From 4894de4996e3b8281d6fc3c68d0caf234cbd1790 Mon Sep 17 00:00:00 2001
From: Norihiro Tanaka <noritnk@kcn.ne.jp>
Date: Thu, 20 Feb 2014 21:58:49 +0900
Subject: [PATCH] Use mbrtowc_cache in DFA engine.

---
 src/dfa.c | 51 ++++++++++++++++++++++++++++++++++++++++++++++++---
 1 file changed, 48 insertions(+), 3 deletions(-)

diff --git a/src/dfa.c b/src/dfa.c
index 8906ed3..c073cf3 100644
--- a/src/dfa.c
+++ b/src/dfa.c
@@ -419,6 +419,13 @@ struct dfa
                                    the dfa.  */
 };
 
+struct mbrtowc_cache {
+    size_t length;
+    wchar_t wchar;
+};
+
+static struct mbrtowc_cache mbrtowc_cache[NOTCHAR];
+
 /* Some macros for user access to dfa internals.  */
 
 /* ACCEPTING returns true if s could possibly be an accepting state of r.  */
@@ -822,12 +829,22 @@ static unsigned char const *buf_end;    /* reference to end in dfaexec.  */
     else					\
       {						\
         wchar_t _wc;				\
-        cur_mb_len = mbrtowc (&_wc, lexptr, lexleft, &mbs); \
+        bool use_cache = false;			\
+        cur_mb_len = mbrtowc_cache[to_uchar (*lexptr)].length; \
+        if (cur_mb_len != (size_t) -2)		\
+          {					\
+            _wc = mbrtowc_cache[to_uchar (*lexptr)].wchar; \
+            use_cache = true;                   \
+          }					\
+        else					\
+          cur_mb_len = mbrtowc (&_wc, lexptr, lexleft, &mbs); \
         if (cur_mb_len <= 0)			\
           {					\
             cur_mb_len = 1;			\
             --lexleft;				\
             (wc) = (c) = to_uchar (*lexptr++);  \
+            if (!use_cache)			\
+              memset (&mbs, 0, sizeof mbs);	\
           }					\
         else					\
           {					\
@@ -3271,8 +3288,19 @@ prepare_wc_buf (const char *begin, const char *end)
     {
       if (remain_bytes == 0)
         {
-          remain_bytes
-            = mbrtowc (inputwcs + i, begin + i, end - begin - i + 1, &mbs);
+          bool use_cache = false;
+          remain_bytes = mbrtowc_cache[to_uchar (begin[i])].length;
+          if (remain_bytes != (size_t) -2)
+            {
+              inputwcs[i] = mbrtowc_cache[to_uchar (begin[i])].wchar;
+              use_cache = true;
+            }
+          else
+            {
+              remain_bytes
+                = mbrtowc (inputwcs + i, begin + i, end - begin - i + 1, &mbs);
+            }
+
           if (remain_bytes < 1
               || remain_bytes == (size_t) -1
               || remain_bytes == (size_t) -2
@@ -3281,6 +3309,8 @@ prepare_wc_buf (const char *begin, const char *end)
               remain_bytes = 0;
               inputwcs[i] = (wchar_t) begin[i];
               mblen_buf[i] = 0;
+              if (!use_cache)
+                memset (&mbs, 0, sizeof mbs);
               if (begin[i] == eol)
                 break;
             }
@@ -3545,11 +3575,26 @@ dfaoptimize (struct dfa *d)
   d->mb_cur_max = 1;
 }
 
+void
+build_mbrtowc_cache (void)
+{
+  int i;
+
+  for (i = CHAR_MIN; i <= CHAR_MAX; ++i)
+    {
+      char c = i;
+      unsigned char uc = i;
+      mbstate_t s = { 0 };
+      mbrtowc_cache[uc].length = mbrtowc (&mbrtowc_cache[uc].wchar, &c, 1, &s);
+    }
+}
+
 /* Parse and analyze a single string of the given length.  */
 void
 dfacomp (char const *s, size_t len, struct dfa *d, int searchflag)
 {
   dfainit (d);
+  build_mbrtowc_cache ();
   dfaparse (s, len, d);
   dfamust (d);
   dfaoptimize (d);
-- 
1.8.5.2