Segfaults in tests about boardrules HOT 9 CLOSED

This stuff is too techie for me, but here is some stuff I found that may make sense to you.

I have seen this popup:
Turn off PHP's garbage collector with zend.enable_gc=0.

Also: travis-ci/travis-ci#764

https://code.google.com/p/sabredav/source/browse/.travis.yml?r=54831d71780c23433daf1d71a2afab528a2a8327#12

symfony/symfony@3f8665a

from boardrules.

If you are interested, please give fixing it a try by following what they did in submitting a PR here (once we find something that works, we'll try to get the fix into the test framework itself).

from boardrules.

@EXreaction So, I switched over the no's acme-demo testing framework, which as you can see in the tests, takes about 30 seconds to test and only 40MB, as opposed to 5 minutes and 400MB with the prior framework.

Unfortunately, the segfaults still randomly occur for this extension.

https://travis-ci.org/VSEphpbb/boardrules/builds/23729638

from boardrules.

http://lickmychip.com/2012/10/18/debug-your-php-segmentation-fault-on-travis-ci/

from boardrules.

Yeah, I've seen that. No idea what to do with it though.

from boardrules.

So I got gdb working and the results of several tests...
https://dl.dropboxusercontent.com/u/68326338/segfault%20logs.zip

Fails at different points all the time (3, 4, 5 were all on the same build), but after the tests are completed and their tearDown functions were called.

from boardrules.

Well it's all gibberish to me, but I do see zend stuff around the seg fault.

Still, the exceptions test has something to do with this too, so I recommend skipping that test for now, until it can be replaced or fixed.

from boardrules.

So, from some more experiences with this, it looks like the seg faults only happen when using the data sets in the data provider that have arrays for the $message in the test_exceptions test function.

Those would be:
array('EXCEPTION_OUT_OF_BOUNDS', '{foo}'),
array('{foo}', 'FIELD_MISSING'),
array('{foo}', 'TOO_LONG'),

of course, I don't understand why these are causing them, but anytime I have these removed from the dataprovider, I don't get segfaults anymore.

from boardrules.

Solved PR #98

from boardrules.