Revision: 4202 Author: misza13 Date: 2007-09-06 17:15:47 +0000 (Thu, 06 Sep 2007)
Log Message: ----------- Fix ImagePage.fileUrl() - image page looks slightly different for .ogg files. Bypassing using negative lookahead (bogus "image link" has a 'class="image"' following it).
Modified Paths: -------------- trunk/pywikipedia/wikipedia.py
Modified: trunk/pywikipedia/wikipedia.py =================================================================== --- trunk/pywikipedia/wikipedia.py 2007-09-05 17:18:41 UTC (rev 4201) +++ trunk/pywikipedia/wikipedia.py 2007-09-06 17:15:47 UTC (rev 4202) @@ -2120,7 +2120,7 @@ # The part after the | is required for copying .ogg files from en:, as they do not # have a "full image link" div. This might change in the future; on commons, there # is a full image link for .ogg and .mid files. - urlR = re.compile(r'<div class="fullImageLink" id="file">.*?<a href="(?P<url>.+?)"|<span class="dangerousLink"><a href="(?P<url2>.+?)"', re.DOTALL) + urlR = re.compile(r'<div class="fullImageLink" id="file">.*?<a href="(?P<url>[^ ]+?)"(?! class="image")|<span class="dangerousLink"><a href="(?P<url2>.+?)"', re.DOTALL) m = urlR.search(self.getImagePageHtml()) try: url = m.group('url') or m.group('url2')