Closed by commit rP8a7a7dcbf1cb.

How are thumbnails generated? Surely not on-demand per request?

(Personally, I'd prefer it if Gmail allowed filtering on custom headers... but this seems unlikely in the near future)

Closed by commit rPHU120a050b9251.

Ok. I'm working on a diff but I might leave some of that stuff out initially.

OK. A couple of thoughts.

Before I go about implementing this, I want to discuss my possible approach to make sure that it is reasonable.

Yeah sure, fair call.

I've opened an issue on the amazon-s3-php-class project with regards to supporting IAM credentials. See https://github.com/tpyo/amazon-s3-php-class/issues/84.

Increasing the priority for this task because it is something that we would really like to use.

My opinion is that the .arclint file should stay as a JSON file. I think that allowing .arclint to be PHP would be very similar to just writing your own linter engine.

I'm curious as to what you mean... can you provide an example?

Interestingly, the macro is one that I stole from here.

Yeah, ok... I let the command run for as long as it needed to. It took 1min 18sec.

Running the convert command manually and it just seems to hang.

I don't think that I know enough about Phabricator to complete this task.

Yeah I know... it's something to do with this (taken from cpplint.py --help):

The reason for this is that cpplint.py doesn't seem to report the second or third warnings when passed input as stdin.

This stuff should generally be fixed by T4749, although some storage stuff may need to be migrated to use edges.

@epriestley, what are your thoughts on this? I'm happy to just close this maniphest if you do not intend to enable lint.cache by default.

I think that the correct fix here is to remove the releeph.installed configuration option and to, instead, allow installation of Releeph from /applications/view/PhabricatorApplicationReleeph/ as with all other applications.

I wasn't aware of this... As I said, this report is coming to me from a colleague. Let me check with them.

Closed by commit rP1503840cd945.

I agree with you on most parts. I do, however, think that accepted-but-blocked differentials should be treated differently from accepted differentials in the UI. In most cases "blocked" can be inferred based on dependencies. It is probably not worth considering other forms of blocking (time based, for example).

✘ Merged into T5255.

◀ Merged tasks: T5242.

Closed by commit rPc2eff7c2163c.

Closed by commit rP1e2a592ceb36.

I'm curious as to how far away this is? D9204 has been accepted but not closed.

In T5226#5, @epriestley wrote:

• If you're in Git, run git log -n1 <hash> --format=%P
• Does that list a single parent twice?
• If so, does examining the commit provide an explanation of why?

Yeah, I should be able to put a bit of work into this.

Closed by commit rP2f668493a0fb.

Yeah I figured it out. flex-sdk was the way to go. It was pretty straight-forward to setup (besides a few cryptic npm errors).

Let me see if I can fix this.

It appears that I was wrong about this... I just tested this out and received the following output:

Yeah, this issue seems to be related to branches that are not tracked by Diffusion. The following modification to PhabricatorRepositoryManagementParentsWorkflow works in the sense that ./bin/repository parents X no longer throws exceptions.

I think this was due to a local modification I had forgotten about.

Closed by commit rARC7b61faa1927e.

◀ Merged tasks: T5206.

✘ Merged into T4567.

Sorry, I wasn't completely honest in the maniphest description... Actually, one user suggested that it should be possible to do so (without ever having done so before) and the other user is likely mistaken.

I have an idea on what this might be... I noticed that when I ran ./bin/repository parents, it seemed to scan more branches than I was tracking in Diffusion. When setting up the repository, I elected to only track the master branch.