Niklas Gustavsson wrote:
On Fri, Jun 27, 2008 at 2:34 PM, Julien Vermillard
<[EMAIL PROTECTED]> wrote:
Yes, very true, the problem is for producing all the artifact, actually
it's a few hours (last time it took me half a day) for generate and
verify everything (signature files, reports, binaries, etc..).
Yeah, it's a pain. However, since we need to review them before the
release, it's probably better for one person to go through the process
rather than all voters.
IMHO, we should do what is necessary to get the candidate release being
testable :
- rev number given
- binary prepared
- branch created
But when it comes to signature, reports, etc, this is a little bit
overkilling. If the release is not done correctly, because we have
produced the wrong signatures, then we do another release. As soon as we
are able to test the binaries, we should be ok.
Perhaps we should do a two stage vote, one for saying current branch is
ok for releasing, then if it's accepted, someone do the tedious release
dry-run job, and a second vote for reviewing the related artifacts.
I don't think we need a formal +1/0/-1 vote for the first decision, a
discussion thread would probably be sufficient.
I agree. Or we would soon vote to launch a vote :)
--
--
cordialement, regards,
Emmanuel Lécharny
www.iktek.com
directory.apache.org
