@epriestley Noticed you'd been doing proximate work; I'm not planning to push this further forward, so please feel free to commandeer this revision.

In T10890#231066, @epriestley wrote:

I don't think this clearly describes a root problem. PHI33 touches on similar issues, and I'll file something vaguely in this realm if anything comes of that.

@epriestley Thanks so much for the detailed guidance! Sorry that my turnaround time hasn't quite justified the work you put into this: my time slots are quite fleeting, as I only pretend to be a web guy from time to time.

Per commit message, this is not at all suitable for merging in its current form.

moar staging area

In T12314#217334, @shandor wrote:

The interaction between subtypes and "required" flag was mentioned briefly in the weaknesses section, but it's a little unclear how this new feature will work with them. With Custom Forms, we are currently unable to use Required custom fields properly because they are still required fields even when hidden in some particular forms. Do Task Subtypes solve this issue naturally, or is it something that still needs to be implemented?

In T550#202580, @epriestley wrote:

Multiple hostname stuff: Today, the SSH hostname of an install and the HTTP hostname do not need to be the same.

We use this in the Phacility cluster: Phacility HTTP instances are at instance.phacility.com, but serve repositories via vault.phacility.com.

The technical reason for this is mostly that ELBs can't accept port 22 traffic -- or couldn't when I set them up. It looks like new VPC-ELBs may be able to accept on 22. We currently use HAProxy to balance SSH traffic and ELBs for HTTP traffic.

We could require that all hostnames be the same and change the cluster to bounce SSH traffic through the ELB to the HAProxy LBs and use new VPC ELBs, or put HAProxy in front, or whatever else. But this is probably more work for us than letting them differ.

There are also some possible arguments for using a different hostname for Conduit CLI SSH access (conduit.whatever.com) that has different DNS-based load balancing behavior than other SSH access.

@epriestley Thanks, that's fixed it! I'd not looked index engines at all before, so wasn't sure if I was just missing something stupid on that side.

OK, this is interesting: having tried reproducing it on a test Phacility instance, it works as expected there. Perhaps one difference is that my local installations don't have Diffusion serving repositories itself over, e.g., SSH? These repositories do not have any URIs other than a remote Git URI which Diffusion is observing.

In T3353#196761, @chad wrote:

Can you do that today with view and edit policy?

In T3353#179075, @epriestley wrote:

• Text/Phriction Inline Comments: See T1894 for followup. I'll update this shortly in light of T7643.

Yep, my bad: I'd fixed it behind that commit, spotted the commits ahead before the rebase, but wasn't actually testing what I thought I was, so thought it was still there in master. (And sorry for the delay, was out all last week.)

[third time's a charm ...]

[no change, just pushing to staging]

@epriestley Here's an anonymised version of a live workboard. I've completely wiped out all identifying data; I thought about summarising the task titles rather than obliterating them, but afraid I lost the will to live at some point.

In T8145#113258, @epriestley wrote:

Does anyone have a use case for this which isn't maintenance notifications?

@epriestley Is there a pathway for reading (not writing) edges which would be open to community contribution? Having arc patch Tnnnn functionality, pulling all the revisions attached to a particular task, would be really attractive for us as we tend to have a larger number of revisions per task.

Also be careful that the 'Depends on:' commit message parsing is strictly additive; if you ever change patch order in any way, your dependencies will get out of whack. If you're doing this, you can modify DifferentialTransactionEditor to reset the dependencies before parsing, but this means you throw away any dependencies manually added through the web UI.

arc amend now succeeds, overriding the global config.

arc amend now works, overriding the global history.immutable configuration.

Unlike previously, this results in arc amend succeeding, rather than complaining.

@avivey Reasonable enough. I'd hoped to be able to use socket paths explicitly, but that would be a separate change to this then. Thanks, and sorry for the noise.

In T4316#49406, @epriestley wrote:

The comment should be merged into the same row if you perform an edit and then leave a comment, as long as it takes you less than 2-3 minutes to write your comment and no intervening activity takes place.

@avivey: Is there any particular rationale to doing this? I'm looking at going in the other direction - completely away from TCP - and this makes that somewhat harder ...

@epriestley Yeah, that would certainly be really useful for some cases; I imagine infrastructurally it'd be a good step towards this as well, along with T10569. I don't think it would help this particular usecase though.

In T4863#152912, @epriestley wrote:

• Diff/Mock: These are fairly straightforward, alhtough I'm not totally convinced that they're valuable. At least in my personal workflow, I don't imagine I'd be likely to ever use them, and I don't recall other requests for them offhand. I'm hesitant about building new hard-coded first-party application stuff, too, since we've made so much progress on getting rid of it. I could build this in a generic way, of course, but that makes it less straightforward.

@epriestley Missed a spot: D16325

In T11364#186934, @epriestley wrote:

Specifically, I'd like to understand what concrete pathway led you to encounter this. A list of hypothetical situations isn't useful in anticipating or recovering from broader problems.

@epriestley Because it allows the reader of this report to reproduce the underlying issue without cloning anything or clobbering their .arcrc. Here are some less contrived scenarios, which are not mutually exclusive:

• the user makes a typo (project lacking .arcconfig, or trying to use Phabricator for the first time themselves)
• the .arcconfig comes from a time before Let's Encrypt (e.g. if you're in the middle of a git bisect that takes you back some distance)
• the admin is just attempting to migrate people to more secure services and does not expect that services will fail (or at least not fail in such an opaque and confusing manner) in the face of a redirect

@benjumanji history.immutable does not actually cover it, but this task is probably the wrong place to begin that entire discussion again.

In T5000#176256, @epriestley wrote:

(I am far less convinced of the value of forking or git push-to-review, except as onboarding tools).

@cspeckmim Right, that was an edge case indeed, but even very active tasks can rapidly pick up a huge number of subscribers/tokens/etc (something I'm seeing on my install). I'd personally aim for a collapsed-but-expandable '[+] 16 users subscribed, 7 tokens awarded', but this is definitely getting deep into design territory.

@epriestley Is there a way to disable T9554's heuristics? With patchsets of 30+ patches which get pushed in a single atomic bundle, we don't want to go back to import mode and lose all notifications.

@epriestley Thanks for the thorough response. I guess it's probably been fixed in intermediate revisions, and as you say, it is all changing pretty dramatically anyway, so no big deal: as it was a newly-created repository, I just scrapped it and created anew.

Now, I do agree with you this interaction is imperfect and could be improved, and your feedback is duly noted, but we're a team of 2 here with 1500 tasks open, so some things are just be imperfect until next time we're in the product.

@chad 'No points' or simply showing nothing is less information. An incomplete word isn't a deliberate choice to show less information, it's how it happened to land. Anyway, seems like you consider it a feature, so I'll leave it alo

@chad If the verbosity is essential, perhaps it should be word-wrapped in the collapsed case, rather than truncated?

@chad OK, fair enough; I guess just close this and I'll build something locally.

From T10891:
When you have the side project profile panels collapsed, then go to create a new milestone, it will have no tasks. The profile panel then shows 'This milestone has no tasks yet', but brutally cut off to 'This mileston-' in the collapsed view.

In T10163#154382, @epriestley wrote:

• D15039: Two effects:
  • If you finish typing a word (e.g., type a space) and there are no results, deactivate since nothing else you could type could possibly cause matches and you're probably not fishing around since you just hit space.

@epriestley Maybe one to consider: you must have a default workboard column for a project; the default workboard column cannot be hidden; milestone workboard columns strictly sort after project workboard columns. The combination of these three means that I have a project which is strictly split into four milestones, but its overall workboard view always includes an empty column before each milestone column. Is it worth filing a task for this?

In T9161#136555, @epriestley wrote:

My personal, completely subjective experience is that Phabricator mail is trivial to deal with, almost all valuable, takes very little of my time, and I want to receive >95% of the mail that I receive, and that I'd barely notice if I had 10x this mail volume.

In T3179#162681, @dolftax wrote:

In T3179#111010, @anda wrote:

Has anybody heard of someone migrating jira issues to phabricator? We're thinking about doing that, and of there were some example scripts for the import to start with, that Spaulding be great.

We probably have around 1500 issues to migrate, with a few different jira project names /labels. Of guess that this would best be translated to a label in phabricator.

Did you find a JIRA -> Phabricator migration script?

@dancunningham @readevalprint Also (and as a reference for anyone else
landing on this page), the actual upstream is at
https://github.com/wikimedia/phabricator-extensions-Sprint - which does
work with current Phabricator. But please do take issues up with them.

@ccope: Yeah, (almost) functionally identical; just a bit slower. The 'almost' is that if you have a branched named T123 in order to automatically link revisions to tasks, your cherry-pick method will succeed (as the branch at time of 'arc diff' is T123), but the rebase method won't (as you'll be on a detached HEAD at the time).

@ccope Try running this over your branch:
git rebase -i --exec 'arc diff -C HEAD HEAD^' origin/master

Yet another set of scripts ...

@chad Thanks; that is indeed our instance. The import from Bugzilla in this case is done by @thiblahute, who can shed some more light on it, but I have some ideas ...

Thanks @couture! Sorry mine vanished: it took quite some time to get the CLA signed and I lost all time after that.

In T4953#105563, @epriestley wrote:

This is unexpected. Please file this as a separate bug with a repro case if you've experienced it.

Specifically, where in your workflow would you run arc diff on changes which might or might not have been sent for review already without knowing?

@staticshock I think I'm seeing the same thing, except not quite. I have a linear patch history, which I submitted to http://phabricator.freedesktop.org (open access) as D1 through D17 (not in any upstream repositories; only in my local tree and Differential). Each patch was created individually with arc diff, per rationale in T7076.

Looking at my Differential dashboard, I suspect this results from the combination of Blocking Others (ordered D13..D17), and Action Required (ordered D17..D13).

@chad Ha, no. It's a pretty terrible hack to manually construct the tags directly from within buildPropertyView; pulling the revision object and then building up each of colour/icon/etc with a bunch of manual switch statements. Lots of copy-and-paste. If using PhabricatorStatusUI on the revisions is the way to go, I'll see how that works tomorrow and if there's anything I need to do with the revision class itself to implement that?