Re: The fate of libc5
On Tue, Jul 11, 2000 at 04:20:21PM +0200, Richard B. Kreckel wrote:
> On Mon, 10 Jul 2000, Ben Collins wrote:
> > Anyone else agree, or can give a real reason why this shouldn't be the
> > case?
> For my own part, I am sometimes using MapleVR3 (which requires me to
> install libc4!) and MapleVR4 (wich needs libc5-compat). Every new version
> of Maple breaks an awful lot of existing software and our folks here doing
> symbolic manipulation really do depend on those old versions. Porting
> 20loc of code somebody else has written to the new version is no fun.
> If it is not a serious problem I vote for leaving it in woody and remove
> it later.
Taking it out of woody does not make it unavailable. As with any obsoleted
programs/libs, the archive.debian.org repository keeps them around.
> BTW: has anybody noticed that because of the frequent libc changes many
> commercial applications are *statically* linked on Linux while they are
> *dynamic* on other Unices? This is contraproductive. (Examples include
> Mathematica4 and the ADSM client software.)
If they would upgrade to glibc 2.0, that would not be a problem. Things
compiled with glibc 2.0 are generally compatible and work with later
versions of glibc. Continuing to support libc5 is like continuing to
support Windows 3.x. It just isn't feasible forever, and developers of
these commercial apps can take a hint. Free Software has moved on, why
can't commercial companies that pay developers do the same?
Anyway, RedHat's next release (rawhide) doesn't even contain libc5, and
RedHat relies on closed-source, proprietary applications more than we
(they have licensing deals with them). I don't see those sorts of
applications as a reason for keeping it around in woody anyway, since it
will always be available in archive.debian.org for those that absolutely
/ Ben Collins -- ...on that fantastic voyage... -- Debian GNU/Linux \
` firstname.lastname@example.org -- email@example.com -- firstname.lastname@example.org '