Sun, Nov 1
@Aklapper Earlier this year you marked the Gadgets 2.0 raodmap as historical. I totally understand reserving the namespace for future use, but being reserved also means we should hear allocation requests. The original plans have apparently fallen through, so the tentative allocation has practically been forfeited. If communities have plans for the namespace, we should seriously consider allowing usurpation. Similarly, if a city administration plans to build a road and allocates no funds or work to it for nearly a decade, I would bet they'd seriously consider a community request to build a park there instead--they were the ones who slept on the plan after all.
Oct 12 2020
Sep 28 2020
Sep 22 2020
I'm confused. Tech News made it seem like this change would cause users with noratelimit to start hitting limits, but it seems like it's just adding the option to see what the limits would be. Am I correct that rate limits still won't apply to those with noratelimit?
Sep 7 2020
We should consider granting to autoconfirmed by default, though I'm not sure it's the right course. Since it only allows deletion of single revision redirects, I doubt that misuse will lead to serious disruption, and with rate limits any problems should be caught before they get out of hand. That said, since revisions are deleted, any problems would have to be resolved by admins because non-sysops wouldn't be able to figure out where the redirect pointed before the page was deleted. For smaller projects with few admins, that might be a bottleneck to avoid.
Aug 19 2020
Jul 19 2020
This is a common problem when FlaggedRevisions is used with an autoexpiry. On enWikipedia editors will often request protection after the PC protection auto-expired and it's difficult for sysops to evaluate the need for further protection because the revisions are no longer flagged. I'll look into whether this can be improved.
Jul 4 2020
Jul 3 2020
@Xaosflux Yep, I've just uploaded a revised patch and the builds are passing. Presuming the CR goes well, it should be ready.
Jun 1 2020
I’m guessing either you’re pushing to gerrit incorrectly or you don’t have rights to amend other people’s patches...
I've got a revised patch ready which doesn't have a merge conflict and does include tests, but I'm having trouble getting it submitted to gerrit; my patches keep getting rejected by remote. I'm going to keep troubleshooting, but for those who want to try their hand, Gergő Tisza's changes need to be moved to includes/Permissions/PermissionManager.php because the modified functions were moved out of Title.php in gerrit patch 495832
May 18 2020
Mar 23 2020
You may want to take a look at the Chess diagram template and its associated lua module (also written by @Kipod). I'm not sure how easy it is to turn lua modules into extensions, but if we want to support static, single position diagrams, it would probably be easier to start there since it already has features such as variable board sizes, fairy chess pieces, and on-board annotations like arrows.
Mar 5 2020
Mar 4 2020
Mar 3 2020
Mar 2 2020
Feb 28 2020
Feb 6 2020
@Jcross Thanks for the triage. That is the goal, yes, however no WMF team is sponsoring this project yet so T244075 has been stalled. You should probably triage this task as lowest priority or stall it (depending on your workflow) until I can talk more with WMF stakeholders. Thanks!
Feb 4 2020
@Jdforrester-WMF No worries! Would you be able to point me to a person or documentation that could explain this more and how to go about resolving it? Needing a WMF staff sponsor wasn't mentioned in the mediawiki documentation, so I'm not entirely sure what to do next. The English Wikipedia community has wanted something like this for years, but every plan to implement it has been stalled. Any information on how to get this unstuck would be appreciated.
Feb 3 2020
@Jdforrester-WMF No staff are sponsoring it that I'm aware of; so far this is a volunteer effort.
Feb 2 2020
Jan 14 2020
Jan 9 2020
@Kipod the changes in 563005 implement some minimal tests for those three cases. The code currently only supports castling in standard board set ups, though presumably that can be changed at some point in the future if/when we can support chess960. You can see on line 152 of ChessParser.php that the source and target squares of the rook are hardcoded, but could be generalized to "other side of the king" in order to support non-standard start positions.
Jan 7 2020
Jan 2 2020
Dec 18 2019
Dec 14 2019
Dec 13 2019
So it seems related to my custom css (w:User:Wugapodes/minerva.css), however it only changes link colors and was working yesterday. Has the interface been edited to add the "visited" class?
Dec 12 2019
You may want to see T240611; the problem for me was a use OutputPage; line which caused these errors.
The exceptions were caused, unsurprisingly, by line 3 of ChessBrowser.php which contained use OutputPage;. PHPunit warned that this statement (and similar) were meaningless. Removing the line entirely resolved the errors.
wgInstantCommons is already set to false by default. I tried setting it to true and the tests still fail.
Related to T240611
Dec 2 2019
I've got tests working for the __construct(), cut(), getChar(), and parseVariation() functions of PgnParser, as well as test for the exceptions they throw.
Dec 1 2019
Probably helpful to read this man page on writing tests for extensions. The necessary tests will have to cover much more than what is stated in the original task, but those are a good place to start.