jenkins-bot merged this change.

View Change

textlib.py: Avoid zero-width matching groups

This is a little trick to circumvent https://bugs.python.org/issue12177 .
The Memory error of re in Python 2.7.2 and 2.7.3 has something to do with
zero-width matching groups.

Here, by using + instead of * in other_chars group we avoid a zero-width match
and to make the group optional again we make the whole positive lookahead and
its group optional.

Bug: T191161
Change-Id: Ibfc8b8f961bdb13284aa5592fd9b7597e47f9d97
---
M pywikibot/textlib.py
1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/pywikibot/textlib.py b/pywikibot/textlib.py
index f959673..694e781 100644
--- a/pywikibot/textlib.py
+++ b/pywikibot/textlib.py
@@ -107,9 +107,9 @@
     \[\[\s*
     (?:%s)  # namespace aliases
     \s*:
-    (?=(?P<filename>
-        [^]|]*
-    ))(?P=filename)
+    ((?=(?P<filename>
+        [^]|]+  # * quantifier may crash on Python 2.7.2 (T191161)
+    ))(?P=filename))?
     (
         \|
         (
@@ -118,9 +118,9 @@
                     \[\[.*?\]\]
                 ))(?P=inner_link)
             )?
-            (?=(?P<other_chars>
-                [^\[\]]*
-            ))(?P=other_chars)
+            ((?=(?P<other_chars>
+                [^\[\]]+  # * quantifier may crash on Python 2.7.2 (T191161)
+            ))(?P=other_chars))?
         |
             (?=(?P<not_wikilink>
                 \[[^]]*\]

To view, visit change 423451. To unsubscribe, visit settings.