In T11650#194907, @epriestley wrote:

Is it possible that this is slightly narrower: "package dominion rules incorrectly consider archived packages"? That seems to fit with the observations, but maybe I'm overlooking something.

Is it possible that this is slightly narrower: "package dominion rules incorrectly consider archived packages"? That seems to fit with the observations, but maybe I'm overlooking something.

In T11152#181357, @flashydave wrote:

Regarding the specific fault it would be interesting to know why arc expects to be within a repository (either for the user repo or for arcanist) in the first place just to print out a version string. Surely if php can execute arc script that should be sufficient?

.remarkup-code td span

D16465 has details on how to move forward here if anyone wants to steal it before @michaeljs1990 gets back to it.

Nyancat

Generally, it's always cheaper for us to code checks defensively per our time. That lightens support which then allows for time for building features. We're just two developers here and we don't get a paycheck for our time or support we offer. Maybe someday :)

Providing support isn't free.

In T11627#194460, @chad wrote:

Features unfortunately aren't free. They cost time to build, test, and provide lifetime support of, that we must pay.

Features unfortunately aren't free. They cost time to build, test, and provide lifetime support of, that we must pay.

In T11627#194423, @epriestley wrote:

This is unique to this particular setup check (open_basedir configured, but not in a way that creates a concrete problem we're able to detect).

I'm inclined to just simplify this check and make any value for open_basedir cause a fatal setup error.

You're right! That probably makes is more feature request, less bug report. (extremely low priority FR)

This is unique to this particular setup check (open_basedir configured, but not in a way that creates a concrete problem we're able to detect).

I don't think this is preview related -- I think PhabricatorEditEngineConfigurationTransaction just doesn't implement hasChangeDetails() and some related methods to generate this link.

Sorry, we don't offer setup and configuration help. See Support Resources for discussion of what we do offer.

i am unable to fix flowing issues,,,
1)MySQL is Using Default Stopword File.
2)MySQL is Using Default Minimum Word Length.
3)MySQL is Using the Default Boolean Syntax.
4)MySQL May Run Slowly.
5)Large File Storage Not Configured..
please help me ,,,,,i try ed your flowing producers......

Are you running apparmor or selinux?

In T11613#194212, @avivey wrote:

It looks like you have all the required extensions, although it's possible that they are only enabled when running php from the command line, and not from Apache; If you can get a php page to run phpinfo(), it will print out the full information from the right context.

Alternatively, if you checkout stable branch of Phabricator, you'll get a better error message.

I am able to reproduce this using another class of errors from https://secure.phabricator.com/diffusion/P/browse/master/src/applications/config/check/PhabricatorPHPPreflightSetupCheck.php, but these are all much more complex to explain.

In any case, I don't think we need the video, thanks.

It looks like you have all the required extensions, although it's possible that they are only enabled when running php from the command line, and not from Apache; If you can get a php page to run phpinfo(), it will print out the full information from the right context.

We found what we think is a repro case, but we don't know if it's the same as your repro case. Only you could tell us that, but I don't think a video would be useful.

Here are my own PHP modules:

Yeah, I think that's extensive enough. And I think that's the only check right now that might blow up, but I didn't follow the recent changes.

maybe that is better?

I think it happens for any of the extensions in https://secure.phabricator.com/diffusion/P/browse/master/src/applications/config/check/PhabricatorExtensionsSetupCheck.php$3

Verified, removing php5-mysql reproduces the error.

oh good, I was in the middle of figuring that out next. I tried installing APC and APCu but still got the error.

yeah, I've removed php-mysql and then I get the same error.

I repro it fine.

In T11613#194139, @avivey wrote:

That might help.
Also, a stack-trace (should be in the webserver error log) might show something.

Installing 14.04.5 VM now.

That might help.
Also, a stack-trace (should be in the webserver error log) might show something.

If you cannot reproduce the issue, I can create a video of a XDebug session leading to the error, so you see the program flow.

Also, the one in IRC says he's got this error with APC loaded, with php 5.6.25, and that reverting to stable fixed this for him.

On Master. Both this report and the one in IRC were about master.

you can't reproduce stable or master? a lot of setup was re-written this week and stable hasn't been cut yet.

I can't reproduce this.

• I've created a new VM, and installed Ubuntu 14.04.5
• I've used apt-get install apache2 php5-cli mysql-server git php5-mysql php5-curl
• I've added apt-get install lamp-server^, for good measure
• I've configured apache2 as per the instructions, and loaded the page
• This is what I got:

• After running ./bin/storage upgrade and reloading, I got this:

Yeah, it sounds like some bug, but I can't reproduce it, either with or without APC, and another user complained about it failing with APC installed.

In T11613#194114, @avivey wrote:

Which versions of php and webserver are you using?

Going to upload a fix myself.

Which versions of php and webserver are you using?