diff options
| author | Stuart Rackham <srackham@methods.co.nz> | 2011-02-10 08:41:25 +1300 |
|---|---|---|
| committer | Stuart Rackham <srackham@methods.co.nz> | 2011-02-10 08:41:25 +1300 |
| commit | 1fa1196e26771824a814ebfefba3aaf2bb488401 (patch) | |
| tree | b72ad17d953c360521c697aa8e06256a2c6e9ada | |
| parent | b398d8c816f7e72d07eb6629f1a92ab066f62345 (diff) | |
| download | asciidoc-1fa1196e26771824a814ebfefba3aaf2bb488401.tar.gz | |
FIXED: Auto-generated section title ids are now unicode aware.
| -rwxr-xr-x | asciidoc.py | 5 | ||||
| -rw-r--r-- | tests/data/testcases.txt | 1 |
2 files changed, 5 insertions, 1 deletions
diff --git a/asciidoc.py b/asciidoc.py index 37ba984..e688f97 100755 --- a/asciidoc.py +++ b/asciidoc.py @@ -2113,7 +2113,10 @@ class Section: NCNameStartChar ::= Letter | '_' NameChar ::= Letter | Digit | '.' | '-' | '_' | ':' """ - base_ident = re.sub(r'[^a-zA-Z0-9]+', '_', title).strip('_').lower() + # Replace non-alpha numeric characters in title with underscores and + # convert to lower case. + base_ident = char_encode(re.sub(r'(?u)\W+', '_', + char_decode(title)).strip('_').lower()) # Prefix the ID name with idprefix attribute or underscore if not # defined. Prefix ensures the ID does not clash with existing IDs. idprefix = document.attributes.get('idprefix','_') diff --git a/tests/data/testcases.txt b/tests/data/testcases.txt index 1e34c18..0281736 100644 --- a/tests/data/testcases.txt +++ b/tests/data/testcases.txt @@ -660,4 +660,5 @@ Lorum ipsum... 这是一个测试 ------------ Double-with character titles. +<<_这是一个测试,link to auto-generated section ID>>. |