Re: [Bugfix] Mergesort used in clist

[Owen Taylor <otaylor redhat com>]

Thomas Schultz <tststs@gmx.de> writes:

> Hi,
> 
> On 2000-08-31, Owen Taylor <otaylor@redhat.com> wrote:
> 
> > This was reported by two people and fixed few weeks ago, and was reported
> > on this list as well as recently as Monday. (Yet, the bug is several
> > years old!) Thanks for catching it, though, anyways.
> 
> Will we see GTK 1.2.9 soon or should I try to work around this in my
> application / tell my users "not to delete the last list entry until the
> next version of GTK"?

1.2.9 will probably be in the next month or so, but not real soon.
You should probably try to work around the bug.
 
> >> I also could not find a "known bugs" page for GTK 1.2.8 - in case this
> >> problem is already fixed, it would certainly have saved me several hours
> >> of debugging. Does something like this exist?
> > Tou could look at the bugtracker. (bugs.gnome.org) Many (if not all)
> > fixed bugs will appear in the closed bug section. Except that the bug
> > tracker is in bad disrepair, and closed bugs are temporarily not being
> > shown on the bug tracking machine so that it doesn't completely die.
> 
> I tried the bugtracker and to be honest, I didn't find it all that useful.
>  - I found the output of the search very user-unfriendly: The overview
>    contained lots of information I was not interested in (like who
>    submitted the bug or which address the mail to close the bug was sent
>    to), but which I needed to read, because the subject was hidden
>    somewhere in the middle of it.
>  - I could not find a way to limit my query to a single package.

Yes, the bugtracker is far from ideal. We are planning to upgrade
to bugzilla at  some point in the near future.

>  - Maybe this is due to the hardware problems you mention, but using
>    various (combinations of) keywords, I could not find any reference to
>    the bug that caused my problem, nor could I get an overview of bugs in
>    the current stable version of GTK that might cause future problems.

I believe that it is the case.

> I also tried to download the archive of the past few months of the GTK
> mailing list, but found that it did not seem to be in valid mbox format
> (mutt could only read 16 of more than 300 postings). After fiddling some
> time to solve this, I noticed that all "References:" and "In-Reply-To:"
> header lines were stripped, making it impossible to reconstruct threads.

We are also migrating that in the new future to a new system which has
real mbox format with all headers; you can see a prototype at:

 http://mail.gnome.org/archives/gtk-devel-list
 
> What am I doing wrong? If my application crashes and I get the impression
> that a GTK problem is involved, how can I check this without spending
> hours?

Well, 99% of the time it isn't a GTK+ problem... you just got unlucky
(at least with stable versions of GTK+.) Asking on gtk-list can be an
effective strategy, especially if you have some code to show.

Sorry for your difficulties,
                                        Owen