FC2 test2 is a BAD joke

[shmuel siegel]

Again we have returned to philosophical issues of the purpose of the
test release. 

According to the wording on the schedule page of Fedora, there is an
expectation that a test release is something that you can run for a few
weeks and find its various bugs. If you can't install, you can't do
that. 

On the other hand, the attitude in this group has been that a test
release is a snapshot which reflects some sense of consistency of the
parts. The failures that are encountered in installing accomplish the
desired goal. The installer is found to not work in enough environments.
The testers are expected to document this issue, use this group to get
around the problem, and then go on to test other issues. In the meantime
they are expected to keep current with rawhide. The failure of the
installer, is just one issue; it should not stand in the way of other
testing since the failure can be circumvented.

I do strongly suggest to the powers that be that read this mailing list,
that the wording on the schedule page be changed to reflect the
expectations of the responders here. I'd also like to see that there at
least an unofficial way to test the installer before test3. It has been
pointed out that RedHat doesn't have all of the hardware configurations
that have been shown to break in this test release. Give the testers
something new to try, i.e., the equivalent of staying in sync with
Rawhide.

On Tue, 2004-03-30 at 15:23, William Hooper wrote:
> Robert P. J. Day said:
> [snip]
> >
> > but all of that assumes you can get the software onto your machine in the
> > first place.  this suggests that, more than anything else, the
> > installation procedure has to be flogged pretty hard before getting out
> > the door to make sure it works.
> 
> Isn't testing the installation procedure the reason there are test
> releases?  A failed test is no less of a test.
> 
> -- 
> William Hooper
>