We've removed this configuration option, obsoleting this issue.

We moved away from project auto-association and I think that's sticking, so there's less of a consistency argument here now.

We haven't seen more requests for this in some time, and dumping the database state is a reasonable solution in the general case (running bin/storage upgrade will always bring an old dump up to date, and we try to migrate and respect settings in almost all cases). If we did see more similar requests, I'd probably favor just documenting that as the recommended approach.

We were never able to reproduce this and no one else has reported it in more than a year.

I'm going to close this under the "No Prototype Requests" rule, and since modern policy support would certainly be part of T3578 if we move forward with that.

I haven't seen more reports of this in the last few weeks and haven't experienced it myself, so I'm presuming it's resolved. Here's a nice chart of outbound mail volume with the Y-axis cleverly omitted so that it looks far more impressive than it actually is:

Oh, this pushed last week or so.

I think we're in pretty good shape here now, some minor possible followups in T7858 and elsewhere.

This feature request doesn't describe a problem. Please read the Contributing Feature Requests document for help filing a good feature request.

I think we should keep it around but make it work, there's no reason the options can't function for logged-out users.

D12549 applies the strategy recursively, although we'll probably hit shell argument length limits before we hit 65K branch heads.

This is rooted in a limitation with Mercurial. I've filed a report upstream:

The change breaks "Track Only", breaks repositories without "Track Only", breaks new imports, runs 75x slower than necessary in the cases above, etc.

D12548 probably fixes this. Let me know how things look after that lands.

I need to go hunt all of these down, this is now statuses=open()&projects=PHID-PROJ-...

Pasteboard.co uses the separate-div-with-contentEditable technique as well:

As I recall, the problem isn't getting the event (that's relatively straightforward), it's reading image data out of it. At least as of a year ago, Safari and Firefox did not let you read image data from a paste.

I think I like #1 the best, but #3 is good, and also #2 is good.

This has never worked, and still wasn't practical to implement when I checked last year (T5411#64865).

For context, it sounds like https://phabricator.wikimedia.org/T96238 is the root issue here, specifically https://phabricator.wikimedia.org/T96238#1234459. In general:

In the general case, we need to let you write this rule:

Yeah, it's nice in a gluey sort of way, I just don't see many gluey use cases for it.

But I don't really see any use cases for this, either.

I'm not really convinced that this is hugely valuable, although it seems like it might plausibly be worth building at some point, maybe.

See also T3725.

good task

We don't currently plan to implement this. Instead, we're going to implement triggers. Triggers would let you change task states or projects as they moved between columns, and then you could query by that derived field.

people type random junk in the next block to celebrate their freedom

This is a bug in Phabricator.

Oh, it's the "authored" prebuilt query specifically.

Does someone have detailed repro steps? Here's what I tried:

It would ideally be nice to do POST over Quicksand too, some day, but we can hold that for v5 or whatever.

phantom back-forward issue

I think we should probably just close T7448 as wontfix, it's just a diagnostic error message with no other negative effects in Firefox only and I don't think the complexity of dealing with explicitly tearing the socket down is worth its weight in code.

Good news: still can't repro.

Oh, I wasn't testing in the column and it's not in my 6 most recent theads anymore.

As usual, I can't reproduce this on my machine. :)

The position: static; in this rule causes the issue:

I have half a diff for this, but project slug lookup is a gigantic mess right now.

Global "Commit" rules in Herald can now add or remove projects, so you could write a rule like this: