- Just found a silly typo error that prevents the correct parsing of categories. On file indywiki.py, line 1068 should be
usock = urllib.urlopen("http://" +language+ ".wikipedia.org/w/api.php?format... (rest of that line)
instead of what it is . This will be fixed on 0.9.8 version. On the meanwhile, I found out that the Nodebox library doesn't get the categories an article belongs for non english Wikipedia sites. I've contacted the authors of this great library to ask if it's a known issue.
- Indywiki needs a quick connection to the internet, otherwise the gui seems it does nothing, or worst that it has hung! There have to be added indications that the program is crawling wikipedia pages, trying to get it's data (and it's NOT DEAD!) ( FIXED ON version 0.9.6 )
- Windows related issues (indywiki displaying 3 or 4 out of 10 images) have been fixed.This was due to qt compiling jpeg handler as a plugin, thus py2exe couldn't include it as a dll and the program once started would search for it at c:\qt\4.3.2\plugins. ( FIXED ON version 0.9 )
- After we get back the results from a search, if we click on an image that belongs to the article we are looking for, the program shouldn't do anything. Instead it clears the text and makes the pointer look busy, thus distracting the user that it is doing something...It should make a search only when the image the user clicks belongs to a different page (a non native image, on indywiki jargon) ( FIXED ON version 0.8.3 )