I've completed some tests here, and this generally seems fine. With the patch I uploaded to GitHub (http://gist.github.com/348454), I was able to get a clean run with the whole itest suite passing. The patch is only for the itest client, so I don't think its show stopping if it isn't committed, but might be nice. The test script (one of David's from a previous release) I used is here if anyone wants to use it: http://people.apache.org/~jgallimore/build.xml<http://people.apache.org/%7Ejgallimore/build.xml>
I've run Rat over the source which looks ok to me, some xml/config files without headers but nothing else. My Rat output is here: http://people.apache.org/~jgallimore/rat-output.log<http://people.apache.org/%7Ejgallimore/rat-output.log> I agree with Kevan and David J - I think we should have a signed source archive as well. Happy to help with any release stuff I can if you decide it needs re-doing, I realise its pretty time-consuming. Jon On Mon, Mar 29, 2010 at 11:59 PM, David Jencks <[email protected]>wrote: > > On Mar 29, 2010, at 2:22 PM, David Blevins wrote: > > > > > On Mar 24, 2010, at 8:51 AM, Kevan Miller wrote: > > > >> > >> On Mar 24, 2010, at 3:45 AM, David Blevins wrote: > >> > >>> Ok, finally got some binaries up. Took a few spins but finally got > something up with good signatures. > >>> > >>> The branch to become a tag: > >>> > >>> http://svn.apache.org/repos/asf/openejb/branches/openejb-3.0.2/ > >>> > >>> The binaries: > >>> > >>> > https://repository.apache.org/content/repositories/orgapacheopenejb-014/ > >>> > >>> So assuming these binaries pass the TCK, here is my +1 vote. > >>> > >>> Vote open for 72 hours or as long as it takes to get some TCK numbers > against the proposed binaries. > >> > >> > >> Source and build look good. Is there a signed source archive? I didn't > find one... > > > > Looks like for 3.0.1 we just relied on the individual signed source > archives created by maven, plus the svn tag. Seems good enough for this > release as well. > > > > I can whip up a tar of the tag if we feel we need it. > > My understanding of current apache policy is that the only actually > required and important artifact in a release is a source bundle sufficient > to build the project from. Everything else is maven candy :-) > > If you use the current apache 7 pom as an ancestor you get this for free. > > Since AFAICT this source bundle is missing from the vote I have to vote -1 > on what is out there now. I'd be pretty uncomfortable with a "whipped up" > source archive that wasn't produced by the normal maven release procedure. > > thanks > david jencks > > > > > -David > > > > > >
