Re: [Plplot-devel] segfaults for tcl examples in the pltcl case but not the tclsh case

[Rafael L. <rla...@us...>]

* Alan W. Irwin <ir...@be...> [2004-01-21 13:17]:

> This new bug for CVS HEAD is clearly generated because of Andrew Ross's
> memory management fixups, but I suspect what is going on is that other parts
> of the code base (in this case the pltcl application) are being sloppy about
> initializing the library and this has now become revealed because of
> Andrew's fixups.
> 
> It's your call, Rafael, about whether these segfaults are a sufficiently
> rare case so we won't worry about it for release or whether we should wait
> until post-release to deal with all the memory management fixups together.

Thanks for your extensive tests and reports, Alan.  I also noticed some
strange behavior with the gnome driver built from CVS HEAD.  My recollection
is vague about this problem hapening in the past and I do not have time now
to explore it further.

I think that these segfault fixups should either be done before the release
or simply postponed.  We could delay the release until the problems are
fixed, but I prefer to stick to the schedule.  We can do a 5.3.1 bug-fixing
release as soon as the bugs are fixed.

My plan now is to generate the RC2 tarball until the next weekend ("release
early, release often").

-- 
Rafael