The Wrapper Project
simo at samba.org
Wed Nov 20 14:26:15 MST 2013
On Thu, 2013-11-21 at 08:38 +1300, Andrew Bartlett wrote:
> On Wed, 2013-11-20 at 19:59 +0100, Jelmer Vernooij wrote:
> > Andreas is proposing moving a chunk of code out of the Samba codebase
> > taking over maintenance of it, removing it from Samba and having Samba
> > depend on that code. 
> > This means that code to which everybody in the Samba team could easily
> > contribute to previously is now harder to edit, harder to follow, and changes
> > to it are no longer necessarily audited.
> My reservations are based around how we develop, but also how we build
> and ship the code as well, because to me it is critical that we continue
> to build and ship these testing wrappers in-tree.
Why doesn't it suffice for them to work ?
We do not ship glibc either, I guess that was a terrible mistake on our
part ? </sarcasm>
> These wrappers were originally developed specifically to make 'make
> test' trivial enough to run that not only all developers could use it,
> but that all our build farm hosts could as well. I'm concerned about
> any changes that make that harder for Samba.
Concern noted, beyond the concern is there any evidence this is the
case ? Because, you know, it is easy to change again should we find out
that this was some terrible mistake.
> I think it's great that other projects can use this technology, but we
> need to ensure that changes made to Samba are the best for Samba,
> because this code is *absolutely critical* to our CI infrastructure.
We have ton of absolutely critical code that would be *better*
maintained by their upstream. The fact a piece of code is critical or
not is not a good reason to have it in samba. glibc is absolutely
critical yet it makes absolutely no sense to have it in samba, the same
may be true for other code.
We have *too much* code already, we can hardly handle what we have.
We should be happy to give away stuff so we can keep our attention on
And this is one of those cases, there are other projects that can
greatly benefit from this infrastructure to the point that Andreas took
the code and already greatly improved it. And they can benefit samba in
return by improving on it and giving back fixes and maintenance, it is
only *fair* that the project resides on "neutral ground" in its own
repo. If the other projects were to reason the same way they would
simply take a copy and go their way and give nothing back. That is not
how you collaborate and share the burden for the benefit of a larger
Let it go, this code is not *so critical* that if lives somewhere else
it is the end of the world.
More information about the samba-technical