On Tuesday 07 November 2000 12:31, you wrote:
>
> I don't have a problem with non-functional software, basically, but what's
> the point of releasing software when you very well know that it does not
> work at all? I mean, sure at your computer you need to have it installed
> to do debugging, but why then release it? I don't understand this....
Probably time to talk about the kde build process and then this may make a
little sense.
KDE must be built in consecuative parts, meaning kdesupport must be built
first, then installed, then kdelibs must be built and installed, then kdebase
must be built and then installed. Then it is possible to test to make sure
kde works.
In order for us to do an install on our build cluster, for example to install
kdesupport, we must upload it to the cooker RPM directory, then install. Most
of the time this prevents serious problems with multiple developers working
on different packages. This is one time it didn't work properly, I did not
find out kde did not run until I got to kdebase. Nothing lost, the packages
where broken already anyways.
Normally no-one would notice, but in this case I decided to let everyone know
as this is not a quick problem to fix and I felt it would not be good to let
people break their desktops.
The first chain of this entire problem was updating gcc from 2.95 to 2.96. At
that point kde was no longer functional in cooker becuase of library
problems. Remember, cooker is our development environment. We try to keep it
working, but sometimes this is just impossible. As always we do not recommend
your using cooker as a production environment.
-Chris
