Where is the talloc repository?
obnox at samba.org
Mon Mar 21 15:33:16 UTC 2016
On 2016-03-18 at 21:51 +0100, Andreas Schneider wrote:
> On Saturday 19 March 2016 08:29:33 Andrew Bartlett wrote:
> > On Tue, 2016-03-15 at 15:50 +0100, Michael Adam wrote:
> > > - According to Andreas, it happens that samba ran against
> > > unreleased talloc / libfoo code. But with the abi checks
> > > in place, how could this actually happen?
> > > ==> need a concrete example.
> > I actually think it is good that git master of Samba can run with
> > unreleased talloc/tdb/ldb changes. It isn't OK to release with those,
> > but otherwise 'technically' we would have to do a core lib release
> > after every git commit, which just seems wrong.
> The problem are not unreleased talloc/tdb/ldb changes.
> The problem is a version bump in the core libs and without a
> tarball release which means Samba requires a version which has
> not been released yet.
Hmm. I don't quite get the point yet...
What is a release? I think the existence of a tarball is not
the decisive factor. The decisive factor for a release is the
existence of the version bump and an according tag in the source
tree. From that the tarball can easily be produced if needed.
So if we had a release in the tree but there was no tarball,
then this is am a mistake, an omission, but not a severe
problem, since it can be fixed in hindsight when spotted by
(e.g.) distro packagers or the like.
Mistakes and omissions happen, but it should not be the rule.
When they happen and someon spots them, we talk about them and
fix them. And try not to let it happen again. (See the failure
to release the fixes that were already in uid_wrapper, exactly
the same kind of thing.)
So what really is the problem with the libs?
Cheers - Michael
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 198 bytes
Desc: not available
More information about the samba-technical