cache UTF8 status for utf8_strlen(); ref #980

2 files changed, 11 insertions, 3 deletions

diff --git a/include/mruby/string.h b/include/mruby/string.h
index a66ba70ff..b94c355fe 100644
--- a/include/mruby/string.h
+++ b/include/mruby/string.h
@@ -81,9 +81,10 @@ mrb_int mrb_str_strlen(mrb_state*, struct RString*);
 #define MRB_STR_SHARED    1
 #define MRB_STR_NOFREE    2
 #define MRB_STR_FROZEN    4
-#define MRB_STR_EMBED     8
-#define MRB_STR_EMBED_LEN_MASK 0x1f0
-#define MRB_STR_EMBED_LEN_SHIFT 4
+#define MRB_STR_NO_UTF    8
+#define MRB_STR_EMBED    16
+#define MRB_STR_EMBED_LEN_MASK 0x3e0
+#define MRB_STR_EMBED_LEN_SHIFT 5
 
 void mrb_gc_free_str(mrb_state*, struct RString*);
 MRB_API void mrb_str_modify(mrb_state*, struct RString*);
diff --git a/src/string.c b/src/string.c
index 2b01cc3a5..8d79defe0 100644
--- a/src/string.c
+++ b/src/string.c
@@ -273,11 +273,17 @@ utf8_strlen(mrb_value str, mrb_int len)
   mrb_int total = 0;
   char* p = RSTRING_PTR(str);
   char* e = p;
+  if (RSTRING(str)->flags & MRB_STR_NO_UTF) {
+    return RSTRING_LEN(str);
+  }
   e += len < 0 ? RSTRING_LEN(str) : len;
   while (p<e) {
     p += utf8len(p, e);
     total++;
   }
+  if (RSTRING_LEN(str) == total) {
+    RSTRING(str)->flags |= MRB_STR_NO_UTF;
+  }
   return total;
 }
 
@@ -652,6 +658,7 @@ MRB_API void
 mrb_str_modify(mrb_state *mrb, struct RString *s)
 {
   check_frozen(mrb, s);
+  s->flags &= ~MRB_STR_NO_UTF;
   if (RSTR_SHARED_P(s)) {
     mrb_shared_string *shared = s->as.heap.aux.shared;