REGRESSION in KDE-3.2.3

[James Richard Tyrer]

Please see bug 83288.

This is another release with at least one serious regression -- a regression 
that most users will find.  The additional problem in comment #1 is probably 
more likely to be found by most users.

The last few releases have all had somewhat serious regressions.

Clearly, we need better pre-release testing.

The problem, as I see it, is that KDE has now become too large for releases to 
be released without regression testing.  What this means is that new releases 
are not always improving the product because of the complexity -- although bugs 
are fixed it appears that new bugs (which are regressions) are being introduced.

I see only one possible solution for this.  We must have some organized 
pre-release testing for regressions.  To start with, all that is needed is for a 
release candidate to be out for sufficient time for these problems to be found 
before the final release.

Then there needs to be some way to identify the bug reports filed against that 
release candidate.  This means that the fact that it is a regression and the 
fact that it is a release candidate need to be parameters in the BugZilla 
database so that these bugs can be easily identified so that the final release 
can be held for them if necessary.

Please notice that you can NOT find the above bug by searching for it.  There is 
no way to search for bugs for KDE-3.2.3 or to search for regressions except to 
do a text search.  That is part of the problem.

Also note how quickly JJ got added to:

	http://bugs.kde.org/

Am I missing something here?

--
JRT