The changelog for PHP 7.3.6 mentions "Fixed possible crashes, because of inconsistent PCRE cache and opcache SHM reset." which might be relevant, but no bug link so who knows.

As a note, we make quite a bit of use of --arcrc-file currently to "handle" multiple people working as root on dev VMs, but we can probably migrate that over to --conduit-token anyway (although we're a way off hitting this, because as much as I want drafts, other people consider T4678: Support changelist-based SVN workflows in Arcanist and Differential critical and we're on a fork because of that.)

My knowledge is centred around broadcast video / IPTV (UDP multicast) rather than HTML5 video, but that seems fairly reasonable. I don't think anything should try and download the file twice unless it is a complete clownshoes implementation that probably has multiple other serious bugs wasting bandwidth. I think the worst case with no type that might be hit here is that the browser could decide to download the entire file on load to figure out the format and duration, rather than starting with byte-range requests, but this is easy to test and unlikely (I'd expect any sane one to always request chunks).

mediainfo is the tool for this, but is really overkill for programmatic use.

Previously, in PHI386, an unusually large repository for the same instance took about 5 hours to git gc.

As a note, pv can rate limit a pipe and is generally very useful all around.

I would accept to have to issue some very long winded command on the console...

Have you tried changing SSH options on your end, such as disabling compression or using weaker encryption ciphers (such as arcfour, no encryption cant be done without agreement of the server)? This can make a significant difference with SSH transfer speeds, and may function as a decent workaround for now.

This is not a demo or test install. Please don't make a mess! You can launch a demo install on Phacility if you want to poke around and try things out.

In T12516#218475, @rabahmeradi wrote:

Is this really the new 'Needs review' icon or I have problem with fonts?

@avivey Trust Email Address is meant to handle that though, is it not? The reported "bug" here is that the email addresses are not being verified when coming from LDAP, not that email verification is required.

You got linked to https://secure.phabricator.com/book/phabricator/article/configuring_preamble/, right?

This is intentional (and not a security issue), the URI contains a security token valid for 30 days (or until the policy is changed).

the -l xxxxx is your configured SSH user, the -i xxxxx is the path to your configured SSH private key (it's written out to a temporary file).

As it says clearly on the home page, this is not a test install - if you want to try something launch a free Phacility test instance.

Are your clocks set correctly? The only sure-fire way for special-snowflake-issue support is Consulting.

This is the Phabricator install for Phabricator development, looks like you're probably in the wrong place.

@davidism You probably don't want to be posting full memory core dumps publicly.

Was there any indication that the form wasn't working / wouldn't have worked?

bin/auth recover is a general-purpose flow not specific to this case.

@epriestley am I good to take a stab here at implementing the T3025-like behaviour?
If user is admin and clientProto != serverProto, display notification.
Approximately 5% familiar with Phabricator's JS side, but always interested in poking around more of the codebase - and I'm guessing most of this is gonna be touched in D15961.

Configuring the Preamble script is important for reverse-proxied configurations (you probably need to set REMOTE_ADDR up as well), and required for many things - which is why it is part of the install guide.

We ran into this on a recent install - the fix was to configure the Preamble script to tell Phabricator it was being accessed over HTTPS (else it filtered out the HTTPS notification server) - check to see if "Connected" or "Disconnected" at the bottom of the notification popup (menu at the top of the screen - see this install). If you just get a grey bar at the bottom without either, you're hitting the same case where the client doesn't think it has a valid notification server.

See Configuring a Preamble Script for instructions on informing Phabricator that the client is connecting over SSL.

In T2714#119487, @Yomi wrote:

Totally off-topic, but how do you pronounce Releeph?

In T4103#97669, @devurandom wrote:

Brings up the next problem for me: How do I query tasks of a certain priority? I cannot even find a documentation of the query language used for the Advanced Search.

Sendgrid han an option to insert an Unsubscribe link into emails, clicking this blocks the email address in the Sendgrid account, it needs to be removed from there.

I'd quite like the option of globally of disabling All Users - much like Public has to be explicitly enabled.

In T5715#4, @epriestley wrote:

This might be a symptom of T2581.

You would need to write a custom solution using Bitbucket and Phabricator's respective APIs, see T3179 for the general case.

With D5854 covering the H1 and H2 headers for markup, it'd be nice if Remarkup could parse Markdown headers below H2, which should reduce conflicts with lists at least.

In T2714#57, @hwinkel wrote:

Is there any Wiki Page describes what releeph will going to offer?

I assume that change hasn't rolled out here yet? This is what I see (I unpinned audit and added feed):

A link to the new launcher would be handy from the sidebar, I was confused for a few moments how to access the other apps before I remembered that they're in search.

Forking other people's pastes might be handy.

I thought I was going mad, but I've reproduced this twice now.

YOU NEED TO CHANGE YOUR PASSWORD

I have no idea why that got closed.

Right, which is exactly why I disagree with renaming Real Name to Display Name... as it's not what is displayed everywhere.

It's not a display name though, that would be confusing. Your username is what's displayed in 99% of the UI.

The remarkup-friendly commit hash (with the preceding 'r' and callsign) is (most likely) shown all over the UI there to make it easy for users to copy and know how to reference them elsewhere - I doubt anyone thinks you need to be constantly reminded what repo you're looking at several times per page.

Is there a specific part you're having trouble with? Almost all of Phabricator is VCS-neutral, and a large number of Arcanist commands (the CLI tool) are as well.

If you enable UTF-8 support in PuTTY (which'll be required for a huge number of things to display correctly, PuTTY's default character encoding is awfully broken), it should fix this.

I think onbeforeunload + a info box at the top of the page (added in onbeforeunload, and kept for future loads of the page) saying there are draft inline comments would cover it and fit in with the UI.

Any PHP errors? (check your web server / fcgid logs)

See T4151: Cleanup and errata for repository hosting.

+1 for Monogram, although Wikipedia adds Cypher to the suggestions - I think Monogram is slightly more obvious though.

Confirmed for the retina set (although it's barely noticeable), seems to be an issue with the highlight positioning.

✘ Merged into T3945.

◀ Merged tasks: T4247.

I don't understand why you would want to do this, what's the use case here?

+1-ing this as well, dealing with 3 installs here and I often look for things on the wrong one.

Sorry, I thought this had already made it into Phabricator. You'll want to follow T2465: Context not available for revisions created on Windows .

Most default policies are configured in the Applications app, only a few have them right now - what do you need that is missing?.

Until that gap is filled

It's currently still quite inconsistent (while there are still apps transitioning to the transactions infrastructure) - in general, I believe that the breakdown is planned to be something like this:

Avatars are displayed for "major" stories, but not for "minor" stories.

I'd write "FreeBSD and Mac OS X are acceptable flavours of Linux" but that will upset some FreeBSD people

If this is pursued, it'd be handy if it showed the URL as well.

Use arc feature instead, which uses arc branch for git and arc bookmark for mercurial (unless you really want permanent branches, in which case carry on, but that should have a huge warning for people that don't expect it).

Clarification: if you set differential.allow-self-accept to true, it sounds like you'll get exactly what you want - where people can accept diffs they submitted.

Right, because you've disabled being able to accept your own revisions by setting differential.allow-self-accept to false.

Abandon Revision is a standard action for authors, the reviewer ones are are accept and request changes - if he isn't seeing those the option is being respected.

Closed by commit rPf8d963a77e5f.