Thanks for the report! This should be fixed in HEAD of master, and promote to stable in a few days. Let me know if I missed anything or you're seeing other issues.

I broke this in D16360, on August 1.

Thanks for the detailed investigation -- glad we've root-caused this and proposed a solution :)

I'm just going to merge this into T5028 since I think that ultimately is essentially the same issue.

thanks Evan, that did the trick!

(T8850 is vaguely related.)

we are using Apache webserver, and yes we are mangling the REMOTE_ADDRESS in preamble as recommended in the configure_preamble article section "adjusting client ips" because we are behind a load balancer.

cluster.instance is only used when serving multiple separate instances from a single cluster (for example, when hosting Phabricator for multiple clients), and should normally be left alone. This setting is used in the Phacility cluster to make x.phacility.com, y.phacility.com, etc., work correctly. This is sort of confusing because we use "cluster" to refer both to instancing and redundancy/scale; both are "clusters" in some sense but we don't generally expect users to care about instancing.

This report lacks reproduction steps, so I'm not sure how to reproduce the issue. We can't fix issues we aren't able to reproduce.

This isn't really resolved, I just decided I didn't care about it. Feel free to change the status to something else if you feel that's more appopriate.

Yes, that's correct.

I just tested the "Replace %s%n%n%b with %B" and it fixed it, Auditors were parsed without any problems.

Okay. Is it correct that if I trust everyone who has access to my instance won't upload malicious files (and are resistant to untrusted people manipulating them to do so), I don't really need a security.alternate-file-domain at all? (That is my reading of https://secure.phabricator.com/book/phabricator/article/configuring_file_domain/ but I'd feel better if you told me whether it's correct. I understand that you are not making any guarantees.)

Yeah, I think this is because of --format=%s. Some discussion in T5028, which is probably most of this.

The windows style was from netbeans, I did the previous ones from command line (if you want I can run netbeans as well to prove it). The title was from when I was strictly using netbeans.

When I run git cat-file -p 69d24cb35c46a4a2fed6a851aa17962d7dc0bd26 | hd, I get only \ns, no \rs at all... same with git log and git log ---format='%B' (git version 2.8.1.210.gb8b4d93 on linux).

I don't have any ideas on how to build a simpler reproduction case offhand.

Okay, so how should I attempt to reproduce this? Any options more reasonable than creating a new EC2 instance, installing all the supporting paraphernalia, then installing Phabricator manually from scratch?

Oh, sorry -- you won't be able to reproduce this particular issue in the Phacility cluster.

To reproduce this, I need to change the security.alternate-file-domain config key to point to a new CloudFront instance I create, but I can't change that key from the web interface on a test instance (as far as I can tell, anyway).

Sorry about that, I didn't know the test git repository existed otherwise I would have done this right after.

You misspelled my username but that's enough for me to go on, thanks.

You should have the commits. You will see the first two didn't trigger Auditors (with auditors in the last line), but the last one did (with Auditors in the first line)

Can you push a commit to rGITTEST on this server which reproduces the problem? Specify "Auditors: epriestley".

The Auditors:alincoln for example, isn't meant to be the whole body.

This may also be a good opportunity to fix T10423 since we'd end up testing the same code.

The first line of a commit is always interpreted as a commit title, so the behavior you describe is expected.

Also, thanks for providing pointers to reproduction cases on this install. Being able to fiddle with those commits locally was helpful in confirming the behavior of -M and -C with paths.

I think there are two issues here:

Can confirm that this affects hosted and other observed repositories as well.

This is because we hate code reuse.

We need detailed reproduction steps which work against a clean environment, as described in Providing Reproduction Steps. We run our entire production cluster through CloudFlare so I'm unable to reproduce the issue as described purely by using CloudFlare.

Okay, I'm still (or again) getting these "The page isn’t redirecting properly" errors when trying to view an uploaded file, or an uploaded image (e.g. my profile picture).

No smoking guns in the access or error logs on admin. The cache/load theory may just be a goose chase, since I don't see any evidence of unusual load around that time.

This is still happening (incidentally just observed because of the build failure due to T11469).

Just observed a failure, see updated description.

Awesome! I appreciate the quick fixes!

• This constraint should be fixed at HEAD of master. We just missed the deploy so it won't go until the cluster until next week (around August 20). Let me know if I missed anything.
• Empty constraints should be fixed at HEAD of master (they now error instead of being ignored).
• buildablePHID vs objectPHID confusion and inconsistencies should be fixed in the future by harbormaster.buildable.search. If you want a skeletal version of this now to reduce API tweaks later, let me know and we can get one in pretty easily.
• Linking to pre-filled forms seems maybe not hugely useful in practice, with some small risk of social engineering attacks? Feel free to file a separate request if you have use cases I'm not thinking of or whatever.

Fix CORGI errors, check if that reveals any obvious indicators of instance information fetch issues.

No disruption is expected and there was no maintenance or known disruption at those times. Nothing jumps out at me immediately in monitoring, either, and I can't reproduce this by casually cloning stuff.

D16396 should make us raise a narrow error for "ids": [], "phids": [], "whateverPHIDs": [], etc., instructing you to pass a nonempty list.

I think fixing "ids":[] is trivial; it just works like that because we use the same Conduit type ("PHIDList") both for constraints and other things which can accept an empty array (like setting subscribers to [] to remove all the subscribers on an object).

(@yelirekim, just a heads up since whatever comes of this might impact you somehow.)

If you don't like the @mention-implies-CC behavior, you can disable it by applying this patch to your install:

In T11471#190079, @chad wrote:

I don't consider this to be a bug and any fix would likely lead to additional user confusion over "magic".

I also don't necessarily consider this a bug (I think a human explicitly @mention-ing a user definitely should add them again, and that quoting someone is pretty close to mentioning them, and that reasonable users could probably expect either behavior, and perhaps that the "add them again" behavior is more obvious or more expected) but T11035 discusses this and some other behaviors which have varying amounts of grey area.

The use case in that linked task is "using Remarkup to demonstrate the behavior of a different markup language", which isn't something we expect Remarkup to support.

The example is pretty descriptive, I'd say. Because of linking to the description of that exact part of the word.

Why do you want to link part of a word?

slightly related: there's some confusing naming in the result for harbormaster.querybuildables:

I don't consider this to be a bug and any fix would likely lead to additional user confusion over "magic".

secure.phabricator.com isn't in the Phacility cluster

Why do you mean by "this instance"?

I got code 500 as I checked out from this instance, was about 23:50 UTC, yesterday. Maybe related? After getting that code, I tried it again after about 10 seconds, then, everything was ok.

I can't reproduce this based on the steps provided. Have you been able to reproduce in a clean environment like a test instance on http://phacility.com ?

Please close - I figured out how to change the username. Not sure why I missed it the first time :(