naruse	2013-07-05 09:54:11 +0900 (Fri, 05 Jul 2013)

  New Revision: 41786

  http://svn.ruby-lang.org/cgi-bin/viewvc.cgi?view=rev&revision=41786

  Log:
    broken utf-8

  Modified files:
    trunk/enc/utf_8.c
    trunk/include/ruby/oniguruma.h
    trunk/regenc.c

Index: include/ruby/oniguruma.h
===================================================================
--- include/ruby/oniguruma.h	(revision 41785)
+++ include/ruby/oniguruma.h	(revision 41786)
@@ -246,11 +246,14 @@ ONIG_EXTERN OnigEncodingType OnigEncodin https://github.com/ruby/ruby/blob/trunk/include/ruby/oniguruma.h#L246
 #define ONIGENC_MBCLEN_CHARFOUND_P(r)           (0 < (r))
 #define ONIGENC_MBCLEN_CHARFOUND_LEN(r)         (r)
 
-#define ONIGENC_CONSTRUCT_MBCLEN_INVALID()      (-1)
-#define ONIGENC_MBCLEN_INVALID_P(r)             ((r) == -1)
+#define ONIGENC_INVALID_NUM                     0x10000
+#define ONIGENC_CONSTRUCT_MBCLEN_INVALID()      (-1-ONIGENC_INVALID_NUM)
+#define ONIGENC_CONSTRUCT_MBCLEN_INVALID2(n)    (-(n)-ONIGENC_INVALID_NUM)
+#define ONIGENC_MBCLEN_INVALID_P(r)             ((r) < -ONIGENC_INVALID_NUM)
+#define ONIGENC_MBCLEN_INVALID_LEN(r)           (-(r)-ONIGENC_INVALID_NUM)
 
 #define ONIGENC_CONSTRUCT_MBCLEN_NEEDMORE(n)    (-1-(n))
-#define ONIGENC_MBCLEN_NEEDMORE_P(r)            ((r) < -1)
+#define ONIGENC_MBCLEN_NEEDMORE_P(r)            (-ONIGENC_INVALID_NUM <= (r) && (r) < -1)
 #define ONIGENC_MBCLEN_NEEDMORE_LEN(r)          (-1-(r))
 
 #define ONIGENC_PRECISE_MBC_ENC_LEN(enc,p,e)   (enc)->precise_mbc_enc_len(p,e,enc)
Index: regenc.c
===================================================================
--- regenc.c	(revision 41785)
+++ regenc.c	(revision 41786)
@@ -55,10 +55,12 @@ extern int https://github.com/ruby/ruby/blob/trunk/regenc.c#L55
 onigenc_mbclen_approximate(const OnigUChar* p,const OnigUChar* e, struct OnigEncodingTypeST* enc)
 {
   int ret = ONIGENC_PRECISE_MBC_ENC_LEN(enc,p,e);
-  if (ONIGENC_MBCLEN_CHARFOUND_P(ret))
-    return ONIGENC_MBCLEN_CHARFOUND_LEN(ret);
-  else if (ONIGENC_MBCLEN_NEEDMORE_P(ret))
+  if (ONIGENC_MBCLEN_NEEDMORE_P(ret))
     return (int)(e-p)+ONIGENC_MBCLEN_NEEDMORE_LEN(ret);
+  else if (ONIGENC_MBCLEN_INVALID_P(ret))
+    return (int)(e-p)+ONIGENC_MBCLEN_INVALID_LEN(ret);
+  else if (ONIGENC_MBCLEN_CHARFOUND_P(ret))
+    return ONIGENC_MBCLEN_CHARFOUND_LEN(ret);
   return 1;
 }
 
Index: enc/utf_8.c
===================================================================
--- enc/utf_8.c	(revision 41785)
+++ enc/utf_8.c	(revision 41786)
@@ -233,12 +233,12 @@ mbc_enc_len(const UChar* p, const UChar* https://github.com/ruby/ruby/blob/trunk/enc/utf_8.c#L233
   if (p == e) return ONIGENC_CONSTRUCT_MBCLEN_NEEDMORE(EncLen_UTF8[firstbyte]-2);
   s = trans[s][*p++];
   if (s < 0) return s == ACCEPT ? ONIGENC_CONSTRUCT_MBCLEN_CHARFOUND(3) :
-                                  ONIGENC_CONSTRUCT_MBCLEN_INVALID();
+                                  ONIGENC_CONSTRUCT_MBCLEN_INVALID2(2);
 
   if (p == e) return ONIGENC_CONSTRUCT_MBCLEN_NEEDMORE(EncLen_UTF8[firstbyte]-3);
   s = trans[s][*p++];
   return s == ACCEPT ? ONIGENC_CONSTRUCT_MBCLEN_CHARFOUND(4) :
-                       ONIGENC_CONSTRUCT_MBCLEN_INVALID();
+                       ONIGENC_CONSTRUCT_MBCLEN_INVALID2(3);
 }
 
 static int

--
ML: ruby-changes@q...
Info: http://www.atdot.net/~ko1/quickml/