Account doesn't seem to exist
Wed, Aug 14
@aaron here's the state of play:
Needs RELEASE-NOTES to complete this task, but core patch CR+2'd as per https://phabricator.wikimedia.org/T191231#5411308
Mon, Aug 12
It should have kept ALL of my settings.
Wed, Aug 7
However, the idea came up that we should prepare a similar maintenance page on our servers for the site in case something major happened with our host and we decided we needed to flip the WikimediaFoundation.org domain back to our servers immediately.
Marking as stalled because this cannot be put live again until some remediation is done, and an incident report written for issues caused on cawiki when pulling data from wikidata
Basically, mediawiki-config shouldn't be sitting behind numerous other patches (such as these MF ones ) while the swat queue is empty
As mentioned in T226056 too... I wonder if we should have some sort of non technical/"laymans" description?
I wonder if there should be a side/parent task for allowing multiple of anything (as appropriate)....
^ to show the re-introduction of the empty "available" section
Will need to get your manager to sign this off
[08:03:04] <Reedy> tgr: If you're bored... https://phabricator.wikimedia.org/T228588 class refactoring in OATHAuth broke 2FA in both our mobile apps [08:03:31] <Reedy> Being questioned about whether the way/stuff we're exposing in clientlogin from authmanager is "right" or we could/should do soemthing else [08:04:17] <tgr> you can provide a custom ID if you want [08:04:47] <Reedy> which I guess is the answer to [08:04:47] <Reedy> >I am wondering why the "id" is an actual PHP class name, rather than a "real" id, like totp, which chould just be provided by the PHP class and therefore would not change in such cases. [08:05:01] <tgr> we haven't really though about clients wanting to depend on a specific ID, the vision was more about clients using the API to get the definitions for some form builder [08:06:46] <tgr> there's AuthenticationRequest::getUniqueId(), although the original intent was to support use cases where you can have multiple instances of the same class (e.g. GoogleLogin where there are multiple Google accounts connected to the user) but there is no harm in using it to provide a fixed string [08:06:54] <tgr> or fake the old value or whatever
Tue, Aug 6
Mon, Aug 5
Sat, Aug 3
Extension wants converting to extension registration before we'd deploy it (should be trivial)
Fri, Aug 2
Ah yeah, forgot about that. Backporting to supported branches is good :)
We don’t bundle MF so we don’t need to
Thu, Aug 1
a new hook was created in OATHAuth to address
Is this actually a bug when the hook in core is lowercased at the start? And I don't think the Hook class does a case insensitive matching...
Wed, Jul 31
I noted on the gerrit patch...
As an aside, I'm poking the various dependancies to add/update their .gitattributes to reduce the number of files we're bringing in with this
Doesn't seem to want to draw me an updated graph :(
Tue, Jul 30
Mon, Jul 29
Sun, Jul 28
Is there something we can do to add (automated) testing for this, or making sure part of the QA of the releases, that this actually works in both versions?
It seems it's broken on Android too :/
Sat, Jul 27
Fri, Jul 26
Wed, Jul 24
Time to move 73 from experimental?