+#define DEBUG_PEEP(str,scan,depth) \
+ DEBUG_OPTIMISE_r({ \
+ SV * const mysv=sv_newmortal(); \
+ regnode *Next = regnext(scan); \
+ regprop(RExC_rx, mysv, scan); \
+ PerlIO_printf(Perl_debug_log, "%*s" str ">%3d: %s [%d]\n", \
+ (int)depth*2, "", REG_NODE_NUM(scan), SvPV_nolen_const(mysv),\
+ Next ? (REG_NODE_NUM(Next)) : 0 ); \
+ });
+
+#define JOIN_EXACT(scan,min,flags) \
+ if (PL_regkind[OP(scan)] == EXACT) \
+ join_exact(pRExC_state,(scan),(min),(flags),NULL,depth+1)
+
+U32
+S_join_exact(pTHX_ RExC_state_t *pRExC_state, regnode *scan, I32 *min, U32 flags,regnode *val, U32 depth) {
+ /* Merge several consecutive EXACTish nodes into one. */
+ regnode *n = regnext(scan);
+ U32 stringok = 1;
+ regnode *next = scan + NODE_SZ_STR(scan);
+ U32 merged = 0;
+ U32 stopnow = 0;
+#ifdef DEBUGGING
+ regnode *stop = scan;
+#endif
+ GET_RE_DEBUG_FLAGS_DECL;
+ DEBUG_PEEP("join",scan,depth);
+
+ /* Skip NOTHING, merge EXACT*. */
+ while (n &&
+ ( PL_regkind[OP(n)] == NOTHING ||
+ (stringok && (OP(n) == OP(scan))))
+ && NEXT_OFF(n)
+ && NEXT_OFF(scan) + NEXT_OFF(n) < I16_MAX) {
+
+ if (OP(n) == TAIL || n > next)
+ stringok = 0;
+ if (PL_regkind[OP(n)] == NOTHING) {
+
+ DEBUG_PEEP("skip:",n,depth);
+ NEXT_OFF(scan) += NEXT_OFF(n);
+ next = n + NODE_STEP_REGNODE;
+#ifdef DEBUGGING
+ if (stringok)
+ stop = n;
+#endif
+ n = regnext(n);
+ }
+ else if (stringok) {
+ const int oldl = STR_LEN(scan);
+ regnode * const nnext = regnext(n);
+
+ DEBUG_PEEP("merg",n,depth);
+
+ merged++;
+ if (oldl + STR_LEN(n) > U8_MAX)
+ break;
+ NEXT_OFF(scan) += NEXT_OFF(n);
+ STR_LEN(scan) += STR_LEN(n);
+ next = n + NODE_SZ_STR(n);
+ /* Now we can overwrite *n : */
+ Move(STRING(n), STRING(scan) + oldl, STR_LEN(n), char);
+#ifdef DEBUGGING
+ stop = next - 1;
+#endif
+ n = nnext;
+ if (stopnow) break;
+ }
+
+#ifdef EXPERIMENTAL_INPLACESCAN
+ if (flags && !NEXT_OFF(n)) {
+ DEBUG_PEEP("atch",val,depth);
+ if (reg_off_by_arg[OP(n)]) {
+ ARG_SET(n, val - n);
+ }
+ else {
+ NEXT_OFF(n) = val - n;
+ }
+ stopnow=1;
+ }
+#endif
+ }
+
+ if (UTF && ( OP(scan) == EXACTF ) && ( STR_LEN(scan) >= 6 ) ) {
+ /*
+ Two problematic code points in Unicode casefolding of EXACT nodes:
+
+ U+0390 - GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS
+ U+03B0 - GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS
+
+ which casefold to
+
+ Unicode UTF-8
+
+ U+03B9 U+0308 U+0301 0xCE 0xB9 0xCC 0x88 0xCC 0x81
+ U+03C5 U+0308 U+0301 0xCF 0x85 0xCC 0x88 0xCC 0x81
+
+ This means that in case-insensitive matching (or "loose matching",
+ as Unicode calls it), an EXACTF of length six (the UTF-8 encoded byte
+ length of the above casefolded versions) can match a target string
+ of length two (the byte length of UTF-8 encoded U+0390 or U+03B0).
+ This would rather mess up the minimum length computation.
+
+ What we'll do is to look for the tail four bytes, and then peek
+ at the preceding two bytes to see whether we need to decrease
+ the minimum length by four (six minus two).
+
+ Thanks to the design of UTF-8, there cannot be false matches:
+ A sequence of valid UTF-8 bytes cannot be a subsequence of
+ another valid sequence of UTF-8 bytes.
+
+ */
+ char * const s0 = STRING(scan), *s, *t;
+ char * const s1 = s0 + STR_LEN(scan) - 1;
+ char * const s2 = s1 - 4;
+ const char t0[] = "\xcc\x88\xcc\x81";
+ const char * const t1 = t0 + 3;
+
+ for (s = s0 + 2;
+ s < s2 && (t = ninstr(s, s1, t0, t1));
+ s = t + 4) {
+ if (((U8)t[-1] == 0xB9 && (U8)t[-2] == 0xCE) ||
+ ((U8)t[-1] == 0x85 && (U8)t[-2] == 0xCF))
+ *min -= 4;
+ }
+ }
+
+#ifdef DEBUGGING
+ /* Allow dumping */
+ n = scan + NODE_SZ_STR(scan);
+ while (n <= stop) {
+ if (PL_regkind[OP(n)] != NOTHING || OP(n) == NOTHING) {
+ OP(n) = OPTIMIZED;
+ NEXT_OFF(n) = 0;
+ }
+ n++;
+ }
+#endif
+ DEBUG_OPTIMISE_r(if (merged){DEBUG_PEEP("finl",scan,depth)});
+ return stopnow;
+}
+
https://perl5.git.perl.org / perl5.git / commitdiff