From 4987551145411af7369f9d7174cba245d1d5a696 Mon Sep 17 00:00:00 2001
From: akr <akr@b2dd03c8-39d4-4d8f-98ff-823fe69b080e>
Date: Fri, 4 Jul 2008 23:50:33 +0000
Subject: * re.c (unescape_nonascii): add has_property argument not to   raise
 error by /\p{Hiragana}\u{3042}/ in EUC-JP script.   (rb_reg_preprocess): use
 has_property argument to make regexp   encoding fixed.

git-svn-id: http://svn.ruby-lang.org/repos/ruby/trunk@17884 b2dd03c8-39d4-4d8f-98ff-823fe69b080e
---
 ChangeLog              |  7 +++++++
 re.c                   | 12 +++++++++---
 test/ruby/test_m17n.rb |  8 ++++++++
 3 files changed, 24 insertions(+), 3 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 49835564e..18417d333 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,10 @@
+Sat Jul  5 08:48:05 2008  Tanaka Akira  <akr@fsij.org>
+
+	* re.c (unescape_nonascii): add has_property argument not to
+	  raise error by /\p{Hiragana}\u{3042}/ in EUC-JP script.
+	  (rb_reg_preprocess): use has_property argument to make regexp
+	  encoding fixed.
+
 Sat Jul  5 08:29:47 2008  Tanaka Akira  <akr@fsij.org>
 
 	* re.c (unescape_nonascii): make regexp fixed_encoding if \p is used.
diff --git a/re.c b/re.c
index 899ede647..2bbf915e9 100644
--- a/re.c
+++ b/re.c
@@ -2085,7 +2085,8 @@ unescape_unicode_bmp(const char **pp, const char *end,
 
 static int
 unescape_nonascii(const char *p, const char *end, rb_encoding *enc,
-        VALUE buf, rb_encoding **encp, onig_errmsg_buffer err)
+        VALUE buf, rb_encoding **encp, int *has_property,
+        onig_errmsg_buffer err)
 {
     char c;
     char smallbuf[2];
@@ -2165,7 +2166,7 @@ unescape_nonascii(const char *p, const char *end, rb_encoding *enc,
 
               case 'p': /* \p{Hiragana} */
                 if (!*encp) {
-                    *encp = enc;
+                    *has_property = 1;
                 }
                 goto escape_asis;
 
@@ -2192,6 +2193,7 @@ rb_reg_preprocess(const char *p, const char *end, rb_encoding *enc,
         rb_encoding **fixed_enc, onig_errmsg_buffer err)
 {
     VALUE buf;
+    int has_property = 0;
 
     buf = rb_str_buf_new(0);
 
@@ -2202,9 +2204,13 @@ rb_reg_preprocess(const char *p, const char *end, rb_encoding *enc,
         rb_enc_associate(buf, enc);
     }
 
-    if (unescape_nonascii(p, end, enc, buf, fixed_enc, err) != 0)
+    if (unescape_nonascii(p, end, enc, buf, fixed_enc, &has_property, err) != 0)
         return Qnil;
 
+    if (has_property && !*fixed_enc) {
+        *fixed_enc = enc;
+    }
+
     if (*fixed_enc) {
         rb_enc_associate(buf, *fixed_enc);
     }
diff --git a/test/ruby/test_m17n.rb b/test/ruby/test_m17n.rb
index c5871898f..b2a67686a 100644
--- a/test/ruby/test_m17n.rb
+++ b/test/ruby/test_m17n.rb
@@ -510,6 +510,14 @@ class TestM17N < Test::Unit::TestCase
     r = /\p{Hiragana}/e
     assert(r.fixed_encoding?)
     assert_match(r, "\xa4\xa2".force_encoding("euc-jp"))
+
+    r = eval('/\u{3042}\p{Hiragana}/'.force_encoding("euc-jp"))
+    assert(r.fixed_encoding?)
+    assert_equal(Encoding::UTF_8, r.encoding)
+
+    r = eval('/\p{Hiragana}\u{3042}/'.force_encoding("euc-jp"))
+    assert(r.fixed_encoding?)
+    assert_equal(Encoding::UTF_8, r.encoding)
   end
 
   def test_regexp_embed_preprocess
-- 
cgit