Re: An alternate view of PEAR QA

From: Date: Sun, 18 Jan 2004 23:33:57 +0000
Subject: Re: An alternate view of PEAR QA
References: 1 2 3 4 5  Groups: php.pear.qa 
Request: Send a blank email to pear-qa+get-375@lists.php.net to get a copy of this message
Pierre-Alain Joye <paj@pearfr.org> wrote: > Let clarify a bit what means testings: > - BC (can be partly automated soon, fully if php5 is used) > - run unit tests > > These two points have never raised on peardev (or too late to act > nicely). Note that they neither raised on pear-qa. Taking the example of Console_Getopt, the release was made (including BC break) without any notification. In such cases, I think the QA approval for stable releases is a great idea and would prevent such things from happening. > Well keeping them update is not that an usefull task if we do not > receive any test result, feedback or whatever, from pear-dev or pear-qa. > Take an example, 1.3b5 is out, there are obviously pending "bugs", only > Jan and a couple of devs reported something. That's not something to > make me think about pear-dev interest in QA (except the classic things > about naming convention ;) ). Yes, I agree -- but I was not referring to QA, I was referring to the discussions on things such as design patterns, etc. which imho would be out of place on pear-qa. > I like the concept of QA approved releases. Many things can be > automatic. But many require feedbacks, from run-tests (wondering if we > can get the pear related tests from php tests (make tests) ). I would like to have a tool where I can automate the tests :-) That way, I would only need to check a page on my server to see if a package had failed. However, I suppose I could do that anyway with a cron job. Speaking of which, I might develop a tool to check the RSS feed for new releases and test them. It would be easiest if it was done on upload, of course. Klaus

« previous php.pear.qa (#375) next »