We should log all the paths used to get to each page, and then show all of them in a big combo box, and when you select one, update your browser's history to match that path!

This seems to have quieted down, now.

Probably a bad idea but maybe I'll revisit this at some point.

I believe this is now resolved. See D20291 for technical details on the main issue and T13266 for a followup.

Also, what is "By Relevance" ?

I also checked for more leftovers in instances/, but it looks like we don't have any non-default paging there.

I've poked around a fair bit locally, although I'm not totally confident I got every combination of paging columns and queries. I think anything that was missed is very likely to produce an explicit error rather than mysterious behavior, so I'm not too worried about it. Stuff to keep an eye on:

I also vaguely think some mailer API historically rejected mail with a "Message-ID" header with an actual error ("You can't set this header, we're going to set it for you.") rather than just ignoring/replacing it (maybe SendGrid?), but I might be misremembering or that behavior may have changed. We could still always set the header and drop it in the Adapter nowadays, this whole thing is just a nest of hornets and kicking it has pretty limited upside.

I was imagining keeping the Message-ID, while having the first email also include the same value for In-Reply-To — I completely agree that removing the Message-ID would not work out. If the first email always included both fields, thus referencing itself as a reply to itself, I suppose there could be MUAs that would go into infinite loops. So, maybe too risky.

I can imagine that cases like this may exist:

The only effect I can imagine of always setting the In-Reply-To would be to encourage threading where it's not currently happening, so removing the behavior and deleting all these little methods would be very attractive to me if it were my project. :)

There's probably some argument here for removing this behavior and assuming we can never generate a "Message-ID", but I'm hesitant to make changes like that since client threading is largely unknowable black magic which we can not test, and the current approach seems to generally produce acceptable results without an overwhelming amount of collateral complexity.

It worked!

Very happy to see that you guys are working on this, thank you. 👍

This is an unconstructive mess that I don't specifically plan to act on. T13069 and other work may eventually resolve this case indirectly.

report back on whether it mitigates

In T13265#245125, @brhodes wrote:

I merely wanted to have both possible resolutions on the table, one of which was to give the first message the ID later used In-Reply-To.

How old were things before the upgrade? (Roughly: less or more than about a year old?)

This previous behavior violated standards — the In-Reply-To identifier is supposed to be the Message-ID of a previously delivered email — but at least it wove a consistent fiction for mail clients.

Have you possibly changed mailers from "SES" to "SES via SMTP" or something like that? The logic here is basically:

See also https://phabricator.wikimedia.org/T113556 downstream.

See PHI1125. Recent versions of JIRA (JIRA 8?) still work with approximately the same instructions, but you have to fill out a modal dialog with about 7 required fields first, and none of the fields have real values. That is, the configuration instructions for recent JIRA are:

I'm going to consider this resolved at this point. The one thing I haven't built is the "task/point count" from the mocks (on the right-hand side of the header, counting points in each group), but I think this is probably only of much use under "Group by Owner", and maybe not even then. I'm open to building this in a future iteration and it's not likely to be very much work, I'd just like to see feedback that we're generally headed in the right direction first and that the element would have real-world use cases.

Currently, you can put a task in a column in these ways:

Currently, expected behavior.

More or less resolved previously in D14918.

D19549 added Spaces support to projects.

Mooted by D20263, which removes the rarely-used drag functionality from global lists.

We could lock this order into "drag to column" interactions only once those are built (i.e., not support "drag to position" -- basically, when you drop a card, we fling it into the correct position, even if that position is at the bottom of the column far off the screen). This will be somewhat confusing but maybe the least-bad of the options.

To summarize the state of the world here: