Do git ls-remote (which, curiously, seems to be significantly faster than git fetch even when git fetch is a no-op).

I think T12393 and this probably have a similar set of root causes.

Ah! That fixed it, thanks @epriestley. I'll let you know once I test out that configuration.

@epriestley I'm in the process of setting up phabricator to help mature the development workflow based on svn here, and make future transitions to git/perforce easier. Changelists would be a natural path here but the workaround for our use case is to just manually specify paths, but it's annoying if you have a locally modified file that you accidentally include in a diff, since you *have* to revert it to remove it from the diff. Using a changelist would let you avoid making the mistake in the first place.

Remarkup is cached, so you'll need to run bin/cache purge --purge-remarkup (or make an edit, or create a new post) to purge the cache.

@epriestley, I wouldn't mind testing that scenario out this weekend (I'll have to change a couple of domain name settings to create the scenario, but that's easily undone).

@stettberger: If there is anything missing from the WMF policy API extension, please feel free to file tasks under https://phabricator.wikimedia.org/tag/wikimedia_phabricator_extensions/ or open a differential revision against https://phabricator.wikimedia.org/source/phab-extensions/. I will be glad to review patches, I can't make any promises about feature requests but reasonable changes will be considered.

Brand new install on another VM, same Database server - same error. So can assume that the bug comes from data stored in database.

You can use any saved search query in Dashboards, the trick is you need to go build and save the query first. You don't construct it in Dashboards.

To define 6 pages yourself: use a text panel, then link to 6 pages in the text. For example, you might use text like this:

Ahh that did the trick - thanks.

The "Policy" task says "implement PolicyInterface", but the interface is already implemented so that Query works. The description should be adjusted to be more accurate.

Something that would be useful to us would be the ability to create, or probably better, assign a task, as it nears a due date. E.g. we may set up a task due in a year's time, and anticipate working on it two months before it is due. It would be great to be able to schedule the task to be assigned on a particular date, and perhaps moved on a workboard. I don't think this would need to be any more complicated than a Herald action that runs at a particular date/time rather than in response to some other action. I'm not familiar with what Calendar can already do, as we don't have prototype applications turned on for our install, so apologies if this is already possible. I just thought I would mention this use case here for future reference.

In T12372#215275, @epriestley wrote:

Great, thanks for testing it! We'll upstream that and I'll file something to put a longer-term fix in place.

"witter.com" <fakeemailtest+163@t>

ah yes this is my very favorite email address of all

Great, thanks for testing it! We'll upstream that and I'll file something to put a longer-term fix in place.

Confirming that the patch above works (tested using send-test --to)!

See Planning

What happened with Phrequent Prototype?
We are looking for a time tracking plugin for Phabircator as well. Is there any plan to have one integrated at some point?

Then, I will use the WMF extension for now. I understand that you won't prioritize this. Just for future reference, I will document the structure we have come up with : I currently employ already a more hierarchical project/subproject structure that can be use over several semesters and makes it possible for me to reference all students at once:

Old-school resigns, accepts, and rejects are still using edges directly (can we get rid of these completely?).

These pathways are still dodging ModularTransactions and thus not getting double writes after D17495:

Specifically, from the error, since X and Y are consistent in the error part and the fetch part, it appears that this process has lost a race to another process which is doing the same fetch.

For now, "Edit Menu" on your home, click "Global" and add the Dashboard. "Default" is the topmost dashboard. You can disable "Home" then. This will be easier after D17492 lands.

How exactly is this currently possible? I am trying to change the default homepage for all users but not managing to do that. Have tried creating a Dashboard, but not sure how I can use that to 'replace' the homepage for everyone.

(Presuming this is resolved since it's been open for a while without feedback.)

Feel free to file a new report if you are able to come up with clear, working reproduction steps that reproduce this issue on a clean install.

Some workarounds in the short term:

This is likely just some of the secure hosts having different configuration. Ideally, we'd get the rCORE deployment stuff aligned to deploy them with it.

We may eventually build this for T7413 but I don't think it's a good use of our time without a strong, narrow motivating use case like that.

This doesn't appear to be moving toward becoming an actionable report.

Any ideas on how we can move forward with this?

These CSS rules:

OK weird, I can't reproduce it anymore. Anyway, I'll add the response body from differential.querydiffs here. It shouldn't contain any sensitive information, but I'll restrict the visibility of the file given that it does contain email addresses.

rLOCATIONS now has two branches:

Sorry, we don't take bug reports as support requests. Have you read and followed Contributing Bug Reports?

Please let us know if we have missed out any settings.

After installing the phabricator we created a new repository.

When the repository is in inactive state the "ssh git clone...." of the repository is shown up.
When we activate the repository and try to see the repository mail screen, the below error is showing up.

Bug reports need complete reproduction instructions and complete version information. See Contributing Bug Reports, Providing Reproduction Steps, and Providing Version Information.

(You can use T12270 to give feedback or report bugs on Badges, specifically)

Please read and follow Contributing Bug Reports. Also, we don't accept bug reports against prototype applications.

Yes, it is not "incorrect" but it is not human-readable format problem same with T8973 either. Chinese / Japanese / Korean unicode should be shown as what it is instead of some unreadable unicode digitals, especially in TRANSLATION functions, otherwise how to translate and confirm that?

Did you try to enable PHP's error logging to get more output?

That's the big problem, the logs say nothing.

Bug reports need complete reproduction instructions and complete version information. See Contributing Bug Reports, Providing Reproduction Steps, and Providing Version Information.

Tried again with node v6.9.4 (current version of node with CentOS7), and still has same problem 😢

Happy to accept a patch for this (deleting that one line of code) if someone wants to claim that they tested the scenario above and it worked correctly.