* .gdbinit (rp): REGEXP handling fixed.

* string.c (rb_str_rindex_m): need not to call rb_enc_check on
  regexp.

* re.c (unescape_escaped_nonascii): try ASCII-8BIT encoding for
  broken strings.

git-svn-id: http://svn.ruby-lang.org/repos/ruby/trunk@19812 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

4 files changed, 25 insertions, 5 deletions

diff --git a/.gdbinit b/.gdbinit
index 96e033fa8..c2a8b28d0 100644
--- a/.gdbinit
+++ b/.gdbinit
@@ -86,11 +86,17 @@ define rp
     print (struct RString *)$arg0
   else
   if ($flags & RUBY_T_MASK) == RUBY_T_REGEXP
+    set $regsrc = ((struct RRegexp*)$arg0)->src
+    set $rsflags = ((struct RBasic*)$regsrc)->flags
     printf "T_REGEXP: "
     set print address off
-    output ((struct RRegexp*)$arg0)->str
+    output (char *)(($rsflags & RUBY_FL_USER1) ? \
+	    ((struct RString*)$regsrc)->as.heap.ptr : \
+	    ((struct RString*)$regsrc)->as.ary)
     set print address on
-    printf " len:%ld ", ((struct RRegexp*)$arg0)->len
+    printf " len:%ld ", ($rsflags & RUBY_FL_USER1) ? \
+            ((struct RString*)$regsrc)->as.heap.len : \
+            (($rsflags & (RUBY_FL_USER2|RUBY_FL_USER3|RUBY_FL_USER4|RUBY_FL_USER5|RUBY_FL_USER6)) >> RUBY_FL_USHIFT+2)
     if $flags & RUBY_FL_USER6
       printf "(none) "
     end
diff --git a/ChangeLog b/ChangeLog
index a96058206..ca156bb39 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -64,6 +64,16 @@ Thu Oct 16 06:20:36 2008  Yukihiro Matsumoto  <matz@ruby-lang.org>
 	* test/ruby/test_transcode.rb (TestTranscode#test_errors):
 	  String#encode now works without any argument. [ruby-dev:36740]
 
+Wed Oct 15 23:48:22 2008  Yukihiro Matsumoto  <matz@ruby-lang.org>
+
+	* .gdbinit (rp): REGEXP handling fixed.
+
+	* string.c (rb_str_rindex_m): need not to call rb_enc_check on
+	  regexp.
+
+	* re.c (unescape_escaped_nonascii): try ASCII-8BIT encoding for
+	  broken strings.
+
 Wed Oct 15 23:11:10 2008  Yukihiro Matsumoto  <matz@ruby-lang.org>
 
 	* lib/delegate.rb (DelegateClass): restored 1.8 behavior for
diff --git a/re.c b/re.c
index fa00e8cea..03d9dc352 100644
--- a/re.c
+++ b/re.c
@@ -1967,8 +1967,12 @@ unescape_escaped_nonascii(const char **pp, const char *end, rb_encoding *enc,
 
     l = rb_enc_precise_mbclen(chbuf, chbuf+chlen, enc);
     if (MBCLEN_INVALID_P(l)) {
-        strcpy(err, "invalid multibyte escape");
-        return -1;
+        if (*encp == 0)
+            enc = *encp = rb_ascii8bit_encoding();
+        else if (*encp != rb_ascii8bit_encoding()) {
+	    strcpy(err, "invalid multibyte escape");
+            return -1;
+        }
     }
     if (1 < chlen || (chbuf[0] & 0x80)) {
         rb_str_buf_cat(buf, chbuf, chlen);
diff --git a/string.c b/string.c
index 859363363..601845c44 100644
--- a/string.c
+++ b/string.c
@@ -2355,7 +2355,7 @@ rb_str_rindex_m(int argc, VALUE *argv, VALUE str)
       case T_REGEXP:
 	/* enc = rb_get_check(str, sub); */
 	pos = str_offset(RSTRING_PTR(str), RSTRING_END(str), pos,
-		     rb_enc_check(str, sub), single_byte_optimizable(str));
+			 STR_ENC_GET(str), single_byte_optimizable(str));
 
 	if (!RREGEXP(sub)->ptr || RREGEXP_SRC_LEN(sub)) {
 	    pos = rb_reg_adjust_startpos(sub, str, pos, 1);