Backlinks for Unicode page name

[theYT <dev@…>]

Pages with unicode (non-ASCII) name are not matched (filtered out) when testing with regex.

Patch file attached. (test with r13390)

[Jun Omae]

It seems backlinks_page should be escaped in regular expression, e.g. Languages/C++.

backlinks/macro.py

@@ -81 +81 @@
         WHERE w1.version = w2.version AND w1.name = w2.name AND
         (w1.text %s)""" % db.like(), ('%' + db.like_escape(backlinks_page) + '%',))
 
+    pattern = re.compile(r'\b%s\b' % re.escape(backlinks_page), re.UNICODE)
     backlinked_pages = []
     for page, text in cursor:
         if page != backlinks_page and page != caller_page and \
-           re.search(r'\b%s\b' % backlinks_page, text):
+           pattern.search(text):
             backlinked_pages.append(page)
 
     return backlinked_pages

[Ryan J Ollos]

In 13928:

7.0dev: Match pages with unicode characters in page name. Fixes #11707.

Patch by Jun Omae.