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 GCC_DIAG_IGNORE_DECL(-Wunused-function);
249 S_croak_memory_wrap(void)
251 Perl_croak_nocontext("%s",PL_memory_wrap);
253 GCC_DIAG_RESTORE_DECL;
255 /* ------------------------------- utf8.h ------------------------------- */
258 =head1 Unicode Support
261 PERL_STATIC_INLINE void
262 S_append_utf8_from_native_byte(const U8 byte, U8** dest)
264 /* Takes an input 'byte' (Latin1 or EBCDIC) and appends it to the UTF-8
265 * encoded string at '*dest', updating '*dest' to include it */
267 PERL_ARGS_ASSERT_APPEND_UTF8_FROM_NATIVE_BYTE;
269 if (NATIVE_BYTE_IS_INVARIANT(byte))
272 *((*dest)++) = UTF8_EIGHT_BIT_HI(byte);
273 *((*dest)++) = UTF8_EIGHT_BIT_LO(byte);
278 =for apidoc valid_utf8_to_uvchr
279 Like C<L</utf8_to_uvchr_buf>>, but should only be called when it is known that
280 the next character in the input UTF-8 string C<s> is well-formed (I<e.g.>,
281 it passes C<L</isUTF8_CHAR>>. Surrogates, non-character code points, and
282 non-Unicode code points are allowed.
288 PERL_STATIC_INLINE UV
289 Perl_valid_utf8_to_uvchr(const U8 *s, STRLEN *retlen)
291 const UV expectlen = UTF8SKIP(s);
292 const U8* send = s + expectlen;
295 PERL_ARGS_ASSERT_VALID_UTF8_TO_UVCHR;
301 /* An invariant is trivially returned */
302 if (expectlen == 1) {
306 /* Remove the leading bits that indicate the number of bytes, leaving just
307 * the bits that are part of the value */
308 uv = NATIVE_UTF8_TO_I8(uv) & UTF_START_MASK(expectlen);
310 /* Now, loop through the remaining bytes, accumulating each into the
311 * working total as we go. (I khw tried unrolling the loop for up to 4
312 * bytes, but there was no performance improvement) */
313 for (++s; s < send; s++) {
314 uv = UTF8_ACCUMULATE(uv, *s);
317 return UNI_TO_NATIVE(uv);
322 =for apidoc is_utf8_invariant_string
324 Returns TRUE if the first C<len> bytes of the string C<s> are the same
325 regardless of the UTF-8 encoding of the string (or UTF-EBCDIC encoding on
326 EBCDIC machines); otherwise it returns FALSE. That is, it returns TRUE if they
327 are UTF-8 invariant. On ASCII-ish machines, all the ASCII characters and only
328 the ASCII characters fit this definition. On EBCDIC machines, the ASCII-range
329 characters are invariant, but so also are the C1 controls.
331 If C<len> is 0, it will be calculated using C<strlen(s)>, (which means if you
332 use this option, that C<s> can't have embedded C<NUL> characters and has to
333 have a terminating C<NUL> byte).
336 C<L</is_utf8_string>>,
337 C<L</is_utf8_string_flags>>,
338 C<L</is_utf8_string_loc>>,
339 C<L</is_utf8_string_loc_flags>>,
340 C<L</is_utf8_string_loclen>>,
341 C<L</is_utf8_string_loclen_flags>>,
342 C<L</is_utf8_fixed_width_buf_flags>>,
343 C<L</is_utf8_fixed_width_buf_loc_flags>>,
344 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
345 C<L</is_strict_utf8_string>>,
346 C<L</is_strict_utf8_string_loc>>,
347 C<L</is_strict_utf8_string_loclen>>,
348 C<L</is_c9strict_utf8_string>>,
349 C<L</is_c9strict_utf8_string_loc>>,
351 C<L</is_c9strict_utf8_string_loclen>>.
357 #define is_utf8_invariant_string(s, len) \
358 is_utf8_invariant_string_loc(s, len, NULL)
361 =for apidoc is_utf8_invariant_string_loc
363 Like C<L</is_utf8_invariant_string>> but upon failure, stores the location of
364 the first UTF-8 variant character in the C<ep> pointer; if all characters are
365 UTF-8 invariant, this function does not change the contents of C<*ep>.
371 PERL_STATIC_INLINE bool
372 S_is_utf8_invariant_string_loc(const U8* const s, STRLEN len, const U8 ** ep)
377 PERL_ARGS_ASSERT_IS_UTF8_INVARIANT_STRING_LOC;
380 len = strlen((const char *)s);
387 /* This looks like 0x010101... */
388 #define PERL_COUNT_MULTIPLIER (~ (UINTMAX_C(0)) / 0xFF)
390 /* This looks like 0x808080... */
391 #define PERL_VARIANTS_WORD_MASK (PERL_COUNT_MULTIPLIER * 0x80)
392 #define PERL_WORDSIZE sizeof(PERL_COUNT_MULTIPLIER)
393 #define PERL_WORD_BOUNDARY_MASK (PERL_WORDSIZE - 1)
395 /* Evaluates to 0 if 'x' is at a word boundary; otherwise evaluates to 1, by
396 * or'ing together the lowest bits of 'x'. Hopefully the final term gets
397 * optimized out completely on a 32-bit system, and its mask gets optimized out
398 * on a 64-bit system */
399 #define PERL_IS_SUBWORD_ADDR(x) (1 & ( PTR2nat(x) \
400 | ( PTR2nat(x) >> 1) \
402 & PERL_WORD_BOUNDARY_MASK) >> 2))))
404 /* Do the word-at-a-time iff there is at least one usable full word. That
405 * means that after advancing to a word boundary, there still is at least a
406 * full word left. The number of bytes needed to advance is 'wordsize -
407 * offset' unless offset is 0. */
408 if ((STRLEN) (send - x) >= PERL_WORDSIZE
410 /* This term is wordsize if subword; 0 if not */
411 + PERL_WORDSIZE * PERL_IS_SUBWORD_ADDR(x)
414 - (PTR2nat(x) & PERL_WORD_BOUNDARY_MASK))
417 /* Process per-byte until reach word boundary. XXX This loop could be
418 * eliminated if we knew that this platform had fast unaligned reads */
419 while (PTR2nat(x) & PERL_WORD_BOUNDARY_MASK) {
420 if (! UTF8_IS_INVARIANT(*x)) {
430 /* Here, we know we have at least one full word to process. Process
431 * per-word as long as we have at least a full word left */
433 if ((* (PERL_UINTMAX_T *) x) & PERL_VARIANTS_WORD_MASK) {
435 /* Found a variant. Just return if caller doesn't want its
441 /* Otherwise fall into final loop to find which byte it is */
445 } while (x + PERL_WORDSIZE <= send);
450 /* Process per-byte */
452 if (! UTF8_IS_INVARIANT(*x)) {
466 #if defined(PERL_CORE) || defined(PERL_EXT)
469 =for apidoc variant_under_utf8_count
471 This function looks at the sequence of bytes between C<s> and C<e>, which are
472 assumed to be encoded in ASCII/Latin1, and returns how many of them would
473 change should the string be translated into UTF-8. Due to the nature of UTF-8,
474 each of these would occupy two bytes instead of the single one in the input
475 string. Thus, this function returns the precise number of bytes the string
476 would expand by when translated to UTF-8.
478 Unlike most of the other functions that have C<utf8> in their name, the input
479 to this function is NOT a UTF-8-encoded string. The function name is slightly
480 I<odd> to emphasize this.
482 This function is internal to Perl because khw thinks that any XS code that
483 would want this is probably operating too close to the internals. Presenting a
484 valid use case could change that.
487 C<L<perlapi/is_utf8_invariant_string>>
489 C<L<perlapi/is_utf8_invariant_string_loc>>,
495 PERL_STATIC_INLINE Size_t
496 S_variant_under_utf8_count(const U8* const s, const U8* const e)
501 PERL_ARGS_ASSERT_VARIANT_UNDER_UTF8_COUNT;
505 if ((STRLEN) (e - x) >= PERL_WORDSIZE
506 + PERL_WORDSIZE * PERL_IS_SUBWORD_ADDR(x)
507 - (PTR2nat(x) & PERL_WORD_BOUNDARY_MASK))
510 /* Process per-byte until reach word boundary. XXX This loop could be
511 * eliminated if we knew that this platform had fast unaligned reads */
512 while (PTR2nat(x) & PERL_WORD_BOUNDARY_MASK) {
513 count += ! UTF8_IS_INVARIANT(*x++);
516 /* Process per-word as long as we have at least a full word left */
517 do { /* Commit 03c1e4ab1d6ee9062fb3f94b0ba31db6698724b1 contains an
518 explanation of how this works */
519 count += ((((* (PERL_UINTMAX_T *) x) & PERL_VARIANTS_WORD_MASK) >> 7)
520 * PERL_COUNT_MULTIPLIER)
521 >> ((PERL_WORDSIZE - 1) * CHARBITS);
523 } while (x + PERL_WORDSIZE <= e);
528 /* Process per-byte */
530 if (! UTF8_IS_INVARIANT(*x)) {
543 #undef PERL_COUNT_MULTIPLIER
544 #undef PERL_WORD_BOUNDARY_MASK
545 #undef PERL_VARIANTS_WORD_MASK
548 =for apidoc is_utf8_string
550 Returns TRUE if the first C<len> bytes of string C<s> form a valid
551 Perl-extended-UTF-8 string; returns FALSE otherwise. If C<len> is 0, it will
552 be calculated using C<strlen(s)> (which means if you use this option, that C<s>
553 can't have embedded C<NUL> characters and has to have a terminating C<NUL>
554 byte). Note that all characters being ASCII constitute 'a valid UTF-8 string'.
556 This function considers Perl's extended UTF-8 to be valid. That means that
557 code points above Unicode, surrogates, and non-character code points are
558 considered valid by this function. Use C<L</is_strict_utf8_string>>,
559 C<L</is_c9strict_utf8_string>>, or C<L</is_utf8_string_flags>> to restrict what
560 code points are considered valid.
563 C<L</is_utf8_invariant_string>>,
564 C<L</is_utf8_invariant_string_loc>>,
565 C<L</is_utf8_string_loc>>,
566 C<L</is_utf8_string_loclen>>,
567 C<L</is_utf8_fixed_width_buf_flags>>,
568 C<L</is_utf8_fixed_width_buf_loc_flags>>,
569 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
574 #define is_utf8_string(s, len) is_utf8_string_loclen(s, len, NULL, NULL)
576 #if defined(PERL_CORE) || defined (PERL_EXT)
579 =for apidoc is_utf8_non_invariant_string
581 Returns TRUE if L<perlapi/is_utf8_invariant_string> returns FALSE for the first
582 C<len> bytes of the string C<s>, but they are, nonetheless, legal Perl-extended
583 UTF-8; otherwise returns FALSE.
585 A TRUE return means that at least one code point represented by the sequence
586 either is a wide character not representable as a single byte, or the
587 representation differs depending on whether the sequence is encoded in UTF-8 or
591 C<L<perlapi/is_utf8_invariant_string>>,
592 C<L<perlapi/is_utf8_string>>
596 This is commonly used to determine if a SV's UTF-8 flag should be turned on.
597 It needn't be if its string is entirely UTF-8 invariant, and it shouldn't be if
598 it otherwise contains invalid UTF-8.
600 It is an internal function because khw thinks that XS code shouldn't be working
601 at this low a level. A valid use case could change that.
605 PERL_STATIC_INLINE bool
606 S_is_utf8_non_invariant_string(const U8* const s, STRLEN len)
608 const U8 * first_variant;
610 PERL_ARGS_ASSERT_IS_UTF8_NON_INVARIANT_STRING;
612 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
616 return is_utf8_string(first_variant, len - (first_variant - s));
622 =for apidoc is_strict_utf8_string
624 Returns TRUE if the first C<len> bytes of string C<s> form a valid
625 UTF-8-encoded string that is fully interchangeable by any application using
626 Unicode rules; otherwise it returns FALSE. If C<len> is 0, it will be
627 calculated using C<strlen(s)> (which means if you use this option, that C<s>
628 can't have embedded C<NUL> characters and has to have a terminating C<NUL>
629 byte). Note that all characters being ASCII constitute 'a valid UTF-8 string'.
631 This function returns FALSE for strings containing any
632 code points above the Unicode max of 0x10FFFF, surrogate code points, or
633 non-character code points.
636 C<L</is_utf8_invariant_string>>,
637 C<L</is_utf8_invariant_string_loc>>,
638 C<L</is_utf8_string>>,
639 C<L</is_utf8_string_flags>>,
640 C<L</is_utf8_string_loc>>,
641 C<L</is_utf8_string_loc_flags>>,
642 C<L</is_utf8_string_loclen>>,
643 C<L</is_utf8_string_loclen_flags>>,
644 C<L</is_utf8_fixed_width_buf_flags>>,
645 C<L</is_utf8_fixed_width_buf_loc_flags>>,
646 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
647 C<L</is_strict_utf8_string_loc>>,
648 C<L</is_strict_utf8_string_loclen>>,
649 C<L</is_c9strict_utf8_string>>,
650 C<L</is_c9strict_utf8_string_loc>>,
652 C<L</is_c9strict_utf8_string_loclen>>.
657 #define is_strict_utf8_string(s, len) is_strict_utf8_string_loclen(s, len, NULL, NULL)
660 =for apidoc is_c9strict_utf8_string
662 Returns TRUE if the first C<len> bytes of string C<s> form a valid
663 UTF-8-encoded string that conforms to
664 L<Unicode Corrigendum #9|http://www.unicode.org/versions/corrigendum9.html>;
665 otherwise it returns FALSE. If C<len> is 0, it will be calculated using
666 C<strlen(s)> (which means if you use this option, that C<s> can't have embedded
667 C<NUL> characters and has to have a terminating C<NUL> byte). Note that all
668 characters being ASCII constitute 'a valid UTF-8 string'.
670 This function returns FALSE for strings containing any code points above the
671 Unicode max of 0x10FFFF or surrogate code points, but accepts non-character
673 L<Corrigendum #9|http://www.unicode.org/versions/corrigendum9.html>.
676 C<L</is_utf8_invariant_string>>,
677 C<L</is_utf8_invariant_string_loc>>,
678 C<L</is_utf8_string>>,
679 C<L</is_utf8_string_flags>>,
680 C<L</is_utf8_string_loc>>,
681 C<L</is_utf8_string_loc_flags>>,
682 C<L</is_utf8_string_loclen>>,
683 C<L</is_utf8_string_loclen_flags>>,
684 C<L</is_utf8_fixed_width_buf_flags>>,
685 C<L</is_utf8_fixed_width_buf_loc_flags>>,
686 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
687 C<L</is_strict_utf8_string>>,
688 C<L</is_strict_utf8_string_loc>>,
689 C<L</is_strict_utf8_string_loclen>>,
690 C<L</is_c9strict_utf8_string_loc>>,
692 C<L</is_c9strict_utf8_string_loclen>>.
697 #define is_c9strict_utf8_string(s, len) is_c9strict_utf8_string_loclen(s, len, NULL, 0)
700 =for apidoc is_utf8_string_flags
702 Returns TRUE if the first C<len> bytes of string C<s> form a valid
703 UTF-8 string, subject to the restrictions imposed by C<flags>;
704 returns FALSE otherwise. If C<len> is 0, it will be calculated
705 using C<strlen(s)> (which means if you use this option, that C<s> can't have
706 embedded C<NUL> characters and has to have a terminating C<NUL> byte). Note
707 that all characters being ASCII constitute 'a valid UTF-8 string'.
709 If C<flags> is 0, this gives the same results as C<L</is_utf8_string>>; if
710 C<flags> is C<UTF8_DISALLOW_ILLEGAL_INTERCHANGE>, this gives the same results
711 as C<L</is_strict_utf8_string>>; and if C<flags> is
712 C<UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE>, this gives the same results as
713 C<L</is_c9strict_utf8_string>>. Otherwise C<flags> may be any
714 combination of the C<UTF8_DISALLOW_I<foo>> flags understood by
715 C<L</utf8n_to_uvchr>>, with the same meanings.
718 C<L</is_utf8_invariant_string>>,
719 C<L</is_utf8_invariant_string_loc>>,
720 C<L</is_utf8_string>>,
721 C<L</is_utf8_string_loc>>,
722 C<L</is_utf8_string_loc_flags>>,
723 C<L</is_utf8_string_loclen>>,
724 C<L</is_utf8_string_loclen_flags>>,
725 C<L</is_utf8_fixed_width_buf_flags>>,
726 C<L</is_utf8_fixed_width_buf_loc_flags>>,
727 C<L</is_utf8_fixed_width_buf_loclen_flags>>,
728 C<L</is_strict_utf8_string>>,
729 C<L</is_strict_utf8_string_loc>>,
730 C<L</is_strict_utf8_string_loclen>>,
731 C<L</is_c9strict_utf8_string>>,
732 C<L</is_c9strict_utf8_string_loc>>,
734 C<L</is_c9strict_utf8_string_loclen>>.
739 PERL_STATIC_INLINE bool
740 S_is_utf8_string_flags(const U8 *s, STRLEN len, const U32 flags)
742 const U8 * first_variant;
744 PERL_ARGS_ASSERT_IS_UTF8_STRING_FLAGS;
745 assert(0 == (flags & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE
746 |UTF8_DISALLOW_PERL_EXTENDED)));
749 len = strlen((const char *)s);
753 return is_utf8_string(s, len);
756 if ((flags & ~UTF8_DISALLOW_PERL_EXTENDED)
757 == UTF8_DISALLOW_ILLEGAL_INTERCHANGE)
759 return is_strict_utf8_string(s, len);
762 if ((flags & ~UTF8_DISALLOW_PERL_EXTENDED)
763 == UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE)
765 return is_c9strict_utf8_string(s, len);
768 if (! is_utf8_invariant_string_loc(s, len, &first_variant)) {
769 const U8* const send = s + len;
770 const U8* x = first_variant;
773 STRLEN cur_len = isUTF8_CHAR_flags(x, send, flags);
774 if (UNLIKELY(! cur_len)) {
786 =for apidoc is_utf8_string_loc
788 Like C<L</is_utf8_string>> but stores the location of the failure (in the
789 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
790 "utf8ness success") in the C<ep> pointer.
792 See also C<L</is_utf8_string_loclen>>.
797 #define is_utf8_string_loc(s, len, ep) is_utf8_string_loclen(s, len, ep, 0)
801 =for apidoc is_utf8_string_loclen
803 Like C<L</is_utf8_string>> but stores the location of the failure (in the
804 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
805 "utf8ness success") in the C<ep> pointer, and the number of UTF-8
806 encoded characters in the C<el> pointer.
808 See also C<L</is_utf8_string_loc>>.
813 PERL_STATIC_INLINE bool
814 Perl_is_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el)
816 const U8 * first_variant;
818 PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN;
821 len = strlen((const char *) s);
824 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
836 const U8* const send = s + len;
837 const U8* x = first_variant;
838 STRLEN outlen = first_variant - s;
841 const STRLEN cur_len = isUTF8_CHAR(x, send);
842 if (UNLIKELY(! cur_len)) {
862 =for apidoc is_strict_utf8_string_loc
864 Like C<L</is_strict_utf8_string>> but stores the location of the failure (in the
865 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
866 "utf8ness success") in the C<ep> pointer.
868 See also C<L</is_strict_utf8_string_loclen>>.
873 #define is_strict_utf8_string_loc(s, len, ep) \
874 is_strict_utf8_string_loclen(s, len, ep, 0)
878 =for apidoc is_strict_utf8_string_loclen
880 Like C<L</is_strict_utf8_string>> but stores the location of the failure (in the
881 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
882 "utf8ness success") in the C<ep> pointer, and the number of UTF-8
883 encoded characters in the C<el> pointer.
885 See also C<L</is_strict_utf8_string_loc>>.
890 PERL_STATIC_INLINE bool
891 S_is_strict_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el)
893 const U8 * first_variant;
895 PERL_ARGS_ASSERT_IS_STRICT_UTF8_STRING_LOCLEN;
898 len = strlen((const char *) s);
901 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
913 const U8* const send = s + len;
914 const U8* x = first_variant;
915 STRLEN outlen = first_variant - s;
918 const STRLEN cur_len = isSTRICT_UTF8_CHAR(x, send);
919 if (UNLIKELY(! cur_len)) {
939 =for apidoc is_c9strict_utf8_string_loc
941 Like C<L</is_c9strict_utf8_string>> but stores the location of the failure (in
942 the case of "utf8ness failure") or the location C<s>+C<len> (in the case of
943 "utf8ness success") in the C<ep> pointer.
945 See also C<L</is_c9strict_utf8_string_loclen>>.
950 #define is_c9strict_utf8_string_loc(s, len, ep) \
951 is_c9strict_utf8_string_loclen(s, len, ep, 0)
955 =for apidoc is_c9strict_utf8_string_loclen
957 Like C<L</is_c9strict_utf8_string>> but stores the location of the failure (in
958 the case of "utf8ness failure") or the location C<s>+C<len> (in the case of
959 "utf8ness success") in the C<ep> pointer, and the number of UTF-8 encoded
960 characters in the C<el> pointer.
962 See also C<L</is_c9strict_utf8_string_loc>>.
967 PERL_STATIC_INLINE bool
968 S_is_c9strict_utf8_string_loclen(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el)
970 const U8 * first_variant;
972 PERL_ARGS_ASSERT_IS_C9STRICT_UTF8_STRING_LOCLEN;
975 len = strlen((const char *) s);
978 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
990 const U8* const send = s + len;
991 const U8* x = first_variant;
992 STRLEN outlen = first_variant - s;
995 const STRLEN cur_len = isC9_STRICT_UTF8_CHAR(x, send);
996 if (UNLIKELY(! cur_len)) {
1016 =for apidoc is_utf8_string_loc_flags
1018 Like C<L</is_utf8_string_flags>> but stores the location of the failure (in the
1019 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
1020 "utf8ness success") in the C<ep> pointer.
1022 See also C<L</is_utf8_string_loclen_flags>>.
1027 #define is_utf8_string_loc_flags(s, len, ep, flags) \
1028 is_utf8_string_loclen_flags(s, len, ep, 0, flags)
1031 /* The above 3 actual functions could have been moved into the more general one
1032 * just below, and made #defines that call it with the right 'flags'. They are
1033 * currently kept separate to increase their chances of getting inlined */
1037 =for apidoc is_utf8_string_loclen_flags
1039 Like C<L</is_utf8_string_flags>> but stores the location of the failure (in the
1040 case of "utf8ness failure") or the location C<s>+C<len> (in the case of
1041 "utf8ness success") in the C<ep> pointer, and the number of UTF-8
1042 encoded characters in the C<el> pointer.
1044 See also C<L</is_utf8_string_loc_flags>>.
1049 PERL_STATIC_INLINE bool
1050 S_is_utf8_string_loclen_flags(const U8 *s, STRLEN len, const U8 **ep, STRLEN *el, const U32 flags)
1052 const U8 * first_variant;
1054 PERL_ARGS_ASSERT_IS_UTF8_STRING_LOCLEN_FLAGS;
1055 assert(0 == (flags & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE
1056 |UTF8_DISALLOW_PERL_EXTENDED)));
1059 len = strlen((const char *) s);
1063 return is_utf8_string_loclen(s, len, ep, el);
1066 if ((flags & ~UTF8_DISALLOW_PERL_EXTENDED)
1067 == UTF8_DISALLOW_ILLEGAL_INTERCHANGE)
1069 return is_strict_utf8_string_loclen(s, len, ep, el);
1072 if ((flags & ~UTF8_DISALLOW_PERL_EXTENDED)
1073 == UTF8_DISALLOW_ILLEGAL_C9_INTERCHANGE)
1075 return is_c9strict_utf8_string_loclen(s, len, ep, el);
1078 if (is_utf8_invariant_string_loc(s, len, &first_variant)) {
1090 const U8* send = s + len;
1091 const U8* x = first_variant;
1092 STRLEN outlen = first_variant - s;
1095 const STRLEN cur_len = isUTF8_CHAR_flags(x, send, flags);
1096 if (UNLIKELY(! cur_len)) {
1115 =for apidoc utf8_distance
1117 Returns the number of UTF-8 characters between the UTF-8 pointers C<a>
1120 WARNING: use only if you *know* that the pointers point inside the
1126 PERL_STATIC_INLINE IV
1127 Perl_utf8_distance(pTHX_ const U8 *a, const U8 *b)
1129 PERL_ARGS_ASSERT_UTF8_DISTANCE;
1131 return (a < b) ? -1 * (IV) utf8_length(a, b) : (IV) utf8_length(b, a);
1135 =for apidoc utf8_hop
1137 Return the UTF-8 pointer C<s> displaced by C<off> characters, either
1138 forward or backward.
1140 WARNING: do not use the following unless you *know* C<off> is within
1141 the UTF-8 data pointed to by C<s> *and* that on entry C<s> is aligned
1142 on the first byte of character or just after the last byte of a character.
1147 PERL_STATIC_INLINE U8 *
1148 Perl_utf8_hop(const U8 *s, SSize_t off)
1150 PERL_ARGS_ASSERT_UTF8_HOP;
1152 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
1153 * the bitops (especially ~) can create illegal UTF-8.
1154 * In other words: in Perl UTF-8 is not just for Unicode. */
1163 while (UTF8_IS_CONTINUATION(*s))
1167 GCC_DIAG_IGNORE_STMT(-Wcast-qual);
1169 GCC_DIAG_RESTORE_STMT;
1173 =for apidoc utf8_hop_forward
1175 Return the UTF-8 pointer C<s> displaced by up to C<off> characters,
1178 C<off> must be non-negative.
1180 C<s> must be before or equal to C<end>.
1182 When moving forward it will not move beyond C<end>.
1184 Will not exceed this limit even if the string is not valid "UTF-8".
1189 PERL_STATIC_INLINE U8 *
1190 Perl_utf8_hop_forward(const U8 *s, SSize_t off, const U8 *end)
1192 PERL_ARGS_ASSERT_UTF8_HOP_FORWARD;
1194 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
1195 * the bitops (especially ~) can create illegal UTF-8.
1196 * In other words: in Perl UTF-8 is not just for Unicode. */
1202 STRLEN skip = UTF8SKIP(s);
1203 if ((STRLEN)(end - s) <= skip) {
1204 GCC_DIAG_IGNORE_STMT(-Wcast-qual);
1206 GCC_DIAG_RESTORE_STMT;
1211 GCC_DIAG_IGNORE_STMT(-Wcast-qual);
1213 GCC_DIAG_RESTORE_STMT;
1217 =for apidoc utf8_hop_back
1219 Return the UTF-8 pointer C<s> displaced by up to C<off> characters,
1222 C<off> must be non-positive.
1224 C<s> must be after or equal to C<start>.
1226 When moving backward it will not move before C<start>.
1228 Will not exceed this limit even if the string is not valid "UTF-8".
1233 PERL_STATIC_INLINE U8 *
1234 Perl_utf8_hop_back(const U8 *s, SSize_t off, const U8 *start)
1236 PERL_ARGS_ASSERT_UTF8_HOP_BACK;
1238 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
1239 * the bitops (especially ~) can create illegal UTF-8.
1240 * In other words: in Perl UTF-8 is not just for Unicode. */
1245 while (off++ && s > start) {
1247 while (UTF8_IS_CONTINUATION(*s) && s > start)
1251 GCC_DIAG_IGNORE_STMT(-Wcast-qual);
1253 GCC_DIAG_RESTORE_STMT;
1257 =for apidoc utf8_hop_safe
1259 Return the UTF-8 pointer C<s> displaced by up to C<off> characters,
1260 either forward or backward.
1262 When moving backward it will not move before C<start>.
1264 When moving forward it will not move beyond C<end>.
1266 Will not exceed those limits even if the string is not valid "UTF-8".
1271 PERL_STATIC_INLINE U8 *
1272 Perl_utf8_hop_safe(const U8 *s, SSize_t off, const U8 *start, const U8 *end)
1274 PERL_ARGS_ASSERT_UTF8_HOP_SAFE;
1276 /* Note: cannot use UTF8_IS_...() too eagerly here since e.g
1277 * the bitops (especially ~) can create illegal UTF-8.
1278 * In other words: in Perl UTF-8 is not just for Unicode. */
1280 assert(start <= s && s <= end);
1283 return utf8_hop_forward(s, off, end);
1286 return utf8_hop_back(s, off, start);
1292 =for apidoc is_utf8_valid_partial_char
1294 Returns 0 if the sequence of bytes starting at C<s> and looking no further than
1295 S<C<e - 1>> is the UTF-8 encoding, as extended by Perl, for one or more code
1296 points. Otherwise, it returns 1 if there exists at least one non-empty
1297 sequence of bytes that when appended to sequence C<s>, starting at position
1298 C<e> causes the entire sequence to be the well-formed UTF-8 of some code point;
1299 otherwise returns 0.
1301 In other words this returns TRUE if C<s> points to a partial UTF-8-encoded code
1304 This is useful when a fixed-length buffer is being tested for being well-formed
1305 UTF-8, but the final few bytes in it don't comprise a full character; that is,
1306 it is split somewhere in the middle of the final code point's UTF-8
1307 representation. (Presumably when the buffer is refreshed with the next chunk
1308 of data, the new first bytes will complete the partial code point.) This
1309 function is used to verify that the final bytes in the current buffer are in
1310 fact the legal beginning of some code point, so that if they aren't, the
1311 failure can be signalled without having to wait for the next read.
1315 #define is_utf8_valid_partial_char(s, e) \
1316 is_utf8_valid_partial_char_flags(s, e, 0)
1320 =for apidoc is_utf8_valid_partial_char_flags
1322 Like C<L</is_utf8_valid_partial_char>>, it returns a boolean giving whether
1323 or not the input is a valid UTF-8 encoded partial character, but it takes an
1324 extra parameter, C<flags>, which can further restrict which code points are
1327 If C<flags> is 0, this behaves identically to
1328 C<L</is_utf8_valid_partial_char>>. Otherwise C<flags> can be any combination
1329 of the C<UTF8_DISALLOW_I<foo>> flags accepted by C<L</utf8n_to_uvchr>>. If
1330 there is any sequence of bytes that can complete the input partial character in
1331 such a way that a non-prohibited character is formed, the function returns
1332 TRUE; otherwise FALSE. Non character code points cannot be determined based on
1333 partial character input. But many of the other possible excluded types can be
1334 determined from just the first one or two bytes.
1339 PERL_STATIC_INLINE bool
1340 S_is_utf8_valid_partial_char_flags(const U8 * const s, const U8 * const e, const U32 flags)
1342 PERL_ARGS_ASSERT_IS_UTF8_VALID_PARTIAL_CHAR_FLAGS;
1344 assert(0 == (flags & ~(UTF8_DISALLOW_ILLEGAL_INTERCHANGE
1345 |UTF8_DISALLOW_PERL_EXTENDED)));
1347 if (s >= e || s + UTF8SKIP(s) <= e) {
1351 return cBOOL(_is_utf8_char_helper(s, e, flags));
1356 =for apidoc is_utf8_fixed_width_buf_flags
1358 Returns TRUE if the fixed-width buffer starting at C<s> with length C<len>
1359 is entirely valid UTF-8, subject to the restrictions given by C<flags>;
1360 otherwise it returns FALSE.
1362 If C<flags> is 0, any well-formed UTF-8, as extended by Perl, is accepted
1363 without restriction. If the final few bytes of the buffer do not form a
1364 complete code point, this will return TRUE anyway, provided that
1365 C<L</is_utf8_valid_partial_char_flags>> returns TRUE for them.
1367 If C<flags> in non-zero, it can be any combination of the
1368 C<UTF8_DISALLOW_I<foo>> flags accepted by C<L</utf8n_to_uvchr>>, and with the
1371 This function differs from C<L</is_utf8_string_flags>> only in that the latter
1372 returns FALSE if the final few bytes of the string don't form a complete code
1377 #define is_utf8_fixed_width_buf_flags(s, len, flags) \
1378 is_utf8_fixed_width_buf_loclen_flags(s, len, 0, 0, flags)
1382 =for apidoc is_utf8_fixed_width_buf_loc_flags
1384 Like C<L</is_utf8_fixed_width_buf_flags>> but stores the location of the
1385 failure in the C<ep> pointer. If the function returns TRUE, C<*ep> will point
1386 to the beginning of any partial character at the end of the buffer; if there is
1387 no partial character C<*ep> will contain C<s>+C<len>.
1389 See also C<L</is_utf8_fixed_width_buf_loclen_flags>>.
1394 #define is_utf8_fixed_width_buf_loc_flags(s, len, loc, flags) \
1395 is_utf8_fixed_width_buf_loclen_flags(s, len, loc, 0, flags)
1399 =for apidoc is_utf8_fixed_width_buf_loclen_flags
1401 Like C<L</is_utf8_fixed_width_buf_loc_flags>> but stores the number of
1402 complete, valid characters found in the C<el> pointer.
1407 PERL_STATIC_INLINE bool
1408 S_is_utf8_fixed_width_buf_loclen_flags(const U8 * const s,
1414 const U8 * maybe_partial;
1416 PERL_ARGS_ASSERT_IS_UTF8_FIXED_WIDTH_BUF_LOCLEN_FLAGS;
1419 ep = &maybe_partial;
1422 /* If it's entirely valid, return that; otherwise see if the only error is
1423 * that the final few bytes are for a partial character */
1424 return is_utf8_string_loclen_flags(s, len, ep, el, flags)
1425 || is_utf8_valid_partial_char_flags(*ep, s + len, flags);
1428 /* ------------------------------- perl.h ----------------------------- */
1431 =head1 Miscellaneous Functions
1433 =for apidoc AiR|bool|is_safe_syscall|const char *pv|STRLEN len|const char *what|const char *op_name
1435 Test that the given C<pv> doesn't contain any internal C<NUL> characters.
1436 If it does, set C<errno> to C<ENOENT>, optionally warn, and return FALSE.
1438 Return TRUE if the name is safe.
1440 Used by the C<IS_SAFE_SYSCALL()> macro.
1445 PERL_STATIC_INLINE bool
1446 S_is_safe_syscall(pTHX_ const char *pv, STRLEN len, const char *what, const char *op_name) {
1447 /* While the Windows CE API provides only UCS-16 (or UTF-16) APIs
1448 * perl itself uses xce*() functions which accept 8-bit strings.
1451 PERL_ARGS_ASSERT_IS_SAFE_SYSCALL;
1455 if (UNLIKELY((null_at = (char *)memchr(pv, 0, len-1)) != NULL)) {
1456 SETERRNO(ENOENT, LIB_INVARG);
1457 Perl_ck_warner(aTHX_ packWARN(WARN_SYSCALLS),
1458 "Invalid \\0 character in %s for %s: %s\\0%s",
1459 what, op_name, pv, null_at+1);
1469 Return true if the supplied filename has a newline character
1470 immediately before the first (hopefully only) NUL.
1472 My original look at this incorrectly used the len from SvPV(), but
1473 that's incorrect, since we allow for a NUL in pv[len-1].
1475 So instead, strlen() and work from there.
1477 This allow for the user reading a filename, forgetting to chomp it,
1480 open my $foo, "$file\0";
1486 PERL_STATIC_INLINE bool
1487 S_should_warn_nl(const char *pv) {
1490 PERL_ARGS_ASSERT_SHOULD_WARN_NL;
1494 return len > 0 && pv[len-1] == '\n';
1499 /* ------------------ pp.c, regcomp.c, toke.c, universal.c ------------ */
1501 #define MAX_CHARSET_NAME_LENGTH 2
1503 PERL_STATIC_INLINE const char *
1504 get_regex_charset_name(const U32 flags, STRLEN* const lenp)
1506 /* Returns a string that corresponds to the name of the regex character set
1507 * given by 'flags', and *lenp is set the length of that string, which
1508 * cannot exceed MAX_CHARSET_NAME_LENGTH characters */
1511 switch (get_regex_charset(flags)) {
1512 case REGEX_DEPENDS_CHARSET: return DEPENDS_PAT_MODS;
1513 case REGEX_LOCALE_CHARSET: return LOCALE_PAT_MODS;
1514 case REGEX_UNICODE_CHARSET: return UNICODE_PAT_MODS;
1515 case REGEX_ASCII_RESTRICTED_CHARSET: return ASCII_RESTRICT_PAT_MODS;
1516 case REGEX_ASCII_MORE_RESTRICTED_CHARSET:
1518 return ASCII_MORE_RESTRICT_PAT_MODS;
1520 /* The NOT_REACHED; hides an assert() which has a rather complex
1521 * definition in perl.h. */
1522 NOT_REACHED; /* NOTREACHED */
1523 return "?"; /* Unknown */
1528 Return false if any get magic is on the SV other than taint magic.
1532 PERL_STATIC_INLINE bool
1533 S_sv_only_taint_gmagic(SV *sv) {
1534 MAGIC *mg = SvMAGIC(sv);
1536 PERL_ARGS_ASSERT_SV_ONLY_TAINT_GMAGIC;
1539 if (mg->mg_type != PERL_MAGIC_taint
1540 && !(mg->mg_flags & MGf_GSKIP)
1541 && mg->mg_virtual->svt_get) {
1544 mg = mg->mg_moremagic;
1550 /* ------------------ cop.h ------------------------------------------- */
1553 /* Enter a block. Push a new base context and return its address. */
1555 PERL_STATIC_INLINE PERL_CONTEXT *
1556 S_cx_pushblock(pTHX_ U8 type, U8 gimme, SV** sp, I32 saveix)
1560 PERL_ARGS_ASSERT_CX_PUSHBLOCK;
1565 cx->blk_gimme = gimme;
1566 cx->blk_oldsaveix = saveix;
1567 cx->blk_oldsp = (I32)(sp - PL_stack_base);
1568 cx->blk_oldcop = PL_curcop;
1569 cx->blk_oldmarksp = (I32)(PL_markstack_ptr - PL_markstack);
1570 cx->blk_oldscopesp = PL_scopestack_ix;
1571 cx->blk_oldpm = PL_curpm;
1572 cx->blk_old_tmpsfloor = PL_tmps_floor;
1574 PL_tmps_floor = PL_tmps_ix;
1575 CX_DEBUG(cx, "PUSH");
1580 /* Exit a block (RETURN and LAST). */
1582 PERL_STATIC_INLINE void
1583 S_cx_popblock(pTHX_ PERL_CONTEXT *cx)
1585 PERL_ARGS_ASSERT_CX_POPBLOCK;
1587 CX_DEBUG(cx, "POP");
1588 /* these 3 are common to cx_popblock and cx_topblock */
1589 PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
1590 PL_scopestack_ix = cx->blk_oldscopesp;
1591 PL_curpm = cx->blk_oldpm;
1593 /* LEAVE_SCOPE() should have made this true. /(?{})/ cheats
1594 * and leaves a CX entry lying around for repeated use, so
1595 * skip for multicall */ \
1596 assert( (CxTYPE(cx) == CXt_SUB && CxMULTICALL(cx))
1597 || PL_savestack_ix == cx->blk_oldsaveix);
1598 PL_curcop = cx->blk_oldcop;
1599 PL_tmps_floor = cx->blk_old_tmpsfloor;
1602 /* Continue a block elsewhere (e.g. NEXT, REDO, GOTO).
1603 * Whereas cx_popblock() restores the state to the point just before
1604 * cx_pushblock() was called, cx_topblock() restores it to the point just
1605 * *after* cx_pushblock() was called. */
1607 PERL_STATIC_INLINE void
1608 S_cx_topblock(pTHX_ PERL_CONTEXT *cx)
1610 PERL_ARGS_ASSERT_CX_TOPBLOCK;
1612 CX_DEBUG(cx, "TOP");
1613 /* these 3 are common to cx_popblock and cx_topblock */
1614 PL_markstack_ptr = PL_markstack + cx->blk_oldmarksp;
1615 PL_scopestack_ix = cx->blk_oldscopesp;
1616 PL_curpm = cx->blk_oldpm;
1618 PL_stack_sp = PL_stack_base + cx->blk_oldsp;
1622 PERL_STATIC_INLINE void
1623 S_cx_pushsub(pTHX_ PERL_CONTEXT *cx, CV *cv, OP *retop, bool hasargs)
1625 U8 phlags = CX_PUSHSUB_GET_LVALUE_MASK(Perl_was_lvalue_sub);
1627 PERL_ARGS_ASSERT_CX_PUSHSUB;
1629 PERL_DTRACE_PROBE_ENTRY(cv);
1630 cx->blk_sub.cv = cv;
1631 cx->blk_sub.olddepth = CvDEPTH(cv);
1632 cx->blk_sub.prevcomppad = PL_comppad;
1633 cx->cx_type |= (hasargs) ? CXp_HASARGS : 0;
1634 cx->blk_sub.retop = retop;
1635 SvREFCNT_inc_simple_void_NN(cv);
1636 cx->blk_u16 = PL_op->op_private & (phlags|OPpDEREF);
1640 /* subsets of cx_popsub() */
1642 PERL_STATIC_INLINE void
1643 S_cx_popsub_common(pTHX_ PERL_CONTEXT *cx)
1647 PERL_ARGS_ASSERT_CX_POPSUB_COMMON;
1648 assert(CxTYPE(cx) == CXt_SUB);
1650 PL_comppad = cx->blk_sub.prevcomppad;
1651 PL_curpad = LIKELY(PL_comppad) ? AvARRAY(PL_comppad) : NULL;
1652 cv = cx->blk_sub.cv;
1653 CvDEPTH(cv) = cx->blk_sub.olddepth;
1654 cx->blk_sub.cv = NULL;
1659 /* handle the @_ part of leaving a sub */
1661 PERL_STATIC_INLINE void
1662 S_cx_popsub_args(pTHX_ PERL_CONTEXT *cx)
1666 PERL_ARGS_ASSERT_CX_POPSUB_ARGS;
1667 assert(CxTYPE(cx) == CXt_SUB);
1668 assert(AvARRAY(MUTABLE_AV(
1669 PadlistARRAY(CvPADLIST(cx->blk_sub.cv))[
1670 CvDEPTH(cx->blk_sub.cv)])) == PL_curpad);
1672 CX_POP_SAVEARRAY(cx);
1673 av = MUTABLE_AV(PAD_SVl(0));
1674 if (UNLIKELY(AvREAL(av)))
1675 /* abandon @_ if it got reified */
1676 clear_defarray(av, 0);
1683 PERL_STATIC_INLINE void
1684 S_cx_popsub(pTHX_ PERL_CONTEXT *cx)
1686 PERL_ARGS_ASSERT_CX_POPSUB;
1687 assert(CxTYPE(cx) == CXt_SUB);
1689 PERL_DTRACE_PROBE_RETURN(cx->blk_sub.cv);
1693 cx_popsub_common(cx);
1697 PERL_STATIC_INLINE void
1698 S_cx_pushformat(pTHX_ PERL_CONTEXT *cx, CV *cv, OP *retop, GV *gv)
1700 PERL_ARGS_ASSERT_CX_PUSHFORMAT;
1702 cx->blk_format.cv = cv;
1703 cx->blk_format.retop = retop;
1704 cx->blk_format.gv = gv;
1705 cx->blk_format.dfoutgv = PL_defoutgv;
1706 cx->blk_format.prevcomppad = PL_comppad;
1709 SvREFCNT_inc_simple_void_NN(cv);
1711 SvREFCNT_inc_void(cx->blk_format.dfoutgv);
1715 PERL_STATIC_INLINE void
1716 S_cx_popformat(pTHX_ PERL_CONTEXT *cx)
1721 PERL_ARGS_ASSERT_CX_POPFORMAT;
1722 assert(CxTYPE(cx) == CXt_FORMAT);
1724 dfout = cx->blk_format.dfoutgv;
1726 cx->blk_format.dfoutgv = NULL;
1727 SvREFCNT_dec_NN(dfout);
1729 PL_comppad = cx->blk_format.prevcomppad;
1730 PL_curpad = LIKELY(PL_comppad) ? AvARRAY(PL_comppad) : NULL;
1731 cv = cx->blk_format.cv;
1732 cx->blk_format.cv = NULL;
1734 SvREFCNT_dec_NN(cv);
1738 PERL_STATIC_INLINE void
1739 S_cx_pusheval(pTHX_ PERL_CONTEXT *cx, OP *retop, SV *namesv)
1741 PERL_ARGS_ASSERT_CX_PUSHEVAL;
1743 cx->blk_eval.retop = retop;
1744 cx->blk_eval.old_namesv = namesv;
1745 cx->blk_eval.old_eval_root = PL_eval_root;
1746 cx->blk_eval.cur_text = PL_parser ? PL_parser->linestr : NULL;
1747 cx->blk_eval.cv = NULL; /* later set by doeval_compile() */
1748 cx->blk_eval.cur_top_env = PL_top_env;
1750 assert(!(PL_in_eval & ~ 0x3F));
1751 assert(!(PL_op->op_type & ~0x1FF));
1752 cx->blk_u16 = (PL_in_eval & 0x3F) | ((U16)PL_op->op_type << 7);
1756 PERL_STATIC_INLINE void
1757 S_cx_popeval(pTHX_ PERL_CONTEXT *cx)
1761 PERL_ARGS_ASSERT_CX_POPEVAL;
1762 assert(CxTYPE(cx) == CXt_EVAL);
1764 PL_in_eval = CxOLD_IN_EVAL(cx);
1765 assert(!(PL_in_eval & 0xc0));
1766 PL_eval_root = cx->blk_eval.old_eval_root;
1767 sv = cx->blk_eval.cur_text;
1768 if (sv && CxEVAL_TXT_REFCNTED(cx)) {
1769 cx->blk_eval.cur_text = NULL;
1770 SvREFCNT_dec_NN(sv);
1773 sv = cx->blk_eval.old_namesv;
1775 cx->blk_eval.old_namesv = NULL;
1776 SvREFCNT_dec_NN(sv);
1781 /* push a plain loop, i.e.
1783 * while (cond) { block }
1784 * for (init;cond;continue) { block }
1785 * This loop can be last/redo'ed etc.
1788 PERL_STATIC_INLINE void
1789 S_cx_pushloop_plain(pTHX_ PERL_CONTEXT *cx)
1791 PERL_ARGS_ASSERT_CX_PUSHLOOP_PLAIN;
1792 cx->blk_loop.my_op = cLOOP;
1796 /* push a true for loop, i.e.
1797 * for var (list) { block }
1800 PERL_STATIC_INLINE void
1801 S_cx_pushloop_for(pTHX_ PERL_CONTEXT *cx, void *itervarp, SV* itersave)
1803 PERL_ARGS_ASSERT_CX_PUSHLOOP_FOR;
1805 /* this one line is common with cx_pushloop_plain */
1806 cx->blk_loop.my_op = cLOOP;
1808 cx->blk_loop.itervar_u.svp = (SV**)itervarp;
1809 cx->blk_loop.itersave = itersave;
1811 cx->blk_loop.oldcomppad = PL_comppad;
1816 PERL_STATIC_INLINE void
1817 S_cx_pushloop_given(pTHX_ PERL_CONTEXT *cx, SV *orig_defsv)
1819 PERL_ARGS_ASSERT_CX_PUSHLOOP_GIVEN;
1821 cx->blk_loop.my_op = cLOOP;
1822 cx->blk_loop.itervar_u.gv = PL_defgv;
1823 cx->blk_loop.itersave = orig_defsv;
1827 /* pop all loop types, including plain */
1829 PERL_STATIC_INLINE void
1830 S_cx_poploop(pTHX_ PERL_CONTEXT *cx)
1832 PERL_ARGS_ASSERT_CX_POPLOOP;
1834 assert(CxTYPE_is_LOOP(cx));
1835 if ( CxTYPE(cx) == CXt_LOOP_ARY
1836 || CxTYPE(cx) == CXt_LOOP_LAZYSV)
1838 /* Free ary or cur. This assumes that state_u.ary.ary
1839 * aligns with state_u.lazysv.cur. See cx_dup() */
1840 SV *sv = cx->blk_loop.state_u.lazysv.cur;
1841 cx->blk_loop.state_u.lazysv.cur = NULL;
1842 SvREFCNT_dec_NN(sv);
1843 if (CxTYPE(cx) == CXt_LOOP_LAZYSV) {
1844 sv = cx->blk_loop.state_u.lazysv.end;
1845 cx->blk_loop.state_u.lazysv.end = NULL;
1846 SvREFCNT_dec_NN(sv);
1849 if (cx->cx_type & (CXp_FOR_PAD|CXp_FOR_GV)) {
1851 SV **svp = (cx)->blk_loop.itervar_u.svp;
1852 if ((cx->cx_type & CXp_FOR_GV))
1853 svp = &GvSV((GV*)svp);
1855 *svp = cx->blk_loop.itersave;
1856 cx->blk_loop.itersave = NULL;
1857 SvREFCNT_dec(cursv);
1862 PERL_STATIC_INLINE void
1863 S_cx_pushwhereso(pTHX_ PERL_CONTEXT *cx)
1865 PERL_ARGS_ASSERT_CX_PUSHWHERESO;
1867 cx->blk_whereso.leave_op = cLOGOP->op_other;
1871 PERL_STATIC_INLINE void
1872 S_cx_popwhereso(pTHX_ PERL_CONTEXT *cx)
1874 PERL_ARGS_ASSERT_CX_POPWHERESO;
1875 assert(CxTYPE(cx) == CXt_WHERESO);
1877 PERL_UNUSED_ARG(cx);
1878 PERL_UNUSED_CONTEXT;
1879 /* currently NOOP */
1883 /* ------------------ util.h ------------------------------------------- */
1886 =head1 Miscellaneous Functions
1890 Returns true if the leading C<len> bytes of the strings C<s1> and C<s2> are the
1892 case-insensitively; false otherwise. Uppercase and lowercase ASCII range bytes
1893 match themselves and their opposite case counterparts. Non-cased and non-ASCII
1894 range bytes match only themselves.
1899 PERL_STATIC_INLINE I32
1900 Perl_foldEQ(const char *s1, const char *s2, I32 len)
1902 const U8 *a = (const U8 *)s1;
1903 const U8 *b = (const U8 *)s2;
1905 PERL_ARGS_ASSERT_FOLDEQ;
1910 if (*a != *b && *a != PL_fold[*b])
1917 PERL_STATIC_INLINE I32
1918 Perl_foldEQ_latin1(const char *s1, const char *s2, I32 len)
1920 /* Compare non-utf8 using Unicode (Latin1) semantics. Does not work on
1921 * MICRO_SIGN, LATIN_SMALL_LETTER_SHARP_S, nor
1922 * LATIN_SMALL_LETTER_Y_WITH_DIAERESIS, and does not check for these. Nor
1923 * does it check that the strings each have at least 'len' characters */
1925 const U8 *a = (const U8 *)s1;
1926 const U8 *b = (const U8 *)s2;
1928 PERL_ARGS_ASSERT_FOLDEQ_LATIN1;
1933 if (*a != *b && *a != PL_fold_latin1[*b]) {
1942 =for apidoc foldEQ_locale
1944 Returns true if the leading C<len> bytes of the strings C<s1> and C<s2> are the
1945 same case-insensitively in the current locale; false otherwise.
1950 PERL_STATIC_INLINE I32
1951 Perl_foldEQ_locale(const char *s1, const char *s2, I32 len)
1954 const U8 *a = (const U8 *)s1;
1955 const U8 *b = (const U8 *)s2;
1957 PERL_ARGS_ASSERT_FOLDEQ_LOCALE;
1962 if (*a != *b && *a != PL_fold_locale[*b])
1969 #if ! defined (HAS_MEMRCHR) && (defined(PERL_CORE) || defined(PERL_EXT))
1971 PERL_STATIC_INLINE void *
1972 S_my_memrchr(const char * s, const char c, const STRLEN len)
1974 /* memrchr(), since many platforms lack it */
1976 const char * t = s + len - 1;
1978 PERL_ARGS_ASSERT_MY_MEMRCHR;
1993 * ex: set ts=8 sts=4 sw=4 et: