Hi Shane,
Sorry should have thought about whether you were subscribed to the
list. It's probably the best place to discuss it, let me know if you
are so I can drop the CC.
On 05/12/2008, at 7:44 AM, Shane Isbell wrote:
I'm in agreement that we shouldn't be pushing .NET assemblies out
just yet.
ok, are you still in agreement that /maven2 is the best eventual
place, or a separate repo on central?
I'm fine with keeping it in "beta" and subject to change for now,
which is essentially what the current repo is, but I think it'd be
good to move it off vmbuild now.
One area I'd like to discuss is tying together keys with group Ids.
So if
someone from Apache (or where ever), compiles the artifact, they own
it and
should use their own groupId for the artifact. A couple of years ago,
compiling a version of NUnit with .NET or Mono would result in an
assembly
that was incompatible with DotGNU, so I had to make some tweaks to
the NUnit
source to get it to work, resulting in a different artifact. I would
not
want to put that up in the org.nunit namespace.
Sure, this is compatible with the current central policy.
We could also do signature verification to group id before allowing
the
upload.
Personally, I like the upper case on the artifact Ids, as this
follows .NET
standard conventions, but it's not something I care about dictating to
people. If a project wants to put it under their own group ID (since
changing the name requires recompiling or at least modification of the
assembly), then I don't really care what they do in terms of naming
the
artifact. I think in the case of the original nunit.framework class
it is
already lowercase. So that's the one that would be used.
Yep, that's what Wendy was suggesting as well.
A couple of other issues we should address: requiring assemblies being
signed before loading into public repo;
Agreed. In the nunit case, they are the original artifacts, and from
what I can tell they are signed.
and looking into what meta-data we
want embedded within the assembly. I could see groupId as a meta-tag
being
very useful.
You mean defining a common meta-tag name that others would use? I
think that makes sense. Requiring it might be limiting though, and re-
signing things to add it in afterwards wouldn't be practical.
Cheers,
Brett
--
Brett Porter
[EMAIL PROTECTED]
http://blogs.exist.com/bporter/
