Currently we have over 1500 active herald rules. Over 1000 of them are global.
Some rules triggers builds on Jenkins. Some rules assigns responsible persons/ccs/projects for tasks. Most rules are global and we allow our users modify them.

Yeah, at least as written. You could tweak the logic or I can swap it for something else if that's easier.

• The way to do multiple builds per package is to use a remarkup custom field and separate them with newlines, correct?
• The herald rule should be "Always" instead of something to do with packages because CustomDifferentialRunPackageBuildsHeraldAction has sufficient logic of it's own?

Is the workboard set to sort by priority when you are doing this?

In D14013#156000, @epriestley wrote:

I am currently waiting for divine inspiration to strike and illuminate a path through T9372.

Did this meet the pain threshold to fix Real Soon or would you prefer a followup task?

Meta email testing.

Virtually all of our audits are triggered by Herald and not Owners, roughly following the small teams guide.

~50% of the preference entries include disabling self actions. There is probably a bit of hand waving to get to that that to a percentage 'all users', or 'all active users'.

This changes behavior a little bit: we previously attempted to exclude, e.g., commits which a package you own needs to audit, but which you have resigned from. This is difficult in general and I think it needs a more comprehensive solution. This shouldn't impact users much, anyway.

14 days seemed like enough time to get some reasonable results. They are.... complicated.

• To my surprise I'm not in the top 10 for delivered email.
• Very roughly, people who get more email seem better at handling it.
• Among the 'too much email it is terrible!' crowd some people turned off virtually all email in phabricator while others have a client side filter. The most vocal complainers have usually succeeded in using the available tools to no get email, but at the cost of creating larger coworkers-ignoring-each-other ones.
• Most people are doing some filtering although perhaps ironically some of the best responders send almost everything to their client.
• For anyone looking at the raw numbers, we have a lot of audit related mail.

The "Unbreak Now!" case is the only one I consider important. Similar behavior might be nice for "Needs Triage" + viewerprojects(). I can't think of any case that does not involve tasks since I expect the common case for audits, reviews, feeds, etc to be that there are results.

Sorry I had several instances of phabricator up comparing results and must have messed up which one was which. As you pointed out viewerprojects() specifically (which really is the one I was interested in) is not available on this install.

In D13957#153462, @cburroughs wrote:

A number of ex-Sun engineers I have meant talk fervently about maintaining First Customer Ship Quality at all times. Sadly there are few public tellings of that tail beyond the pale shadow of a wiki page but the impression I have gotten is (1) The crisis that lead into their version of "do not break the build" was a career defining event for many of the engineers (2) The story was told and retold so it was seared into the souls of the next generation of engineers. Probably that means they don't break the build anymore, but it is even more important if they come away with a visceral feeling for the product, not any one project, is what matters ormistakes will happen; negligence cannot.

• clever++, or is that ++clever?

I have not seen this a second time.

I tried a bunch of clever/terser ways but they all ran into 'Unsafe Usage of Dynamic String' or phutil_passthru quoting even the empty string.

When you say it "does not appear to do anything", do you mean that the click is completely swallowed? Or the menu closes but nothing else happens? Or the menu closes and the input is updated (this is the expected behavior), but you aren't taken anywhere and expected to be taken somewhere?

Hmm, also nothing locally or in /mail with ANSR in the subject.

Using Drydock, this will be feasible for hosted and external repositories? (Assuming phabricator has a credential with push access.)

In D13957#152702, @epriestley wrote:

• "Storytelling", which is sort of what I was originally interested in in that task: open-ended questions with no right answer where the question is more like a writing prompt and the answers are somewhat narrative. This is how I used Quora in ~2011, and also roughly how Everything2 (a niche community I was part of many years ago) works. This is a use case I find personally interesting, but I'm increasingly unsure it has any place in the product. In this use case, the value of the answer is as much from reading a piece of writing as from getting the answer to a question.

/me is from a java shop and jumps in.

@joshuaspence Did your stab draw blood?

For binding builds to packages to work I think they need to overlap. For example: Run the 'Super Sweet Selinium Spaces Tests' if src/applications/spaces OR src/infrastructure changes, but src/infrastructure is also a trigger for other packages.

The IRC support was a significant positive factor when we chose Phabricator. I dug up the old evaluation doc and right at the top had "Fanatical upstream support". The fact that it is IRC on freenode was nice since that signals a thing people use (instead of some astroturf fake free support ala google apps). That isn't to say that it will continue to be a worthwhile tradeoff to do all of that support, but I am very glad you did when we were new!

In T7338#125216, @epriestley wrote:

From T8781, monitoring instance daemon queue lengths would be particularly helpful in identifying at least 1-2 issues we've hit.

FWIW The particular commit that was causing trouble eventually (!?!) was processed. Since I could reproduce multiple times times with 4x the RAM this feels spooky. Maybe the memory consumption is non-deterministic? Or somehow very different when running in the daemon instead of the cli?

FWIW Creating an hidden empty file in a directoy seems to be the convention for making git create a directory in the working copy on checkout.

In T7013#130406, @cburroughs wrote:

• As an admin being able to see people's email settings would be nice, but probably has thorny security issues and I think I can just go figure this out from the DB anyway without too much trouble anyway.

I agree that too much email is often a mostly cultural problem. I'm personally only moderately annoyed by email volume, and usually only in cases where a bunch of things happened at around the same time. But I also have 30k unread emails sorted from mailing lists to read on train trips and I suspect most would view that as a bizarre edge case. Batching felt like it might be a salve of sorts for legitimately annoying cases, but I'm not particularly attached to it. Anecdotally some users have reported things along the lines of 'I have used a bunch of other systems and phabricator is the only one I have felt overwhelmed by email with' (I know that feedback is too general to be of much use here.)

In T7013#92738, @epriestley wrote:

No, this refers to real-time notifications only. Users will see no difference in behavior (except an imperceptible delay in some lower-priority real-time notifications). This only improves scalability.

We do not currently plan to ever batch email.

Foodcritic has a few quirks in its interaction with Arcanist

(max 2)

If I give this ticket two tokens do I get two badges?

With my 'I have a corporate install' hat I think 'Alternate Short Domain' is needed for the 'Shortener' case to work out in practice. phabricator.mycompany.com is already longer than an entire t.co url. Those two features are probably enough to replace our current internal shortener thingy.

We have not had any poltergeist sightings recently.

I just realized the examples use the WoW rarity scheme ❤️

For posterity:

• For rough ballparks, a gzipped sql dump of our db is 2.1 GiB.
• On our test instance (fork of the prod db from a few weeks ago) running the migration normally took 30-45 minutes.
• For the production migration we disabled fsync (zfs set sync=disabled tank/lxc/foo) and it took 4m10.926s.

Is it expected that these migrations will take a long time (like > 30 minutes)?

Do you expect handling alert triage to be a good fit for Nuance?

No longer intend to run this in production.

As an Administrator it would be nice if there was a way to tell which version of arcanist was being used. Right now I guess based on depricated conduit calls.

(Wait I may be crazy.)

The client is doing validation of the response including presence of fields and their type.

This is a related wishlist-y thing but it would be really cool if:

• The wiki was a repo
• It could be mirrored in a minimal read-only way (which could mean a lot of things, maybe it renders remarkup but otherwise just links everything?) but sync with the 'main' instance somehow (master with a repo mirror?). Honestly just the raw markup in a repo would probably be good enough to make me switch.

Oh Cowboy Commits giddyup

In T7829#116214, @epriestley wrote:

• We should move to giving the bot a list of channels to join, and a list of handlers to run in each channel -- not a list of handlers to run at top level, and then other top-level config for each handler about channel behaviors. This also resolves issues like T7800.

Is the vision for Diviner that it will become as full featured as something like sphinx (the place to write Serious Docs for all my software projects) or just moderately useful for things that are not Phabricator?

Just keeping patch stack up to date.