Re: AW: [PHP-QA] New run-tests code

On Wed, 9 Sep 2009, Schürmann Sebastian wrote: > Von: Derick Rethans [mailto:derick@php.net] > > > > On Wed, 9 Sep 2009, zoe wrote: > > > > > Stefan, Georg and I have been discussing making a small change in the > > > new run-tests.php code that would make the implementation different > > > from the current run-tests.php. > > > > > > At the moment when you use run-tests.php the test results can be > > > affected by your php.ini file. This causes endless confusion for > > > people that don't realise that they need to use the -n option to > > > ignore their php.ini. We'd like to make ignoring the user's php.ini > > > file the default behaviour in the new version. Does anyone see a > > > problem with this? > > > > Yes, I do. We picked the original behavior so that we can make our > > tests work in every environment. It provides a better spread of code > > and settings in the wild. > > As much as I understand this decision (we made the same for our > internal codebase), as much I do know how much is gained through > finding a good configuration that is fixed to the test and is samesame > over any system you test on (we changed it later). I guess there would > be variants of tests which have the same code, but conditional > execution just for the INI aka the configuration. No, we find what settings are influencing the tests and set the correct --INI-- setting mostly, sometimes it means a test gets split in two to test for both variants. > Can you give us a good example of code that supports what you are > saying? I think the idea is good, but we need the pain in the a** > examples. These will help this discussion. Well, it's mostly *required* for shared extensions, and especially Zend extensions that can not even be compiled in. I don't have other examples on the top of my head. regards, Derick -- http://derickrethans.nl | http://ezcomponents.org | http://xdebug.org twitter: @derickr