One more word: there was a warning section in result that it has been
truncated. I don't really believe this is connected to the problem, but
also can't exclude it at the moment.
If so, is there a way to iterate (gently) past $wgAPIMaxResultSize
Okay. Found the answer:
Set revCount=50 and it takes 25 minutes before returning any results.
When it does return, it returns them all at once, which means it held them
somewhere (memory presumably) while accumulating them. This should
probably be refactored into a generator that does not fetch more from the
API until it has yielded all (or most) of what it has already.
Is this the result of throttling perhaps? Seems likely. Is there a way
to disable throttling for a single call to fullVersionHistory and instead
enforce it between calls to fullVersionHistory for different pages?
21753 200416 2012-02-07T17:25:58Z
21754 200435 2012-02-07T17:45:33Z
21755 200344 2012-02-07T18:10:33Z
21756 200344 2012-02-08T22:42:47Z
finished!
real 25m48.924s
user 1m19.839s
sys 1m4.913s
_______________________________________________
Pywikipedia-l mailing list
[email protected]
https://lists.wikimedia.org/mailman/listinfo/pywikipedia-l
