This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Adding zlib to GCC
>>>>> "Alexandre" == Alexandre Oliva <aoliva@redhat.com> writes:
Alexandre> On Aug 29, 2000, Marc Espie <espie@quatramaran.ens.fr>
Alexandre> wrote:
>> If there is a common repository, I don't really see agreement
>> as to synch'ed releases between these.
Alexandre> There's no need for synchronization. GDB and binutils,
Alexandre> for example, that have large common parts and live in
Alexandre> the same CVS repo, are not released at the same time.
Alexandre> At some point, the maintainers of each separate project
Alexandre> decides it's time to create a branch for a release,
Alexandre> stabilizes it and releases it. Meanwhile, all other
Alexandre> projects proceed their development in the trunk.
It's really not that simple, and I agree with Marc Espie.
The problem is that you only want to make that branch if the
components are relatively stable. You can't branch GDB, for instance,
if BFD is crashing all over the place at the moment. The more stuff
we pile into one repository, the more interdependent we become.
The whole point of component architectures is to provide stable,
independent components. Something like zlib should be a different
component, and it should be used unmodified in GCC. If it needs
modification, we should work to put together a new zlib release, and
we should use that release. Trust me -- I'm finding it very hard to
get a GCC 3.0 release done, largely because I'm waiting for people to
stabilize various pieces of the tree, and complete little bits of
implementation. The more intertwined things are, the harder that will
be.
There's a good recursion-theoretic analogy here, but I won't try that
one...
--
Mark Mitchell mark@codesourcery.com
CodeSourcery, LLC http://www.codesourcery.com