Manybubbles added a comment.
Are we sure the fix has been deployed? Given the timing it'd have to have been
SWAT deployed to make it. I haven't checked if it was scheduled or if there is
a backport patch for it.
TASK DETAIL
https://phabricator.wikimedia.org/T75601
REPLY HANDLER ACTIONS
Tobi_WMDE_SW added a comment.
https://www.wikidata.org/wiki/Special:Version says that version 1.1.3 is
deployed. That should include the fix according to
https://github.com/DataValues/Geo#113-2014-11-19. Also note that the line where
the error happens is slightly different now.
TASK DETAIL
JanZerebecki added a comment.
From my reading of the code that should not happen, because its compared with
= 0 beforehand:
https://github.com/DataValues/Geo/blob/e8d28c92e9ce02305ca712db0d47052c51138d2b/src/Formatters/GeoCoordinateFormatter.php#L116
The branch wmf/1.25wmf8 of core.git contains
Manybubbles added a comment.
I'm not sure this is it but formatLatLongValue is public and offers a path
to call this method with a 0 precision. Funny that people always complain
that Java exceptions are so wordy but a real (back|stack)trace is exact why
you want here.
TASK DETAIL