(perl #108276) add wrappers for deferred op processing

[perl5.git] / utf8.c

diff --git a/utf8.c b/utf8.c
index 9095cda..fc4a0c1 100644 (file)
--- a/utf8.c
+++ b/utf8.c
@@ -3184,14 +3184,14 @@ S_warn_on_first_deprecated_use(pTHX_ const char * const name,
 
             if (instr(file, "mathoms.c")) {
                 Perl_warner(aTHX_ WARN_DEPRECATED,
-                            "In %s, line %d, starting in Perl v5.30, %s()"
+                            "In %s, line %d, starting in Perl v5.32, %s()"
                             " will be removed.  Avoid this message by"
                             " converting to use %s().\n",
                             file, line, name, alternative);
             }
             else {
                 Perl_warner(aTHX_ WARN_DEPRECATED,
-                            "In %s, line %d, starting in Perl v5.30, %s() will"
+                            "In %s, line %d, starting in Perl v5.32, %s() will"
                             " require an additional parameter.  Avoid this"
                             " message by converting to use %s().\n",
                             file, line, name, alternative);
@@ -5471,7 +5471,13 @@ L<http://www.unicode.org/unicode/reports/tr21/> (Case Mappings).
  *                          that effect.  However, if the caller knows what
  *                          it's doing, it can pass this flag to indicate that,
  *                          and the assertion is skipped.
- *  FOLDEQ_S2_ALREADY_FOLDED  Similarly.
+ *  FOLDEQ_S2_ALREADY_FOLDED  Similar to FOLDEQ_S1_ALREADY_FOLDED, but applies
+ *                          to s2, and s2 doesn't have to be UTF-8 encoded.
+ *                          This introduces an asymmetry to save a few branches
+ *                          in a loop.  Currently, this is not a problem, as
+ *                          never are both inputs pre-folded.  Simply call this
+ *                          function with the pre-folded one as the second
+ *                          string.
  *  FOLDEQ_S2_FOLDS_SANE
  */
 I32
@@ -5518,6 +5524,9 @@ Perl_foldEQ_utf8_flags(pTHX_ const char *s1, char **pe1, UV l1, bool u1,
             flags_for_folder |= FOLD_FLAGS_LOCALE;
         }
     }
+    if (flags & FOLDEQ_UTF8_NOMIX_ASCII) {
+        flags_for_folder |= FOLD_FLAGS_NOMIX_ASCII;
+    }
 
     if (pe1) {
         e1 = *(U8**)pe1;
@@ -5602,9 +5611,23 @@ Perl_foldEQ_utf8_flags(pTHX_ const char *s1, char **pe1, UV l1, bool u1,
 
         if (n2 == 0) {    /* Same for s2 */
            if (flags & FOLDEQ_S2_ALREADY_FOLDED) {
-               f2 = (U8 *) p2;
-                assert(u2);
-               n2 = UTF8SKIP(f2);
+
+                /* Point to the already-folded character.  But for non-UTF-8
+                 * variants, convert to UTF-8 for the algorithm below */
+               if (UTF8_IS_INVARIANT(*p2)) {
+                    f2 = (U8 *) p2;
+                    n2 = 1;
+                }
+                else if (u2) {
+                    f2 = (U8 *) p2;
+                    n2 = UTF8SKIP(f2);
+                }
+                else {
+                    foldbuf2[0] = UTF8_EIGHT_BIT_HI(*p2);
+                    foldbuf2[1] = UTF8_EIGHT_BIT_LO(*p2);
+                    f2 = foldbuf2;
+                    n2 = 2;
+                }
            }
            else {
                 if (isASCII(*p2) && ! (flags & FOLDEQ_LOCALE)) {