[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: sid: libc6-2.2.5-4 kills vmware workstation 3.0

On Tue, Apr 09, 2002 at 08:45:16PM +0200, Jeroen Dekkers wrote:
> On Tue, Apr 09, 2002 at 01:21:49PM -0500, Colin Watson wrote:
> > On Tue, Apr 09, 2002 at 07:29:14PM +0200, Jeroen Dekkers wrote:
> > > Vmware isn't even in Debian. This is truely a problem of vmware
> > > itself. IMHO this isn't something for debian-devel. Or do you want to
> > > make debian-devel a list where all Debian users can come with their
> > > problems running buggy non-free software?
> > 
> > What if they turned out to be caused by bugs in our free software?
> > Telling them to go away then would be foolish, since we want to know
> > about bugs, no matter how they were caused.
> > 
> > Apparently this breakage was caused by a change in glibc. As a general
> > rule, changes in the C library should not break any software, whether
> > free or non-free. Sometimes this is not the case (e.g. StarOffice's use
> > of private glibc symbols a few years ago), but bugs should be
> > investigated rather than casually dismissed.
> Did you *read* the thread? The cause of the problem was already found,
> it was a vmware bug, vmware already provided patches but not for the
> version Donald was using.

I was rather under the impression that we were talking about any
discussion of problems with non-free software, regardless of the cause.
It certainly sounded that way to me.

Even your improved reply says:

> This problem is very common for non-free software.

... which really doesn't seem all that relevant apart from sounding
good; hell, the change in nice()'s return value appears to be a problem
for start-stop-daemon in dpkg, see #141500, and a minor problem with X,
see #140012. The nice() interface *did* change without versioning - it's
true that programs that relied on the old behaviour were buggy, but
there are plenty of such programs in Debian main and that is something
Debian developers should be aware of. Patting ourselves on the back is
great when it's justified, but I think it's somewhat counterproductive
when it isn't.

Now, it's true that one generally can't fix the non-free stuff when it
breaks in this way, nor can Debian. But claiming that the *problem* is
common in non-free software, implying that it is not common in free
software, is simply not true.


Colin Watson                                  [cjwatson@flatline.org.uk]

To UNSUBSCRIBE, email to debian-devel-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

Reply to: