# HG changeset patch
# User FUJIWARA Katsunori <foozy@lares.dti.ne.jp>
# Date 2011-10-31 12:06:18
# Node ID 87bb6b7644f6b9376d06d6c5b8a7d36ff24a537e
# Parent  d7bfbc92a1c0b5d878febf44e1a1f1dc83b1259a

minirst: use unicode string as intermediate form for replacement

# this change redones part of 521c8e0c93bf, backed out by 0ad0ebe67815

Some character encodings use ASCII characters other than
control/alphabet/digit as a part of multi-bytes characters, so direct
replacing with such characters on strings in local encoding causes
invalid byte sequences.

[mpm: test changed to simple doctest]

diff --git a/mercurial/minirst.py b/mercurial/minirst.py
--- a/mercurial/minirst.py
+++ b/mercurial/minirst.py
@@ -23,9 +23,27 @@ import util, encoding
 from i18n import _
 
 def replace(text, substs):
+    '''
+    Apply a list of (find, replace) pairs to a text.
+
+    >>> replace("foo bar", [('f', 'F'), ('b', 'B')])
+    'Foo Bar'
+    >>> encoding.encoding = 'latin1'
+    >>> replace('\\x81\\\\', [('\\\\', '/')])
+    '\\x81/'
+    >>> encoding.encoding = 'shiftjis'
+    >>> replace('\\x81\\\\', [('\\\\', '/')])
+    '\\x81\\\\'
+    '''
+
+    # some character encodings (cp932 for Japanese, at least) use
+    # ASCII characters other than control/alphabet/digit as a part of
+    # multi-bytes characters, so direct replacing with such characters
+    # on strings in local encoding causes invalid byte sequences.
+    utext = text.decode(encoding.encoding)
     for f, t in substs:
-        text = text.replace(f, t)
-    return text
+        utext = utext.replace(f, t)
+    return utext.encode(encoding.encoding)
 
 _blockre = re.compile(r"\n(?:\s*\n)+")
 
diff --git a/tests/test-doctest.py b/tests/test-doctest.py
--- a/tests/test-doctest.py
+++ b/tests/test-doctest.py
@@ -36,3 +36,6 @@ doctest.testmod(hgext.convert.cvsps)
 
 import mercurial.revset
 doctest.testmod(mercurial.revset)
+
+import mercurial.minirst
+doctest.testmod(mercurial.minirst)