That did it. Thanks so much

Will rebase on D10304 once that lands / the locking issue is solved.

Cool, thanks for checking!

confirmed with new code.

Awesome. So what about the rest? I'm happy with keeping this in our branch (I think it's never led to a merge problem), but I also think it's super useful to at least have documented somewhere if we don't want to put this into the main branch (it took me a day to figure out what the right solution is (after talking to some of my colleagues), so if somebody else hits this they'll probably also be wasting a lot of time on it; figuring out why mail is rejected is about the least fun thing one can work on).

Screenshot?

Add myself as a follower.

Closed by commit rP5449200972db (authored by @epriestley).

Okay. The logic currently skips password resets if that's disabled, but offhand that doesn't make sense to me. Let me dig through the changelog and see if I can figure out why we're doing that.

Yes.

I pulled the quoted-printable bit into D10303, since another install hit this issue. Hopefully this doesn't break like 3 other mailers. :/

Testing the new locking behaviour now, will report back when I've verified that it behaves correctly.

...I have no idea what makes for a good high chair. :/

Closed by commit rP6f85c22faffd.

Closed by commit rP6f85c22faffd (authored by @hach-que).

Closed by commit rP94cdddc211d1 (authored by @epriestley).

Had a user actually hit this.

Closed by commit rPeacfb5b489ce (authored by @chad).

Include the working directory fix in D10293 against the new EC2 allocator as well

Worked fine for me locally.

Closed by commit rPe8ece70ee0f2 (authored by @epriestley).

Update based on feedback

Use PhabricatorGlobalLock. Will test this when I'm in the office since I don't want to be calling AWS APIs on a sketchy mobile internet connection.

Closed by commit rPdf7fb098451c (authored by @hach-que).

Oh okay, that explains why I had all those issues :)

Use $lock->lock(123); to wait 123 seconds for the lock.

Is there a way we can detect if there is no HarbormasterBuildWorker in the queue, but the build has commands against it, and only offer the button in those circumstances? I don't want to confuse people with a button that does nothing 99% of the time.

So from memory, I started with something like:

That is, I would expect that PhutilGlobalLock is the correct lock type to use here and will be far simpler than using row locks.

You mentioned elsewhere that you moved away from PhutilGlobalLock to beginReadLocking() -- can you walk me through that?

wow those baby chairs are super expensive

I don't think we should pursue this.

Closed by commit rP0057cf17e9d0.

Closed by commit rP0057cf17e9d0.

Closed by commit rP0057cf17e9d0 (authored by @chad).

Also, @btrahan, which highchair should I go with. Leaning towards the one that looks like it came from Ork.

Is account.editable disabled in your configuration?

Specifically, I think this issue is related to the enormous length of the command. The command is supposed to look like:

Do you have ~10,000 branch heads or bookmarks in this repository?

The current structure has implicit keys "0", "1", "2", etc, so you can just make those implicit keys explicit. For example, change this:

• By convention, prefer to write this like other key patches (see below).
• Prefer to provide an explicit key name.
• Provide a database name explicitly.

Are you using PhabricatorDraft, explicitly? Can you show me that code? I don't think we write PhabricatorDraft objects implicitly right now.

Closed by commit rPe8c51cd934da.

Closed by commit rPe8c51cd934da (authored by @epriestley).

It looks like I have to pay for a "Windows Azure Subscription" (I guess?) in order to create an OAuth application, although I'm having trouble even doing that (the form has a section called "Payment Information" but won't let me enter any, and the "Sign Up" button is greyed out).

... and it looks like --cacert is old news: http://curl.haxx.se/mail/archive-2013-10/0036.html. So I'm not sure how arc worked on Mavericks unless the certs registered with the keychain changed in 10.10 DP6.

Even setting CURLOPT_SSL_VERIFYPEER to false doesn't make the SSL error go away. I was able to get arc temporarily working again by not setting CURLOPT_CAINFO.

Nothing related was mentioned in DP6's release note: https://developer.apple.com/library/prerelease/mac/releasenotes/General/rn-osx-10.10/index.html

arc was fine up to beta 5. But something changed in beta 6...

I found Yosemite completely unusable for many reasons.

The security risk is that if a markdown implementation does not escape HTML correctly, an attacker with access to commit to any repository (or create a repository) can push a README file which renders with Javascript in it, then trick a user into viewing the page. After the user clicks the link, the Javascript can read their CSRF tokens and make requests which impersonate them.

git version 1.9.4.msysgit.0, on windows (git bash).

Possibly a windows / "\r" issue

Yes, I am.

• derp