On 7/30/07, André Schnabel <[EMAIL PROTECTED]> wrote: > > Suggestions > > * For uniqueness of qatesttool - I think you should add the date. > > > I thought about this - but did not do so for the moment. The problem is > to decide "what version do we need". If we have no answer, we will end > up with a lot of archives. At the other hand, it is rather easy to tell > on what date the archive has been created - the day befor you download > (or at the same day).
The scripts are very linked to the milestones. They have to bother to which bugs are fixed or introduced. So... having a set of scripts newer than your build might be disastrous (i.e., it might to think some bug was fixed and point a lot of errors to you). We should provide scripts that are likely to work with some milestone build and we could add the milestone info to its naming. We could upload the HEAD everyday too, but a date would be nice to have, because it makes easier to the script developers to know what revision was used for that set (even if that info can be obtained on the header of each script). > If we really need a fixed set of scripts, I'd rather go with CVS-Tags .. > or wait for other suggestions (we will need such suggestions anyway, as > "soon" as we start to collect resultfiles from the community). We need the scripts from other branches than HEAD. The ideal would be to have a certified version from our automation team. This would avoid bothering with errors from the scripts while testing.
