Some states:

• shouldBroadcast() has been removed; getShouldBroadcast() now means the same thing.

• getHasBroadcast(), setHasBroadcast(), and PROPERTY_HAS_BROADCAST are now getShouldBroadcast(), etc.

• Implementing BuildableInterface now requires implementing newBuildableEngine().
• Implementing BuildableInterface no longer requires implementing getHarbormasterPublishablePHID(). Publishing indirection is now handled by BuildableEngine.
• PhabricatorTransactions::TYPE_BUILDABLE no longer exists. Instead, applications should implement application-specific modular transactions to report build status information. DiffusionCommitBuildableTransaction may be useful as a template.

Duplicate of T8936?

As with @michaeloa's case above, this is a blocker to porting repos of ours over (from Gitlab EE, in our case) for testing purposes.

I'd like to analyze data on the $5 tier with an eye to removing it. Anecdotally, this doesn't seem to be doing anything good for us ... It also feels like we have a lot of $5 test instances.

The UI and allocation logic for shards need tweaking.

When this eventually moves forward, it's probably also appropriate to remove the syntax-highlighter.engine config option. I think this has essentially zero use in the wild and we have better modern patterns to approach the problem (without extreme-niche config options) if there are real-world use cases.

I confirm the Maxime Corteel patch to be working with PHPUnit 7.0.1 too.

GHC recently upgraded its Phabricator installation and has been rather unpleasantly affected by this feature. We have multi-hour CI jobs, which often queue to a day or so of builds. Making contributors wait this long before receiving feedback is turning out to be rather problematic.

Oh, and I found the beast to get a portal to Saqawal's Roost and killed Saqawal and got the boots, but the boots and set in general seem kinda bad so they're sitting in one of my six tabs full of unique items that I will never ever sell or do anything with.

D19269 makes Harbormaster objects destructible. I've made Build destroy BuildMessage (which it doesn't receive, so this never destroys anything) but not BuildCommand (which it does receive), so that when this change is made (and it starts receiving BuildMessage instead of BuildCommand) things will just work. Until then, BuildCommand objects are technically left hanging, but destroying builds is exceptionally rare so I'm just assuming that this will probably be fixed before anyone notices.

Hey we actually have the same issue-- folks on the team would like to edit "owning team" from within the action menu. The problem with having to edit the task is that the UI for task editing is less discoverable and requires more clicks than the action menu -- folks are used to being able to accomplish what they need from the action menu so it frustrates them when they can't. This isn't a blocking impediment to our workflows or anything but it would be a nice-to-have UI improvement.

I think this is sort of less bad/weird now, at least. You get a tokenizer instead, which is still empty, but you're prompted to "Type a repository name..." and the UI isn't this odd empty select thing anymore.

D19266 adds some additional Support Pact adventure mode content. Some other ideas for improving this mode:

I've barely had a chance to play recently but I know everyone is longing for an update so: I made it to level 88 (which I think is the highest level I've reached on any character), and completed all the white maps, and haven't died yet.

See PHI496. Multiple users have now reached the maximum level ("Heavy Wizard") on their Support characters, and we're overdue for an expansion pack to add new adventure content.

Much of this survives in T13110 but the rest is resolved.

okay. thanks!

See Planning.

Hello @epriestley, could you please provide any update on this task?
Won'tFix or InQueue ?

After clicking 17,000 emails I successfully transferred everything to AWS, with some minor caveats:

I moved SSL and registration for phurl.io to AWS in T13113 so this is just a transfer issue now.

oh wow it actually worked 😱

AWS also supports DNS-based authorization now, which reduces the need for all the MX juggling.

I was able to MX phurl.io and get an SSL authorization link working. I moved phurl.io SSL to AWS ACM so this (certificate expiration) shouldn't happen again.

(See also T7125, but this particular problem has conceptually been solved for every domain except phurl.io by using AWS certificates.)

See also T12917 and a brief mention in T13062.

Changeset rPfb4ce851c4ae: Add a PDF document "rendering" engine allows to view PDFs from the Files application. However, {Fxxx, ...} references to a PDF file in Phriction still only allow to download the file or open a lightbox with only an icon (again with a download link).

im leaning towards ethical. someone traded me a 3 ex helmet for 40c worth of regrets. I looked up how much the helm was and demanded they trade me 40c for their helm back xD

the best part of being so unethical

Actually, it seems like rel="noreferrer" fixes this. This is bizarre so maybe this is a problem with a spooky ghost haunting my computer?

I really like this, Remarkup idea, I think there are many text based files that people add to reviews which would be easily transformed to something more readable by being transfer directly to Remarkup, even for example something as simple as CSV,