RE: [PHP-QA] changing make test for automatic testing

At 16:22 29.01.2003, Derick Rethans wrote:

On Wed, 29 Jan 2003, Kai Schroder wrote:

Handing out keys is a pretty nifty idea, but I think we should offer some kind of registration on the QA-website for people who plan to regularly submit testruns and don't object having their system identified by a unique id. This user-key should be distinguishable to a qa-teammember-key though.

Of course. Even I'm not sure if people without key should be able to submit at all. Most of these testruns done by non-qa people are pretty worthless anyway.

ACK. We have a lot of Linux boxes with 10 or 12 standard modules in test results, but round about the half of existing modules are never tested. If we hand out machine keys, we can faster find out, which source changes are broken and we "know" the people behind the machine/email. Other point is, that we want rate the importance of a bug by counting failed test results. If you don't have a unique machine key, you can decrease the importance by running make test duplicate on your machine (seen on bugs.php.net more as one time). In my opinion we should differ between trusted machines (with key) and untrusted in statistics. Anonymous tests should only be visible for QA team members. Yup, I agree with that. Derick

Yes, i can only aggree.