
2890       this, but various areas of the perl core do need to support it.
2891
2892       And this isn't the whole story.  Starting in Perl v5.12, strings that
2893       aren't encoded in UTF-8 may also be treated as Unicode under various
2894       conditions (see "ASCII Rules versus Unicode Rules" in perlunicode).
2895       This is only really a problem for characters whose ordinals are between
2896       128 and 255, and their behavior varies under ASCII versus Unicode rules
2897       in ways that your code cares about (see "The "Unicode Bug"" in
2898       perlunicode).  There is no published API for dealing with this, as it
2899       is subject to change, but you can look at the code for "pp_lc" in pp.c
2900       for an example as to how it's currently done.
2901
2902   How do I convert a string to UTF-8?
2903       If you're mixing UTF-8 and non-UTF-8 strings, it is necessary to
2904       upgrade the non-UTF-8 strings to UTF-8.  If you've got an SV, the
2905       easiest way to do this is:
2906
2907           sv_utf8_upgrade(sv);
2908
2909       However, you must not do this, for example:
2910
2911           if (!SvUTF8(left))
2912               sv_utf8_upgrade(left);
2913
2914       If you do this in a binary operator, you will actually change one of
2915       the strings that came into the operator, and, while it shouldn't be
2916       noticeable by the end user, it can cause problems in deficient code.
2917
2918       Instead, "bytes_to_utf8" will give you a UTF-8-encoded copy of its
2919       string argument.  This is useful for having the data available for
2920       comparisons and so on, without harming the original SV.  There's also
2921       "utf8_to_bytes" to go the other way, but naturally, this will fail if
2922       the string contains any characters above 255 that can't be represented
2923       in a single byte.
2924
2925   How do I compare strings?
2926       "sv_cmp" in perlapi and "sv_cmp_flags" in perlapi do a lexigraphic
2927       comparison of two SV's, and handle UTF-8ness properly.  Note, however,
2928       that Unicode specifies a much fancier mechanism for collation,
2929       available via the Unicode::Collate module.
2930
2931       To just compare two strings for equality/non-equality, you can just use
2932       "memEQ()" and "memNE()" as usual, except the strings must be both UTF-8
2933       or not UTF-8 encoded.
2934
2935       To compare two strings case-insensitively, use "foldEQ_utf8()" (the
2936       strings don't have to have the same UTF-8ness).
2937
2938   Is there anything else I need to know?
2939       Not really.  Just remember these things:
2940
2941       •  There's no way to tell if a "char *" or "U8 *" string is UTF-8 or
2942          not.  But you can tell if an SV is to be treated as UTF-8 by calling
2943          "DO_UTF8" on it, after stringifying it with "SvPV" or a similar
2944          macro.  And, you can tell if SV is actually UTF-8 (even if it is not
2945          to be treated as such) by looking at its "SvUTF8" flag (again after
2946          stringifying it).  Don't forget to set the flag if something should
2947          be UTF-8.  Treat the flag as part of the PV, even though it's not --
2948          if you pass on the PV to somewhere, pass on the flag too.
2949
2950       •  If a string is UTF-8, always use "utf8_to_uvchr_buf" to get at the
2951          value, unless "UTF8_IS_INVARIANT(*s)" in which case you can use *s.
2952
2953       •  When writing a character UV to a UTF-8 string, always use
2954          "uvchr_to_utf8", unless "UVCHR_IS_INVARIANT(uv))" in which case you
2955          can use "*s = uv".
2956
2957       •  Mixing UTF-8 and non-UTF-8 strings is tricky.  Use "bytes_to_utf8"
2958          to get a new string which is UTF-8 encoded, and then combine them.
2959

Custom Operators

2961       Custom operator support is an experimental feature that allows you to
2962       define your own ops.  This is primarily to allow the building of
2963       interpreters for other languages in the Perl core, but it also allows
2964       optimizations through the creation of "macro-ops" (ops which perform
2965       the functions of multiple ops which are usually executed together, such
2966       as "gvsv, gvsv, add".)
2967
2968       This feature is implemented as a new op type, "OP_CUSTOM".  The Perl
2969       core does not "know" anything special about this op type, and so it
2970       will not be involved in any optimizations.  This also means that you
2971       can define your custom ops to be any op structure -- unary, binary,
2972       list and so on -- you like.
2973
2974       It's important to know what custom operators won't do for you.  They
2975       won't let you add new syntax to Perl, directly.  They won't even let
2976       you add new keywords, directly.  In fact, they won't change the way
2977       Perl compiles a program at all.  You have to do those changes yourself,
2978       after Perl has compiled the program.  You do this either by
2979       manipulating the op tree using a "CHECK" block and the "B::Generate"
2980       module, or by adding a custom peephole optimizer with the "optimize"
2981       module.
2982
2983       When you do this, you replace ordinary Perl ops with custom ops by
2984       creating ops with the type "OP_CUSTOM" and the "op_ppaddr" of your own
2985       PP function.  This should be defined in XS code, and should look like
2986       the PP ops in "pp_*.c".  You are responsible for ensuring that your op
2987       takes the appropriate number of values from the stack, and you are
2988       responsible for adding stack marks if necessary.
2989
2990       You should also "register" your op with the Perl interpreter so that it
2991       can produce sensible error and warning messages.  Since it is possible
2992       to have multiple custom ops within the one "logical" op type
2993       "OP_CUSTOM", Perl uses the value of "o->op_ppaddr" to determine which
2994       custom op it is dealing with.  You should create an "XOP" structure for
2995       each ppaddr you use, set the properties of the custom op with
2996       "XopENTRY_set", and register the structure against the ppaddr using
2997       "Perl_custom_op_register".  A trivial example might look like:
2998
2999           static XOP my_xop;
3000           static OP *my_pp(pTHX);
3001
3002           BOOT:
3003               XopENTRY_set(&my_xop, xop_name, "myxop");
3004               XopENTRY_set(&my_xop, xop_desc, "Useless custom op");
3005               Perl_custom_op_register(aTHX_ my_pp, &my_xop);
3006
3007       The available fields in the structure are:
3008
3009       xop_name
3010           A short name for your op.  This will be included in some error
3011           messages, and will also be returned as "$op->name" by the B module,
3012           so it will appear in the output of module like B::Concise.
3013
3014       xop_desc
3015           A short description of the function of the op.
3016
3017       xop_class
3018           Which of the various *OP structures this op uses.  This should be
3019           one of the "OA_*" constants from op.h, namely
3020
3021           OA_BASEOP
3022           OA_UNOP
3023           OA_BINOP
3024           OA_LOGOP
3025           OA_LISTOP
3026           OA_PMOP
3027           OA_SVOP
3028           OA_PADOP
3029           OA_PVOP_OR_SVOP
3030               This should be interpreted as '"PVOP"' only.  The "_OR_SVOP" is
3031               because the only core "PVOP", "OP_TRANS", can sometimes be a
3032               "SVOP" instead.
3033
3034           OA_LOOP
3035           OA_COP
3036
3037           The other "OA_*" constants should not be used.
3038
3039       xop_peep
3040           This member is of type "Perl_cpeep_t", which expands to "void
3041           (*Perl_cpeep_t)(aTHX_ OP *o, OP *oldop)".  If it is set, this
3042           function will be called from "Perl_rpeep" when ops of this type are
3043           encountered by the peephole optimizer.  o is the OP that needs
3044           optimizing; oldop is the previous OP optimized, whose "op_next"
3045           points to o.
3046
3047       "B::Generate" directly supports the creation of custom ops by name.
3048

Stacks

3050       Descriptions above occasionally refer to "the stack", but there are in
3051       fact many stack-like data structures within the perl interpreter. When
3052       otherwise unqualified, "the stack" usually refers to the value stack.
3053
3054       The various stacks have different purposes, and operate in slightly
3055       different ways. Their differences are noted below.
3056
3057   Value Stack
3058       This stack stores the values that regular perl code is operating on,
3059       usually intermediate values of expressions within a statement. The
3060       stack itself is formed of an array of SV pointers.
3061
3062       The base of this stack is pointed to by the interpreter variable
3063       "PL_stack_base", of type "SV **".
3064
3065       The head of the stack is "PL_stack_sp", and points to the most
3066       recently-pushed item.
3067
3068       Items are pushed to the stack by using the "PUSHs()" macro or its
3069       variants described above; "XPUSHs()", "mPUSHs()", "mXPUSHs()" and the
3070       typed versions. Note carefully that the non-"X" versions of these
3071       macros do not check the size of the stack and assume it to be big
3072       enough. These must be paired with a suitable check of the stack's size,
3073       such as the "EXTEND" macro to ensure it is large enough. For example
3074
3075           EXTEND(SP, 4);
3076           mPUSHi(10);
3077           mPUSHi(20);
3078           mPUSHi(30);
3079           mPUSHi(40);
3080
3081       This is slightly more performant than making four separate checks in
3082       four separate "mXPUSHi()" calls.
3083
3084       As a further performance optimisation, the various "PUSH" macros all
3085       operate using a local variable "SP", rather than the interpreter-global
3086       variable "PL_stack_sp". This variable is declared by the "dSP" macro -
3087       though it is normally implied by XSUBs and similar so it is rare you
3088       have to consider it directly. Once declared, the "PUSH" macros will
3089       operate only on this local variable, so before invoking any other perl
3090       core functions you must use the "PUTBACK" macro to return the value
3091       from the local "SP" variable back to the interpreter variable.
3092       Similarly, after calling a perl core function which may have had reason
3093       to move the stack or push/pop values to it, you must use the "SPAGAIN"
3094       macro which refreshes the local "SP" value back from the interpreter
3095       one.
3096
3097       Items are popped from the stack by using the "POPs" macro or its typed
3098       versions, There is also a macro "TOPs" that inspects the topmost item
3099       without removing it.
3100
3101       Note specifically that SV pointers on the value stack do not contribute
3102       to the overall reference count of the xVs being referred to. If newly-
3103       created xVs are being pushed to the stack you must arrange for them to
3104       be destroyed at a suitable time; usually by using one of the "mPUSH*"
3105       macros or "sv_2mortal()" to mortalise the xV.
3106
3107   Mark Stack
3108       The value stack stores individual perl scalar values as temporaries
3109       between expressions. Some perl expressions operate on entire lists; for
3110       that purpose we need to know where on the stack each list begins. This
3111       is the purpose of the mark stack.
3112
3113       The mark stack stores integers as I32 values, which are the height of
3114       the value stack at the time before the list began; thus the mark itself
3115       actually points to the value stack entry one before the list. The list
3116       itself starts at "mark + 1".
3117
3118       The base of this stack is pointed to by the interpreter variable
3119       "PL_markstack", of type "I32 *".
3120
3121       The head of the stack is "PL_markstack_ptr", and points to the most
3122       recently-pushed item.
3123
3124       Items are pushed to the stack by using the "PUSHMARK()" macro. Even
3125       though the stack itself stores (value) stack indices as integers, the
3126       "PUSHMARK" macro should be given a stack pointer directly; it will
3127       calculate the index offset by comparing to the "PL_stack_sp" variable.
3128       Thus almost always the code to perform this is
3129
3130           PUSHMARK(SP);
3131
3132       Items are popped from the stack by the "POPMARK" macro. There is also a
3133       macro "TOPMARK" that inspects the topmost item without removing it.
3134       These macros return I32 index values directly. There is also the
3135       "dMARK" macro which declares a new SV double-pointer variable, called
3136       "mark", which points at the marked stack slot; this is the usual macro
3137       that C code will use when operating on lists given on the stack.
3138
3139       As noted above, the "mark" variable itself will point at the most
3140       recently pushed value on the value stack before the list begins, and so
3141       the list itself starts at "mark + 1". The values of the list may be
3142       iterated by code such as
3143
3144           for(SV **svp = mark + 1; svp <= PL_stack_sp; svp++) {
3145             SV *item = *svp;
3146             ...
3147           }
3148
3149       Note specifically in the case that the list is already empty, "mark"
3150       will equal "PL_stack_sp".
3151
3152       Because the "mark" variable is converted to a pointer on the value
3153       stack, extra care must be taken if "EXTEND" or any of the "XPUSH"
3154       macros are invoked within the function, because the stack may need to
3155       be moved to extend it and so the existing pointer will now be invalid.
3156       If this may be a problem, a possible solution is to track the mark
3157       offset as an integer and track the mark itself later on after the stack
3158       had been moved.
3159
3160           I32 markoff = POPMARK;
3161
3162           ...
3163
3164           SP **mark = PL_stack_base + markoff;
3165
3166   Temporaries Stack
3167       As noted above, xV references on the main value stack do not contribute
3168       to the reference count of an xV, and so another mechanism is used to
3169       track when temporary values which live on the stack must be released.
3170       This is the job of the temporaries stack.
3171
3172       The temporaries stack stores pointers to xVs whose reference counts
3173       will be decremented soon.
3174
3175       The base of this stack is pointed to by the interpreter variable
3176       "PL_tmps_stack", of type "SV **".
3177
3178       The head of the stack is indexed by "PL_tmps_ix", an integer which
3179       stores the index in the array of the most recently-pushed item.
3180
3181       There is no public API to directly push items to the temporaries stack.
3182       Instead, the API function "sv_2mortal()" is used to mortalize an xV,
3183       adding its address to the temporaries stack.
3184
3185       Likewise, there is no public API to read values from the temporaries
3186       stack.  Instead. the macros "SAVETMPS" and "FREETPMS" are used. The
3187       "SAVETMPS" macro establishes the base levels of the temporaries stack,
3188       by capturing the current value of "PL_tmps_ix" into "PL_tmps_floor" and
3189       saving the previous value to the save stack. Thereafter, whenever
3190       "FREETMPS" is invoked all of the temporaries that have been pushed
3191       since that level are reclaimed.
3192
3193       While it is common to see these two macros in pairs within an "ENTER"/
3194       "LEAVE" pair, it is not necessary to match them. It is permitted to
3195       invoke "FREETMPS" multiple times since the most recent "SAVETMPS"; for
3196       example in a loop iterating over elements of a list. While you can
3197       invoke "SAVETMPS" multiple times within a scope pair, it is unlikely to
3198       be useful. Subsequent invocations will move the temporaries floor
3199       further up, thus effectively trapping the existing temporaries to only
3200       be released at the end of the scope.
3201
3202   Save Stack
3203       The save stack is used by perl to implement the "local" keyword and
3204       other similar behaviours; any cleanup operations that need to be
3205       performed when leaving the current scope. Items pushed to this stack
3206       generally capture the current value of some internal variable or state,
3207       which will be restored when the scope is unwound due to leaving,
3208       "return", "die", "goto" or other reasons.
3209
3210       Whereas other perl internal stacks store individual items all of the
3211       same type (usually SV pointers or integers), the items pushed to the
3212       save stack are formed of many different types, having multiple fields
3213       to them. For example, the "SAVEt_INT" type needs to store both the
3214       address of the "int" variable to restore, and the value to restore it
3215       to. This information could have been stored using fields of a "struct",
3216       but would have to be large enough to store three pointers in the
3217       largest case, which would waste a lot of space in most of the smaller
3218       cases.
3219
3220       Instead, the stack stores information in a variable-length encoding of
3221       "ANY" structures. The final value pushed is stored in the "UV" field
3222       which encodes the kind of item held by the preceeding items; the count
3223       and types of which will depend on what kind of item is being stored.
3224       The kind field is pushed last because that will be the first field to
3225       be popped when unwinding items from the stack.
3226
3227       The base of this stack is pointed to by the interpreter variable
3228       "PL_savestack", of type "ANY *".
3229
3230       The head of the stack is indexed by "PL_savestack_ix", an integer which
3231       stores the index in the array at which the next item should be pushed.
3232       (Note that this is different to most other stacks, which reference the
3233       most recently-pushed item).
3234
3235       Items are pushed to the save stack by using the various "SAVE...()"
3236       macros.  Many of these macros take a variable and store both its
3237       address and current value on the save stack, ensuring that value gets
3238       restored on scope exit.
3239
3240           SAVEI8(i8)
3241           SAVEI16(i16)
3242           SAVEI32(i32)
3243           SAVEINT(i)
3244           ...
3245
3246       There are also a variety of other special-purpose macros which save
3247       particular types or values of interest. "SAVETMPS" has already been
3248       mentioned above.  Others include "SAVEFREEPV" which arranges for a PV
3249       (i.e. a string buffer) to be freed, or "SAVEDESTRUCTOR" which arranges
3250       for a given function pointer to be invoked on scope exit. A full list
3251       of such macros can be found in scope.h.
3252
3253       There is no public API for popping individual values or items from the
3254       save stack. Instead, via the scope stack, the "ENTER" and "LEAVE" pair
3255       form a way to start and stop nested scopes. Leaving a nested scope via
3256       "LEAVE" will restore all of the saved values that had been pushed since
3257       the most recent "ENTER".
3258
3259   Scope Stack
3260       As with the mark stack to the value stack, the scope stack forms a pair
3261       with the save stack. The scope stack stores the height of the save
3262       stack at which nested scopes begin, and allows the save stack to be
3263       unwound back to that point when the scope is left.
3264
3265       When perl is built with debugging enabled, there is a second part to
3266       this stack storing human-readable string names describing the type of
3267       stack context. Each push operation saves the name as well as the height
3268       of the save stack, and each pop operation checks the topmost name with
3269       what is expected, causing an assertion failure if the name does not
3270       match.
3271
3272       The base of this stack is pointed to by the interpreter variable
3273       "PL_scopestack", of type "I32 *". If enabled, the scope stack names are
3274       stored in a separate array pointed to by "PL_scopestack_name", of type
3275       "const char **".
3276
3277       The head of the stack is indexed by "PL_scopestack_ix", an integer
3278       which stores the index of the array or arrays at which the next item
3279       should be pushed. (Note that this is different to most other stacks,
3280       which reference the most recently-pushed item).
3281
3282       Values are pushed to the scope stack using the "ENTER" macro, which
3283       begins a new nested scope. Any items pushed to the save stack are then
3284       restored at the next nested invocation of the "LEAVE" macro.
3285

Dynamic Scope and the Context Stack

3287       Note: this section describes a non-public internal API that is subject
3288       to change without notice.
3289
3290   Introduction to the context stack
3291       In Perl, dynamic scoping refers to the runtime nesting of things like
3292       subroutine calls, evals etc, as well as the entering and exiting of
3293       block scopes. For example, the restoring of a "local"ised variable is
3294       determined by the dynamic scope.
3295
3296       Perl tracks the dynamic scope by a data structure called the context
3297       stack, which is an array of "PERL_CONTEXT" structures, and which is
3298       itself a big union for all the types of context. Whenever a new scope
3299       is entered (such as a block, a "for" loop, or a subroutine call), a new
3300       context entry is pushed onto the stack. Similarly when leaving a block
3301       or returning from a subroutine call etc. a context is popped. Since the
3302       context stack represents the current dynamic scope, it can be searched.
3303       For example, "next LABEL" searches back through the stack looking for a
3304       loop context that matches the label; "return" pops contexts until it
3305       finds a sub or eval context or similar; "caller" examines sub contexts
3306       on the stack.
3307
3308       Each context entry is labelled with a context type, "cx_type". Typical
3309       context types are "CXt_SUB", "CXt_EVAL" etc., as well as "CXt_BLOCK"
3310       and "CXt_NULL" which represent a basic scope (as pushed by "pp_enter")
3311       and a sort block. The type determines which part of the context union
3312       are valid.
3313
3314       The main division in the context struct is between a substitution scope
3315       ("CXt_SUBST") and block scopes, which are everything else. The former
3316       is just used while executing "s///e", and won't be discussed further
3317       here.
3318
3319       All the block scope types share a common base, which corresponds to
3320       "CXt_BLOCK". This stores the old values of various scope-related
3321       variables like "PL_curpm", as well as information about the current
3322       scope, such as "gimme". On scope exit, the old variables are restored.
3323
3324       Particular block scope types store extra per-type information. For
3325       example, "CXt_SUB" stores the currently executing CV, while the various
3326       for loop types might hold the original loop variable SV. On scope exit,
3327       the per-type data is processed; for example the CV has its reference
3328       count decremented, and the original loop variable is restored.
3329
3330       The macro "cxstack" returns the base of the current context stack,
3331       while "cxstack_ix" is the index of the current frame within that stack.
3332
3333       In fact, the context stack is actually part of a stack-of-stacks
3334       system; whenever something unusual is done such as calling a "DESTROY"
3335       or tie handler, a new stack is pushed, then popped at the end.
3336
3337       Note that the API described here changed considerably in perl 5.24;
3338       prior to that, big macros like "PUSHBLOCK" and "POPSUB" were used; in
3339       5.24 they were replaced by the inline static functions described below.
3340       In addition, the ordering and detail of how these macros/function work
3341       changed in many ways, often subtly. In particular they didn't handle
3342       saving the savestack and temps stack positions, and required additional
3343       "ENTER", "SAVETMPS" and "LEAVE" compared to the new functions. The old-
3344       style macros will not be described further.
3345
3346   Pushing contexts
3347       For pushing a new context, the two basic functions are "cx =
3348       cx_pushblock()", which pushes a new basic context block and returns its
3349       address, and a family of similar functions with names like
3350       "cx_pushsub(cx)" which populate the additional type-dependent fields in
3351       the "cx" struct. Note that "CXt_NULL" and "CXt_BLOCK" don't have their
3352       own push functions, as they don't store any data beyond that pushed by
3353       "cx_pushblock".
3354
3355       The fields of the context struct and the arguments to the "cx_*"
3356       functions are subject to change between perl releases, representing
3357       whatever is convenient or efficient for that release.
3358
3359       A typical context stack pushing can be found in "pp_entersub"; the
3360       following shows a simplified and stripped-down example of a non-XS
3361       call, along with comments showing roughly what each function does.
3362
3363        dMARK;
3364        U8 gimme      = GIMME_V;
3365        bool hasargs  = cBOOL(PL_op->op_flags & OPf_STACKED);
3366        OP *retop     = PL_op->op_next;
3367        I32 old_ss_ix = PL_savestack_ix;
3368        CV *cv        = ....;
3369
3370        /* ... make mortal copies of stack args which are PADTMPs here ... */
3371
3372        /* ... do any additional savestack pushes here ... */
3373
3374        /* Now push a new context entry of type 'CXt_SUB'; initially just
3375         * doing the actions common to all block types: */
3376
3377        cx = cx_pushblock(CXt_SUB, gimme, MARK, old_ss_ix);
3378
3379            /* this does (approximately):
3380                CXINC;              /* cxstack_ix++ (grow if necessary) */
3381                cx = CX_CUR();      /* and get the address of new frame */
3382                cx->cx_type        = CXt_SUB;
3383                cx->blk_gimme      = gimme;
3384                cx->blk_oldsp      = MARK - PL_stack_base;
3385                cx->blk_oldsaveix  = old_ss_ix;
3386                cx->blk_oldcop     = PL_curcop;
3387                cx->blk_oldmarksp  = PL_markstack_ptr - PL_markstack;
3388                cx->blk_oldscopesp = PL_scopestack_ix;
3389                cx->blk_oldpm      = PL_curpm;
3390                cx->blk_old_tmpsfloor = PL_tmps_floor;
3391
3392                PL_tmps_floor        = PL_tmps_ix;
3393            */
3394
3395
3396        /* then update the new context frame with subroutine-specific info,
3397         * such as the CV about to be executed: */
3398
3399        cx_pushsub(cx, cv, retop, hasargs);
3400
3401            /* this does (approximately):
3402                cx->blk_sub.cv          = cv;
3403                cx->blk_sub.olddepth    = CvDEPTH(cv);
3404                cx->blk_sub.prevcomppad = PL_comppad;
3405                cx->cx_type            |= (hasargs) ? CXp_HASARGS : 0;
3406                cx->blk_sub.retop       = retop;
3407                SvREFCNT_inc_simple_void_NN(cv);
3408            */
3409
3410       Note that "cx_pushblock()" sets two new floors: for the args stack (to
3411       "MARK") and the temps stack (to "PL_tmps_ix"). While executing at this
3412       scope level, every "nextstate" (amongst others) will reset the args and
3413       tmps stack levels to these floors. Note that since "cx_pushblock" uses
3414       the current value of "PL_tmps_ix" rather than it being passed as an
3415       arg, this dictates at what point "cx_pushblock" should be called. In
3416       particular, any new mortals which should be freed only on scope exit
3417       (rather than at the next "nextstate") should be created first.
3418
3419       Most callers of "cx_pushblock" simply set the new args stack floor to
3420       the top of the previous stack frame, but for "CXt_LOOP_LIST" it stores
3421       the items being iterated over on the stack, and so sets "blk_oldsp" to
3422       the top of these items instead. Note that, contrary to its name,
3423       "blk_oldsp" doesn't always represent the value to restore "PL_stack_sp"
3424       to on scope exit.
3425
3426       Note the early capture of "PL_savestack_ix" to "old_ss_ix", which is
3427       later passed as an arg to "cx_pushblock". In the case of "pp_entersub",
3428       this is because, although most values needing saving are stored in
3429       fields of the context struct, an extra value needs saving only when the
3430       debugger is running, and it doesn't make sense to bloat the struct for
3431       this rare case. So instead it is saved on the savestack. Since this
3432       value gets calculated and saved before the context is pushed, it is
3433       necessary to pass the old value of "PL_savestack_ix" to "cx_pushblock",
3434       to ensure that the saved value gets freed during scope exit.  For most
3435       users of "cx_pushblock", where nothing needs pushing on the save stack,
3436       "PL_savestack_ix" is just passed directly as an arg to "cx_pushblock".
3437
3438       Note that where possible, values should be saved in the context struct
3439       rather than on the save stack; it's much faster that way.
3440
3441       Normally "cx_pushblock" should be immediately followed by the
3442       appropriate "cx_pushfoo", with nothing between them; this is because if
3443       code in-between could die (e.g. a warning upgraded to fatal), then the
3444       context stack unwinding code in "dounwind" would see (in the example
3445       above) a "CXt_SUB" context frame, but without all the subroutine-
3446       specific fields set, and crashes would soon ensue.
3447
3448       Where the two must be separate, initially set the type to "CXt_NULL" or
3449       "CXt_BLOCK", and later change it to "CXt_foo" when doing the
3450       "cx_pushfoo". This is exactly what "pp_enteriter" does, once it's
3451       determined which type of loop it's pushing.
3452
3453   Popping contexts
3454       Contexts are popped using "cx_popsub()" etc. and "cx_popblock()". Note
3455       however, that unlike "cx_pushblock", neither of these functions
3456       actually decrement the current context stack index; this is done
3457       separately using "CX_POP()".
3458
3459       There are two main ways that contexts are popped. During normal
3460       execution as scopes are exited, functions like "pp_leave",
3461       "pp_leaveloop" and "pp_leavesub" process and pop just one context using
3462       "cx_popfoo" and "cx_popblock". On the other hand, things like
3463       "pp_return" and "next" may have to pop back several scopes until a sub
3464       or loop context is found, and exceptions (such as "die") need to pop
3465       back contexts until an eval context is found. Both of these are
3466       accomplished by "dounwind()", which is capable of processing and
3467       popping all contexts above the target one.
3468
3469       Here is a typical example of context popping, as found in "pp_leavesub"
3470       (simplified slightly):
3471
3472        U8 gimme;
3473        PERL_CONTEXT *cx;
3474        SV **oldsp;
3475        OP *retop;
3476
3477        cx = CX_CUR();
3478
3479        gimme = cx->blk_gimme;
3480        oldsp = PL_stack_base + cx->blk_oldsp; /* last arg of previous frame */
3481
3482        if (gimme == G_VOID)
3483            PL_stack_sp = oldsp;
3484        else
3485            leave_adjust_stacks(oldsp, oldsp, gimme, 0);
3486
3487        CX_LEAVE_SCOPE(cx);
3488        cx_popsub(cx);
3489        cx_popblock(cx);
3490        retop = cx->blk_sub.retop;
3491        CX_POP(cx);
3492
3493        return retop;
3494
3495       The steps above are in a very specific order, designed to be the
3496       reverse order of when the context was pushed. The first thing to do is
3497       to copy and/or protect any return arguments and free any temps in the
3498       current scope. Scope exits like an rvalue sub normally return a mortal
3499       copy of their return args (as opposed to lvalue subs). It is important
3500       to make this copy before the save stack is popped or variables are
3501       restored, or bad things like the following can happen:
3502
3503           sub f { my $x =...; $x }  # $x freed before we get to copy it
3504           sub f { /(...)/;    $1 }  # PL_curpm restored before $1 copied
3505
3506       Although we wish to free any temps at the same time, we have to be
3507       careful not to free any temps which are keeping return args alive; nor
3508       to free the temps we have just created while mortal copying return
3509       args. Fortunately, "leave_adjust_stacks()" is capable of making mortal
3510       copies of return args, shifting args down the stack, and only
3511       processing those entries on the temps stack that are safe to do so.
3512
3513       In void context no args are returned, so it's more efficient to skip
3514       calling "leave_adjust_stacks()". Also in void context, a "nextstate" op
3515       is likely to be imminently called which will do a "FREETMPS", so
3516       there's no need to do that either.
3517
3518       The next step is to pop savestack entries: "CX_LEAVE_SCOPE(cx)" is just
3519       defined as "LEAVE_SCOPE(cx->blk_oldsaveix)". Note that during the
3520       popping, it's possible for perl to call destructors, call "STORE" to
3521       undo localisations of tied vars, and so on. Any of these can die or
3522       call "exit()". In this case, "dounwind()" will be called, and the
3523       current context stack frame will be re-processed. Thus it is vital that
3524       all steps in popping a context are done in such a way to support
3525       reentrancy.  The other alternative, of decrementing "cxstack_ix" before
3526       processing the frame, would lead to leaks and the like if something
3527       died halfway through, or overwriting of the current frame.
3528
3529       "CX_LEAVE_SCOPE" itself is safely re-entrant: if only half the
3530       savestack items have been popped before dying and getting trapped by
3531       eval, then the "CX_LEAVE_SCOPE"s in "dounwind" or "pp_leaveeval" will
3532       continue where the first one left off.
3533
3534       The next step is the type-specific context processing; in this case
3535       "cx_popsub". In part, this looks like:
3536
3537           cv = cx->blk_sub.cv;
3538           CvDEPTH(cv) = cx->blk_sub.olddepth;
3539           cx->blk_sub.cv = NULL;
3540           SvREFCNT_dec(cv);
3541
3542       where its processing the just-executed CV. Note that before it
3543       decrements the CV's reference count, it nulls the "blk_sub.cv". This
3544       means that if it re-enters, the CV won't be freed twice. It also means
3545       that you can't rely on such type-specific fields having useful values
3546       after the return from "cx_popfoo".
3547
3548       Next, "cx_popblock" restores all the various interpreter vars to their
3549       previous values or previous high water marks; it expands to:
3550
3551           PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
3552           PL_scopestack_ix = cx->blk_oldscopesp;
3553           PL_curpm         = cx->blk_oldpm;
3554           PL_curcop        = cx->blk_oldcop;
3555           PL_tmps_floor    = cx->blk_old_tmpsfloor;
3556
3557       Note that it doesn't restore "PL_stack_sp"; as mentioned earlier, which
3558       value to restore it to depends on the context type (specifically "for
3559       (list) {}"), and what args (if any) it returns; and that will already
3560       have been sorted out earlier by "leave_adjust_stacks()".
3561
3562       Finally, the context stack pointer is actually decremented by
3563       "CX_POP(cx)".  After this point, it's possible that that the current
3564       context frame could be overwritten by other contexts being pushed.
3565       Although things like ties and "DESTROY" are supposed to work within a
3566       new context stack, it's best not to assume this. Indeed on debugging
3567       builds, "CX_POP(cx)" deliberately sets "cx" to null to detect code that
3568       is still relying on the field values in that context frame. Note in the
3569       "pp_leavesub()" example above, we grab "blk_sub.retop" before calling
3570       "CX_POP".
3571
3572   Redoing contexts
3573       Finally, there is "cx_topblock(cx)", which acts like a
3574       super-"nextstate" as regards to resetting various vars to their base
3575       values. It is used in places like "pp_next", "pp_redo" and "pp_goto"
3576       where rather than exiting a scope, we want to re-initialise the scope.
3577       As well as resetting "PL_stack_sp" like "nextstate", it also resets
3578       "PL_markstack_ptr", "PL_scopestack_ix" and "PL_curpm". Note that it
3579       doesn't do a "FREETMPS".
3580

Slab-based operator allocation

3582       Note: this section describes a non-public internal API that is subject
3583       to change without notice.
3584
3585       Perl's internal error-handling mechanisms implement "die" (and its
3586       internal equivalents) using longjmp. If this occurs during lexing,
3587       parsing or compilation, we must ensure that any ops allocated as part
3588       of the compilation process are freed. (Older Perl versions did not
3589       adequately handle this situation: when failing a parse, they would leak
3590       ops that were stored in C "auto" variables and not linked anywhere
3591       else.)
3592
3593       To handle this situation, Perl uses op slabs that are attached to the
3594       currently-compiling CV. A slab is a chunk of allocated memory. New ops
3595       are allocated as regions of the slab. If the slab fills up, a new one
3596       is created (and linked from the previous one). When an error occurs and
3597       the CV is freed, any ops remaining are freed.
3598
3599       Each op is preceded by two pointers: one points to the next op in the
3600       slab, and the other points to the slab that owns it. The next-op
3601       pointer is needed so that Perl can iterate over a slab and free all its
3602       ops. (Op structures are of different sizes, so the slab's ops can't
3603       merely be treated as a dense array.)  The slab pointer is needed for
3604       accessing a reference count on the slab: when the last op on a slab is
3605       freed, the slab itself is freed.
3606
3607       The slab allocator puts the ops at the end of the slab first. This will
3608       tend to allocate the leaves of the op tree first, and the layout will
3609       therefore hopefully be cache-friendly. In addition, this means that
3610       there's no need to store the size of the slab (see below on why slabs
3611       vary in size), because Perl can follow pointers to find the last op.
3612
3613       It might seem possible eliminate slab reference counts altogether, by
3614       having all ops implicitly attached to "PL_compcv" when allocated and
3615       freed when the CV is freed. That would also allow "op_free" to skip
3616       "FreeOp" altogether, and thus free ops faster. But that doesn't work in
3617       those cases where ops need to survive beyond their CVs, such as re-
3618       evals.
3619
3620       The CV also has to have a reference count on the slab. Sometimes the
3621       first op created is immediately freed. If the reference count of the
3622       slab reaches 0, then it will be freed with the CV still pointing to it.
3623
3624       CVs use the "CVf_SLABBED" flag to indicate that the CV has a reference
3625       count on the slab. When this flag is set, the slab is accessible via
3626       "CvSTART" when "CvROOT" is not set, or by subtracting two pointers
3627       "(2*sizeof(I32 *))" from "CvROOT" when it is set. The alternative to
3628       this approach of sneaking the slab into "CvSTART" during compilation
3629       would be to enlarge the "xpvcv" struct by another pointer. But that
3630       would make all CVs larger, even though slab-based op freeing is
3631       typically of benefit only for programs that make significant use of
3632       string eval.
3633
3634       When the "CVf_SLABBED" flag is set, the CV takes responsibility for
3635       freeing the slab. If "CvROOT" is not set when the CV is freed or
3636       undeffed, it is assumed that a compilation error has occurred, so the
3637       op slab is traversed and all the ops are freed.
3638
3639       Under normal circumstances, the CV forgets about its slab (decrementing
3640       the reference count) when the root is attached. So the slab reference
3641       counting that happens when ops are freed takes care of freeing the
3642       slab. In some cases, the CV is told to forget about the slab
3643       ("cv_forget_slab") precisely so that the ops can survive after the CV
3644       is done away with.
3645
3646       Forgetting the slab when the root is attached is not strictly
3647       necessary, but avoids potential problems with "CvROOT" being written
3648       over. There is code all over the place, both in core and on CPAN, that
3649       does things with "CvROOT", so forgetting the slab makes things more
3650       robust and avoids potential problems.
3651
3652       Since the CV takes ownership of its slab when flagged, that flag is
3653       never copied when a CV is cloned, as one CV could free a slab that
3654       another CV still points to, since forced freeing of ops ignores the
3655       reference count (but asserts that it looks right).
3656
3657       To avoid slab fragmentation, freed ops are marked as freed and attached
3658       to the slab's freed chain (an idea stolen from DBM::Deep). Those freed
3659       ops are reused when possible. Not reusing freed ops would be simpler,
3660       but it would result in significantly higher memory usage for programs
3661       with large "if (DEBUG) {...}" blocks.
3662
3663       "SAVEFREEOP" is slightly problematic under this scheme. Sometimes it
3664       can cause an op to be freed after its CV. If the CV has forcibly freed
3665       the ops on its slab and the slab itself, then we will be fiddling with
3666       a freed slab. Making "SAVEFREEOP" a no-op doesn't help, as sometimes an
3667       op can be savefreed when there is no compilation error, so the op would
3668       never be freed. It holds a reference count on the slab, so the whole
3669       slab would leak. So "SAVEFREEOP" now sets a special flag on the op
3670       ("->op_savefree"). The forced freeing of ops after a compilation error
3671       won't free any ops thus marked.
3672
3673       Since many pieces of code create tiny subroutines consisting of only a
3674       few ops, and since a huge slab would be quite a bit of baggage for
3675       those to carry around, the first slab is always very small. To avoid
3676       allocating too many slabs for a single CV, each subsequent slab is
3677       twice the size of the previous.
3678
3679       Smartmatch expects to be able to allocate an op at run time, run it,
3680       and then throw it away. For that to work the op is simply malloced when
3681       PL_compcv hasn't been set up. So all slab-allocated ops are marked as
3682       such ("->op_slabbed"), to distinguish them from malloced ops.
3683

AUTHORS

3685       Until May 1997, this document was maintained by Jeff Okamoto
3686       <okamoto@corp.hp.com>.  It is now maintained as part of Perl itself by
3687       the Perl 5 Porters <perl5-porters@perl.org>.
3688
3689       With lots of help and suggestions from Dean Roehrich, Malcolm Beattie,
3690       Andreas Koenig, Paul Hudson, Ilya Zakharevich, Paul Marquess, Neil
3691       Bowers, Matthew Green, Tim Bunce, Spider Boardman, Ulrich Pfeifer,
3692       Stephen McCamant, and Gurusamy Sarathy.
3693

SEE ALSO

3695       perlapi, perlintern, perlxs, perlembed
3696
3697
3698
3699perl v5.32.1                      2021-05-31                       PERLGUTS(1)
Impressum