https://travis-ci.org/emanuele45/Dialogo/builds/13287431
YAY!! :D
That means we can add MOAR tests and let github/travis take care of them! :D
Very awesome !!! I see a lot of projects that use Travis, just have never had the time to set anything up. We still need to add more test in more areas, but as we do this will help us break fewer things O:-)
There is still postgre that doesn't work at the moment for the way the "installer" works, I'll try to fix it later today. :)
And added to the main repo! :D
https://github.com/elkarte/Elkarte/pull/959
It tests for both mysql and postgre with php 5.2, 5.3, 5.4 and 5.5.
When the beta will be out I'll investigate how to do the tests on SQLite too (I think I have to checkout the other repo and move some files around, not much more I hope).
ETA: it's still broken, sorry... :-[
Adding some more:
https://travis-ci.org/emanuele45/Dialogo/builds/13343659&url_id=5ca76ac97d9bcaa67cb11f4e8db47cf794849a4b
The build is still failing, but now the setup of the environment (including installation of apache) and the installation of Elk are actually working! :P
I have to work on the errors now. O:-)
This time it really works! :D
https://travis-ci.org/emanuele45/Dialogo/builds/13388825
Tomorrow I'll make the commits nicer and I'll send the PR.
This is so cool ;D
Question of the day:
$_GET['action'] = array('test1', 'test2');
$_GET['action'] = (string) $_GET['action'];
Any idea why php 5.4 and 5.5 returns an error, while php 5.3 does not?
5.3: https://travis-ci.org/emanuele45/Dialogo/jobs/13418891
5.4: https://travis-ci.org/emanuele45/Dialogo/jobs/13418890
5.5: https://travis-ci.org/emanuele45/Dialogo/jobs/13418889
Ummm trick question, maybe because php 5.3 is buggy :P ... (is that just from the test or do we need an array_map ? )
Nope, no need for array_map or alike, it is just test that checks for the error in case it is an array.
Yeah, the buggy is the first thing I thought, but it looks weird because seems a rather
big error... lol
The only difference I was able to find is this:
Though we are using error_reporting(E_STRICT | E_ALL), so they should act the same...
headscratch
I was ready to blame travis or simpletest for the problems with postgre, and instead it is a legit bug!! xD
haha, already paid off it looks like. I'll have to look more into this. Sounds cool to have some automated testing if I'm understanding this right.
Yep.
It already found... 5 bugs I think, mainly related to postgre (anyone wants to guess if we test more on mysql or postgre? :P).