use Sunday quick search algorithm instead of Boyer-Moor-Horspool

diff --git a/src/regcomp.c b/src/regcomp.c
index fc2b48f653f24ad04b86474c8795008d98a7151b..08dc3d613d02b1308a712ac9ce4c02d8b74c1281 100644 (file)
--- a/src/regcomp.c
+++ b/src/regcomp.c
@@ -4552,6 +4552,24 @@ setup_tree(Node* node, regex_t* reg, int state, ScanEnv* env)
   return r;
 }
 
+#ifdef USE_SUNDAY_QUICK_SEARCH_ALGORITHM
+static int
+set_bmh_skip(UChar* s, UChar* end, OnigEncoding enc ARG_UNUSED, UChar skip[])
+{
+  int i, len;
+
+  len = (int )(end - s);
+  if (len + 1 >= ONIG_CHAR_TABLE_SIZE)
+    return ONIGERR_PARSER_BUG;
+
+  for (i = 0; i < ONIG_CHAR_TABLE_SIZE; i++) skip[i] = (UChar )(len + 1);
+
+  for (i = 0; i < len; i++)
+    skip[s[i]] = len - i;
+
+  return 0;
+}
+#else
 /* set skip map for Boyer-Moore-Horspool search */
 static int
 set_bmh_skip(UChar* s, UChar* end, OnigEncoding enc ARG_UNUSED, UChar skip[])
@@ -4571,6 +4589,8 @@ set_bmh_skip(UChar* s, UChar* end, OnigEncoding enc ARG_UNUSED, UChar skip[])
     return ONIGERR_PARSER_BUG;
   }
 }
+#endif /* USE_SUNDAY_QUICK_SEARCH_ALGORITHM */
+
 
 #define OPT_EXACT_MAXLEN   24
 

diff --git a/src/regexec.c b/src/regexec.c
index ce60cf85df8cdfb663e1fa5c610ecf263520f1de..48886bf58daf2a47973b733b9e5fe7bf2dd762b3 100644 (file)
--- a/src/regexec.c
+++ b/src/regexec.c
@@ -4112,6 +4112,79 @@ slow_search_backward_ic(OnigEncoding enc, int case_fold_flag,
   return (UChar* )NULL;
 }
 
+#ifdef USE_SUNDAY_QUICK_SEARCH_ALGORITHM
+
+static UChar*
+bmh_search_forward(regex_t* reg, const UChar* target, const UChar* target_end,
+                   const UChar* text, const UChar* text_end,
+                   const UChar* text_range)
+{
+  const UChar *s, *se, *t, *p, *end;
+  const UChar *tail;
+  int skip, tlen1;
+
+#ifdef ONIG_DEBUG_SEARCH
+  fprintf(stderr,
+          "bmh_search_forward: text: %p, text_end: %p, text_range: %p\n",
+          text, text_end, text_range);
+#endif
+
+  tail = target_end - 1;
+  tlen1 = (int )(tail - target);
+  end = text_range;
+  if (end + tlen1 > text_end)
+    end = text_end - tlen1;
+
+  s = text;
+
+  while (s < end) {
+    p = se = s + tlen1;
+    t = tail;
+    while (*p == *t) {
+      if (t == target) return (UChar* )s;
+      p--; t--;
+    }
+    if (se + 1 >= text_end) break;
+    skip = reg->map[*(se + 1)];
+    t = s;
+    do {
+      s += enclen(reg->enc, s);
+    } while ((s - t) < skip && s < end);
+  }
+
+  return (UChar* )NULL;
+}
+
+static UChar*
+bmh_search(regex_t* reg, const UChar* target, const UChar* target_end,
+           const UChar* text, const UChar* text_end, const UChar* text_range)
+{
+  const UChar *s, *t, *p, *end;
+  const UChar *tail;
+
+  end = text_range + (target_end - target);
+  if (end > text_end)
+    end = text_end;
+
+  tail = target_end - 1;
+  s = text + (tail - target);
+
+  while (s < end) {
+    p = s;
+    t = tail;
+    while (*p == *t) {
+      if (t == target) return (UChar* )p;
+      p--; t--;
+    }
+    if (s + 1 >= text_end) break;
+    s += reg->map[*(s + 1)];
+  }
+
+  return (UChar* )NULL;
+}
+
+#else
+
 static UChar*
 bmh_search_forward(regex_t* reg, const UChar* target, const UChar* target_end,
                    const UChar* text, const UChar* text_end,
@@ -4178,6 +4251,7 @@ bmh_search(regex_t* reg, const UChar* target, const UChar* target_end,
 
   return (UChar* )NULL;
 }
+#endif /* USE_SUNDAY_QUICK_SEARCH_ALGORITHM */
 
 static UChar*
 map_search(OnigEncoding enc, UChar map[],

diff --git a/src/regint.h b/src/regint.h
index 1782dbe26ab7c5a1fc7949fba7c54260f310d653..50a9e36afbe3914a9d38df9f91d523ff8f03e84c 100644 (file)
--- a/src/regint.h
+++ b/src/regint.h
@@ -62,6 +62,7 @@
 #define USE_INSISTENT_CHECK_CAPTURES_IN_EMPTY_REPEAT    /* /(?:()|())*\2/ */
 #define USE_NEWLINE_AT_END_OF_STRING_HAS_EMPTY_LINE     /* /\n$/ =~ "\n" */
 #define USE_WARNING_REDUNDANT_NESTED_REPEAT_OPERATOR
+#define USE_SUNDAY_QUICK_SEARCH_ALGORITHM
 
 #define USE_RETRY_LIMIT_IN_MATCH