#include "invlist_inline.h"
#include "unicode_constants.h"
-#define B_ON_NON_UTF8_LOCALE_IS_WRONG \
- "Use of \\b{} or \\B{} for non-UTF-8 locale is wrong. Assuming a UTF-8 locale"
+static const char b_utf8_locale_required[] =
+ "Use of \\b{} or \\B{} for non-UTF-8 locale is wrong."
+ " Assuming a UTF-8 locale";
+
+#define CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_BOUND \
+ STMT_START { \
+ if (! IN_UTF8_CTYPE_LOCALE) { \
+ Perl_ck_warner(aTHX_ packWARN(WARN_LOCALE), \
+ b_utf8_locale_required); \
+ } \
+ } STMT_END
-static const char utf8_locale_required[] =
+static const char sets_utf8_locale_required[] =
"Use of (?[ ]) for non-UTF-8 locale is wrong. Assuming a UTF-8 locale";
+#define CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_SETS(n) \
+ STMT_START { \
+ if (! IN_UTF8_CTYPE_LOCALE && ANYOFL_UTF8_LOCALE_REQD(FLAGS(n))) { \
+ Perl_ck_warner(aTHX_ packWARN(WARN_LOCALE), \
+ sets_utf8_locale_required); \
+ } \
+ } STMT_END
+
#ifdef DEBUGGING
/* At least one required character in the target string is expressible only in
* UTF-8. */
goto target; \
} STMT_END
-#define HAS_NONLATIN1_FOLD_CLOSURE(i) _HAS_NONLATIN1_FOLD_CLOSURE_ONLY_FOR_USE_BY_REGCOMP_DOT_C_AND_REGEXEC_DOT_C(i)
-
#ifndef STATIC
#define STATIC static
#endif
const UV total_elems = paren_elems_to_push + REGCP_OTHER_ELEMS;
const UV elems_shifted = total_elems << SAVE_TIGHT_SHIFT;
I32 p;
- GET_RE_DEBUG_FLAGS_DECL;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
PERL_ARGS_ASSERT_REGCPPUSH;
{
UV i;
U32 paren;
- GET_RE_DEBUG_FLAGS_DECL;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
PERL_ARGS_ASSERT_REGCPPOP;
* Ideally this could be replaced by a just an array of function pointers
* to the C library functions that implement the macros this calls.
* However, to compile, the precise function signatures are required, and
- * these may vary from platform to to platform. To avoid having to figure
+ * these may vary from platform to platform. To avoid having to figure
* out what those all are on each platform, I (khw) am using this method,
* which adds an extra layer of function call overhead (unless the C
* optimizer strips it away). But we don't particularly care about
* rules, ignoring any locale. So use the Unicode function if this class
* requires an inversion list, and use the Unicode macro otherwise. */
- dVAR;
PERL_ARGS_ASSERT_ISFOO_UTF8_LC;
span_word |= span_word << 4;
/* That reduces the problem to what this function solves */
- return s + _variant_byte_number(span_word);
+ return s + variant_byte_number(span_word);
#endif
masked &= PERL_VARIANTS_WORD_MASK;
/* This reduces the problem to that solved by this function */
- s += _variant_byte_number(masked);
+ s += variant_byte_number(masked);
return s;
} while (s + PERL_WORDSIZE <= send);
masked |= masked << 1;
masked |= masked << 2;
masked |= masked << 4;
- return s + _variant_byte_number(masked);
+ return s + variant_byte_number(masked);
#endif
RXi_GET_DECL(prog,progi);
regmatch_info reginfo_buf; /* create some info to pass to find_byclass */
regmatch_info *const reginfo = ®info_buf;
- GET_RE_DEBUG_FLAGS_DECL;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
PERL_ARGS_ASSERT_RE_INTUIT_START;
PERL_UNUSED_ARG(flags);
const U8* const str = (U8*)STRING(progi->regstclass);
/* XXX this value could be pre-computed */
- const int cl_l = (PL_regkind[OP(progi->regstclass)] == EXACT
+ const SSize_t cl_l = (PL_regkind[OP(progi->regstclass)] == EXACT
? (reginfo->is_utf8_pat
- ? utf8_distance(str + STR_LEN(progi->regstclass), str)
- : STR_LEN(progi->regstclass))
+ ? (SSize_t)utf8_distance(str + STR_LEN(progi->regstclass), str)
+ : (SSize_t)STR_LEN(progi->regstclass))
: 1);
char * endpos;
char *s;
dump_exec_pos(li,s,(reginfo->strend),(reginfo->strbeg), \
startpos, doutf8, depth)
-#define REXEC_FBC_SCAN(UTF8, CODE) \
+#define REXEC_FBC_UTF8_SCAN(CODE) \
+ STMT_START { \
+ while (s < strend) { \
+ CODE \
+ s += UTF8_SAFE_SKIP(s, reginfo->strend); \
+ } \
+ } STMT_END
+
+#define REXEC_FBC_NON_UTF8_SCAN(CODE) \
STMT_START { \
while (s < strend) { \
CODE \
- s += ((UTF8) \
- ? UTF8_SAFE_SKIP(s, reginfo->strend) \
- : 1); \
+ s++; \
} \
} STMT_END
-#define REXEC_FBC_CLASS_SCAN(UTF8, COND) \
+#define REXEC_FBC_UTF8_CLASS_SCAN(COND) \
STMT_START { \
while (s < strend) { \
- REXEC_FBC_CLASS_SCAN_GUTS(UTF8, COND) \
+ REXEC_FBC_UTF8_CLASS_SCAN_GUTS(COND) \
} \
} STMT_END
-#define REXEC_FBC_CLASS_SCAN_GUTS(UTF8, COND) \
+#define REXEC_FBC_NON_UTF8_CLASS_SCAN(COND) \
+ STMT_START { \
+ while (s < strend) { \
+ REXEC_FBC_NON_UTF8_CLASS_SCAN_GUTS(COND) \
+ } \
+ } STMT_END
+
+#define REXEC_FBC_UTF8_CLASS_SCAN_GUTS(COND) \
if (COND) { \
FBC_CHECK_AND_TRY \
- s += ((UTF8) ? UTF8_SAFE_SKIP(s, reginfo->strend) : 1);\
+ s += UTF8_SAFE_SKIP(s, reginfo->strend); \
previous_occurrence_end = s; \
} \
else { \
- s += ((UTF8) ? UTF8SKIP(s) : 1); \
+ s += UTF8SKIP(s); \
}
-#define REXEC_FBC_CSCAN(CONDUTF8,COND) \
- if (utf8_target) { \
- REXEC_FBC_CLASS_SCAN(1, CONDUTF8); \
+#define REXEC_FBC_NON_UTF8_CLASS_SCAN_GUTS(COND) \
+ if (COND) { \
+ FBC_CHECK_AND_TRY \
+ s++; \
+ previous_occurrence_end = s; \
} \
else { \
- REXEC_FBC_CLASS_SCAN(0, COND); \
+ s++; \
}
/* We keep track of where the next character should start after an occurrence
}
-/* This differs from the above macros in that it calls a function which returns
- * the next occurrence of the thing being looked for in 's'; and 'strend' if
- * there is no such occurrence. */
-#define REXEC_FBC_FIND_NEXT_SCAN(UTF8, f) \
+/* These differ from the above macros in that they call a function which
+ * returns the next occurrence of the thing being looked for in 's'; and
+ * 'strend' if there is no such occurrence. */
+#define REXEC_FBC_UTF8_FIND_NEXT_SCAN(f) \
+ while (s < strend) { \
+ s = (f); \
+ if (s >= strend) { \
+ break; \
+ } \
+ \
+ FBC_CHECK_AND_TRY \
+ s += UTF8SKIP(s); \
+ previous_occurrence_end = s; \
+ }
+
+#define REXEC_FBC_NON_UTF8_FIND_NEXT_SCAN(f) \
while (s < strend) { \
s = (f); \
if (s >= strend) { \
} \
\
FBC_CHECK_AND_TRY \
- s += (UTF8) ? UTF8SKIP(s) : 1; \
+ s++; \
previous_occurrence_end = s; \
}
} \
}
-/* The three macros below are slightly different versions of the same logic.
+/* The four macros below are slightly different versions of the same logic.
*
* The first is for /a and /aa when the target string is UTF-8. This can only
- * match ascii, but it must advance based on UTF-8. The other two handle the
- * non-UTF-8 and the more generic UTF-8 cases. In all three, we are looking
- * for the boundary (or non-boundary) between a word and non-word character.
- * The utf8 and non-utf8 cases have the same logic, but the details must be
- * different. Find the "wordness" of the character just prior to this one, and
- * compare it with the wordness of this one. If they differ, we have a
- * boundary. At the beginning of the string, pretend that the previous
+ * match ascii, but it must advance based on UTF-8. The other three handle
+ * the non-UTF-8 and the more generic UTF-8 cases. In all four, we are
+ * looking for the boundary (or non-boundary) between a word and non-word
+ * character. The utf8 and non-utf8 cases have the same logic, but the details
+ * must be different. Find the "wordness" of the character just prior to this
+ * one, and compare it with the wordness of this one. If they differ, we have
+ * a boundary. At the beginning of the string, pretend that the previous
* character was a new-line.
*
* All these macros uncleanly have side-effects with each other and outside
* see if this tentative match actually works, and if so, to quit the loop
* here. And vice-versa if we are looking for a non-boundary.
*
- * 'tmp' below in the next three macros in the REXEC_FBC_SCAN and
- * REXEC_FBC_SCAN loops is a loop invariant, a bool giving the return of
+ * 'tmp' below in the next four macros in the REXEC_FBC_UTF8_SCAN and
+ * REXEC_FBC_UTF8_SCAN loops is a loop invariant, a bool giving the return of
* TEST_NON_UTF8(s-1). To see this, note that that's what it is defined to be
* at entry to the loop, and to get to the IF_FAIL branch, tmp must equal
* TEST_NON_UTF8(s), and in the opposite branch, IF_SUCCESS, tmp is that
#define FBC_UTF8_A(TEST_NON_UTF8, IF_SUCCESS, IF_FAIL) \
tmp = (s != reginfo->strbeg) ? UCHARAT(s - 1) : '\n'; \
tmp = TEST_NON_UTF8(tmp); \
- REXEC_FBC_SCAN(1, /* 1=>is-utf8; advances s while s < strend */ \
+ REXEC_FBC_UTF8_SCAN( /* advances s while s < strend */ \
if (tmp == ! TEST_NON_UTF8((U8) *s)) { \
tmp = !tmp; \
IF_SUCCESS; /* Is a boundary if values for s-1 and s differ */ \
/* Like FBC_UTF8_A, but TEST_UV is a macro which takes a UV as its input, and
* TEST_UTF8 is a macro that for the same input code points returns identically
- * to TEST_UV, but takes a pointer to a UTF-8 encoded string instead */
+ * to TEST_UV, but takes a pointer to a UTF-8 encoded string instead (and an
+ * end pointer as well) */
#define FBC_UTF8(TEST_UV, TEST_UTF8, IF_SUCCESS, IF_FAIL) \
if (s == reginfo->strbeg) { \
tmp = '\n'; \
0, UTF8_ALLOW_DEFAULT); \
} \
tmp = TEST_UV(tmp); \
- REXEC_FBC_SCAN(1, /* 1=>is-utf8; advances s while s < strend */ \
+ REXEC_FBC_UTF8_SCAN(/* advances s while s < strend */ \
if (tmp == ! (TEST_UTF8((U8 *) s, (U8 *) reginfo->strend))) { \
tmp = !tmp; \
IF_SUCCESS; \
} \
);
-/* Like the above two macros. UTF8_CODE is the complete code for handling
- * UTF-8. Common to the BOUND and NBOUND cases, set-up by the FBC_BOUND, etc
- * macros below */
-#define FBC_BOUND_COMMON(UTF8_CODE, TEST_NON_UTF8, IF_SUCCESS, IF_FAIL) \
- if (utf8_target) { \
- UTF8_CODE \
- } \
- else { /* Not utf8 */ \
- tmp = (s != reginfo->strbeg) ? UCHARAT(s - 1) : '\n'; \
- tmp = TEST_NON_UTF8(tmp); \
- REXEC_FBC_SCAN(0, /* 0=>not-utf8; advances s while s < strend */ \
- if (tmp == ! TEST_NON_UTF8((U8) *s)) { \
- IF_SUCCESS; \
- tmp = !tmp; \
- } \
- else { \
- IF_FAIL; \
- } \
- ); \
- } \
+/* Like the above two macros, for a UTF-8 target string. UTF8_CODE is the
+ * complete code for handling UTF-8. Common to the BOUND and NBOUND cases,
+ * set-up by the FBC_BOUND, etc macros below */
+#define FBC_BOUND_COMMON_UTF8(UTF8_CODE, TEST_NON_UTF8, IF_SUCCESS, IF_FAIL) \
+ UTF8_CODE; \
/* Here, things have been set up by the previous code so that tmp is the \
- * return of TEST_NON_UTF(s-1) or TEST_UTF8(s-1) (depending on the \
- * utf8ness of the target). We also have to check if this matches against \
- * the EOS, which we treat as a \n (which is the same value in both UTF-8 \
- * or non-UTF8, so can use the non-utf8 test condition even for a UTF-8 \
- * string */ \
+ * return of TEST_NON_UTF8(s-1). We also have to check if this matches \
+ * against the EOS, which we treat as a \n */ \
if (tmp == ! TEST_NON_UTF8('\n')) { \
IF_SUCCESS; \
} \
IF_FAIL; \
}
+/* Same as the macro above, but the target isn't UTF-8 */
+#define FBC_BOUND_COMMON_NON_UTF8(TEST_NON_UTF8, IF_SUCCESS, IF_FAIL) \
+ tmp = (s != reginfo->strbeg) ? UCHARAT(s - 1) : '\n'; \
+ tmp = TEST_NON_UTF8(tmp); \
+ REXEC_FBC_NON_UTF8_SCAN(/* advances s while s < strend */ \
+ if (tmp == ! TEST_NON_UTF8(UCHARAT(s))) { \
+ IF_SUCCESS; \
+ tmp = !tmp; \
+ } \
+ else { \
+ IF_FAIL; \
+ } \
+ ); \
+ /* Here, things have been set up by the previous code so that tmp is \
+ * the return of TEST_NON_UTF8(s-1). We also have to check if this \
+ * matches against the EOS, which we treat as a \n */ \
+ if (tmp == ! TEST_NON_UTF8('\n')) { \
+ IF_SUCCESS; \
+ } \
+ else { \
+ IF_FAIL; \
+ }
+
/* This is the macro to use when we want to see if something that looks like it
* could match, actually does, and if so exits the loop. It needs to be used
* only for bounds checking macros, as it allows for matching beyond the end of
* The TEST_FOO parameters are for operating on different forms of input, but
* all should be ones that return identically for the same underlying code
* points */
-#define FBC_BOUND(TEST_NON_UTF8, TEST_UV, TEST_UTF8) \
- FBC_BOUND_COMMON( \
- FBC_UTF8(TEST_UV, TEST_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER), \
+
+#define FBC_BOUND_UTF8(TEST_NON_UTF8, TEST_UV, TEST_UTF8) \
+ FBC_BOUND_COMMON_UTF8( \
+ FBC_UTF8(TEST_UV, TEST_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER), \
TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER)
-#define FBC_BOUND_A(TEST_NON_UTF8) \
- FBC_BOUND_COMMON( \
- FBC_UTF8_A(TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER), \
- TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER)
+#define FBC_BOUND_NON_UTF8(TEST_NON_UTF8) \
+ FBC_BOUND_COMMON_NON_UTF8(TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER)
+
+#define FBC_BOUND_A_UTF8(TEST_NON_UTF8) \
+ FBC_BOUND_COMMON_UTF8( \
+ FBC_UTF8_A(TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER),\
+ TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER)
+
+#define FBC_BOUND_A_NON_UTF8(TEST_NON_UTF8) \
+ FBC_BOUND_COMMON_NON_UTF8(TEST_NON_UTF8, REXEC_FBC_TRYIT, PLACEHOLDER)
-#define FBC_NBOUND(TEST_NON_UTF8, TEST_UV, TEST_UTF8) \
- FBC_BOUND_COMMON( \
- FBC_UTF8(TEST_UV, TEST_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT), \
- TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT)
+#define FBC_NBOUND_UTF8(TEST_NON_UTF8, TEST_UV, TEST_UTF8) \
+ FBC_BOUND_COMMON_UTF8( \
+ FBC_UTF8(TEST_UV, TEST_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT), \
+ TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT)
-#define FBC_NBOUND_A(TEST_NON_UTF8) \
- FBC_BOUND_COMMON( \
- FBC_UTF8_A(TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT), \
+#define FBC_NBOUND_NON_UTF8(TEST_NON_UTF8) \
+ FBC_BOUND_COMMON_NON_UTF8(TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT)
+
+#define FBC_NBOUND_A_UTF8(TEST_NON_UTF8) \
+ FBC_BOUND_COMMON_UTF8( \
+ FBC_UTF8_A(TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT), \
TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT)
+#define FBC_NBOUND_A_NON_UTF8(TEST_NON_UTF8) \
+ FBC_BOUND_COMMON_NON_UTF8(TEST_NON_UTF8, PLACEHOLDER, REXEC_FBC_TRYIT)
+
#ifdef DEBUGGING
static IV
S_get_break_val_cp_checked(SV* const invlist, const UV cp_in) {
S_find_byclass(pTHX_ regexp * prog, const regnode *c, char *s,
const char *strend, regmatch_info *reginfo)
{
- dVAR;
/* TRUE if x+ need not match at just the 1st pos of run of x's */
const I32 doevery = (prog->intflags & PREGf_SKIP) == 0;
PERL_ARGS_ASSERT_FIND_BYCLASS;
- /* We know what class it must start with. */
- switch (OP(c)) {
- case ANYOFPOSIXL:
- case ANYOFL:
+ /* We know what class it must start with. The case statements below have
+ * encoded the OP, and the UTF8ness of the target ('t8' for is UTF-8; 'tb'
+ * for it isn't; 'b' stands for byte), and the UTF8ness of the pattern
+ * ('p8' and 'pb'. */
+ switch (with_tp_UTF8ness(OP(c), utf8_target, is_utf8_pat)) {
+
+ case ANYOFPOSIXL_t8_pb:
+ case ANYOFPOSIXL_t8_p8:
+ case ANYOFL_t8_pb:
+ case ANYOFL_t8_p8:
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
+ CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_SETS(c);
- if (ANYOFL_UTF8_LOCALE_REQD(FLAGS(c)) && ! IN_UTF8_CTYPE_LOCALE) {
- Perl_ck_warner(aTHX_ packWARN(WARN_LOCALE), utf8_locale_required);
- }
+ /* FALLTHROUGH */
+
+ case ANYOFD_t8_pb:
+ case ANYOFD_t8_p8:
+ case ANYOF_t8_pb:
+ case ANYOF_t8_p8:
+ REXEC_FBC_UTF8_CLASS_SCAN(
+ reginclass(prog, c, (U8*)s, (U8*) strend, 1 /* is utf8 */));
+ break;
+
+ case ANYOFPOSIXL_tb_pb:
+ case ANYOFPOSIXL_tb_p8:
+ case ANYOFL_tb_pb:
+ case ANYOFL_tb_p8:
+ _CHECK_AND_WARN_PROBLEMATIC_LOCALE;
+ CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_SETS(c);
/* FALLTHROUGH */
- case ANYOFD:
- case ANYOF:
- if (utf8_target) {
- REXEC_FBC_CLASS_SCAN(1, /* 1=>is-utf8 */
- reginclass(prog, c, (U8*)s, (U8*) strend, utf8_target));
- }
- else if (ANYOF_FLAGS(c) & ~ ANYOF_MATCHES_ALL_ABOVE_BITMAP) {
+
+ case ANYOFD_tb_pb:
+ case ANYOFD_tb_p8:
+ case ANYOF_tb_pb:
+ case ANYOF_tb_p8:
+ if (ANYOF_FLAGS(c) & ~ ANYOF_MATCHES_ALL_ABOVE_BITMAP) {
/* We know that s is in the bitmap range since the target isn't
* UTF-8, so what happens for out-of-range values is not relevant,
* so exclude that from the flags */
- REXEC_FBC_CLASS_SCAN(0, reginclass(prog,c, (U8*)s, (U8*)s+1, 0));
+ REXEC_FBC_NON_UTF8_CLASS_SCAN(reginclass(prog,c, (U8*)s, (U8*)s+1,
+ 0));
}
else {
- REXEC_FBC_CLASS_SCAN(0, ANYOF_BITMAP_TEST(c, *((U8*)s)));
+ REXEC_FBC_NON_UTF8_CLASS_SCAN(ANYOF_BITMAP_TEST(c, *((U8*)s)));
}
break;
- case ANYOFM: /* ARG() is the base byte; FLAGS() the mask byte */
- /* UTF-8ness doesn't matter because only matches UTF-8 invariants, so
- * use 0 */
- REXEC_FBC_FIND_NEXT_SCAN(0,
- (char *) find_next_masked((U8 *) s, (U8 *) strend,
- (U8) ARG(c), FLAGS(c)));
+ case ANYOFM_tb_pb: /* ARG() is the base byte; FLAGS() the mask byte */
+ case ANYOFM_tb_p8:
+ REXEC_FBC_NON_UTF8_FIND_NEXT_SCAN(
+ (char *) find_next_masked((U8 *) s, (U8 *) strend,
+ (U8) ARG(c), FLAGS(c)));
break;
- case NANYOFM: /* UTF-8ness does matter because can match UTF-8 variants.
- */
- REXEC_FBC_FIND_NEXT_SCAN(utf8_target,
- (char *) find_span_end_mask((U8 *) s, (U8 *) strend,
- (U8) ARG(c), FLAGS(c)));
+ case ANYOFM_t8_pb:
+ case ANYOFM_t8_p8:
+ /* UTF-8ness doesn't matter because only matches UTF-8 invariants. But
+ * we do anyway for performance reasons, as otherwise we would have to
+ * examine all the continuation characters */
+ REXEC_FBC_UTF8_FIND_NEXT_SCAN(
+ (char *) find_next_masked((U8 *) s, (U8 *) strend,
+ (U8) ARG(c), FLAGS(c)));
break;
- case ANYOFH:
- if (utf8_target) { /* Can't possibly match a non-UTF-8 target */
- REXEC_FBC_CLASS_SCAN(TRUE,
- ( (U8) NATIVE_UTF8_TO_I8(*s) >= ANYOF_FLAGS(c)
- && reginclass(prog, c, (U8*)s, (U8*) strend, utf8_target)));
- }
+ case NANYOFM_tb_pb:
+ case NANYOFM_tb_p8:
+ REXEC_FBC_NON_UTF8_FIND_NEXT_SCAN(
+ (char *) find_span_end_mask((U8 *) s, (U8 *) strend,
+ (U8) ARG(c), FLAGS(c)));
break;
- case ANYOFHb:
- if (utf8_target) { /* Can't possibly match a non-UTF-8 target */
+ case NANYOFM_t8_pb:
+ case NANYOFM_t8_p8: /* UTF-8ness does matter because can match UTF-8
+ variants. */
+ REXEC_FBC_UTF8_FIND_NEXT_SCAN(
+ (char *) find_span_end_mask((U8 *) s, (U8 *) strend,
+ (U8) ARG(c), FLAGS(c)));
+ break;
- /* We know what the first byte of any matched string should be */
+ /* These nodes all require at least one code point to be in UTF-8 to
+ * match */
+ case ANYOFH_tb_pb:
+ case ANYOFH_tb_p8:
+ case ANYOFHb_tb_pb:
+ case ANYOFHb_tb_p8:
+ case ANYOFHr_tb_pb:
+ case ANYOFHr_tb_p8:
+ case ANYOFHs_tb_pb:
+ case ANYOFHs_tb_p8:
+ case EXACTFLU8_tb_pb:
+ case EXACTFLU8_tb_p8:
+ case EXACTFU_REQ8_tb_pb:
+ case EXACTFU_REQ8_tb_p8:
+ break;
+
+ case ANYOFH_t8_pb:
+ case ANYOFH_t8_p8:
+ REXEC_FBC_UTF8_CLASS_SCAN(
+ ( (U8) NATIVE_UTF8_TO_I8(*s) >= ANYOF_FLAGS(c)
+ && reginclass(prog, c, (U8*)s, (U8*) strend, 1 /* is utf8 */)));
+ break;
+
+ case ANYOFHb_t8_pb:
+ case ANYOFHb_t8_p8:
+ {
+ /* We know what the first byte of any matched string should be. */
U8 first_byte = FLAGS(c);
REXEC_FBC_FIND_NEXT_UTF8_BYTE_SCAN(first_byte,
- reginclass(prog, c, (U8*)s, (U8*) strend, utf8_target));
+ reginclass(prog, c, (U8*)s, (U8*) strend, 1 /* is utf8 */));
}
break;
- case ANYOFHr:
- if (utf8_target) { /* Can't possibly match a non-UTF-8 target */
- REXEC_FBC_CLASS_SCAN(TRUE,
- ( inRANGE((U8) NATIVE_UTF8_TO_I8(*s),
- LOWEST_ANYOF_HRx_BYTE(ANYOF_FLAGS(c)),
- HIGHEST_ANYOF_HRx_BYTE(ANYOF_FLAGS(c)))
- && reginclass(prog, c, (U8*)s, (U8*) strend, utf8_target)));
- }
+ case ANYOFHr_t8_pb:
+ case ANYOFHr_t8_p8:
+ REXEC_FBC_UTF8_CLASS_SCAN(
+ ( inRANGE(NATIVE_UTF8_TO_I8(*s),
+ LOWEST_ANYOF_HRx_BYTE(ANYOF_FLAGS(c)),
+ HIGHEST_ANYOF_HRx_BYTE(ANYOF_FLAGS(c)))
+ && reginclass(prog, c, (U8*)s, (U8*) strend,
+ 1 /* is utf8 */)));
break;
- case EXACTFAA_NO_TRIE: /* This node only generated for non-utf8 patterns */
- assert(! is_utf8_pat);
- /* FALLTHROUGH */
- case EXACTFAA:
- if (is_utf8_pat) {
- utf8_fold_flags = FOLDEQ_UTF8_NOMIX_ASCII
- |FOLDEQ_S2_ALREADY_FOLDED|FOLDEQ_S2_FOLDS_SANE;
- goto do_exactf_utf8;
- }
- else if (utf8_target) {
+ case ANYOFHs_t8_pb:
+ case ANYOFHs_t8_p8:
+ REXEC_FBC_UTF8_CLASS_SCAN(
+ ( strend -s >= FLAGS(c)
+ && memEQ(s, ((struct regnode_anyofhs *) c)->string, FLAGS(c))
+ && reginclass(prog, c, (U8*)s, (U8*) strend, 1 /* is utf8 */)));
+ break;
- /* Here, and elsewhere in this file, the reason we can't consider a
- * non-UTF-8 pattern already folded in the presence of a UTF-8
- * target is because any MICRO SIGN in the pattern won't be folded.
- * Since the fold of the MICRO SIGN requires UTF-8 to represent, we
- * can consider a non-UTF-8 pattern folded when matching a
- * non-UTF-8 target */
- utf8_fold_flags = FOLDEQ_UTF8_NOMIX_ASCII;
- goto do_exactf_utf8;
+ case ANYOFR_tb_pb:
+ case ANYOFR_tb_p8:
+ REXEC_FBC_NON_UTF8_CLASS_SCAN(withinCOUNT((U8) *s,
+ ANYOFRbase(c), ANYOFRdelta(c)));
+ break;
+
+ case ANYOFR_t8_pb:
+ case ANYOFR_t8_p8:
+ REXEC_FBC_UTF8_CLASS_SCAN(
+ ( NATIVE_UTF8_TO_I8(*s) >= ANYOF_FLAGS(c)
+ && withinCOUNT(utf8_to_uvchr_buf((U8 *) s,
+ (U8 *) strend,
+ NULL),
+ ANYOFRbase(c), ANYOFRdelta(c))));
+ break;
+
+ case ANYOFRb_tb_pb:
+ case ANYOFRb_tb_p8:
+ REXEC_FBC_NON_UTF8_CLASS_SCAN(withinCOUNT((U8) *s,
+ ANYOFRbase(c), ANYOFRdelta(c)));
+ break;
+
+ case ANYOFRb_t8_pb:
+ case ANYOFRb_t8_p8:
+ { /* We know what the first byte of any matched string should be */
+ U8 first_byte = FLAGS(c);
+
+ REXEC_FBC_FIND_NEXT_UTF8_BYTE_SCAN(first_byte,
+ withinCOUNT(utf8_to_uvchr_buf((U8 *) s,
+ (U8 *) strend,
+ NULL),
+ ANYOFRbase(c), ANYOFRdelta(c)));
}
+ break;
+
+ case EXACTFAA_tb_pb:
/* Latin1 folds are not affected by /a, except it excludes the sharp s,
* which these functions don't handle anyway */
folder = foldEQ_latin1_s2_folded;
goto do_exactf_non_utf8;
- case EXACTF: /* This node only generated for non-utf8 patterns */
- assert(! is_utf8_pat);
- if (utf8_target) {
- goto do_exactf_utf8;
- }
+ case EXACTF_tb_pb:
fold_array = PL_fold;
folder = foldEQ;
goto do_exactf_non_utf8;
- case EXACTFL:
+ case EXACTFL_tb_pb:
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
- if (is_utf8_pat || utf8_target || IN_UTF8_CTYPE_LOCALE) {
+
+ if (IN_UTF8_CTYPE_LOCALE) {
utf8_fold_flags = FOLDEQ_LOCALE;
goto do_exactf_utf8;
}
+
fold_array = PL_fold_locale;
folder = foldEQ_locale;
goto do_exactf_non_utf8;
- case EXACTFUP: /* Problematic even though pattern isn't UTF-8. Use
- full functionality normally not done except for
- UTF-8 */
- assert(! is_utf8_pat);
- goto do_exactf_utf8;
-
- case EXACTFLU8:
- if (! utf8_target) { /* All code points in this node require
- UTF-8 to express. */
- break;
- }
- utf8_fold_flags = FOLDEQ_LOCALE | FOLDEQ_S2_ALREADY_FOLDED
- | FOLDEQ_S2_FOLDS_SANE;
- goto do_exactf_utf8;
-
- case EXACTFU_ONLY8:
- if (! utf8_target) {
- break;
- }
- assert(is_utf8_pat);
- utf8_fold_flags = FOLDEQ_S2_ALREADY_FOLDED;
- goto do_exactf_utf8;
-
- case EXACTFU:
- if (is_utf8_pat || utf8_target) {
- utf8_fold_flags = FOLDEQ_S2_ALREADY_FOLDED;
- goto do_exactf_utf8;
- }
-
- /* Any 'ss' in the pattern should have been replaced by regcomp,
- * so we don't have to worry here about this single special case
- * in the Latin1 range */
+ case EXACTFU_tb_pb:
+ /* Any 'ss' in the pattern should have been replaced by regcomp, so we
+ * don't have to worry here about this single special case in the
+ * Latin1 range */
fold_array = PL_fold_latin1;
folder = foldEQ_latin1_s2_folded;
/* FALLTHROUGH */
- do_exactf_non_utf8: /* Neither pattern nor string are UTF8, and there
- are no glitches with fold-length differences
- between the target string and pattern */
+ do_exactf_non_utf8: /* Neither pattern nor string are UTF8, and there
+ are no glitches with fold-length differences
+ between the target string and pattern */
- /* The idea in the non-utf8 EXACTF* cases is to first find the
- * first character of the EXACTF* node and then, if necessary,
+ /* The idea in the non-utf8 EXACTF* cases is to first find the first
+ * character of the EXACTF* node and then, if necessary,
* case-insensitively compare the full text of the node. c1 is the
* first character. c2 is its fold. This logic will not work for
- * Unicode semantics and the german sharp ss, which hence should
- * not be compiled into a node that gets here. */
- pat_string = STRING(c);
- ln = STR_LEN(c); /* length to match in octets/bytes */
-
- /* We know that we have to match at least 'ln' bytes (which is the
- * same as characters, since not utf8). If we have to match 3
- * characters, and there are only 2 availabe, we know without
- * trying that it will fail; so don't start a match past the
- * required minimum number from the far end */
+ * Unicode semantics and the german sharp ss, which hence should not be
+ * compiled into a node that gets here. */
+ pat_string = STRINGs(c);
+ ln = STR_LENs(c); /* length to match in octets/bytes */
+
+ /* We know that we have to match at least 'ln' bytes (which is the same
+ * as characters, since not utf8). If we have to match 3 characters,
+ * and there are only 2 availabe, we know without trying that it will
+ * fail; so don't start a match past the required minimum number from
+ * the far end */
e = HOP3c(strend, -((SSize_t)ln), s);
if (e < s)
break;
}
break;
- do_exactf_utf8:
- {
- unsigned expansion;
-
- /* If one of the operands is in utf8, we can't use the simpler folding
- * above, due to the fact that many different characters can have the
- * same fold, or portion of a fold, or different- length fold */
- pat_string = STRING(c);
- ln = STR_LEN(c); /* length to match in octets/bytes */
- pat_end = pat_string + ln;
- lnc = is_utf8_pat /* length to match in characters */
- ? utf8_length((U8 *) pat_string, (U8 *) pat_end)
- : ln;
-
- /* We have 'lnc' characters to match in the pattern, but because of
- * multi-character folding, each character in the target can match
- * up to 3 characters (Unicode guarantees it will never exceed
- * this) if it is utf8-encoded; and up to 2 if not (based on the
- * fact that the Latin 1 folds are already determined, and the
- * only multi-char fold in that range is the sharp-s folding to
- * 'ss'. Thus, a pattern character can match as little as 1/3 of a
- * string character. Adjust lnc accordingly, rounding up, so that
- * if we need to match at least 4+1/3 chars, that really is 5. */
- expansion = (utf8_target) ? UTF8_MAX_FOLD_CHAR_EXPAND : 2;
- lnc = (lnc + expansion - 1) / expansion;
-
- /* As in the non-UTF8 case, if we have to match 3 characters, and
- * only 2 are left, it's guaranteed to fail, so don't start a
- * match that would require us to go beyond the end of the string
- */
- e = HOP3c(strend, -((SSize_t)lnc), s);
-
- /* XXX Note that we could recalculate e to stop the loop earlier,
- * as the worst case expansion above will rarely be met, and as we
- * go along we would usually find that e moves further to the left.
- * This would happen only after we reached the point in the loop
- * where if there were no expansion we should fail. Unclear if
- * worth the expense */
-
- while (s <= e) {
- char *my_strend= (char *)strend;
- if (foldEQ_utf8_flags(s, &my_strend, 0, utf8_target,
- pat_string, NULL, ln, is_utf8_pat, utf8_fold_flags)
- && (reginfo->intuit || regtry(reginfo, &s)) )
- {
- goto got_it;
- }
- s += (utf8_target) ? UTF8_SAFE_SKIP(s, reginfo->strend) : 1;
- }
- break;
- }
+ case EXACTFAA_tb_p8:
+ case EXACTFAA_t8_p8:
+ utf8_fold_flags = FOLDEQ_UTF8_NOMIX_ASCII
+ |FOLDEQ_S2_ALREADY_FOLDED
+ |FOLDEQ_S2_FOLDS_SANE;
+ goto do_exactf_utf8;
- case BOUNDL:
- _CHECK_AND_WARN_PROBLEMATIC_LOCALE;
- if (FLAGS(c) != TRADITIONAL_BOUND) {
- if (! IN_UTF8_CTYPE_LOCALE) {
- Perl_ck_warner(aTHX_ packWARN(WARN_LOCALE),
- B_ON_NON_UTF8_LOCALE_IS_WRONG);
- }
- goto do_boundu;
- }
+ case EXACTFAA_NO_TRIE_tb_pb:
+ case EXACTFAA_NO_TRIE_t8_pb:
+ case EXACTFAA_t8_pb:
- FBC_BOUND(isWORDCHAR_LC, isWORDCHAR_LC_uvchr, isWORDCHAR_LC_utf8_safe);
- break;
+ /* Here, and elsewhere in this file, the reason we can't consider a
+ * non-UTF-8 pattern already folded in the presence of a UTF-8 target
+ * is because any MICRO SIGN in the pattern won't be folded. Since the
+ * fold of the MICRO SIGN requires UTF-8 to represent, we can consider
+ * a non-UTF-8 pattern folded when matching a non-UTF-8 target */
+ utf8_fold_flags = FOLDEQ_UTF8_NOMIX_ASCII;
+ goto do_exactf_utf8;
- case NBOUNDL:
+ case EXACTFL_tb_p8:
+ case EXACTFL_t8_pb:
+ case EXACTFL_t8_p8:
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
- if (FLAGS(c) != TRADITIONAL_BOUND) {
- if (! IN_UTF8_CTYPE_LOCALE) {
- Perl_ck_warner(aTHX_ packWARN(WARN_LOCALE),
- B_ON_NON_UTF8_LOCALE_IS_WRONG);
+ utf8_fold_flags = FOLDEQ_LOCALE;
+ goto do_exactf_utf8;
+
+ case EXACTFLU8_t8_pb:
+ case EXACTFLU8_t8_p8:
+ utf8_fold_flags = FOLDEQ_LOCALE | FOLDEQ_S2_ALREADY_FOLDED
+ | FOLDEQ_S2_FOLDS_SANE;
+ goto do_exactf_utf8;
+
+ case EXACTFU_REQ8_t8_p8:
+ utf8_fold_flags = FOLDEQ_S2_ALREADY_FOLDED;
+ goto do_exactf_utf8;
+
+ case EXACTFU_tb_p8:
+ case EXACTFU_t8_pb:
+ case EXACTFU_t8_p8:
+ utf8_fold_flags = FOLDEQ_S2_ALREADY_FOLDED;
+ goto do_exactf_utf8;
+
+ /* The following are problematic even though pattern isn't UTF-8. Use
+ * full functionality normally not done except for UTF-8. */
+ case EXACTF_t8_pb:
+ case EXACTFUP_tb_pb:
+ case EXACTFUP_t8_pb:
+
+ do_exactf_utf8:
+ {
+ unsigned expansion;
+
+ /* If one of the operands is in utf8, we can't use the simpler
+ * folding above, due to the fact that many different characters
+ * can have the same fold, or portion of a fold, or different-
+ * length fold */
+ pat_string = STRINGs(c);
+ ln = STR_LENs(c); /* length to match in octets/bytes */
+ pat_end = pat_string + ln;
+ lnc = is_utf8_pat /* length to match in characters */
+ ? utf8_length((U8 *) pat_string, (U8 *) pat_end)
+ : ln;
+
+ /* We have 'lnc' characters to match in the pattern, but because of
+ * multi-character folding, each character in the target can match
+ * up to 3 characters (Unicode guarantees it will never exceed
+ * this) if it is utf8-encoded; and up to 2 if not (based on the
+ * fact that the Latin 1 folds are already determined, and the only
+ * multi-char fold in that range is the sharp-s folding to 'ss'.
+ * Thus, a pattern character can match as little as 1/3 of a string
+ * character. Adjust lnc accordingly, rounding up, so that if we
+ * need to match at least 4+1/3 chars, that really is 5. */
+ expansion = (utf8_target) ? UTF8_MAX_FOLD_CHAR_EXPAND : 2;
+ lnc = (lnc + expansion - 1) / expansion;
+
+ /* As in the non-UTF8 case, if we have to match 3 characters, and
+ * only 2 are left, it's guaranteed to fail, so don't start a match
+ * that would require us to go beyond the end of the string */
+ e = HOP3c(strend, -((SSize_t)lnc), s);
+
+ /* XXX Note that we could recalculate e to stop the loop earlier,
+ * as the worst case expansion above will rarely be met, and as we
+ * go along we would usually find that e moves further to the left.
+ * This would happen only after we reached the point in the loop
+ * where if there were no expansion we should fail. Unclear if
+ * worth the expense */
+
+ while (s <= e) {
+ char *my_strend= (char *)strend;
+ if ( foldEQ_utf8_flags(s, &my_strend, 0, utf8_target,
+ pat_string, NULL, ln, is_utf8_pat,
+ utf8_fold_flags)
+ && (reginfo->intuit || regtry(reginfo, &s)) )
+ {
+ goto got_it;
+ }
+ s += (utf8_target) ? UTF8_SAFE_SKIP(s, reginfo->strend) : 1;
}
- goto do_nboundu;
}
-
- FBC_NBOUND(isWORDCHAR_LC, isWORDCHAR_LC_uvchr, isWORDCHAR_LC_utf8_safe);
break;
- case BOUND: /* regcomp.c makes sure that this only has the traditional \b
- meaning */
+ case BOUNDA_tb_pb:
+ case BOUNDA_tb_p8:
+ case BOUND_tb_pb: /* /d without utf8 target is /a */
+ case BOUND_tb_p8:
+ /* regcomp.c makes sure that these only have the traditional \b
+ * meaning. */
assert(FLAGS(c) == TRADITIONAL_BOUND);
- FBC_BOUND(isWORDCHAR, isWORDCHAR_uni, isWORDCHAR_utf8_safe);
+ FBC_BOUND_A_NON_UTF8(isWORDCHAR_A);
break;
- case BOUNDA: /* regcomp.c makes sure that this only has the traditional \b
- meaning */
+ case BOUNDA_t8_pb: /* What /a matches is same under UTF-8 */
+ case BOUNDA_t8_p8:
+ /* regcomp.c makes sure that these only have the traditional \b
+ * meaning. */
assert(FLAGS(c) == TRADITIONAL_BOUND);
- FBC_BOUND_A(isWORDCHAR_A);
+ FBC_BOUND_A_UTF8(isWORDCHAR_A);
break;
- case NBOUND: /* regcomp.c makes sure that this only has the traditional \b
- meaning */
+ case NBOUNDA_tb_pb:
+ case NBOUNDA_tb_p8:
+ case NBOUND_tb_pb: /* /d without utf8 target is /a */
+ case NBOUND_tb_p8:
+ /* regcomp.c makes sure that these only have the traditional \b
+ * meaning. */
assert(FLAGS(c) == TRADITIONAL_BOUND);
- FBC_NBOUND(isWORDCHAR, isWORDCHAR_uni, isWORDCHAR_utf8_safe);
+ FBC_NBOUND_A_NON_UTF8(isWORDCHAR_A);
break;
- case NBOUNDA: /* regcomp.c makes sure that this only has the traditional \b
- meaning */
+ case NBOUNDA_t8_pb: /* What /a matches is same under UTF-8 */
+ case NBOUNDA_t8_p8:
+ /* regcomp.c makes sure that these only have the traditional \b
+ * meaning. */
assert(FLAGS(c) == TRADITIONAL_BOUND);
- FBC_NBOUND_A(isWORDCHAR_A);
+ FBC_NBOUND_A_UTF8(isWORDCHAR_A);
break;
- case NBOUNDU:
+ case NBOUNDU_tb_pb:
+ case NBOUNDU_tb_p8:
if ((bound_type) FLAGS(c) == TRADITIONAL_BOUND) {
- FBC_NBOUND(isWORDCHAR_L1, isWORDCHAR_uni, isWORDCHAR_utf8_safe);
+ FBC_NBOUND_NON_UTF8(isWORDCHAR_L1);
+ break;
+ }
+
+ to_complement = 1;
+ goto do_boundu_non_utf8;
+
+ case NBOUNDL_tb_pb:
+ case NBOUNDL_tb_p8:
+ _CHECK_AND_WARN_PROBLEMATIC_LOCALE;
+ if (FLAGS(c) == TRADITIONAL_BOUND) {
+ FBC_NBOUND_NON_UTF8(isWORDCHAR_LC);
break;
}
- do_nboundu:
+ CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_BOUND;
to_complement = 1;
- /* FALLTHROUGH */
+ goto do_boundu_non_utf8;
- case BOUNDU:
- do_boundu:
- switch((bound_type) FLAGS(c)) {
- case TRADITIONAL_BOUND:
- FBC_BOUND(isWORDCHAR_L1, isWORDCHAR_uni, isWORDCHAR_utf8_safe);
- break;
- case GCB_BOUND:
- if (s == reginfo->strbeg) {
- if (reginfo->intuit || regtry(reginfo, &s))
- {
- goto got_it;
- }
+ case BOUNDL_tb_pb:
+ case BOUNDL_tb_p8:
+ _CHECK_AND_WARN_PROBLEMATIC_LOCALE;
+ if (FLAGS(c) == TRADITIONAL_BOUND) {
+ FBC_BOUND_NON_UTF8(isWORDCHAR_LC);
+ break;
+ }
- /* Didn't match. Try at the next position (if there is one) */
- s += (utf8_target) ? UTF8_SAFE_SKIP(s, reginfo->strend) : 1;
- if (UNLIKELY(s >= reginfo->strend)) {
- break;
- }
- }
+ CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_BOUND;
- if (utf8_target) {
- GCB_enum before = getGCB_VAL_UTF8(
- reghop3((U8*)s, -1,
- (U8*)(reginfo->strbeg)),
- (U8*) reginfo->strend);
- while (s < strend) {
- GCB_enum after = getGCB_VAL_UTF8((U8*) s,
- (U8*) reginfo->strend);
- if ( (to_complement ^ isGCB(before,
- after,
- (U8*) reginfo->strbeg,
- (U8*) s,
- utf8_target))
- && (reginfo->intuit || regtry(reginfo, &s)))
- {
- goto got_it;
- }
- before = after;
- s += UTF8_SAFE_SKIP(s, reginfo->strend);
- }
- }
- else { /* Not utf8. Everything is a GCB except between CR and
- LF */
- while (s < strend) {
- if ((to_complement ^ ( UCHARAT(s - 1) != '\r'
- || UCHARAT(s) != '\n'))
- && (reginfo->intuit || regtry(reginfo, &s)))
- {
- goto got_it;
- }
- s++;
- }
- }
+ goto do_boundu_non_utf8;
+
+ case BOUNDU_tb_pb:
+ case BOUNDU_tb_p8:
+ if ((bound_type) FLAGS(c) == TRADITIONAL_BOUND) {
+ FBC_BOUND_NON_UTF8(isWORDCHAR_L1);
+ break;
+ }
- /* And, since this is a bound, it can match after the final
- * character in the string */
- if ( reginfo->intuit
- || (s <= reginfo->strend && regtry(reginfo, &s)))
+ do_boundu_non_utf8:
+ if (s == reginfo->strbeg) {
+ if (reginfo->intuit || regtry(reginfo, &s))
+ {
+ goto got_it;
+ }
+
+ /* Didn't match. Try at the next position (if there is one) */
+ s++;
+ if (UNLIKELY(s >= reginfo->strend)) {
+ break;
+ }
+ }
+
+ switch((bound_type) FLAGS(c)) {
+ case TRADITIONAL_BOUND: /* Should have already been handled */
+ assert(0);
+ break;
+
+ case GCB_BOUND:
+ /* Not utf8. Everything is a GCB except between CR and LF */
+ while (s < strend) {
+ if ((to_complement ^ ( UCHARAT(s - 1) != '\r'
+ || UCHARAT(s) != '\n'))
+ && (reginfo->intuit || regtry(reginfo, &s)))
{
goto got_it;
}
- break;
+ s++;
+ }
- case LB_BOUND:
- if (s == reginfo->strbeg) {
- if (reginfo->intuit || regtry(reginfo, &s)) {
+ break;
+
+ case LB_BOUND:
+ {
+ LB_enum before = getLB_VAL_CP((U8) *(s -1));
+ while (s < strend) {
+ LB_enum after = getLB_VAL_CP((U8) *s);
+ if (to_complement ^ isLB(before,
+ after,
+ (U8*) reginfo->strbeg,
+ (U8*) s,
+ (U8*) reginfo->strend,
+ 0 /* target not utf8 */ )
+ && (reginfo->intuit || regtry(reginfo, &s)))
+ {
goto got_it;
}
- s += (utf8_target) ? UTF8_SAFE_SKIP(s, reginfo->strend) : 1;
- if (UNLIKELY(s >= reginfo->strend)) {
- break;
- }
+ before = after;
+ s++;
}
+ }
- if (utf8_target) {
- LB_enum before = getLB_VAL_UTF8(reghop3((U8*)s,
- -1,
- (U8*)(reginfo->strbeg)),
- (U8*) reginfo->strend);
- while (s < strend) {
- LB_enum after = getLB_VAL_UTF8((U8*) s, (U8*) reginfo->strend);
- if (to_complement ^ isLB(before,
- after,
- (U8*) reginfo->strbeg,
- (U8*) s,
- (U8*) reginfo->strend,
- utf8_target)
- && (reginfo->intuit || regtry(reginfo, &s)))
- {
- goto got_it;
- }
- before = after;
- s += UTF8_SAFE_SKIP(s, reginfo->strend);
+ break;
+
+ case SB_BOUND:
+ {
+ SB_enum before = getSB_VAL_CP((U8) *(s -1));
+ while (s < strend) {
+ SB_enum after = getSB_VAL_CP((U8) *s);
+ if ((to_complement ^ isSB(before,
+ after,
+ (U8*) reginfo->strbeg,
+ (U8*) s,
+ (U8*) reginfo->strend,
+ 0 /* target not utf8 */ ))
+ && (reginfo->intuit || regtry(reginfo, &s)))
+ {
+ goto got_it;
}
+ before = after;
+ s++;
}
- else { /* Not utf8. */
- LB_enum before = getLB_VAL_CP((U8) *(s -1));
- while (s < strend) {
- LB_enum after = getLB_VAL_CP((U8) *s);
- if (to_complement ^ isLB(before,
- after,
- (U8*) reginfo->strbeg,
- (U8*) s,
- (U8*) reginfo->strend,
- utf8_target)
- && (reginfo->intuit || regtry(reginfo, &s)))
- {
- goto got_it;
- }
- before = after;
- s++;
+ }
+
+ break;
+
+ case WB_BOUND:
+ {
+ WB_enum previous = WB_UNKNOWN;
+ WB_enum before = getWB_VAL_CP((U8) *(s -1));
+ while (s < strend) {
+ WB_enum after = getWB_VAL_CP((U8) *s);
+ if ((to_complement ^ isWB(previous,
+ before,
+ after,
+ (U8*) reginfo->strbeg,
+ (U8*) s,
+ (U8*) reginfo->strend,
+ 0 /* target not utf8 */ ))
+ && (reginfo->intuit || regtry(reginfo, &s)))
+ {
+ goto got_it;
}
+ previous = before;
+ before = after;
+ s++;
}
+ }
+ }
- if ( reginfo->intuit
- || (s <= reginfo->strend && regtry(reginfo, &s)))
- {
- goto got_it;
- }
+ /* Here are at the final position in the target string, which is a
+ * boundary by definition, so matches, depending on other constraints.
+ * */
+ if ( reginfo->intuit
+ || (s <= reginfo->strend && regtry(reginfo, &s)))
+ {
+ goto got_it;
+ }
+
+ break;
+
+ case BOUNDL_t8_pb:
+ case BOUNDL_t8_p8:
+ _CHECK_AND_WARN_PROBLEMATIC_LOCALE;
+ if (FLAGS(c) == TRADITIONAL_BOUND) {
+ FBC_BOUND_UTF8(isWORDCHAR_LC, isWORDCHAR_LC_uvchr,
+ isWORDCHAR_LC_utf8_safe);
+ break;
+ }
+
+ CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_BOUND;
+
+ to_complement = 1;
+ goto do_boundu_utf8;
+
+ case NBOUNDL_t8_pb:
+ case NBOUNDL_t8_p8:
+ _CHECK_AND_WARN_PROBLEMATIC_LOCALE;
+ if (FLAGS(c) == TRADITIONAL_BOUND) {
+ FBC_NBOUND_UTF8(isWORDCHAR_LC, isWORDCHAR_LC_uvchr,
+ isWORDCHAR_LC_utf8_safe);
+ break;
+ }
+
+ CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_BOUND;
+
+ to_complement = 1;
+ goto do_boundu_utf8;
+
+ case NBOUND_t8_pb:
+ case NBOUND_t8_p8:
+ /* regcomp.c makes sure that these only have the traditional \b
+ * meaning. */
+ assert(FLAGS(c) == TRADITIONAL_BOUND);
+
+ /* FALLTHROUGH */
+
+ case NBOUNDU_t8_pb:
+ case NBOUNDU_t8_p8:
+ if ((bound_type) FLAGS(c) == TRADITIONAL_BOUND) {
+ FBC_NBOUND_UTF8(isWORDCHAR_L1, isWORDCHAR_uni,
+ isWORDCHAR_utf8_safe);
+ break;
+ }
+ to_complement = 1;
+ goto do_boundu_utf8;
+
+ case BOUND_t8_pb:
+ case BOUND_t8_p8:
+ /* regcomp.c makes sure that these only have the traditional \b
+ * meaning. */
+ assert(FLAGS(c) == TRADITIONAL_BOUND);
+
+ /* FALLTHROUGH */
+
+ case BOUNDU_t8_pb:
+ case BOUNDU_t8_p8:
+ if ((bound_type) FLAGS(c) == TRADITIONAL_BOUND) {
+ FBC_BOUND_UTF8(isWORDCHAR_L1, isWORDCHAR_uni, isWORDCHAR_utf8_safe);
+ break;
+ }
+
+ do_boundu_utf8:
+ if (s == reginfo->strbeg) {
+ if (reginfo->intuit || regtry(reginfo, &s))
+ {
+ goto got_it;
+ }
+
+ /* Didn't match. Try at the next position (if there is one) */
+ s += UTF8_SAFE_SKIP(s, reginfo->strend);
+ if (UNLIKELY(s >= reginfo->strend)) {
break;
+ }
+ }
- case SB_BOUND:
- if (s == reginfo->strbeg) {
- if (reginfo->intuit || regtry(reginfo, &s)) {
- goto got_it;
- }
- s += (utf8_target) ? UTF8_SAFE_SKIP(s, reginfo->strend) : 1;
- if (UNLIKELY(s >= reginfo->strend)) {
- break;
- }
- }
+ switch((bound_type) FLAGS(c)) {
+ case TRADITIONAL_BOUND: /* Should have already been handled */
+ assert(0);
+ break;
- if (utf8_target) {
- SB_enum before = getSB_VAL_UTF8(reghop3((U8*)s,
- -1,
- (U8*)(reginfo->strbeg)),
- (U8*) reginfo->strend);
- while (s < strend) {
- SB_enum after = getSB_VAL_UTF8((U8*) s,
- (U8*) reginfo->strend);
- if ((to_complement ^ isSB(before,
- after,
- (U8*) reginfo->strbeg,
- (U8*) s,
- (U8*) reginfo->strend,
- utf8_target))
- && (reginfo->intuit || regtry(reginfo, &s)))
- {
- goto got_it;
- }
- before = after;
- s += UTF8_SAFE_SKIP(s, reginfo->strend);
- }
- }
- else { /* Not utf8. */
- SB_enum before = getSB_VAL_CP((U8) *(s -1));
- while (s < strend) {
- SB_enum after = getSB_VAL_CP((U8) *s);
- if ((to_complement ^ isSB(before,
+ case GCB_BOUND:
+ {
+ GCB_enum before = getGCB_VAL_UTF8(
+ reghop3((U8*)s, -1,
+ (U8*)(reginfo->strbeg)),
+ (U8*) reginfo->strend);
+ while (s < strend) {
+ GCB_enum after = getGCB_VAL_UTF8((U8*) s,
+ (U8*) reginfo->strend);
+ if ( (to_complement ^ isGCB(before,
after,
(U8*) reginfo->strbeg,
(U8*) s,
- (U8*) reginfo->strend,
- utf8_target))
- && (reginfo->intuit || regtry(reginfo, &s)))
- {
- goto got_it;
- }
- before = after;
- s++;
+ 1 /* target is utf8 */ ))
+ && (reginfo->intuit || regtry(reginfo, &s)))
+ {
+ goto got_it;
}
+ before = after;
+ s += UTF8_SAFE_SKIP(s, reginfo->strend);
}
+ }
+ break;
- /* Here are at the final position in the target string. The SB
- * value is always true here, so matches, depending on other
- * constraints */
- if ( reginfo->intuit
- || (s <= reginfo->strend && regtry(reginfo, &s)))
- {
- goto got_it;
+ case LB_BOUND:
+ {
+ LB_enum before = getLB_VAL_UTF8(reghop3((U8*)s,
+ -1,
+ (U8*)(reginfo->strbeg)),
+ (U8*) reginfo->strend);
+ while (s < strend) {
+ LB_enum after = getLB_VAL_UTF8((U8*) s,
+ (U8*) reginfo->strend);
+ if (to_complement ^ isLB(before,
+ after,
+ (U8*) reginfo->strbeg,
+ (U8*) s,
+ (U8*) reginfo->strend,
+ 1 /* target is utf8 */ )
+ && (reginfo->intuit || regtry(reginfo, &s)))
+ {
+ goto got_it;
+ }
+ before = after;
+ s += UTF8_SAFE_SKIP(s, reginfo->strend);
}
+ }
- break;
+ break;
- case WB_BOUND:
- if (s == reginfo->strbeg) {
- if (reginfo->intuit || regtry(reginfo, &s)) {
+ case SB_BOUND:
+ {
+ SB_enum before = getSB_VAL_UTF8(reghop3((U8*)s,
+ -1,
+ (U8*)(reginfo->strbeg)),
+ (U8*) reginfo->strend);
+ while (s < strend) {
+ SB_enum after = getSB_VAL_UTF8((U8*) s,
+ (U8*) reginfo->strend);
+ if ((to_complement ^ isSB(before,
+ after,
+ (U8*) reginfo->strbeg,
+ (U8*) s,
+ (U8*) reginfo->strend,
+ 1 /* target is utf8 */ ))
+ && (reginfo->intuit || regtry(reginfo, &s)))
+ {
goto got_it;
}
- s += (utf8_target) ? UTF8_SAFE_SKIP(s, reginfo->strend) : 1;
- if (UNLIKELY(s >= reginfo->strend)) {
- break;
- }
+ before = after;
+ s += UTF8_SAFE_SKIP(s, reginfo->strend);
}
+ }
- if (utf8_target) {
- /* We are at a boundary between char_sub_0 and char_sub_1.
- * We also keep track of the value for char_sub_-1 as we
- * loop through the line. Context may be needed to make a
- * determination, and if so, this can save having to
- * recalculate it */
- WB_enum previous = WB_UNKNOWN;
- WB_enum before = getWB_VAL_UTF8(
- reghop3((U8*)s,
- -1,
- (U8*)(reginfo->strbeg)),
- (U8*) reginfo->strend);
- while (s < strend) {
- WB_enum after = getWB_VAL_UTF8((U8*) s,
- (U8*) reginfo->strend);
- if ((to_complement ^ isWB(previous,
- before,
- after,
- (U8*) reginfo->strbeg,
- (U8*) s,
- (U8*) reginfo->strend,
- utf8_target))
- && (reginfo->intuit || regtry(reginfo, &s)))
- {
- goto got_it;
- }
- previous = before;
- before = after;
- s += UTF8_SAFE_SKIP(s, reginfo->strend);
- }
- }
- else { /* Not utf8. */
- WB_enum previous = WB_UNKNOWN;
- WB_enum before = getWB_VAL_CP((U8) *(s -1));
- while (s < strend) {
- WB_enum after = getWB_VAL_CP((U8) *s);
- if ((to_complement ^ isWB(previous,
- before,
- after,
- (U8*) reginfo->strbeg,
- (U8*) s,
- (U8*) reginfo->strend,
- utf8_target))
- && (reginfo->intuit || regtry(reginfo, &s)))
- {
- goto got_it;
- }
- previous = before;
- before = after;
- s++;
+ break;
+
+ case WB_BOUND:
+ {
+ /* We are at a boundary between char_sub_0 and char_sub_1.
+ * We also keep track of the value for char_sub_-1 as we
+ * loop through the line. Context may be needed to make a
+ * determination, and if so, this can save having to
+ * recalculate it */
+ WB_enum previous = WB_UNKNOWN;
+ WB_enum before = getWB_VAL_UTF8(
+ reghop3((U8*)s,
+ -1,
+ (U8*)(reginfo->strbeg)),
+ (U8*) reginfo->strend);
+ while (s < strend) {
+ WB_enum after = getWB_VAL_UTF8((U8*) s,
+ (U8*) reginfo->strend);
+ if ((to_complement ^ isWB(previous,
+ before,
+ after,
+ (U8*) reginfo->strbeg,
+ (U8*) s,
+ (U8*) reginfo->strend,
+ 1 /* target is utf8 */ ))
+ && (reginfo->intuit || regtry(reginfo, &s)))
+ {
+ goto got_it;
}
+ previous = before;
+ before = after;
+ s += UTF8_SAFE_SKIP(s, reginfo->strend);
}
+ }
+ }
- if ( reginfo->intuit
- || (s <= reginfo->strend && regtry(reginfo, &s)))
- {
- goto got_it;
- }
+ /* Here are at the final position in the target string, which is a
+ * boundary by definition, so matches, depending on other constraints.
+ * */
+
+ if ( reginfo->intuit
+ || (s <= reginfo->strend && regtry(reginfo, &s)))
+ {
+ goto got_it;
}
break;
- case LNBREAK:
- REXEC_FBC_CSCAN(is_LNBREAK_utf8_safe(s, strend),
- is_LNBREAK_latin1_safe(s, strend)
- );
+ case LNBREAK_t8_pb:
+ case LNBREAK_t8_p8:
+ REXEC_FBC_UTF8_CLASS_SCAN(is_LNBREAK_utf8_safe(s, strend));
+ break;
+
+ case LNBREAK_tb_pb:
+ case LNBREAK_tb_p8:
+ REXEC_FBC_NON_UTF8_CLASS_SCAN(is_LNBREAK_latin1_safe(s, strend));
break;
- /* The argument to all the POSIX node types is the class number to pass to
- * _generic_isCC() to build a mask for searching in PL_charclass[] */
+ /* The argument to all the POSIX node types is the class number to pass
+ * to _generic_isCC() to build a mask for searching in PL_charclass[] */
- case NPOSIXL:
+ case NPOSIXL_t8_pb:
+ case NPOSIXL_t8_p8:
to_complement = 1;
/* FALLTHROUGH */
- case POSIXL:
+ case POSIXL_t8_pb:
+ case POSIXL_t8_p8:
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
- REXEC_FBC_CSCAN(to_complement ^ cBOOL(isFOO_utf8_lc(FLAGS(c), (U8 *) s, (U8 *) strend)),
- to_complement ^ cBOOL(isFOO_lc(FLAGS(c), *s)));
+ REXEC_FBC_UTF8_CLASS_SCAN(
+ to_complement ^ cBOOL(isFOO_utf8_lc(FLAGS(c), (U8 *) s,
+ (U8 *) strend)));
break;
- case NPOSIXD:
+ case NPOSIXL_tb_pb:
+ case NPOSIXL_tb_p8:
to_complement = 1;
/* FALLTHROUGH */
- case POSIXD:
- if (utf8_target) {
- goto posix_utf8;
- }
- goto posixa;
-
- case NPOSIXA:
- if (utf8_target) {
- /* The complement of something that matches only ASCII matches all
- * non-ASCII, plus everything in ASCII that isn't in the class. */
- REXEC_FBC_CLASS_SCAN(1, ! isASCII_utf8_safe(s, strend)
- || ! _generic_isCC_A(*s, FLAGS(c)));
- break;
- }
+ case POSIXL_tb_pb:
+ case POSIXL_tb_p8:
+ _CHECK_AND_WARN_PROBLEMATIC_LOCALE;
+ REXEC_FBC_NON_UTF8_CLASS_SCAN(
+ to_complement ^ cBOOL(isFOO_lc(FLAGS(c), *s)));
+ break;
- to_complement = 1;
- goto posixa;
+ case NPOSIXA_t8_pb:
+ case NPOSIXA_t8_p8:
+ /* The complement of something that matches only ASCII matches all
+ * non-ASCII, plus everything in ASCII that isn't in the class. */
+ REXEC_FBC_UTF8_CLASS_SCAN( ! isASCII_utf8_safe(s, strend)
+ || ! _generic_isCC_A(*s, FLAGS(c)));
+ break;
- case POSIXA:
+ case POSIXA_t8_pb:
+ case POSIXA_t8_p8:
/* Don't need to worry about utf8, as it can match only a single
* byte invariant character. But we do anyway for performance reasons,
* as otherwise we would have to examine all the continuation
* characters */
- if (utf8_target) {
- REXEC_FBC_CLASS_SCAN(1, _generic_isCC_A(*s, FLAGS(c)));
- break;
- }
+ REXEC_FBC_UTF8_CLASS_SCAN(_generic_isCC_A(*s, FLAGS(c)));
+ break;
+
+ case NPOSIXD_tb_pb:
+ case NPOSIXD_tb_p8:
+ case NPOSIXA_tb_pb:
+ case NPOSIXA_tb_p8:
+ to_complement = 1;
+ /* FALLTHROUGH */
- posixa:
- REXEC_FBC_CLASS_SCAN(0, /* 0=>not-utf8 */
+ case POSIXD_tb_pb:
+ case POSIXD_tb_p8:
+ case POSIXA_tb_pb:
+ case POSIXA_tb_p8:
+ REXEC_FBC_NON_UTF8_CLASS_SCAN(
to_complement ^ cBOOL(_generic_isCC_A(*s, FLAGS(c))));
break;
- case NPOSIXU:
+ case NPOSIXU_tb_pb:
+ case NPOSIXU_tb_p8:
to_complement = 1;
/* FALLTHROUGH */
- case POSIXU:
- if (! utf8_target) {
- REXEC_FBC_CLASS_SCAN(0, /* 0=>not-utf8 */
+ case POSIXU_tb_pb:
+ case POSIXU_tb_p8:
+ REXEC_FBC_NON_UTF8_CLASS_SCAN(
to_complement ^ cBOOL(_generic_isCC(*s,
FLAGS(c))));
- }
- else {
+ break;
- posix_utf8:
- classnum = (_char_class_number) FLAGS(c);
- switch (classnum) {
- default:
- REXEC_FBC_CLASS_SCAN(1, /* 1=>is-utf8 */
+ case NPOSIXD_t8_pb:
+ case NPOSIXD_t8_p8:
+ case NPOSIXU_t8_pb:
+ case NPOSIXU_t8_p8:
+ to_complement = 1;
+ /* FALLTHROUGH */
+
+ case POSIXD_t8_pb:
+ case POSIXD_t8_p8:
+ case POSIXU_t8_pb:
+ case POSIXU_t8_p8:
+ classnum = (_char_class_number) FLAGS(c);
+ switch (classnum) {
+ default:
+ REXEC_FBC_UTF8_CLASS_SCAN(
to_complement ^ cBOOL(_invlist_contains_cp(
- PL_XPosix_ptrs[classnum],
- utf8_to_uvchr_buf((U8 *) s,
+ PL_XPosix_ptrs[classnum],
+ utf8_to_uvchr_buf((U8 *) s,
(U8 *) strend,
NULL))));
- break;
- case _CC_ENUM_SPACE:
- REXEC_FBC_CLASS_SCAN(1, /* 1=>is-utf8 */
+ break;
+
+ case _CC_ENUM_SPACE:
+ REXEC_FBC_UTF8_CLASS_SCAN(
to_complement ^ cBOOL(isSPACE_utf8_safe(s, strend)));
- break;
+ break;
- case _CC_ENUM_BLANK:
- REXEC_FBC_CLASS_SCAN(1,
+ case _CC_ENUM_BLANK:
+ REXEC_FBC_UTF8_CLASS_SCAN(
to_complement ^ cBOOL(isBLANK_utf8_safe(s, strend)));
- break;
+ break;
- case _CC_ENUM_XDIGIT:
- REXEC_FBC_CLASS_SCAN(1,
- to_complement ^ cBOOL(isXDIGIT_utf8_safe(s, strend)));
- break;
+ case _CC_ENUM_XDIGIT:
+ REXEC_FBC_UTF8_CLASS_SCAN(
+ to_complement ^ cBOOL(isXDIGIT_utf8_safe(s, strend)));
+ break;
- case _CC_ENUM_VERTSPACE:
- REXEC_FBC_CLASS_SCAN(1,
- to_complement ^ cBOOL(isVERTWS_utf8_safe(s, strend)));
- break;
+ case _CC_ENUM_VERTSPACE:
+ REXEC_FBC_UTF8_CLASS_SCAN(
+ to_complement ^ cBOOL(isVERTWS_utf8_safe(s, strend)));
+ break;
- case _CC_ENUM_CNTRL:
- REXEC_FBC_CLASS_SCAN(1,
+ case _CC_ENUM_CNTRL:
+ REXEC_FBC_UTF8_CLASS_SCAN(
to_complement ^ cBOOL(isCNTRL_utf8_safe(s, strend)));
- break;
- }
+ break;
}
break;
- case AHOCORASICKC:
- case AHOCORASICK:
+ case AHOCORASICKC_tb_pb:
+ case AHOCORASICKC_tb_p8:
+ case AHOCORASICKC_t8_pb:
+ case AHOCORASICKC_t8_p8:
+ case AHOCORASICK_tb_pb:
+ case AHOCORASICK_tb_p8:
+ case AHOCORASICK_t8_pb:
+ case AHOCORASICK_t8_p8:
{
DECL_TRIE_TYPE(c);
/* what trie are we using right now */
reg_ac_data *aho = (reg_ac_data*)progi->data->data[ ARG( c ) ];
- reg_trie_data *trie = (reg_trie_data*)progi->data->data[ aho->trie ];
+ reg_trie_data *trie = (reg_trie_data*)progi->data->data[aho->trie];
HV *widecharmap = MUTABLE_HV(progi->data->data[ aho->trie + 1 ]);
const char *last_start = strend - trie->minlen;
U8 *bitmap=NULL;
- GET_RE_DEBUG_FLAGS_DECL;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
/* We can't just allocate points here. We need to wrap it in
* an SV so it gets freed properly if there is a croak while
if( state==1 ) {
if ( bitmap ) {
DEBUG_TRIE_EXECUTE_r(
- if ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
- dump_exec_pos( (char *)uc, c, strend, real_start,
+ if ( uc <= (U8*)last_start
+ && !BITMAP_TEST(bitmap,*uc) )
+ {
+ dump_exec_pos( (char *)uc, c, strend,
+ real_start,
(char *)uc, utf8_target, 0 );
Perl_re_printf( aTHX_
" Scanning for legal start char...\n");
}
);
if (utf8_target) {
- while ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
+ while ( uc <= (U8*)last_start
+ && !BITMAP_TEST(bitmap,*uc) )
+ {
uc += UTF8SKIP(uc);
}
} else {
- while ( uc <= (U8*)last_start && !BITMAP_TEST(bitmap,*uc) ) {
+ while ( uc <= (U8*)last_start
+ && ! BITMAP_TEST(bitmap,*uc) )
+ {
uc++;
}
}
}
if ( word ) {
- U8 *lpos= points[ (pointpos - trie->wordinfo[word].len) % maxlen ];
+ U8 *lpos= points[ (pointpos - trie->wordinfo[word].len)
+ % maxlen ];
if (!leftmost || lpos < leftmost) {
DEBUG_r(accepted_word=word);
leftmost= lpos;
DEBUG_TRIE_EXECUTE_r({
if (failed)
- dump_exec_pos( (char *)uc, c, strend, real_start,
+ dump_exec_pos((char *)uc, c, strend, real_start,
s, utf8_target, 0 );
Perl_re_printf( aTHX_
"%sState: %4" UVxf ", word=%" UVxf,
}
}
if ( aho->states[ state ].wordnum ) {
- U8 *lpos = points[ (pointpos - trie->wordinfo[aho->states[ state ].wordnum].len) % maxlen ];
+ U8 *lpos = points[ (pointpos
+ - trie->wordinfo[aho->states[ state ]
+ .wordnum].len) % maxlen ];
if (!leftmost || lpos < leftmost) {
DEBUG_r(accepted_word=aho->states[ state ].wordnum);
leftmost = lpos;
if (leftmost) {
s = (char*)leftmost;
DEBUG_TRIE_EXECUTE_r({
- Perl_re_printf( aTHX_ "Matches word #%" UVxf " at position %" IVdf ". Trying full pattern...\n",
+ Perl_re_printf( aTHX_ "Matches word #%" UVxf
+ " at position %" IVdf ". Trying full"
+ " pattern...\n",
(UV)accepted_word, (IV)(s - real_start)
);
});
s = HOPc(s,1);
}
DEBUG_TRIE_EXECUTE_r({
- Perl_re_printf( aTHX_ "Pattern failed. Looking for new start point...\n");
+ Perl_re_printf( aTHX_
+ "Pattern failed. Looking for new start"
+ " point...\n");
});
} else {
DEBUG_TRIE_EXECUTE_r(
LEAVE;
}
break;
- default:
+
+ case EXACTFU_REQ8_t8_pb:
+ case EXACTFUP_tb_p8:
+ case EXACTFUP_t8_p8:
+ case EXACTF_tb_p8:
+ case EXACTF_t8_p8: /* This node only generated for non-utf8 patterns */
+ case EXACTFAA_NO_TRIE_tb_p8:
+ case EXACTFAA_NO_TRIE_t8_p8: /* This node only generated for non-utf8
+ patterns */
+ assert(0);
+
+ default:
Perl_croak(aTHX_ "panic: unknown regstclass %d", (int)OP(c));
- }
+ } /* End of switch on node type */
+
return 0;
+
got_it:
return s;
}
regmatch_info *const reginfo = ®info_buf;
regexp_paren_pair *swap = NULL;
I32 oldsave;
- GET_RE_DEBUG_FLAGS_DECL;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
PERL_ARGS_ASSERT_REGEXEC_FLAGS;
PERL_UNUSED_ARG(data);
if (!startpos ||
((flags & REXEC_FAIL_ON_UNDERFLOW) && startpos < stringarg))
{
- DEBUG_r(Perl_re_printf( aTHX_
+ DEBUG_GPOS_r(Perl_re_printf( aTHX_
"fail: ganch-gofs before earliest possible start\n"));
return 0;
}
minlen = prog->minlen;
if ((startpos + minlen) > strend || startpos < strbeg) {
- DEBUG_r(Perl_re_printf( aTHX_
- "Regex match can't succeed, so not even tried\n"));
+ DEBUG_EXECUTE_r(Perl_re_printf( aTHX_
+ "Regex match can't succeed, so not even tried\n"));
return 0;
}
to_utf8_substr(prog);
}
ch = SvPVX_const(prog->anchored_utf8)[0];
- REXEC_FBC_SCAN(0, /* 0=>not-utf8 */
+ REXEC_FBC_UTF8_SCAN(
if (*s == ch) {
DEBUG_EXECUTE_r( did_match = 1 );
if (regtry(reginfo, &s)) goto got_it;
}
}
ch = SvPVX_const(prog->anchored_substr)[0];
- REXEC_FBC_SCAN(0, /* 0=>not-utf8 */
+ REXEC_FBC_NON_UTF8_SCAN(
if (*s == ch) {
DEBUG_EXECUTE_r( did_match = 1 );
if (regtry(reginfo, &s)) goto got_it;
U32 depth = 0; /* used by REGCP_SET */
#endif
RXi_GET_DECL(prog,progi);
- GET_RE_DEBUG_FLAGS_DECL;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
PERL_ARGS_ASSERT_REGTRY;
UV c1 = (UV)CHRTEST_NOT_A_CP_1;
UV c2 = (UV)CHRTEST_NOT_A_CP_2;
bool use_chrtest_void = FALSE;
- const bool is_utf8_pat = reginfo->is_utf8_pat;
+ const bool utf8_pat = reginfo->is_utf8_pat;
/* Used when we have both utf8 input and utf8 output, to avoid converting
* to/from code points */
bool utf8_has_been_setup = FALSE;
- dVAR;
U8 *pat = (U8*)STRING(text_node);
U8 folded[UTF8_MAX_FOLD_CHAR_EXPAND * UTF8_MAXBYTES_CASE + 1] = { '\0' };
+ const U8 op = OP(text_node);
- if ( OP(text_node) == EXACT
- || OP(text_node) == EXACT_ONLY8
- || OP(text_node) == EXACTL)
- {
+ if (! isEXACTFish(OP(text_node))) {
/* In an exact node, only one thing can be matched, that first
* character. If both the pat and the target are UTF-8, we can just
* copy the input to the output, avoiding finding the code point of
* that character */
- if (!is_utf8_pat) {
- assert(OP(text_node) != EXACT_ONLY8);
+ if (! utf8_pat) {
+ assert(! isEXACT_REQ8(OP(text_node)));
c2 = c1 = *pat;
}
else if (utf8_target) {
Copy(pat, c2_utf8, UTF8SKIP(pat), U8);
utf8_has_been_setup = TRUE;
}
- else if (OP(text_node) == EXACT_ONLY8) {
+ else if (isEXACT_REQ8(OP(text_node))) {
return FALSE; /* Can only match UTF-8 target */
}
else {
}
}
else { /* an EXACTFish node */
- U8 *pat_end = pat + STR_LEN(text_node);
+ U8 *pat_end = pat + STR_LENs(text_node);
/* An EXACTFL node has at least some characters unfolded, because what
* they match is not known until now. So, now is the time to fold
* fold. But, in such a pattern only locale-problematic characters
* aren't folded, so we can skip this completely if the first character
* in the node isn't one of the tricky ones */
- if (OP(text_node) == EXACTFL) {
+ if (op == EXACTFL) {
- if (! is_utf8_pat) {
+ if (! utf8_pat) {
if (IN_UTF8_CTYPE_LOCALE && *pat == LATIN_SMALL_LETTER_SHARP_S)
{
folded[0] = folded[1] = 's';
}
}
- if ( ( is_utf8_pat && is_MULTI_CHAR_FOLD_utf8_safe(pat, pat_end))
- || (!is_utf8_pat && is_MULTI_CHAR_FOLD_latin1_safe(pat, pat_end)))
+ if ( ( utf8_pat && is_MULTI_CHAR_FOLD_utf8_safe(pat, pat_end))
+ || (!utf8_pat && is_MULTI_CHAR_FOLD_latin1_safe(pat, pat_end)))
{
/* Multi-character folds require more context to sort out. Also
* PL_utf8_foldclosures used below doesn't handle them, so have to
use_chrtest_void = TRUE;
}
else { /* an EXACTFish node which doesn't begin with a multi-char fold */
- c1 = is_utf8_pat ? valid_utf8_to_uvchr(pat, NULL) : *pat;
+ c1 = utf8_pat ? valid_utf8_to_uvchr(pat, NULL) : *pat;
if ( UNLIKELY(PL_in_utf8_turkic_locale)
- && OP(text_node) == EXACTFL
+ && op == EXACTFL
&& UNLIKELY( c1 == 'i' || c1 == 'I'
|| c1 == LATIN_CAPITAL_LETTER_I_WITH_DOT_ABOVE
|| c1 == LATIN_SMALL_LETTER_DOTLESS_I))
}
}
else if (c1 > 255) {
- const unsigned int * remaining_folds;
- unsigned int first_fold;
+ const U32 * remaining_folds;
+ U32 first_fold;
/* Look up what code points (besides c1) fold to c1; e.g.,
* [ 'K', KELVIN_SIGN ] both fold to 'k'. */
* circumstances. If it isn't, it means the only legal
* match of c1 is itself. */
if ( c2 < 256
- && ( ( OP(text_node) == EXACTFL
+ && ( ( op == EXACTFL
&& ! IN_UTF8_CTYPE_LOCALE)
- || (( OP(text_node) == EXACTFAA
- || OP(text_node) == EXACTFAA_NO_TRIE)
+ || (( op == EXACTFAA
+ || op == EXACTFAA_NO_TRIE)
&& (isASCII(c1) || isASCII(c2)))))
{
c2 = c1;
else /* Here, c1 is <= 255 */
if ( utf8_target
&& HAS_NONLATIN1_FOLD_CLOSURE(c1)
- && ( ! (OP(text_node) == EXACTFL && ! IN_UTF8_CTYPE_LOCALE))
- && ( ( OP(text_node) != EXACTFAA
- && OP(text_node) != EXACTFAA_NO_TRIE)
+ && ( ! (op == EXACTFL && ! IN_UTF8_CTYPE_LOCALE))
+ && ( ( op != EXACTFAA
+ && op != EXACTFAA_NO_TRIE)
|| ! isASCII(c1)))
{
/* Here, there could be something above Latin1 in the target
}
else { /* Here nothing above Latin1 can fold to the pattern
character */
- switch (OP(text_node)) {
+ switch (op) {
case EXACTFL: /* /l rules */
c2 = PL_fold_locale[c1];
case EXACTF: /* This node only generated for non-utf8
patterns */
- assert(! is_utf8_pat);
+ assert(! utf8_pat);
if (! utf8_target) { /* /d rules */
c2 = PL_fold[c1];
break;
* EXACTFAA as nothing in Latin1 folds to ASCII */
case EXACTFAA_NO_TRIE: /* This node only generated for
non-utf8 patterns */
- assert(! is_utf8_pat);
+ assert(! utf8_pat);
/* FALLTHROUGH */
case EXACTFAA:
case EXACTFUP:
case EXACTFU:
c2 = PL_fold_latin1[c1];
break;
- case EXACTFU_ONLY8:
+ case EXACTFU_REQ8:
return FALSE;
NOT_REACHED; /* NOTREACHED */
default:
- Perl_croak(aTHX_ "panic: Unexpected op %u", OP(text_node));
+ Perl_croak(aTHX_ "panic: Unexpected op %u", op);
NOT_REACHED; /* NOTREACHED */
}
}
S_isGCB(pTHX_ const GCB_enum before, const GCB_enum after, const U8 * const strbeg, const U8 * const curpos, const bool utf8_target)
{
/* returns a boolean indicating if there is a Grapheme Cluster Boundary
- * between the inputs. See http://www.unicode.org/reports/tr29/. */
+ * between the inputs. See https://www.unicode.org/reports/tr29/. */
PERL_ARGS_ASSERT_ISGCB;
}
while (prev == GCB_Extend);
- return prev != GCB_XPG_XX;
+ return prev != GCB_ExtPict_XX;
}
default:
STATIC GCB_enum
S_backup_one_GCB(pTHX_ const U8 * const strbeg, U8 ** curpos, const bool utf8_target)
{
- dVAR;
GCB_enum gcb;
PERL_ARGS_ASSERT_BACKUP_ONE_GCB;
STATIC LB_enum
S_advance_one_LB(pTHX_ U8 ** curpos, const U8 * const strend, const bool utf8_target)
{
- dVAR;
LB_enum lb;
STATIC LB_enum
S_backup_one_LB(pTHX_ const U8 * const strbeg, U8 ** curpos, const bool utf8_target)
{
- dVAR;
LB_enum lb;
PERL_ARGS_ASSERT_BACKUP_ONE_LB;
const bool utf8_target)
{
/* returns a boolean indicating if there is a Sentence Boundary Break
- * between the inputs. See http://www.unicode.org/reports/tr29/ */
+ * between the inputs. See https://www.unicode.org/reports/tr29/ */
U8 * lpos = (U8 *) curpos;
bool has_para_sep = FALSE;
STATIC SB_enum
S_advance_one_SB(pTHX_ U8 ** curpos, const U8 * const strend, const bool utf8_target)
{
- dVAR;
SB_enum sb;
PERL_ARGS_ASSERT_ADVANCE_ONE_SB;
STATIC SB_enum
S_backup_one_SB(pTHX_ const U8 * const strbeg, U8 ** curpos, const bool utf8_target)
{
- dVAR;
SB_enum sb;
PERL_ARGS_ASSERT_BACKUP_ONE_SB;
const bool utf8_target,
const bool skip_Extend_Format)
{
- dVAR;
WB_enum wb;
PERL_ARGS_ASSERT_ADVANCE_ONE_WB;
STATIC WB_enum
S_backup_one_WB(pTHX_ WB_enum * previous, const U8 * const strbeg, U8 ** curpos, const bool utf8_target)
{
- dVAR;
WB_enum wb;
PERL_ARGS_ASSERT_BACKUP_ONE_WB;
/* Macros for regmatch(), using its internal variables */
#define NEXTCHR_EOS -10 /* nextchr has fallen off the end */
-#define NEXTCHR_IS_EOS (nextchr < 0)
+#define NEXTCHR_IS_EOS (nextbyte < 0)
#define SET_nextchr \
- nextchr = ((locinput < reginfo->strend) ? UCHARAT(locinput) : NEXTCHR_EOS)
+ nextbyte = ((locinput < reginfo->strend) ? UCHARAT(locinput) : NEXTCHR_EOS)
#define SET_locinput(p) \
locinput = (p); \
rest of the pattern. Variable and state names reflect this convention.
The states in the main switch are the union of ops and failure/success of
-substates associated with with that op. For example, IFMATCH is the op
+substates associated with that op. For example, IFMATCH is the op
that does lookahead assertions /(?=A)B/ and so the IFMATCH state means
'execute IFMATCH'; while IFMATCH_A is a state saying that we have just
successfully matched A and IFMATCH_A_fail is a state saying that we have
STATIC SSize_t
S_regmatch(pTHX_ regmatch_info *reginfo, char *startpos, regnode *prog)
{
- dVAR;
const bool utf8_target = reginfo->is_utf8_target;
const U32 uniflags = UTF8_ALLOW_DEFAULT;
REGEXP *rex_sv = reginfo->prog;
char *pushinput; /* where to continue after a PUSH */
char *pusheol; /* where to stop matching (loceol) after a PUSH */
U8 *pushsr0; /* save starting pos of script run */
- I32 nextchr; /* is always set to UCHARAT(locinput), or -1 at EOS */
+ PERL_INT_FAST16_T nextbyte; /* is always set to UCHARAT(locinput), or -1
+ at EOS */
bool result = 0; /* return value of S_regmatch */
U32 depth = 0; /* depth of backtrack stack */
#endif
#ifdef DEBUGGING
- GET_RE_DEBUG_FLAGS_DECL;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
#endif
/* protect against undef(*^R) */
st = PL_regmatch_state;
- /* Note that nextchr is a byte even in UTF */
+ /* Note that nextbyte is a byte even in UTF */
SET_nextchr;
scan = prog;
to_complement = 0;
SET_nextchr;
- assert(nextchr < 256 && (nextchr >= 0 || nextchr == NEXTCHR_EOS));
+ assert(nextbyte < 256 && (nextbyte >= 0 || nextbyte == NEXTCHR_EOS));
switch (state_num) {
case SBOL: /* /^../ and /\A../ */
NOT_REACHED; /* NOTREACHED */
case MEOL: /* /..$/m */
- if (!NEXTCHR_IS_EOS && nextchr != '\n')
+ if (!NEXTCHR_IS_EOS && nextbyte != '\n')
sayNO;
break;
case SEOL: /* /..$/ */
- if (!NEXTCHR_IS_EOS && nextchr != '\n')
+ if (!NEXTCHR_IS_EOS && nextbyte != '\n')
sayNO;
if (reginfo->strend - locinput > 1)
sayNO;
case REG_ANY: /* /./ */
if ( NEXTCHR_IS_EOS
|| locinput >= loceol
- || nextchr == '\n')
+ || nextbyte == '\n')
{
sayNO;
}
*/
if ( ! NEXTCHR_IS_EOS
&& locinput < loceol
- && ! ANYOF_BITMAP_TEST(scan, nextchr))
+ && ! ANYOF_BITMAP_TEST(scan, nextbyte))
{
DEBUG_EXECUTE_r(
Perl_re_exec_indentf( aTHX_ "%sTRIE: failed to match trie start class...%s\n",
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
if (utf8_target
&& ! NEXTCHR_IS_EOS
- && UTF8_IS_ABOVE_LATIN1(nextchr)
+ && UTF8_IS_ABOVE_LATIN1(nextbyte)
&& scan->flags == EXACTL)
{
/* We only output for EXACTL, as we let the folder
if ( trie->bitmap
&& ( NEXTCHR_IS_EOS
|| locinput >= loceol
- || ! TRIE_BITMAP_TEST(trie, nextchr)))
+ || ! TRIE_BITMAP_TEST(trie, nextbyte)))
{
if (trie->states[ state ].wordnum) {
DEBUG_EXECUTE_r(
}
#undef ST
+ case LEXACT_REQ8:
+ if (! utf8_target) {
+ sayNO;
+ }
+ /* FALLTHROUGH */
+
+ case LEXACT:
+ {
+ char *s;
+
+ s = STRINGl(scan);
+ ln = STR_LENl(scan);
+ goto join_short_long_exact;
+
case EXACTL: /* /abc/l */
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
_CHECK_AND_OUTPUT_WIDE_LOCALE_UTF8_MSG(locinput, reginfo->strend);
}
goto do_exact;
- case EXACT_ONLY8:
+ case EXACT_REQ8:
if (! utf8_target) {
sayNO;
}
/* FALLTHROUGH */
- case EXACT: { /* /abc/ */
- char *s;
+
+ case EXACT: /* /abc/ */
do_exact:
- s = STRING(scan);
- ln = STR_LEN(scan);
+ s = STRINGs(scan);
+ ln = STR_LENs(scan);
+
+ join_short_long_exact:
if (utf8_target != is_utf8_pat) {
/* The target and the pattern have differing utf8ness. */
char *l = locinput;
/* The target and the pattern have the same utf8ness. */
/* Inline the first character, for speed. */
if ( loceol - locinput < ln
- || UCHARAT(s) != nextchr
+ || UCHARAT(s) != nextbyte
|| (ln > 1 && memNE(s, locinput, ln)))
{
sayNO;
fold_array = PL_fold_latin1;
goto do_exactf;
- case EXACTFU_ONLY8: /* /abc/iu with something in /abc/ > 255 */
+ case EXACTFU_REQ8: /* /abc/iu with something in /abc/ > 255 */
if (! utf8_target) {
sayNO;
}
fold_utf8_flags = 0;
do_exactf:
- s = STRING(scan);
- ln = STR_LEN(scan);
+ s = STRINGs(scan);
+ ln = STR_LENs(scan);
if ( utf8_target
|| is_utf8_pat
}
/* Neither the target nor the pattern are utf8 */
- if (UCHARAT(s) != nextchr
+ if (UCHARAT(s) != nextbyte
&& !NEXTCHR_IS_EOS
- && UCHARAT(s) != fold_array[nextchr])
+ && UCHARAT(s) != fold_array[nextbyte])
{
sayNO;
}
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
if (FLAGS(scan) != TRADITIONAL_BOUND) {
- if (! IN_UTF8_CTYPE_LOCALE) {
- Perl_ck_warner(aTHX_ packWARN(WARN_LOCALE),
- B_ON_NON_UTF8_LOCALE_IS_WRONG);
- }
+ CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_BOUND;
goto boundu;
}
if (locinput == reginfo->strbeg)
b1 = isWORDCHAR_LC('\n');
else {
- b1 = isWORDCHAR_LC_utf8_safe(reghop3((U8*)locinput, -1,
- (U8*)(reginfo->strbeg)),
- (U8*)(reginfo->strend));
+ U8 *p = reghop3((U8*)locinput, -1,
+ (U8*)(reginfo->strbeg));
+ b1 = isWORDCHAR_LC_utf8_safe(p, (U8*)(reginfo->strend));
}
b2 = (NEXTCHR_IS_EOS)
? isWORDCHAR_LC('\n')
: isWORDCHAR_LC(UCHARAT(locinput - 1));
b2 = (NEXTCHR_IS_EOS)
? isWORDCHAR_LC('\n')
- : isWORDCHAR_LC(nextchr);
+ : isWORDCHAR_LC(nextbyte);
}
if (to_complement ^ (b1 == b2)) {
sayNO;
: isWORDCHAR_A(UCHARAT(locinput - 1));
b2 = (NEXTCHR_IS_EOS)
? isWORDCHAR_A('\n')
- : isWORDCHAR_A(nextchr);
+ : isWORDCHAR_A(nextbyte);
if (to_complement ^ (b1 == b2)) {
sayNO;
}
case TRADITIONAL_BOUND:
{
bool b1, b2;
- b1 = (locinput == reginfo->strbeg)
- ? 0 /* isWORDCHAR_L1('\n') */
- : isWORDCHAR_utf8_safe(
- reghop3((U8*)locinput,
- -1,
- (U8*)(reginfo->strbeg)),
- (U8*) reginfo->strend);
+ if (locinput == reginfo->strbeg) {
+ b1 = 0 /* isWORDCHAR_L1('\n') */;
+ }
+ else {
+ U8 *p = reghop3((U8*)locinput, -1,
+ (U8*)(reginfo->strbeg));
+
+ b1 = isWORDCHAR_utf8_safe(p, (U8*) reginfo->strend);
+ }
b2 = (NEXTCHR_IS_EOS)
? 0 /* isWORDCHAR_L1('\n') */
: isWORDCHAR_utf8_safe((U8*)locinput,
: isWORDCHAR_L1(UCHARAT(locinput - 1));
b2 = (NEXTCHR_IS_EOS)
? 0 /* isWORDCHAR_L1('\n') */
- : isWORDCHAR_L1(nextchr);
+ : isWORDCHAR_L1(nextbyte);
match = cBOOL(b1 != b2);
break;
}
case ANYOFPOSIXL:
case ANYOFL: /* /[abc]/l */
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
+ CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_SETS(scan);
- if (ANYOFL_UTF8_LOCALE_REQD(FLAGS(scan)) && ! IN_UTF8_CTYPE_LOCALE)
- {
- Perl_ck_warner(aTHX_ packWARN(WARN_LOCALE), utf8_locale_required);
- }
/* FALLTHROUGH */
case ANYOFD: /* /[abc]/d */
case ANYOF: /* /[abc]/ */
case ANYOFH:
if ( ! utf8_target
|| NEXTCHR_IS_EOS
- || ANYOF_FLAGS(scan) > NATIVE_UTF8_TO_I8((U8) *locinput)
+ || ANYOF_FLAGS(scan) > NATIVE_UTF8_TO_I8(*locinput)
|| ! reginclass(rex, scan, (U8*)locinput, (U8*) loceol,
utf8_target))
{
goto increment_locinput;
break;
+ case ANYOFHs:
+ if ( ! utf8_target
+ || NEXTCHR_IS_EOS
+ || loceol - locinput < FLAGS(scan)
+ || memNE(locinput, ((struct regnode_anyofhs *) scan)->string, FLAGS(scan))
+ || ! reginclass(rex, scan, (U8*)locinput, (U8*) loceol,
+ utf8_target))
+ {
+ sayNO;
+ }
+ goto increment_locinput;
+ break;
+
+ case ANYOFR:
+ if (NEXTCHR_IS_EOS) {
+ sayNO;
+ }
+
+ if (utf8_target) {
+ if ( ANYOF_FLAGS(scan) > NATIVE_UTF8_TO_I8(*locinput)
+ || ! withinCOUNT(utf8_to_uvchr_buf((U8 *) locinput,
+ (U8 *) reginfo->strend,
+ NULL),
+ ANYOFRbase(scan), ANYOFRdelta(scan)))
+ {
+ sayNO;
+ }
+ }
+ else {
+ if (! withinCOUNT((U8) *locinput,
+ ANYOFRbase(scan), ANYOFRdelta(scan)))
+ {
+ sayNO;
+ }
+ }
+ goto increment_locinput;
+ break;
+
+ case ANYOFRb:
+ if (NEXTCHR_IS_EOS) {
+ sayNO;
+ }
+
+ if (utf8_target) {
+ if ( ANYOF_FLAGS(scan) != (U8) *locinput
+ || ! withinCOUNT(utf8_to_uvchr_buf((U8 *) locinput,
+ (U8 *) reginfo->strend,
+ NULL),
+ ANYOFRbase(scan), ANYOFRdelta(scan)))
+ {
+ sayNO;
+ }
+ }
+ else {
+ if (! withinCOUNT((U8) *locinput,
+ ANYOFRbase(scan), ANYOFRdelta(scan)))
+ {
+ sayNO;
+ }
+ }
+ goto increment_locinput;
+ break;
+
/* The argument (FLAGS) to all the POSIX node types is the class number
* */
/* Use isFOO_lc() for characters within Latin1. (Note that
* UTF8_IS_INVARIANT works even on non-UTF-8 strings, or else
* wouldn't be invariant) */
- if (UTF8_IS_INVARIANT(nextchr) || ! utf8_target) {
- if (! (to_complement ^ cBOOL(isFOO_lc(FLAGS(scan), (U8) nextchr)))) {
+ if (UTF8_IS_INVARIANT(nextbyte) || ! utf8_target) {
+ if (! (to_complement ^ cBOOL(isFOO_lc(FLAGS(scan), (U8) nextbyte)))) {
sayNO;
}
/* Here is a UTF-8 variant code point below 256 and the target is
* UTF-8 */
if (! (to_complement ^ cBOOL(isFOO_lc(FLAGS(scan),
- EIGHT_BIT_UTF8_TO_NATIVE(nextchr,
+ EIGHT_BIT_UTF8_TO_NATIVE(nextbyte,
*(locinput + 1))))))
{
sayNO;
}
/* All UTF-8 variants match */
- if (! UTF8_IS_INVARIANT(nextchr)) {
+ if (! UTF8_IS_INVARIANT(nextbyte)) {
goto increment_locinput;
}
join_nposixa:
- if (! (to_complement ^ cBOOL(_generic_isCC_A(nextchr,
+ if (! (to_complement ^ cBOOL(_generic_isCC_A(nextbyte,
FLAGS(scan)))))
{
sayNO;
/* Use _generic_isCC() for characters within Latin1. (Note that
* UTF8_IS_INVARIANT works even on non-UTF-8 strings, or else
* wouldn't be invariant) */
- if (UTF8_IS_INVARIANT(nextchr) || ! utf8_target) {
- if (! (to_complement ^ cBOOL(_generic_isCC(nextchr,
+ if (UTF8_IS_INVARIANT(nextbyte) || ! utf8_target) {
+ if (! (to_complement ^ cBOOL(_generic_isCC(nextbyte,
FLAGS(scan)))))
{
sayNO;
}
else if (UTF8_IS_NEXT_CHAR_DOWNGRADEABLE(locinput, reginfo->strend)) {
if (! (to_complement
- ^ cBOOL(_generic_isCC(EIGHT_BIT_UTF8_TO_NATIVE(nextchr,
+ ^ cBOOL(_generic_isCC(EIGHT_BIT_UTF8_TO_NATIVE(nextbyte,
*(locinput + 1)),
FLAGS(scan)))))
{
/* Match either CR LF or '.', as all the other possibilities
* require utf8 */
locinput++; /* Match the . or CR */
- if (nextchr == '\r' /* And if it was CR, and the next is LF,
+ if (nextbyte == '\r' /* And if it was CR, and the next is LF,
match the LF */
&& locinput < loceol
&& UCHARAT(locinput) == '\n')
/* Not utf8: Inline the first character, for speed. */
if ( ! NEXTCHR_IS_EOS
&& locinput < loceol
- && UCHARAT(s) != nextchr
+ && UCHARAT(s) != nextbyte
&& ( type == REF
- || UCHARAT(s) != fold_array[nextchr]))
+ || UCHARAT(s) != fold_array[nextbyte]))
{
sayNO;
}
rex->recurse_locinput[arg]= locinput;
DEBUG_r({
- GET_RE_DEBUG_FLAGS_DECL;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
DEBUG_STACK_r({
Perl_re_exec_indentf( aTHX_
"entering GOSUB, prev_recurse_locinput=%p recurse_locinput[%d]=%p\n",
/* NOTREACHED */
case EVAL: /* /(?{...})B/ /(??{A})B/ and /(?(?{...})X|Y)B/ */
- if (cur_eval && cur_eval->locinput==locinput) {
+ if (logical == 2 && cur_eval && cur_eval->locinput==locinput) {
if ( ++nochange_depth > max_nochange_depth )
Perl_croak(aTHX_ "EVAL without pos change exceeded limit in regex");
} else {
depth, (IV)ST.count)
);
if (! NEXTCHR_IS_EOS && ST.c1 != CHRTEST_VOID) {
- if (! UTF8_IS_INVARIANT(nextchr) && utf8_target) {
+ if (! UTF8_IS_INVARIANT(nextbyte) && utf8_target) {
/* (We can use memEQ and memNE in this file without
* having to worry about one being shorter than the
goto reenter_switch;
}
}
- else if (nextchr != ST.c1 && nextchr != ST.c2) {
+ else if (nextbyte != ST.c1 && nextbyte != ST.c2) {
/* simulate B failing */
DEBUG_OPTIMISE_r(
Perl_re_exec_indentf( aTHX_ "CURLYM Fast bail next target=0x%X c1=0x%X c2=0x%X\n",
depth,
- (int) nextchr, ST.c1, ST.c2)
+ (int) nextbyte, ST.c1, ST.c2)
);
state_num = CURLYM_B_fail;
goto reenter_switch;
increment_locinput:
assert(!NEXTCHR_IS_EOS);
if (utf8_target) {
- locinput += PL_utf8skip[nextchr];
+ locinput += PL_utf8skip[nextbyte];
/* locinput is allowed to go 1 char off the end (signifying
* EOS), but not 2+ */
if (locinput > loceol)
/* push a new regex state, then continue at scan */
{
regmatch_state *newst;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
- DEBUG_STACK_r({
+ DEBUG_r( /* DEBUG_STACK_r */
+ if (DEBUG_v_TEST || RE_DEBUG_FLAG(RE_DEBUG_EXTRA_STACK)) {
regmatch_state *cur = st;
regmatch_state *curyes = yes_state;
U32 i;
if (curyes == cur)
curyes = cur->u.yes.prev_yes_state;
}
- } else
+ } else {
DEBUG_STATE_pp("push")
- );
+ });
depth++;
st->locinput = locinput;
st->loceol = loceol;
S_regrepeat(pTHX_ regexp *prog, char **startposp, const regnode *p,
char * loceol, regmatch_info *const reginfo, I32 max _pDEPTH)
{
- dVAR;
char *scan; /* Pointer to current position in target string */
I32 c;
char *this_eol = loceol; /* potentially adjusted version. */
else
scan = this_eol;
break;
+
+ case LEXACT_REQ8:
+ if (! utf8_target) {
+ break;
+ }
+ /* FALLTHROUGH */
+
+ case LEXACT:
+ {
+ U8 * string;
+ Size_t str_len;
+
+ string = (U8 *) STRINGl(p);
+ str_len = STR_LENl(p);
+ goto join_short_long_exact;
+
case EXACTL:
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
if (utf8_target && UTF8_IS_ABOVE_LATIN1(*scan)) {
}
goto do_exact;
- case EXACT_ONLY8:
+ case EXACT_REQ8:
if (! utf8_target) {
break;
}
/* FALLTHROUGH */
case EXACT:
do_exact:
- assert(STR_LEN(p) == reginfo->is_utf8_pat ? UTF8SKIP(STRING(p)) : 1);
+ string = (U8 *) STRINGs(p);
+ str_len = STR_LENs(p);
+
+ join_short_long_exact:
+ assert(str_len == reginfo->is_utf8_pat ? UTF8SKIP(string) : 1);
- c = (U8)*STRING(p);
+ c = *string;
/* Can use a simple find if the pattern char to match on is invariant
* under UTF-8, or both target and pattern aren't UTF-8. Note that we
* string EQ */
while (hardcount < max
&& scan < this_eol
- && (scan_char_len = UTF8SKIP(scan)) <= STR_LEN(p)
- && memEQ(scan, STRING(p), scan_char_len))
+ && (scan_char_len = UTF8SKIP(scan)) <= str_len
+ && memEQ(scan, string, scan_char_len))
{
scan += scan_char_len;
hardcount++;
/* Target isn't utf8; convert the character in the UTF-8
* pattern to non-UTF8, and do a simple find */
- c = EIGHT_BIT_UTF8_TO_NATIVE(c, *(STRING(p) + 1));
+ c = EIGHT_BIT_UTF8_TO_NATIVE(c, *(string + 1));
scan = (char *) find_span_end((U8 *) scan, (U8 *) this_eol, (U8) c);
} /* else pattern char is above Latin1, can't possibly match the
non-UTF-8 target */
}
}
break;
+ }
case EXACTFAA_NO_TRIE: /* This node only generated for non-utf8 patterns */
assert(! reginfo->is_utf8_pat);
| FOLDEQ_S2_FOLDS_SANE;
goto do_exactf;
- case EXACTFU_ONLY8:
+ case EXACTFU_REQ8:
if (! utf8_target) {
break;
}
int c1, c2;
U8 c1_utf8[UTF8_MAXBYTES+1], c2_utf8[UTF8_MAXBYTES+1];
- assert(STR_LEN(p) == reginfo->is_utf8_pat ? UTF8SKIP(STRING(p)) : 1);
+ assert(STR_LENs(p) == reginfo->is_utf8_pat ? UTF8SKIP(STRINGs(p)) : 1);
if (S_setup_EXACTISH_ST_c1_c2(aTHX_ p, &c1, c1_utf8, &c2, c2_utf8,
reginfo))
if (c1 == CHRTEST_VOID) {
/* Use full Unicode fold matching */
char *tmpeol = loceol;
- STRLEN pat_len = reginfo->is_utf8_pat ? UTF8SKIP(STRING(p)) : 1;
+ STRLEN pat_len = reginfo->is_utf8_pat ? UTF8SKIP(STRINGs(p)) : 1;
while (hardcount < max
&& foldEQ_utf8_flags(scan, &tmpeol, 0, utf8_target,
- STRING(p), NULL, pat_len,
+ STRINGs(p), NULL, pat_len,
reginfo->is_utf8_pat, utf8_flags))
{
scan = tmpeol;
case ANYOFPOSIXL:
case ANYOFL:
_CHECK_AND_WARN_PROBLEMATIC_LOCALE;
+ CHECK_AND_WARN_NON_UTF8_CTYPE_LOCALE_IN_SETS(p);
- if (ANYOFL_UTF8_LOCALE_REQD(FLAGS(p)) && ! IN_UTF8_CTYPE_LOCALE) {
- Perl_ck_warner(aTHX_ packWARN(WARN_LOCALE), utf8_locale_required);
- }
/* FALLTHROUGH */
case ANYOFD:
case ANYOF:
if (utf8_target) { /* ANYOFH only can match UTF-8 targets */
while ( hardcount < max
&& scan < this_eol
- && NATIVE_UTF8_TO_I8((U8) *scan) >= ANYOF_FLAGS(p)
+ && NATIVE_UTF8_TO_I8(*scan) >= ANYOF_FLAGS(p)
&& reginclass(prog, p, (U8*)scan, (U8*) this_eol, TRUE))
{
scan += UTF8SKIP(scan);
if (utf8_target) { /* ANYOFH only can match UTF-8 targets */
while ( hardcount < max
&& scan < this_eol
- && inRANGE((U8) NATIVE_UTF8_TO_I8(*scan),
+ && inRANGE(NATIVE_UTF8_TO_I8(*scan),
LOWEST_ANYOF_HRx_BYTE(ANYOF_FLAGS(p)),
HIGHEST_ANYOF_HRx_BYTE(ANYOF_FLAGS(p)))
- && NATIVE_UTF8_TO_I8((U8) *scan) >= ANYOF_FLAGS(p)
+ && NATIVE_UTF8_TO_I8(*scan) >= ANYOF_FLAGS(p)
+ && reginclass(prog, p, (U8*)scan, (U8*) this_eol, TRUE))
+ {
+ scan += UTF8SKIP(scan);
+ hardcount++;
+ }
+ }
+ break;
+
+ case ANYOFHs:
+ if (utf8_target) { /* ANYOFH only can match UTF-8 targets */
+ while ( hardcount < max
+ && scan + FLAGS(p) < this_eol
+ && memEQ(scan, ((struct regnode_anyofhs *) p)->string, FLAGS(p))
&& reginclass(prog, p, (U8*)scan, (U8*) this_eol, TRUE))
{
scan += UTF8SKIP(scan);
}
break;
+ case ANYOFR:
+ if (utf8_target) {
+ while ( hardcount < max
+ && scan < this_eol
+ && NATIVE_UTF8_TO_I8(*scan) >= ANYOF_FLAGS(p)
+ && withinCOUNT(utf8_to_uvchr_buf((U8 *) scan,
+ (U8 *) this_eol,
+ NULL),
+ ANYOFRbase(p), ANYOFRdelta(p)))
+ {
+ scan += UTF8SKIP(scan);
+ hardcount++;
+ }
+ }
+ else {
+ while ( hardcount < max
+ && scan < this_eol
+ && withinCOUNT((U8) *scan, ANYOFRbase(p), ANYOFRdelta(p)))
+ {
+ scan++;
+ hardcount++;
+ }
+ }
+ break;
+
+ case ANYOFRb:
+ if (utf8_target) {
+ while ( hardcount < max
+ && scan < this_eol
+ && (U8) *scan == ANYOF_FLAGS(p)
+ && withinCOUNT(utf8_to_uvchr_buf((U8 *) scan,
+ (U8 *) this_eol,
+ NULL),
+ ANYOFRbase(p), ANYOFRdelta(p)))
+ {
+ scan += UTF8SKIP(scan);
+ hardcount++;
+ }
+ }
+ else {
+ while ( hardcount < max
+ && scan < this_eol
+ && withinCOUNT((U8) *scan, ANYOFRbase(p), ANYOFRdelta(p)))
+ {
+ scan++;
+ hardcount++;
+ }
+ }
+ break;
+
/* The argument (FLAGS) to all the POSIX node types is the class number */
case NPOSIXL:
}
break;
- case BOUNDL:
- case NBOUNDL:
- _CHECK_AND_WARN_PROBLEMATIC_LOCALE;
- /* FALLTHROUGH */
- case BOUND:
- case BOUNDA:
- case BOUNDU:
- case EOS:
- case GPOS:
- case KEEPS:
- case NBOUND:
- case NBOUNDA:
- case NBOUNDU:
- case OPFAIL:
- case SBOL:
- case SEOL:
- /* These are all 0 width, so match right here or not at all. */
- break;
-
default:
Perl_croak(aTHX_ "panic: regrepeat() called with unrecognized node type %d='%s'", OP(p), PL_reg_name[OP(p)]);
NOT_REACHED; /* NOTREACHED */
*startposp = scan;
DEBUG_r({
- GET_RE_DEBUG_FLAGS_DECL;
+ DECLARE_AND_GET_RE_DEBUG_FLAGS;
DEBUG_EXECUTE_r({
SV * const prop = sv_newmortal();
regprop(prog, prop, p, reginfo, NULL);
STATIC bool
S_reginclass(pTHX_ regexp * const prog, const regnode * const n, const U8* const p, const U8* const p_end, const bool utf8_target)
{
- dVAR;
- const char flags = (inRANGE(OP(n), ANYOFH, ANYOFHr))
+ const char flags = (inRANGE(OP(n), ANYOFH, ANYOFHs))
? 0
: ANYOF_FLAGS(n);
bool match = FALSE;
}
else if (flags & ANYOF_LOCALE_FLAGS) {
if ( (flags & ANYOFL_FOLD)
- && c < sizeof(PL_fold_locale)
+ && c < 256
&& ANYOF_BITMAP_TEST(n, PL_fold_locale[c]))
{
match = TRUE;
&& IN_UTF8_CTYPE_LOCALE)))
{
SV* only_utf8_locale = NULL;
- SV * const definition = _get_regclass_nonbitmap_data(prog, n, TRUE,
- 0, &only_utf8_locale, NULL);
+ SV * const definition =
+#if !defined(PERL_IN_XSUB_RE) || defined(PLUGGABLE_RE_EXTENSION)
+ get_regclass_nonbitmap_data(prog, n, TRUE, 0,
+ &only_utf8_locale, NULL);
+#else
+ get_re_gclass_nonbitmap_data(prog, n, TRUE, 0,
+ &only_utf8_locale, NULL);
+#endif
if (definition) {
U8 utf8_buffer[2];
U8 * utf8_p;
/* this regexp is also owned by the new PL_reg_curpm, which
will try to free it. */
av_push(PL_regex_padav, repointer);
- PL_reg_curpm->op_pmoffset = av_tindex(PL_regex_padav);
+ PL_reg_curpm->op_pmoffset = av_top_index(PL_regex_padav);
PL_regex_pad = AvARRAY(PL_regex_padav);
}
#endif
#ifndef PERL_IN_XSUB_RE
bool
-Perl__is_grapheme(pTHX_ const U8 * strbeg, const U8 * s, const U8 * strend, const UV cp)
+Perl_is_grapheme(pTHX_ const U8 * strbeg, const U8 * s, const U8 * strend, const UV cp)
{
/* Temporary helper function for toke.c. Verify that the code point 'cp'
* is a stand-alone grapheme. The UTF-8 for 'cp' begins at position 's' in
* the larger string bounded by 'strbeg' and 'strend'.
*
- * 'cp' needs to be assigned (if not a future version of the Unicode
+ * 'cp' needs to be assigned (if not, a future version of the Unicode
* Standard could make it something that combines with adjacent characters,
* so code using it would then break), and there has to be a GCB break
* before and after the character. */
- dVAR;
GCB_enum cp_gcb_val, prev_cp_gcb_val, next_cp_gcb_val;
const U8 * prev_cp_start;
- PERL_ARGS_ASSERT__IS_GRAPHEME;
+ PERL_ARGS_ASSERT_IS_GRAPHEME;
if ( UNLIKELY(UNICODE_IS_SUPER(cp))
|| UNLIKELY(UNICODE_IS_NONCHAR(cp)))
}
/*
-=head1 Unicode Support
+=for apidoc_section $unicode
=for apidoc isSCRIPT_RUN
* characters for at least one language in the Unicode Common Locale Data
* Repository [CLDR]. */
- dVAR;
/* Things that match /\d/u */
SV * decimals_invlist = PL_XPosix_ptrs[_CC_DIGIT];
/* If is within the range [+0 .. +9] of the script's zero, it also is a
* digit in that script. We can skip the rest of this code for this
* character. */
- if (UNLIKELY( zero_of_run
- && cp >= zero_of_run
- && cp - zero_of_run <= 9))
- {
+ if (UNLIKELY(zero_of_run && withinCOUNT(cp, zero_of_run, 9))) {
continue;
}
* several scripts, and the intersection is not empty. However, if the
* character is a decimal digit, it could still mean failure if it is
* from the wrong sequence of 10. So, we need to look at if it's a
- * digit. We've already handled the 10 decimal digits, and the next
+ * digit. We've already handled the 10 digits [0-9], and the next
* lowest one is this one: */
if (cp < FIRST_NON_ASCII_DECIMAL_DIGIT) {
continue; /* Not a digit; this character is part of the run */
if ( script_of_char >= 0
&& (zero_of_char = script_zeros[script_of_char]))
{
- if ( cp < zero_of_char
- || cp > zero_of_char + 9)
- {
+ if (! withinCOUNT(cp, zero_of_char, 9)) {
continue; /* Not a digit; this character is part of the run
*/
}