Sat, Apr 20
Changing server state like this (user options) should not be possible solely using GET parameters.
This is going to be so abused for political statements and by LTAs once they find its availability.
Wed, Apr 17
Tue, Apr 16
Um, this is mostly intentional. We *want* people to see new messages posted on the village pump, etc.
Mon, Apr 15
Would an increase in the size of stuff being stored in APC by the ExtensionRegistry be related? rMWd1666a89fff6: resourceloader: Support TestModules registration via extension.json probably caused extension.json stuffs to get even bigger. We currently don't optimize what we store at all (there's definitely some low hanging fruit).
one page is Commons:Deletion_requests/Files_uploaded_by_Nabila.selim in the main namespace (0),
Somewhat of a dupe of T187045: Decide on future of Extension:ShortUrl on Wikimedia Wikis. I think the main open question is a technical one, do we a) add a read only mode to ShortUrl, and keep it deployed forever, or b) move the logic for short url into UrlShortener, so we only have one thing deployed and routing stuff.
Sun, Apr 14
Should we have a CDN purge when we create new short codes just in case the 404 was cached?
@Ladsgroup is there anything else left to do here?
IIRC apc has a history of causing weird bugs in CI with stuff not being invalidated properly. I don't know of those bugs still exist, but if you want a persistent cache to test with, you should use HashBagOStuff.
In general I think that's a good idea. Two things I think we'll need to do:
We should already be deduplicating HTTPS/HTTP.
Fri, Apr 12
This is not enough for some Wikidata queries that main contain up to 4000 characters. My suggestion is to increase the limit to 10K and see if there are some use cases of longer URLs.
Thu, Apr 11
Because phan requires PHP >=7.0 (other dev deps require >=5.6.99) and until recently also required specific versions of php-ast, which would have make composer install really inconvenient for a lot of people.
Wed, Apr 10
I haven't verified myself, but if this is an issue, it's UBN.
Tue, Apr 9
I'm going to defer to @Ladsgroup since he's taken over the main deployment, but I do think we went a bit overboard on the easter eggs. My initial plan was to have no more than 10 joke/easter egg style ones, and then allocate "b" -> wikibooks, "c" -> commons, etc.
Mon, Apr 8
I always thought the cat was paying homage to MonoBook's book.
Sun, Apr 7
Sat, Apr 6
And filed https://github.com/symfony/symfony/issues/30877 with symfony.
OK, I haven't seen any OOMs after deploying that change. I also submitted https://github.com/phan/phan/pull/2634 ("Avoid repeatedly calculating the terminal's width") upstream based on hashar's analysis.
Fri, Apr 5
Thu, Apr 4
My previous attempt at this: https://gerrit.wikimedia.org/r/c/mediawiki/core/+/307913
Mon, Apr 1
tools.ldap@tools-sgebastion-07:~/www/python/src$ git remote -v origin https://phabricator.wikimedia.org/source/tool-ldap.git (fetch) origin https://phabricator.wikimedia.org/source/tool-ldap.git (push) tools.ldap@tools-sgebastion-07:~/www/python/src$ git remote set-url origin https://gerrit.wikimedia.org/r/labs/tools/ldap tools.ldap@tools-sgebastion-07:~/www/python/src$ git remote -v origin https://gerrit.wikimedia.org/r/labs/tools/ldap (fetch) origin https://gerrit.wikimedia.org/r/labs/tools/ldap (push) tools.ldap@tools-sgebastion-07:~/www/python/src$ git pull remote: Counting objects: 4, done remote: Finding sources: 100% (3/3) remote: Getting sizes: 100% (3/3) remote: Compressing objects: 100% (290/290) remote: Total 3 (delta 1), reused 0 (delta 0) Unpacking objects: 100% (3/3), done. From https://gerrit.wikimedia.org/r/labs/tools/ldap d797c2b..30de8d5 master -> origin/master Updating d797c2b..30de8d5 Fast-forward .gitreview | 6 ++++++ 1 file changed, 6 insertions(+) create mode 100644 .gitreview
Pushed and deployed.
Please. It would be nice if all of my tools could be moved en-masse. I suppose it would be worthwhile to script it?
I don't have arcanist set up anymore, can you email me or pastebin this as a .patch file? git format-patch HEAD~1 should do it. IIRC the suggested "Patch without arc" doesn't retain authorship information.
Disabling phan isn't an option due to the amount of issues it regularly catches. Worst case we can revert the upgrade but that's going to be tricky....
Fri, Mar 29
commit 2603d119922583d92f0392df9c1ddb9c57ac1562 (HEAD -> master, origin/master, origin/HEAD) Author: Kunal Mehta <email@example.com> Date: Thu Mar 28 18:50:15 2019 -0700
Thu, Mar 28
So when I set this up in January, I assumed hound would do the git pulls by itself. I guess it doesn't, and we need a cronjob to do so. Pulled it manually for now.
Wed, Mar 27
Tue, Mar 26
Mon, Mar 25
This is not a duplicate of T219012.
Sun, Mar 24
Filed an issue upstream for #3: https://github.com/doxygen/doxygen/issues/6900
AIUI the wiki was never formally deleted (just closed), the domain was redirected and the wiki continued to exist until these errors started cropping up and people began cleaning it up one by one.
Thanks @daniel. So I created a wmde-mediawiki group (https://gerrit.wikimedia.org/r/admin/groups/1597) with everyone listed except for Jeroen. But... due to recent permissions shuffles I can't actually grant the proper rights (T219086), so stay tuned...
Sat, Mar 23
Mar 23 2019