On Mon, Apr 11, 2011 at 5:29 PM, Cristian Rodríguez
El 11/04/11 18:11, Christian Boltz escribió:
Hello,
on Montag, 11. April 2011, jdd wrote:
anyway, the testers/users have no present way to figure what applications are very well tested and wich are not.
<BOfH mode> There is a way: "If it's totally broken, nobody tested it"
But the chance was I had just to build a dvd and could test them with real data.
That's probably the most important point: People test what they use/need. They don't test something they will never need or use. This is especially true for community (aka "non-paid") testers. (But you would have to pay a lot to make me testing beer *g*)
Heh :) In any case, totally broken stuff should never get published in the first place, packages usually come with a test suite, that should be run during build.
One of the GSoC projects is to work on spec-cleaner. Is it feasible to update the spec for specfiles and get a %test section added. Then have rpmlint complain if it empty. I'm not sure where spec-cleaner would fit in, but maybe it could at least ensure "cleaned" specfiles had a empty %test section. Just seeing that might be enough of a prompt to get more self-tests into specfiles. Greg -- To unsubscribe, e-mail: opensuse-factory+unsubscribe@opensuse.org For additional commands, e-mail: opensuse-factory+help@opensuse.org