User Details
- User Since
- Feb 20 2019, 9:17 PM (57 w, 6 d)
- Availability
- Available
- LDAP User
- Tsevener
- MediaWiki User
- TSevener (WMF) [ Global Accounts ]
Today
Yesterday
Mon, Mar 30
Fri, Mar 27
Adding another OTRS ticket - https://ticket.wikimedia.org/otrs/index.pl?Action=AgentTicketZoom;TicketID=11341140
Adding another OTRS ticket.
Thu, Mar 26
Wed, Mar 25
Preliminary contract thoughts. Takes the 2019 fundraising object, nulls out not applicable keys, and adds additionalInfo object for config data.
Mon, Mar 23
If we happen to already persist this offline and it's just a matter of it not being wired up properly, go ahead and fix this.
Fri, Mar 20
Did some testing of migration in Prod. This is looking good, no code changes needed.
Used this to test Prod in general, cleaned out one class (https://github.com/wikimedia/wikipedia-ios/pull/3495) and filed an Infrastructure bug (https://phabricator.wikimedia.org/T248197).
Update - I do still see Javascript files downloaded over and over again in Prod, but now it at least returns 304 because the ETag is working properly. We still should investigate.
@SNowick_WMF this should be ready to test in TestFlight build #1731.
Thu, Mar 19
I've done some manual performance testing with 1000 saved articles and added any obvious bugs to the backlog. This could probably use some more manual testing with 1000 articles across a variety of languages, mine were just in EN. After that we can write performance tests as mentioned in the description. Unassigning myself to reflect that I'm not currently working on it but hoping to pick it back up tomorrow.
Tue, Mar 17
Pulled this in because it seems like a quick fix, but it did exist before mobile-html rewrite.
Since we're saving the header separately, we can just remove this Mime type setting on the file.
Update - this could be fine, I see now that it is slowly decreasing each time I return to the screen because SavedArticlesFetcher is deleting them one at a time.
I think it might be worth holding off on investigating this until the mobile-html endpoints are ready in Prod. Etag seems to work differently between labs and prod so that could affect results.
Mon, Mar 16
Here's a previous PR where we've had to try fixing this. Might be helpful as a starting place when investigating.
Fri, Mar 13
Thu, Mar 12
Updated with some initial thoughts to make this ticket actionable, but I imagine more will be added to as we dig into this.