Re: glibc 2.1.91 based distributions :-(.

[Andreas Jaeger <aj at suse dot de>]

>>>>> Jes Sorensen writes:

>>>>> "Andreas" == Andreas Jaeger <aj@suse.de> writes:
Andreas> Hi Uli,

Andreas> we should release the current glibc version now as final
Andreas> version.  I've just read another announcement where a beta
Andreas> distribution is announced that uses glibc 2.1.91.  It seems
Andreas> that glibc 2.1.91 is very stable and as bug free as 2.1.3.

Jes> Pardon me but why is this a problem as long as it is a beta
Jes> distribution? It's a pretty good way for getting the library tested
Jes> and the distribution vendors really want glibc-2.2 in their next
Jes> releases if they possibly can.

The problems I see are mainly that people think this is a beta test of
a distribution - and of glibc.  It might hinder development of glibc
because there's no [1] big warning that e.g. glibc is itself beta and
subject to change.  glibc will definitly change in an incompatible
way, e.g. the RPC-IPv6 code has to be fixed.  

People might take features of glibc for granted and will complain to
us, the glibc developers, if we change things.  Instead they should
complain to the distributors for releasing an *alpha* version of
glibc.

This has caused already in the past some problems, e.g. the usage
of an early test release of glibc 2.0 on PowerPC and Alpha (we changed
a serious bug and introduced an incompatibility in the final 2.0).

I do know that this means lots of testers and testing - which I do
appreciate.

Andreas

Footnotes: 
[1]  I only read the announcements of two beta distributions, I didn't
     try the beta distributions.
-- 
 Andreas Jaeger
  SuSE Labs aj@suse.de
   private aj@arthur.inka.de
    http://www.suse.de/~aj