3 * Copyright (C) 2012 by Larry Wall and others
5 * You may distribute under the terms of either the GNU General Public
6 * License or the Artistic License, as specified in the README file.
8 * This file is a home for static inline functions that cannot go in other
9 * headers files, because they depend on proto.h (included after most other
10 * headers) or struct definitions.
12 * Each section names the header file that the functions "belong" to.
15 /* ------------------------------- av.h ------------------------------- */
17 PERL_STATIC_INLINE SSize_t
18 S_av_top_index(pTHX_ AV *av)
20 PERL_ARGS_ASSERT_AV_TOP_INDEX;
21 assert(SvTYPE(av) == SVt_PVAV);
26 /* ------------------------------- cv.h ------------------------------- */
28 PERL_STATIC_INLINE GV *
32 ? Perl_cvgv_from_hek(aTHX_ sv)
33 : ((XPVCV*)MUTABLE_PTR(SvANY(sv)))->xcv_gv_u.xcv_gv;
36 PERL_STATIC_INLINE I32 *
37 S_CvDEPTHp(const CV * const sv)
39 assert(SvTYPE(sv) == SVt_PVCV || SvTYPE(sv) == SVt_PVFM);
40 return &((XPVCV*)SvANY(sv))->xcv_depth;
44 CvPROTO returns the prototype as stored, which is not necessarily what
45 the interpreter should be using. Specifically, the interpreter assumes
46 that spaces have been stripped, which has been the case if the prototype
47 was added by toke.c, but is generally not the case if it was added elsewhere.
48 Since we can't enforce the spacelessness at assignment time, this routine
49 provides a temporary copy at parse time with spaces removed.
50 I<orig> is the start of the original buffer, I<len> is the length of the
51 prototype and will be updated when this returns.
55 PERL_STATIC_INLINE char *
56 S_strip_spaces(pTHX_ const char * orig, STRLEN * const len)
60 tmpsv = newSVpvn_flags(orig, *len, SVs_TEMP);
68 *len = tmps - SvPVX(tmpsv);
73 /* ------------------------------- mg.h ------------------------------- */
75 #if defined(PERL_CORE) || defined(PERL_EXT)
76 /* assumes get-magic and stringification have already occurred */
77 PERL_STATIC_INLINE STRLEN
78 S_MgBYTEPOS(pTHX_ MAGIC *mg, SV *sv, const char *s, STRLEN len)
80 assert(mg->mg_type == PERL_MAGIC_regex_global);
81 assert(mg->mg_len != -1);
82 if (mg->mg_flags & MGf_BYTES || !DO_UTF8(sv))
83 return (STRLEN)mg->mg_len;
85 const STRLEN pos = (STRLEN)mg->mg_len;
86 /* Without this check, we may read past the end of the buffer: */
87 if (pos > sv_or_pv_len_utf8(sv, s, len)) return len+1;
88 return sv_or_pv_pos_u2b(sv, s, pos, NULL);
93 /* ------------------------------- pad.h ------------------------------ */
95 #if defined(PERL_IN_PAD_C) || defined(PERL_IN_OP_C)
96 PERL_STATIC_INLINE bool
97 PadnameIN_SCOPE(const PADNAME * const pn, const U32 seq)
99 /* is seq within the range _LOW to _HIGH ?
100 * This is complicated by the fact that PL_cop_seqmax
101 * may have wrapped around at some point */
102 if (COP_SEQ_RANGE_LOW(pn) == PERL_PADSEQ_INTRO)
103 return FALSE; /* not yet introduced */
105 if (COP_SEQ_RANGE_HIGH(pn) == PERL_PADSEQ_INTRO) {
106 /* in compiling scope */
108 (seq > COP_SEQ_RANGE_LOW(pn))
109 ? (seq - COP_SEQ_RANGE_LOW(pn) < (U32_MAX >> 1))
110 : (COP_SEQ_RANGE_LOW(pn) - seq > (U32_MAX >> 1))
115 (COP_SEQ_RANGE_LOW(pn) > COP_SEQ_RANGE_HIGH(pn))
117 ( seq > COP_SEQ_RANGE_LOW(pn)
118 || seq <= COP_SEQ_RANGE_HIGH(pn))
120 : ( seq > COP_SEQ_RANGE_LOW(pn)
121 && seq <= COP_SEQ_RANGE_HIGH(pn))
128 /* ------------------------------- pp.h ------------------------------- */
130 PERL_STATIC_INLINE I32
133 DEBUG_s(DEBUG_v(PerlIO_printf(Perl_debug_log,
134 "MARK top %p %" IVdf "\n",
136 (IV)*PL_markstack_ptr)));
137 return *PL_markstack_ptr;
140 PERL_STATIC_INLINE I32
143 DEBUG_s(DEBUG_v(PerlIO_printf(Perl_debug_log,
144 "MARK pop %p %" IVdf "\n",
145 (PL_markstack_ptr-1),
146 (IV)*(PL_markstack_ptr-1))));
147 assert((PL_markstack_ptr > PL_markstack) || !"MARK underflow");
148 return *PL_markstack_ptr--;
151 /* ----------------------------- regexp.h ----------------------------- */
153 PERL_STATIC_INLINE struct regexp *
154 S_ReANY(const REGEXP * const re)
156 XPV* const p = (XPV*)SvANY(re);
157 assert(isREGEXP(re));
158 return SvTYPE(re) == SVt_PVLV ? p->xpv_len_u.xpvlenu_rx
159 : (struct regexp *)p;
162 /* ------------------------------- sv.h ------------------------------- */
164 PERL_STATIC_INLINE SV *
165 S_SvREFCNT_inc(SV *sv)
167 if (LIKELY(sv != NULL))
171 PERL_STATIC_INLINE SV *
172 S_SvREFCNT_inc_NN(SV *sv)
177 PERL_STATIC_INLINE void
178 S_SvREFCNT_inc_void(SV *sv)
180 if (LIKELY(sv != NULL))
183 PERL_STATIC_INLINE void
184 S_SvREFCNT_dec(pTHX_ SV *sv)
186 if (LIKELY(sv != NULL)) {
187 U32 rc = SvREFCNT(sv);
189 SvREFCNT(sv) = rc - 1;
191 Perl_sv_free2(aTHX_ sv, rc);
195 PERL_STATIC_INLINE void
196 S_SvREFCNT_dec_NN(pTHX_ SV *sv)
198 U32 rc = SvREFCNT(sv);
200 SvREFCNT(sv) = rc - 1;
202 Perl_sv_free2(aTHX_ sv, rc);
205 PERL_STATIC_INLINE void
209 if (SvOBJECT(SvRV(sv))) HvAMAGIC_on(SvSTASH(SvRV(sv)));
211 PERL_STATIC_INLINE void
214 if (SvROK(sv) && SvOBJECT(SvRV(sv)))
215 HvAMAGIC_off(SvSTASH(SvRV(sv)));
218 PERL_STATIC_INLINE U32
219 S_SvPADSTALE_on(SV *sv)
221 assert(!(SvFLAGS(sv) & SVs_PADTMP));
222 return SvFLAGS(sv) |= SVs_PADSTALE;
224 PERL_STATIC_INLINE U32
225 S_SvPADSTALE_off(SV *sv)
227 assert(!(SvFLAGS(sv) & SVs_PADTMP));
228 return SvFLAGS(sv) &= ~SVs_PADSTALE;
230 #if defined(PERL_CORE) || defined (PERL_EXT)
231 PERL_STATIC_INLINE STRLEN
232 S_sv_or_pv_pos_u2b(pTHX_ SV *sv, const char *pv, STRLEN pos, STRLEN *lenp)
234 PERL_ARGS_ASSERT_SV_OR_PV_POS_U2B;
236 U8 *hopped = utf8_hop((U8 *)pv, pos);
237 if (lenp) *lenp = (STRLEN)(utf8_hop(hopped, *lenp) - hopped);
238 return (STRLEN)(hopped - (U8 *)pv);
240 return sv_pos_u2b_flags(sv,pos,lenp,SV_CONST_RETURN);
244 /* ------------------------------- handy.h ------------------------------- */
246 /* saves machine code for a common noreturn idiom typically used in Newx*() */
247 #ifdef GCC_DIAG_PRAGMA
248 GCC_DIAG_IGNORE(-Wunused-function) /* Intentionally left semicolonless. */
251 S_croak_memory_wrap(void)
253 Perl_croak_nocontext("%s",PL_memory_wrap);
255 #ifdef GCC_DIAG_PRAGMA
256 GCC_DIAG_RESTORE /* Intentionally left semicolonless. */
259 /* ------------------------------- utf8.h ------------------------------- */
262 =head1 Unicode Support
265 PERL_STATIC_INLINE void
266 S_append_utf8_from_native_byte(const U8 byte, U8** dest)
268 /* Takes an input 'byte' (Latin1 or EBCDIC) and appends it to the UTF-8
269 * encoded string at '*dest', updating '*dest' to include it */
271 PERL_ARGS_ASSERT_APPEND_UTF8_FROM_NATIVE_BYTE;
273 if (NATIVE_BYTE_IS_INVARIANT(byte))
276 *((*dest)++) = UTF8_EIGHT_BIT_HI(byte);
277 *((*dest)++) = UTF8_EIGHT_BIT_LO(byte);
282 =for apidoc valid_utf8_to_uvchr
283 Like C<L</utf8_to_uvchr_buf>>, but should only be called when it is known that
284 the next character in the input UTF-8 string C<s> is well-formed (I<e.g.>,
285 it passes C<L</isUTF8_CHAR>>. Surrogates, non-character code points, and
286 non-Unicode code points are allowed.
292 PERL_STATIC_INLINE UV
293 Perl_valid_utf8_to_uvchr(const U8 *s, STRLEN *retlen)
295 const UV expectlen = UTF8SKIP(s);
296 const U8* send = s + expectlen;
299 PERL_ARGS_ASSERT_VALID_UTF8_TO_UVCHR;
305 /* An invariant is trivially returned */
306 if (expectlen == 1) {
310 /* Remove the leading bits that indicate the number of bytes, leaving just
311 * the bits that are part of the value */
312 uv = NATIVE_UTF8_TO_I8(uv) & UTF_START_MASK(expectlen);
314 /* Now, loop through the remaining bytes, accumulating each into the
315 * working total as we go. (I khw tried unrolling the loop for up to 4
316 * bytes, but there was no performance improvement) */
317 for (++s; s < send; s++) {
318 uv = UTF8_ACCUMULATE(uv, *s);
321 return UNI_TO_NATIVE(uv);
326 =for apidoc is_utf8_invariant_string
328 Returns TRUE if the first C<len> bytes of the string C<s> are the same
329 regardless of the UTF-8 encoding of the string (or UTF-EBCDIC encoding on
330 EBCDIC machines); otherwise it returns FALSE. That is, it returns TRUE if they
331 are UTF-8 invariant. On ASCII-ish machines, all the ASCII characters and only
332 the ASCII characters fit this definition. On EBCDIC machines, the ASCII-range
333 characters are invariant, but so also are the C1 controls.
335 If C<len> is 0, it will be calculated using C<strlen(s)>, (which means if you
336 use this option, that C<s> can't have embedded C<NUL> characters and has to
337 have a terminating C<NUL> byte).
340 C<L</is_utf8_string>>,
341 C<L</is_utf8_string_flags>>,
342 C<L</is_utf8_string_loc>>,
343 C<L</is_utf8_string_loc_flags>>,
344 C<L</is_utf8_string_loclen>>,
345 C<L</is_utf8_string_loclen_flags>>,
346 C<L</is_utf8_fixed_width_buf_flags>>,
347 C<L</is_utf8_fixed_width_buf_loc_flags>>,
348 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
349 C<L</is_strict_utf8_string>>,
350 C<L</is_strict_utf8_string_loc>>,
351 C<L</is_strict_utf8_string_loclen>>,
352 C<L</is_c9strict_utf8_string>>,
353 C<L</is_c9strict_utf8_string_loc>>,
355 C<L</is_c9strict_utf8_string_loclen>>.
361 #define is_utf8_invariant_string(s, len) \
362 is_utf8_invariant_string_loc(s, len, NULL)
365 =for apidoc is_utf8_invariant_string_loc
367 Like C<L</is_utf8_invariant_string>> but upon failure, stores the location of
368 the first UTF-8 variant character in the C<ep> pointer; if all characters are
369 UTF-8 invariant, this function does not change the contents of C<*ep>.
375 PERL_STATIC_INLINE bool
376 S_is_utf8_invariant_string_loc(const U8* const s, STRLEN len, const U8 ** ep)
381 PERL_ARGS_ASSERT_IS_UTF8_INVARIANT_STRING_LOC;
384 len = strlen((const char *)s);
390 /* Try to get the widest word on this platform */
391 # ifdef HAS_LONG_LONG
392 # define PERL_WORDCAST unsigned long long
393 # define PERL_WORDSIZE LONGLONGSIZE
395 # define PERL_WORDCAST UV
396 # define PERL_WORDSIZE UVSIZE
399 # if PERL_WORDSIZE == 4
400 # define PERL_VARIANTS_WORD_MASK 0x80808080
401 # define PERL_WORD_BOUNDARY_MASK 0x3
402 # elif PERL_WORDSIZE == 8
403 # define PERL_VARIANTS_WORD_MASK UINT64_C(0x8080808080808080)
404 # define PERL_WORD_BOUNDARY_MASK 0x7
406 # error Unexpected word size
409 /* Process per-byte until reach word boundary. XXX This loop could be
410 * eliminated if we knew that this platform had fast unaligned reads */
411 while (x < send && (PTR2nat(x) & PERL_WORD_BOUNDARY_MASK)) {
412 if (! UTF8_IS_INVARIANT(*x)) {
422 /* Process per-word as long as we have at least a full word left */
423 while (x + PERL_WORDSIZE <= send) {
424 if ((* (PERL_WORDCAST *) x) & PERL_VARIANTS_WORD_MASK) {
426 /* Found a variant. Just return if caller doesn't want its exact
432 /* Otherwise fall into final loop to find which byte it is */
438 # undef PERL_WORDCAST
439 # undef PERL_WORDSIZE
440 # undef PERL_WORD_BOUNDARY_MASK
441 # undef PERL_VARIANTS_WORD_MASK
444 /* Process per-byte */
446 if (! UTF8_IS_INVARIANT(*x)) {
461 =for apidoc is_utf8_string
463 Returns TRUE if the first C<len> bytes of string C<s> form a valid
464 Perl-extended-UTF-8 string; returns FALSE otherwise. If C<len> is 0, it will
465 be calculated using C<strlen(s)> (which means if you use this option, that C<s>
466 can't have embedded C<NUL> characters and has to have a terminating C<NUL>
467 byte). Note that all characters being ASCII constitute 'a valid UTF-8 string'.
469 This function considers Perl's extended UTF-8 to be valid. That means that
470 code points above Unicode, surrogates, and non-character code points are
471 considered valid by this function. Use C<L</is_strict_utf8_string>>,
472 C<L</is_c9strict_utf8_string>>, or C<L</is_utf8_string_flags>> to restrict what
473 code points are considered valid.
476 C<L</is_utf8_invariant_string>>,
477 C<L</is_utf8_invariant_string_loc>>,
478 C<L</is_utf8_string_loc>>,
479 C<L</is_utf8_string_loclen>>,
480 C<L</is_utf8_fixed_width_buf_flags>>,
481 C<L</is_utf8_fixed_width_buf_loc_flags>>,
482 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
487 #define is_utf8_string(s, len) is_utf8_string_loclen(s, len, NULL, NULL)
489 #if defined(PERL_CORE) || defined (PERL_EXT)
492 =for apidoc is_utf8_non_invariant_string
494 Returns TRUE if L<perlapi/is_utf8_invariant_string> returns FALSE for the first
495 C<len> bytes of the string C<s>, but they are, nonetheless, legal Perl-extended
496 UTF-8; otherwise returns FALSE.
498 A TRUE return means that at least one code point represented by the sequence
499 either is a wide character not representable as a single byte, or the
500 representation differs depending on whether the sequence is encoded in UTF-8 or
504 C<L<perlapi/is_utf8_invariant_string>>,
505 C<L<perlapi/is_utf8_string>>
509 This is commonly used to determine if a SV's UTF-8 flag should be turned on.
510 It needn't be if its string is entirely UTF-8 invariant, and it shouldn't be if
511 it otherwise contains invalid UTF-8.
513 It is an internal function because khw thinks that XS code shouldn't be working
514 at this low a level. A valid use case could change that.
518 PERL_STATIC_INLINE bool
519 S_is_utf8_non_invariant_string(const U8* const s, STRLEN len)
521 const U8 * first_variant;
523 PERL_ARGS_ASSERT_IS_UTF8_NON_INVARIANT_STRING;
525 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
529 return is_utf8_string(first_variant, len - (first_variant - s));
535 =for apidoc is_strict_utf8_string
537 Returns TRUE if the first C<len> bytes of string C<s> form a valid
538 UTF-8-encoded string that is fully interchangeable by any application using
539 Unicode rules; otherwise it returns FALSE. If C<len> is 0, it will be
540 calculated using C<strlen(s)> (which means if you use this option, that C<s>
541 can't have embedded C<NUL> characters and has to have a terminating C<NUL>
542 byte). Note that all characters being ASCII constitute 'a valid UTF-8 string'.
544 This function returns FALSE for strings containing any
545 code points above the Unicode max of 0x10FFFF, surrogate code points, or
546 non-character code points.
549 C<L</is_utf8_invariant_string>>,
550 C<L</is_utf8_invariant_string_loc>>,
551 C<L</is_utf8_string>>,
552 C<L</is_utf8_string_flags>>,
553 C<L</is_utf8_string_loc>>,
554 C<L</is_utf8_string_loc_flags>>,
555 C<L</is_utf8_string_loclen>>,
556 C<L</is_utf8_string_loclen_flags>>,
557 C<L</is_utf8_fixed_width_buf_flags>>,
558 C<L</is_utf8_fixed_width_buf_loc_flags>>,
559 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
560 C<L</is_strict_utf8_string_loc>>,
561 C<L</is_strict_utf8_string_loclen>>,
562 C<L</is_c9strict_utf8_string>>,
563 C<L</is_c9strict_utf8_string_loc>>,
565 C<L</is_c9strict_utf8_string_loclen>>.
570 #define is_strict_utf8_string(s, len) is_strict_utf8_string_loclen(s, len, NULL, NULL)
573 =for apidoc is_c9strict_utf8_string
575 Returns TRUE if the first C<len> bytes of string C<s> form a valid
576 UTF-8-encoded string that conforms to
577 L<Unicode Corrigendum #9|http://www.unicode.org/versions/corrigendum9.html>;
578 otherwise it returns FALSE. If C<len> is 0, it will be calculated using
579 C<strlen(s)> (which means if you use this option, that C<s> can't have embedded
580 C<NUL> characters and has to have a terminating C<NUL> byte). Note that all
581 characters being ASCII constitute 'a valid UTF-8 string'.
583 This function returns FALSE for strings containing any code points above the
584 Unicode max of 0x10FFFF or surrogate code points, but accepts non-character
586 L<Corrigendum #9|http://www.unicode.org/versions/corrigendum9.html>.
589 C<L</is_utf8_invariant_string>>,
590 C<L</is_utf8_invariant_string_loc>>,
591 C<L</is_utf8_string>>,
592 C<L</is_utf8_string_flags>>,
593 C<L</is_utf8_string_loc>>,
594 C<L</is_utf8_string_loc_flags>>,
595 C<L</is_utf8_string_loclen>>,
596 C<L</is_utf8_string_loclen_flags>>,
597 C<L</is_utf8_fixed_width_buf_flags>>,
598 C<L</is_utf8_fixed_width_buf_loc_flags>>,
599 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
600 C<L</is_strict_utf8_string>>,
601 C<L</is_strict_utf8_string_loc>>,
602 C<L</is_strict_utf8_string_loclen>>,
603 C<L</is_c9strict_utf8_string_loc>>,
605 C<L</is_c9strict_utf8_string_loclen>>.
610 #define is_c9strict_utf8_string(s, len) is_c9strict_utf8_string_loclen(s, len, NULL, 0)
613 =for apidoc is_utf8_string_flags
615 Returns TRUE if the first C<len> bytes of string C<s> form a valid
616 UTF-8 string, subject to the restrictions imposed by C<flags>;
617 returns FALSE otherwise. If C<len> is 0, it will be calculated
618 using C<strlen(s)> (which means if you use this option, that C<s> can't have
619 embedded C<NUL> characters and has to have a terminating C<NUL> byte). Note
620 that all characters being ASCII constitute 'a valid UTF-8 string'.
622 If C<flags> is 0, this gives the same results as C<L</is_utf8_string>>; if
623 C<flags> is C<UTF8_DISALLOW_ILLEGAL_INTERCHANGE>, this gives the same results
624 as C<L</is_strict_utf8_string>>; and if C<flags> is
625 C<UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE>, this gives the same results as
626 C<L</is_c9strict_utf8_string>>. Otherwise C<flags> may be any
627 combination of the C<UTF8_DISALLOW_I<foo>> flags understood by
628 C<L</utf8n_to_uvchr>>, with the same meanings.
631 C<L</is_utf8_invariant_string>>,
632 C<L</is_utf8_invariant_string_loc>>,
633 C<L</is_utf8_string>>,
634 C<L</is_utf8_string_loc>>,
635 C<L</is_utf8_string_loc_flags>>,
636 C<L</is_utf8_string_loclen>>,
637 C<L</is_utf8_string_loclen_flags>>,
638 C<L</is_utf8_fixed_width_buf_flags>>,
639 C<L</is_utf8_fixed_width_buf_loc_flags>>,
640 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
641 C<L</is_strict_utf8_string>>,
642 C<L</is_strict_utf8_string_loc>>,
643 C<L</is_strict_utf8_string_loclen>>,
644 C<L</is_c9strict_utf8_string>>,
645 C<L</is_c9strict_utf8_string_loc>>,
647 C<L</is_c9strict_utf8_string_loclen>>.
652 PERL_STATIC_INLINE bool
653 S_is_utf8_string_flags(const U8 *s, STRLEN len, const U32 flags)
655 const U8 * first_variant;
657 PERL_ARGS_ASSERT_IS_UTF8_STRING_FLAGS;
658 assert(0 == (flags & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE
659 |UTF8_DISALLOW_PERL_EXTENDED)));
662 len = strlen((const char *)s);
666 return is_utf8_string(s, len);
669 if ((flags & ~UTF8_DISALLOW_PERL_EXTENDED)
670 == UTF8_DISALLOW_ILLEGAL_INTERCHANGE)
672 return is_strict_utf8_string(s, len);
675 if ((flags & ~UTF8_DISALLOW_PERL_EXTENDED)
676 == UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE)
678 return is_c9strict_utf8_string(s, len);
681 if (! is_utf8_invariant_string_loc(s, len, &first_variant)) {
682 const U8* const send = s + len;
683 const U8* x = first_variant;
686 STRLEN cur_len = isUTF8_CHAR_flags(x, send, flags);
687 if (UNLIKELY(! cur_len)) {
699 =for apidoc is_utf8_string_loc
701 Like C<L</is_utf8_string>> but stores the location of the failure (in the
702 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
703 "utf8ness success") in the C<ep> pointer.
705 See also C<L</is_utf8_string_loclen>>.
710 #define is_utf8_string_loc(s, len, ep) is_utf8_string_loclen(s, len, ep, 0)
714 =for apidoc is_utf8_string_loclen
716 Like C<L</is_utf8_string>> but stores the location of the failure (in the
717 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
718 "utf8ness success") in the C<ep> pointer, and the number of UTF-8
719 encoded characters in the C<el> pointer.
721 See also C<L</is_utf8_string_loc>>.
726 PERL_STATIC_INLINE bool
727 Perl_is_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el)
729 const U8 * first_variant;
731 PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN;
734 len = strlen((const char *) s);
737 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
749 const U8* const send = s + len;
750 const U8* x = first_variant;
751 STRLEN outlen = first_variant - s;
754 const STRLEN cur_len = isUTF8_CHAR(x, send);
755 if (UNLIKELY(! cur_len)) {
775 =for apidoc is_strict_utf8_string_loc
777 Like C<L</is_strict_utf8_string>> but stores the location of the failure (in the
778 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
779 "utf8ness success") in the C<ep> pointer.
781 See also C<L</is_strict_utf8_string_loclen>>.
786 #define is_strict_utf8_string_loc(s, len, ep) \
787 is_strict_utf8_string_loclen(s, len, ep, 0)
791 =for apidoc is_strict_utf8_string_loclen
793 Like C<L</is_strict_utf8_string>> but stores the location of the failure (in the
794 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
795 "utf8ness success") in the C<ep> pointer, and the number of UTF-8
796 encoded characters in the C<el> pointer.
798 See also C<L</is_strict_utf8_string_loc>>.
803 PERL_STATIC_INLINE bool
804 S_is_strict_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el)
806 const U8 * first_variant;
808 PERL_ARGS_ASSERT_IS_STRICT_UTF8_STRING_LOCLEN;
811 len = strlen((const char *) s);
814 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
826 const U8* const send = s + len;
827 const U8* x = first_variant;
828 STRLEN outlen = first_variant - s;
831 const STRLEN cur_len = isSTRICT_UTF8_CHAR(x, send);
832 if (UNLIKELY(! cur_len)) {
852 =for apidoc is_c9strict_utf8_string_loc
854 Like C<L</is_c9strict_utf8_string>> but stores the location of the failure (in
855 the case of "utf8ness failure") or the location C<s>+C<len> (in the case of
856 "utf8ness success") in the C<ep> pointer.
858 See also C<L</is_c9strict_utf8_string_loclen>>.
863 #define is_c9strict_utf8_string_loc(s, len, ep) \
864 is_c9strict_utf8_string_loclen(s, len, ep, 0)
868 =for apidoc is_c9strict_utf8_string_loclen
870 Like C<L</is_c9strict_utf8_string>> but stores the location of the failure (in
871 the case of "utf8ness failure") or the location C<s>+C<len> (in the case of
872 "utf8ness success") in the C<ep> pointer, and the number of UTF-8 encoded
873 characters in the C<el> pointer.
875 See also C<L</is_c9strict_utf8_string_loc>>.
880 PERL_STATIC_INLINE bool
881 S_is_c9strict_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el)
883 const U8 * first_variant;
885 PERL_ARGS_ASSERT_IS_C9STRICT_UTF8_STRING_LOCLEN;
888 len = strlen((const char *) s);
891 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
903 const U8* const send = s + len;
904 const U8* x = first_variant;
905 STRLEN outlen = first_variant - s;
908 const STRLEN cur_len = isC9_STRICT_UTF8_CHAR(x, send);
909 if (UNLIKELY(! cur_len)) {
929 =for apidoc is_utf8_string_loc_flags
931 Like C<L</is_utf8_string_flags>> but stores the location of the failure (in the
932 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
933 "utf8ness success") in the C<ep> pointer.
935 See also C<L</is_utf8_string_loclen_flags>>.
940 #define is_utf8_string_loc_flags(s, len, ep, flags) \
941 is_utf8_string_loclen_flags(s, len, ep, 0, flags)
944 /* The above 3 actual functions could have been moved into the more general one
945 * just below, and made #defines that call it with the right 'flags'. They are
946 * currently kept separate to increase their chances of getting inlined */
950 =for apidoc is_utf8_string_loclen_flags
952 Like C<L</is_utf8_string_flags>> but stores the location of the failure (in the
953 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
954 "utf8ness success") in the C<ep> pointer, and the number of UTF-8
955 encoded characters in the C<el> pointer.
957 See also C<L</is_utf8_string_loc_flags>>.
962 PERL_STATIC_INLINE bool
963 S_is_utf8_string_loclen_flags(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el, const U32 flags)
965 const U8 * first_variant;
967 PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN_FLAGS;
968 assert(0 == (flags & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE
969 |UTF8_DISALLOW_PERL_EXTENDED)));
972 len = strlen((const char *) s);
976 return is_utf8_string_loclen(s, len, ep, el);
979 if ((flags & ~UTF8_DISALLOW_PERL_EXTENDED)
980 == UTF8_DISALLOW_ILLEGAL_INTERCHANGE)
982 return is_strict_utf8_string_loclen(s, len, ep, el);
985 if ((flags & ~UTF8_DISALLOW_PERL_EXTENDED)
986 == UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE)
988 return is_c9strict_utf8_string_loclen(s, len, ep, el);
991 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
1003 const U8* send = s + len;
1004 const U8* x = first_variant;
1005 STRLEN outlen = first_variant - s;
1008 const STRLEN cur_len = isUTF8_CHAR_flags(x, send, flags);
1009 if (UNLIKELY(! cur_len)) {
1028 =for apidoc utf8_distance
1030 Returns the number of UTF-8 characters between the UTF-8 pointers C<a>
1033 WARNING: use only if you *know* that the pointers point inside the
1039 PERL_STATIC_INLINE IV
1040 Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b)
1042 PERL_ARGS_ASSERT_UTF8_DISTANCE;
1044 return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a);
1048 =for apidoc utf8_hop
1050 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
1051 forward or backward.
1053 WARNING: do not use the following unless you *know* C<off> is within
1054 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
1055 on the first byte of character or just after the last byte of a character.
1060 PERL_STATIC_INLINE U8 *
1061 Perl_utf8_hop(const U8 *s, SSize_t off)
1063 PERL_ARGS_ASSERT_UTF8_HOP;
1065 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
1066 * the bitops (especially ~) can create illegal UTF-8.
1067 * In other words: in Perl UTF-8 is not just for Unicode. */
1076 while (UTF8_IS_CONTINUATION(*s))
1080 GCC_DIAG_IGNORE(-Wcast-qual);
1086 =for apidoc utf8_hop_forward
1088 Return the UTF-8 pointer C<s> displaced by up to C<off> characters,
1091 C<off> must be non-negative.
1093 C<s> must be before or equal to C<end>.
1095 When moving forward it will not move beyond C<end>.
1097 Will not exceed this limit even if the string is not valid "UTF-8".
1102 PERL_STATIC_INLINE U8 *
1103 Perl_utf8_hop_forward(const U8 *s, SSize_t off, const U8 *end)
1105 PERL_ARGS_ASSERT_UTF8_HOP_FORWARD;
1107 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
1108 * the bitops (especially ~) can create illegal UTF-8.
1109 * In other words: in Perl UTF-8 is not just for Unicode. */
1115 STRLEN skip = UTF8SKIP(s);
1116 if ((STRLEN)(end - s) <= skip) {
1117 GCC_DIAG_IGNORE(-Wcast-qual);
1124 GCC_DIAG_IGNORE(-Wcast-qual);
1130 =for apidoc utf8_hop_back
1132 Return the UTF-8 pointer C<s> displaced by up to C<off> characters,
1135 C<off> must be non-positive.
1137 C<s> must be after or equal to C<start>.
1139 When moving backward it will not move before C<start>.
1141 Will not exceed this limit even if the string is not valid "UTF-8".
1146 PERL_STATIC_INLINE U8 *
1147 Perl_utf8_hop_back(const U8 *s, SSize_t off, const U8 *start)
1149 PERL_ARGS_ASSERT_UTF8_HOP_BACK;
1151 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
1152 * the bitops (especially ~) can create illegal UTF-8.
1153 * In other words: in Perl UTF-8 is not just for Unicode. */
1158 while (off++ && s > start) {
1160 while (UTF8_IS_CONTINUATION(*s) && s > start)
1164 GCC_DIAG_IGNORE(-Wcast-qual);
1170 =for apidoc utf8_hop_safe
1172 Return the UTF-8 pointer C<s> displaced by up to C<off> characters,
1173 either forward or backward.
1175 When moving backward it will not move before C<start>.
1177 When moving forward it will not move beyond C<end>.
1179 Will not exceed those limits even if the string is not valid "UTF-8".
1184 PERL_STATIC_INLINE U8 *
1185 Perl_utf8_hop_safe(const U8 *s, SSize_t off, const U8 *start, const U8 *end)
1187 PERL_ARGS_ASSERT_UTF8_HOP_SAFE;
1189 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
1190 * the bitops (especially ~) can create illegal UTF-8.
1191 * In other words: in Perl UTF-8 is not just for Unicode. */
1193 assert(start <= s && s <= end);
1196 return utf8_hop_forward(s, off, end);
1199 return utf8_hop_back(s, off, start);
1205 =for apidoc is_utf8_valid_partial_char
1207 Returns 0 if the sequence of bytes starting at C<s> and looking no further than
1208 S<C<e - 1>> is the UTF-8 encoding, as extended by Perl, for one or more code
1209 points. Otherwise, it returns 1 if there exists at least one non-empty
1210 sequence of bytes that when appended to sequence C<s>, starting at position
1211 C<e> causes the entire sequence to be the well-formed UTF-8 of some code point;
1212 otherwise returns 0.
1214 In other words this returns TRUE if C<s> points to a partial UTF-8-encoded code
1217 This is useful when a fixed-length buffer is being tested for being well-formed
1218 UTF-8, but the final few bytes in it don't comprise a full character; that is,
1219 it is split somewhere in the middle of the final code point's UTF-8
1220 representation. (Presumably when the buffer is refreshed with the next chunk
1221 of data, the new first bytes will complete the partial code point.) This
1222 function is used to verify that the final bytes in the current buffer are in
1223 fact the legal beginning of some code point, so that if they aren't, the
1224 failure can be signalled without having to wait for the next read.
1228 #define is_utf8_valid_partial_char(s, e) \
1229 is_utf8_valid_partial_char_flags(s, e, 0)
1233 =for apidoc is_utf8_valid_partial_char_flags
1235 Like C<L</is_utf8_valid_partial_char>>, it returns a boolean giving whether
1236 or not the input is a valid UTF-8 encoded partial character, but it takes an
1237 extra parameter, C<flags>, which can further restrict which code points are
1240 If C<flags> is 0, this behaves identically to
1241 C<L</is_utf8_valid_partial_char>>. Otherwise C<flags> can be any combination
1242 of the C<UTF8_DISALLOW_I<foo>> flags accepted by C<L</utf8n_to_uvchr>>. If
1243 there is any sequence of bytes that can complete the input partial character in
1244 such a way that a non-prohibited character is formed, the function returns
1245 TRUE; otherwise FALSE. Non character code points cannot be determined based on
1246 partial character input. But many of the other possible excluded types can be
1247 determined from just the first one or two bytes.
1252 PERL_STATIC_INLINE bool
1253 S_is_utf8_valid_partial_char_flags(const U8 * const s, const U8 * const e, const U32 flags)
1255 PERL_ARGS_ASSERT_IS_UTF8_VALID_PARTIAL_CHAR_FLAGS;
1257 assert(0 == (flags & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE
1258 |UTF8_DISALLOW_PERL_EXTENDED)));
1260 if (s >= e || s + UTF8SKIP(s) <= e) {
1264 return cBOOL(_is_utf8_char_helper(s, e, flags));
1269 =for apidoc is_utf8_fixed_width_buf_flags
1271 Returns TRUE if the fixed-width buffer starting at C<s> with length C<len>
1272 is entirely valid UTF-8, subject to the restrictions given by C<flags>;
1273 otherwise it returns FALSE.
1275 If C<flags> is 0, any well-formed UTF-8, as extended by Perl, is accepted
1276 without restriction. If the final few bytes of the buffer do not form a
1277 complete code point, this will return TRUE anyway, provided that
1278 C<L</is_utf8_valid_partial_char_flags>> returns TRUE for them.
1280 If C<flags> in non-zero, it can be any combination of the
1281 C<UTF8_DISALLOW_I<foo>> flags accepted by C<L</utf8n_to_uvchr>>, and with the
1284 This function differs from C<L</is_utf8_string_flags>> only in that the latter
1285 returns FALSE if the final few bytes of the string don't form a complete code
1290 #define is_utf8_fixed_width_buf_flags(s, len, flags) \
1291 is_utf8_fixed_width_buf_loclen_flags(s, len, 0, 0, flags)
1295 =for apidoc is_utf8_fixed_width_buf_loc_flags
1297 Like C<L</is_utf8_fixed_width_buf_flags>> but stores the location of the
1298 failure in the C<ep> pointer. If the function returns TRUE, C<*ep> will point
1299 to the beginning of any partial character at the end of the buffer; if there is
1300 no partial character C<*ep> will contain C<s>+C<len>.
1302 See also C<L</is_utf8_fixed_width_buf_loclen_flags>>.
1307 #define is_utf8_fixed_width_buf_loc_flags(s, len, loc, flags) \
1308 is_utf8_fixed_width_buf_loclen_flags(s, len, loc, 0, flags)
1312 =for apidoc is_utf8_fixed_width_buf_loclen_flags
1314 Like C<L</is_utf8_fixed_width_buf_loc_flags>> but stores the number of
1315 complete, valid characters found in the C<el> pointer.
1320 PERL_STATIC_INLINE bool
1321 S_is_utf8_fixed_width_buf_loclen_flags(const U8 * const s,
1327 const U8 * maybe_partial;
1329 PERL_ARGS_ASSERT_IS_UTF8_FIXED_WIDTH_BUF_LOCLEN_FLAGS;
1332 ep = &maybe_partial;
1335 /* If it's entirely valid, return that; otherwise see if the only error is
1336 * that the final few bytes are for a partial character */
1337 return is_utf8_string_loclen_flags(s, len, ep, el, flags)
1338 || is_utf8_valid_partial_char_flags(*ep, s + len, flags);
1341 /* ------------------------------- perl.h ----------------------------- */
1344 =head1 Miscellaneous Functions
1346 =for apidoc AiR|bool|is_safe_syscall|const char *pv|STRLEN len|const char *what|const char *op_name
1348 Test that the given C<pv> doesn't contain any internal C<NUL> characters.
1349 If it does, set C<errno> to C<ENOENT>, optionally warn, and return FALSE.
1351 Return TRUE if the name is safe.
1353 Used by the C<IS_SAFE_SYSCALL()> macro.
1358 PERL_STATIC_INLINE bool
1359 S_is_safe_syscall(pTHX_ const char *pv, STRLEN len, const char *what, const char *op_name) {
1360 /* While the Windows CE API provides only UCS-16 (or UTF-16) APIs
1361 * perl itself uses xce*() functions which accept 8-bit strings.
1364 PERL_ARGS_ASSERT_IS_SAFE_SYSCALL;
1368 if (UNLIKELY((null_at = (char *)memchr(pv, 0, len-1)) != NULL)) {
1369 SETERRNO(ENOENT, LIB_INVARG);
1370 Perl_ck_warner(aTHX_ packWARN(WARN_SYSCALLS),
1371 "Invalid \\0 character in %s for %s: %s\\0%s",
1372 what, op_name, pv, null_at+1);
1382 Return true if the supplied filename has a newline character
1383 immediately before the first (hopefully only) NUL.
1385 My original look at this incorrectly used the len from SvPV(), but
1386 that's incorrect, since we allow for a NUL in pv[len-1].
1388 So instead, strlen() and work from there.
1390 This allow for the user reading a filename, forgetting to chomp it,
1393 open my $foo, "$file\0";
1399 PERL_STATIC_INLINE bool
1400 S_should_warn_nl(const char *pv) {
1403 PERL_ARGS_ASSERT_SHOULD_WARN_NL;
1407 return len > 0 && pv[len-1] == '\n';
1412 /* ------------------ pp.c, regcomp.c, toke.c, universal.c ------------ */
1414 #define MAX_CHARSET_NAME_LENGTH 2
1416 PERL_STATIC_INLINE const char *
1417 get_regex_charset_name(const U32 flags, STRLEN* const lenp)
1419 /* Returns a string that corresponds to the name of the regex character set
1420 * given by 'flags', and *lenp is set the length of that string, which
1421 * cannot exceed MAX_CHARSET_NAME_LENGTH characters */
1424 switch (get_regex_charset(flags)) {
1425 case REGEX_DEPENDS_CHARSET: return DEPENDS_PAT_MODS;
1426 case REGEX_LOCALE_CHARSET: return LOCALE_PAT_MODS;
1427 case REGEX_UNICODE_CHARSET: return UNICODE_PAT_MODS;
1428 case REGEX_ASCII_RESTRICTED_CHARSET: return ASCII_RESTRICT_PAT_MODS;
1429 case REGEX_ASCII_MORE_RESTRICTED_CHARSET:
1431 return ASCII_MORE_RESTRICT_PAT_MODS;
1433 /* The NOT_REACHED; hides an assert() which has a rather complex
1434 * definition in perl.h. */
1435 NOT_REACHED; /* NOTREACHED */
1436 return "?"; /* Unknown */
1441 Return false if any get magic is on the SV other than taint magic.
1445 PERL_STATIC_INLINE bool
1446 S_sv_only_taint_gmagic(SV *sv) {
1447 MAGIC *mg = SvMAGIC(sv);
1449 PERL_ARGS_ASSERT_SV_ONLY_TAINT_GMAGIC;
1452 if (mg->mg_type != PERL_MAGIC_taint
1453 && !(mg->mg_flags & MGf_GSKIP)
1454 && mg->mg_virtual->svt_get) {
1457 mg = mg->mg_moremagic;
1463 /* ------------------ cop.h ------------------------------------------- */
1466 /* Enter a block. Push a new base context and return its address. */
1468 PERL_STATIC_INLINE PERL_CONTEXT *
1469 S_cx_pushblock(pTHX_ U8 type, U8 gimme, SV** sp, I32 saveix)
1473 PERL_ARGS_ASSERT_CX_PUSHBLOCK;
1478 cx->blk_gimme = gimme;
1479 cx->blk_oldsaveix = saveix;
1480 cx->blk_oldsp = (I32)(sp - PL_stack_base);
1481 cx->blk_oldcop = PL_curcop;
1482 cx->blk_oldmarksp = (I32)(PL_markstack_ptr - PL_markstack);
1483 cx->blk_oldscopesp = PL_scopestack_ix;
1484 cx->blk_oldpm = PL_curpm;
1485 cx->blk_old_tmpsfloor = PL_tmps_floor;
1487 PL_tmps_floor = PL_tmps_ix;
1488 CX_DEBUG(cx, "PUSH");
1493 /* Exit a block (RETURN and LAST). */
1495 PERL_STATIC_INLINE void
1496 S_cx_popblock(pTHX_ PERL_CONTEXT *cx)
1498 PERL_ARGS_ASSERT_CX_POPBLOCK;
1500 CX_DEBUG(cx, "POP");
1501 /* these 3 are common to cx_popblock and cx_topblock */
1502 PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
1503 PL_scopestack_ix = cx->blk_oldscopesp;
1504 PL_curpm = cx->blk_oldpm;
1506 /* LEAVE_SCOPE() should have made this true. /(?{})/ cheats
1507 * and leaves a CX entry lying around for repeated use, so
1508 * skip for multicall */ \
1509 assert( (CxTYPE(cx) == CXt_SUB && CxMULTICALL(cx))
1510 || PL_savestack_ix == cx->blk_oldsaveix);
1511 PL_curcop = cx->blk_oldcop;
1512 PL_tmps_floor = cx->blk_old_tmpsfloor;
1515 /* Continue a block elsewhere (e.g. NEXT, REDO, GOTO).
1516 * Whereas cx_popblock() restores the state to the point just before
1517 * cx_pushblock() was called, cx_topblock() restores it to the point just
1518 * *after* cx_pushblock() was called. */
1520 PERL_STATIC_INLINE void
1521 S_cx_topblock(pTHX_ PERL_CONTEXT *cx)
1523 PERL_ARGS_ASSERT_CX_TOPBLOCK;
1525 CX_DEBUG(cx, "TOP");
1526 /* these 3 are common to cx_popblock and cx_topblock */
1527 PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
1528 PL_scopestack_ix = cx->blk_oldscopesp;
1529 PL_curpm = cx->blk_oldpm;
1531 PL_stack_sp = PL_stack_base + cx->blk_oldsp;
1535 PERL_STATIC_INLINE void
1536 S_cx_pushsub(pTHX_ PERL_CONTEXT *cx, CV *cv, OP *retop, bool hasargs)
1538 U8 phlags = CX_PUSHSUB_GET_LVALUE_MASK(Perl_was_lvalue_sub);
1540 PERL_ARGS_ASSERT_CX_PUSHSUB;
1542 PERL_DTRACE_PROBE_ENTRY(cv);
1543 cx->blk_sub.cv = cv;
1544 cx->blk_sub.olddepth = CvDEPTH(cv);
1545 cx->blk_sub.prevcomppad = PL_comppad;
1546 cx->cx_type |= (hasargs) ? CXp_HASARGS : 0;
1547 cx->blk_sub.retop = retop;
1548 SvREFCNT_inc_simple_void_NN(cv);
1549 cx->blk_u16 = PL_op->op_private & (phlags|OPpDEREF);
1553 /* subsets of cx_popsub() */
1555 PERL_STATIC_INLINE void
1556 S_cx_popsub_common(pTHX_ PERL_CONTEXT *cx)
1560 PERL_ARGS_ASSERT_CX_POPSUB_COMMON;
1561 assert(CxTYPE(cx) == CXt_SUB);
1563 PL_comppad = cx->blk_sub.prevcomppad;
1564 PL_curpad = LIKELY(PL_comppad) ? AvARRAY(PL_comppad) : NULL;
1565 cv = cx->blk_sub.cv;
1566 CvDEPTH(cv) = cx->blk_sub.olddepth;
1567 cx->blk_sub.cv = NULL;
1572 /* handle the @_ part of leaving a sub */
1574 PERL_STATIC_INLINE void
1575 S_cx_popsub_args(pTHX_ PERL_CONTEXT *cx)
1579 PERL_ARGS_ASSERT_CX_POPSUB_ARGS;
1580 assert(CxTYPE(cx) == CXt_SUB);
1581 assert(AvARRAY(MUTABLE_AV(
1582 PadlistARRAY(CvPADLIST(cx->blk_sub.cv))[
1583 CvDEPTH(cx->blk_sub.cv)])) == PL_curpad);
1585 CX_POP_SAVEARRAY(cx);
1586 av = MUTABLE_AV(PAD_SVl(0));
1587 if (UNLIKELY(AvREAL(av)))
1588 /* abandon @_ if it got reified */
1589 clear_defarray(av, 0);
1596 PERL_STATIC_INLINE void
1597 S_cx_popsub(pTHX_ PERL_CONTEXT *cx)
1599 PERL_ARGS_ASSERT_CX_POPSUB;
1600 assert(CxTYPE(cx) == CXt_SUB);
1602 PERL_DTRACE_PROBE_RETURN(cx->blk_sub.cv);
1606 cx_popsub_common(cx);
1610 PERL_STATIC_INLINE void
1611 S_cx_pushformat(pTHX_ PERL_CONTEXT *cx, CV *cv, OP *retop, GV *gv)
1613 PERL_ARGS_ASSERT_CX_PUSHFORMAT;
1615 cx->blk_format.cv = cv;
1616 cx->blk_format.retop = retop;
1617 cx->blk_format.gv = gv;
1618 cx->blk_format.dfoutgv = PL_defoutgv;
1619 cx->blk_format.prevcomppad = PL_comppad;
1622 SvREFCNT_inc_simple_void_NN(cv);
1624 SvREFCNT_inc_void(cx->blk_format.dfoutgv);
1628 PERL_STATIC_INLINE void
1629 S_cx_popformat(pTHX_ PERL_CONTEXT *cx)
1634 PERL_ARGS_ASSERT_CX_POPFORMAT;
1635 assert(CxTYPE(cx) == CXt_FORMAT);
1637 dfout = cx->blk_format.dfoutgv;
1639 cx->blk_format.dfoutgv = NULL;
1640 SvREFCNT_dec_NN(dfout);
1642 PL_comppad = cx->blk_format.prevcomppad;
1643 PL_curpad = LIKELY(PL_comppad) ? AvARRAY(PL_comppad) : NULL;
1644 cv = cx->blk_format.cv;
1645 cx->blk_format.cv = NULL;
1647 SvREFCNT_dec_NN(cv);
1651 PERL_STATIC_INLINE void
1652 S_cx_pusheval(pTHX_ PERL_CONTEXT *cx, OP *retop, SV *namesv)
1654 PERL_ARGS_ASSERT_CX_PUSHEVAL;
1656 cx->blk_eval.retop = retop;
1657 cx->blk_eval.old_namesv = namesv;
1658 cx->blk_eval.old_eval_root = PL_eval_root;
1659 cx->blk_eval.cur_text = PL_parser ? PL_parser->linestr : NULL;
1660 cx->blk_eval.cv = NULL; /* later set by doeval_compile() */
1661 cx->blk_eval.cur_top_env = PL_top_env;
1663 assert(!(PL_in_eval & ~ 0x3F));
1664 assert(!(PL_op->op_type & ~0x1FF));
1665 cx->blk_u16 = (PL_in_eval & 0x3F) | ((U16)PL_op->op_type << 7);
1669 PERL_STATIC_INLINE void
1670 S_cx_popeval(pTHX_ PERL_CONTEXT *cx)
1674 PERL_ARGS_ASSERT_CX_POPEVAL;
1675 assert(CxTYPE(cx) == CXt_EVAL);
1677 PL_in_eval = CxOLD_IN_EVAL(cx);
1678 assert(!(PL_in_eval & 0xc0));
1679 PL_eval_root = cx->blk_eval.old_eval_root;
1680 sv = cx->blk_eval.cur_text;
1681 if (sv && CxEVAL_TXT_REFCNTED(cx)) {
1682 cx->blk_eval.cur_text = NULL;
1683 SvREFCNT_dec_NN(sv);
1686 sv = cx->blk_eval.old_namesv;
1688 cx->blk_eval.old_namesv = NULL;
1689 SvREFCNT_dec_NN(sv);
1694 /* push a plain loop, i.e.
1696 * while (cond) { block }
1697 * for (init;cond;continue) { block }
1698 * This loop can be last/redo'ed etc.
1701 PERL_STATIC_INLINE void
1702 S_cx_pushloop_plain(pTHX_ PERL_CONTEXT *cx)
1704 PERL_ARGS_ASSERT_CX_PUSHLOOP_PLAIN;
1705 cx->blk_loop.my_op = cLOOP;
1709 /* push a true for loop, i.e.
1710 * for var (list) { block }
1713 PERL_STATIC_INLINE void
1714 S_cx_pushloop_for(pTHX_ PERL_CONTEXT *cx, void *itervarp, SV* itersave)
1716 PERL_ARGS_ASSERT_CX_PUSHLOOP_FOR;
1718 /* this one line is common with cx_pushloop_plain */
1719 cx->blk_loop.my_op = cLOOP;
1721 cx->blk_loop.itervar_u.svp = (SV**)itervarp;
1722 cx->blk_loop.itersave = itersave;
1724 cx->blk_loop.oldcomppad = PL_comppad;
1729 /* pop all loop types, including plain */
1731 PERL_STATIC_INLINE void
1732 S_cx_poploop(pTHX_ PERL_CONTEXT *cx)
1734 PERL_ARGS_ASSERT_CX_POPLOOP;
1736 assert(CxTYPE_is_LOOP(cx));
1737 if ( CxTYPE(cx) == CXt_LOOP_ARY
1738 || CxTYPE(cx) == CXt_LOOP_LAZYSV)
1740 /* Free ary or cur. This assumes that state_u.ary.ary
1741 * aligns with state_u.lazysv.cur. See cx_dup() */
1742 SV *sv = cx->blk_loop.state_u.lazysv.cur;
1743 cx->blk_loop.state_u.lazysv.cur = NULL;
1744 SvREFCNT_dec_NN(sv);
1745 if (CxTYPE(cx) == CXt_LOOP_LAZYSV) {
1746 sv = cx->blk_loop.state_u.lazysv.end;
1747 cx->blk_loop.state_u.lazysv.end = NULL;
1748 SvREFCNT_dec_NN(sv);
1751 if (cx->cx_type & (CXp_FOR_PAD|CXp_FOR_GV)) {
1753 SV **svp = (cx)->blk_loop.itervar_u.svp;
1754 if ((cx->cx_type & CXp_FOR_GV))
1755 svp = &GvSV((GV*)svp);
1757 *svp = cx->blk_loop.itersave;
1758 cx->blk_loop.itersave = NULL;
1759 SvREFCNT_dec(cursv);
1764 PERL_STATIC_INLINE void
1765 S_cx_pushwhen(pTHX_ PERL_CONTEXT *cx)
1767 PERL_ARGS_ASSERT_CX_PUSHWHEN;
1769 cx->blk_givwhen.leave_op = cLOGOP->op_other;
1773 PERL_STATIC_INLINE void
1774 S_cx_popwhen(pTHX_ PERL_CONTEXT *cx)
1776 PERL_ARGS_ASSERT_CX_POPWHEN;
1777 assert(CxTYPE(cx) == CXt_WHEN);
1779 PERL_UNUSED_ARG(cx);
1780 PERL_UNUSED_CONTEXT;
1781 /* currently NOOP */
1785 PERL_STATIC_INLINE void
1786 S_cx_pushgiven(pTHX_ PERL_CONTEXT *cx, SV *orig_defsv)
1788 PERL_ARGS_ASSERT_CX_PUSHGIVEN;
1790 cx->blk_givwhen.leave_op = cLOGOP->op_other;
1791 cx->blk_givwhen.defsv_save = orig_defsv;
1795 PERL_STATIC_INLINE void
1796 S_cx_popgiven(pTHX_ PERL_CONTEXT *cx)
1800 PERL_ARGS_ASSERT_CX_POPGIVEN;
1801 assert(CxTYPE(cx) == CXt_GIVEN);
1803 sv = GvSV(PL_defgv);
1804 GvSV(PL_defgv) = cx->blk_givwhen.defsv_save;
1805 cx->blk_givwhen.defsv_save = NULL;
1809 /* ------------------ util.h ------------------------------------------- */
1812 =head1 Miscellaneous Functions
1816 Returns true if the leading C<len> bytes of the strings C<s1> and C<s2> are the
1818 case-insensitively; false otherwise. Uppercase and lowercase ASCII range bytes
1819 match themselves and their opposite case counterparts. Non-cased and non-ASCII
1820 range bytes match only themselves.
1825 PERL_STATIC_INLINE I32
1826 Perl_foldEQ(const char *s1, const char *s2, I32 len)
1828 const U8 *a = (const U8 *)s1;
1829 const U8 *b = (const U8 *)s2;
1831 PERL_ARGS_ASSERT_FOLDEQ;
1836 if (*a != *b && *a != PL_fold[*b])
1843 PERL_STATIC_INLINE I32
1844 Perl_foldEQ_latin1(const char *s1, const char *s2, I32 len)
1846 /* Compare non-utf8 using Unicode (Latin1) semantics. Does not work on
1847 * MICRO_SIGN, LATIN_SMALL_LETTER_SHARP_S, nor
1848 * LATIN_SMALL_LETTER_Y_WITH_DIAERESIS, and does not check for these. Nor
1849 * does it check that the strings each have at least 'len' characters */
1851 const U8 *a = (const U8 *)s1;
1852 const U8 *b = (const U8 *)s2;
1854 PERL_ARGS_ASSERT_FOLDEQ_LATIN1;
1859 if (*a != *b && *a != PL_fold_latin1[*b]) {
1868 =for apidoc foldEQ_locale
1870 Returns true if the leading C<len> bytes of the strings C<s1> and C<s2> are the
1871 same case-insensitively in the current locale; false otherwise.
1876 PERL_STATIC_INLINE I32
1877 Perl_foldEQ_locale(const char *s1, const char *s2, I32 len)
1880 const U8 *a = (const U8 *)s1;
1881 const U8 *b = (const U8 *)s2;
1883 PERL_ARGS_ASSERT_FOLDEQ_LOCALE;
1888 if (*a != *b && *a != PL_fold_locale[*b])
1895 #if ! defined (HAS_MEMRCHR) && (defined(PERL_CORE) || defined(PERL_EXT))
1897 PERL_STATIC_INLINE void *
1898 S_my_memrchr(const char * s, const char c, const STRLEN len)
1900 /* memrchr(), since many platforms lack it */
1902 const char * t = s + len - 1;
1904 PERL_ARGS_ASSERT_MY_MEMRCHR;
1919 * ex: set ts=8 sts=4 sw=4 et: