I'm obviously low on AM brian power today.

Heh, that was going to be my suggestion in the original discourse post, but I figured it was too simplistic. Glad to hear we're on the same page. Thanks everyone!

That sounds much easier than all the solutions I came up with.

Is there any reason that the rule "if any menu items have the same link target as the request URI, highlight the first one" wouldn't work?

Or maybe there is a simple way to get the link to highlight, I couldn't think of one though.

A link doesn't stay highlighted if clicked on. Like dashboards, I can understand wanting multiple board views from the sidebar.

• What's the root problem?
• Why can't they use links?
• How is this different from T12374?

Closing in favor of Discourse as the main user hangout spot for now.

We should probably just remove this workflow, it's not clear to me that there's ever any reason for anyone to run bin/files purge in the modern codebase.

Oh, I didn't even expect that option to be configurable on Phacility given it's a security related setting. I'll turn it on and do some speed tests next week to see if I get any measurable difference in cloning.

Have you enabled diffusion.allow-http-auth?

When you click Clone in the Phacility UI on a repository, it doesn't show any HTTPS URLs. It's possible it works if you copy the URL from the address bar, but the UI in Phacility itself doesn't give any kind of indication that it will work.

In the past there used to be a Browse button at the repository homepage in Diffusion. Clicking on it would show the file tree for the current branch and some extra buttons. You could then click Search to grep the file contents of the current branch.

Feel free to use our Discourse for general admin questions (as opposed to our working tasks). Thank you!

@ofbeaton Thanks for the tips. Sounds better than the approach we are using to slightly modify the code base.

Just keeping an eye on this: growth on these shards is slow and both are still below the 80% caution threshold, so I'm punting this for another week.

@jessjohnson We use puppet to deploy phabricator stable from a copy on a local file server, then have it run a shell script that applies our patches using git apply on .patch files. Lastly we restart services.

@ofbeaton Just curious what's your strategy on applying patches as part of your deployment for phabricator? Do you use a find/replace strategy? Apply diffs?

I'm seeing these errors failing regularly:

I've got a reproduction case for this now: https://secure.phabricator.com/differential/diff/44008/

Most browsers can only give us a timezone offset, like "UTC-7", not a locale name like "America/Los_Angeles". We already attempt to read this value (D15962, T3025#176547) but when I wrote that stuff it was only available in Chrome and not necessarily reliable. We do try to guess the default zone selection based on the information we get from the browser, but currently ask the user to confirm it in all cases.

I was thinking that the first time they log in, we just take the browser timezone and set the user config to that.

@epriestly I'd lost that connection, so reachability, it might appear.

@kfsone Are you aware of arc diff --only (briefly mentioned above -- "just generate a diff to look at in the web UI, without other slow/complicated steps")? How do you picture arc selfreview differing from arc diff --only?

@epriestley I think you under-estimate the potential and value of your own product.

The simplest way to approach this is probably to do double writes: start writing the new tables, switch readers over, stop writing the old tables.

This could also just go on the existing page at the bottom. We tell people to copy/paste it when submitting bug reports, but it's probably fine if they're including more information, especially since there is a nice walled garden moat around my personally having to deal with it.

See also PHI22.

T12941 is a third case of a user with no mysql or mysqli.

Let me know if that diff works.

I honestly can't get the cool looking macos dropdown on firefox no matter what I try.

The particulars in T8918 are slightly different, but the root issue (showing too much stuff to users with partial sessions) is the same.

Agreed. I haven't experienced the problem since I upgraded, so I think it was related to an earlier fix, even if it wasn't the identified fix (which should have already been in my install when I did have the problems). There's nothing that needs to be addressed here.

Mooooooo

This is vague and may not actually identify a problem; I don't plan to pursue it.

We've closed registration on this install, mooting this.

Dropping these doesn't seem to have caused any major pushback.

This presumably counts as "not much interest" or should move to Discourse now or something, maybe.

This seems to mostly be a matter of taste, and we already have behavior similar to the behavior of git, which seems reasonable.

We don't have reproduction steps for this and thus can't move forward.

This doesn't affect us, but more fuel on the fire -- XSS in example code which ships with PHPMailer (we don't distribute this code):

We appear to have survived this.

Per above, this is not a bug.

No one but me cares.

We kinda did some of this, it can live on in subtasks.

This appears to lack reproduction steps, so we can't move forward.

We aren't going to implement a bin/phd start-missing-daemon command.