Re: [jetty-dev] 7.0.0.M0 progress

[Jesse McConnell <jesse.mcconnell@xxxxxxxxx>]

Ya, we'll drop this staged release and roll another one, I fixed up an
import upperbounding osgi issue that jeff noticed on another thread...

looking reasonable from what I have seen on this so far..

jesse

--
jesse mcconnell
jesse.mcconnell@xxxxxxxxx



On Wed, Apr 1, 2009 at 14:41, Jesse McConnell <jesse.mcconnell@xxxxxxxxx> wrote:
> I have deployed the 7.0.0.M0 version to the staging location..
>
> http://oss.repository.sonatype.org/content/repositories/jetty-stage-003/
>
> So it is clear this is likely _NOT_ what will be released, it is an
> intermediary staging so we can get comfortable with what it will look
> like and to check out the osgi manifests, etc etc.
>
> Anywho...wanted to link it to folks to review.  The traditional
> download artifact is located in o.e.j.jetty-distribution
>
> cheers!
>
> jesse
>
> --
> jesse mcconnell
> jesse.mcconnell@xxxxxxxxx
>
>
>
> On Wed, Apr 1, 2009 at 13:23, Jesse McConnell <jesse.mcconnell@xxxxxxxxx> wrote:
>>> Jan has checked most of the OSGi meta data except webapp and xml,
>>> so can you have a look at those.
>>
>> I went back over this and ended up axing out most everything..here is why..
>>
>> * updated the bundle plugin to 2.0.0
>> * part of the reasoning behind our flat dir structure + segmenting of
>> packaging in our jars was so that these things could be automatically
>> generated for us, ergo no more having to fiddle with this goop
>> * in the jetty-project pom, the plugin is configured with the
>> duplication that was in each of these
>> * checked the MANIFEST against that of an equinox one and we have the
>> missing Bundle-* items in there now
>>
>> Now, regarding the import and export packaging..I read up on the
>> workings of this from the bundle plugin documentation as afaick we are
>> fine with the defaults.  We would only need to be masking out internal
>> implementation packages if we were exporting some sort of api.  The
>> automatically generated Import and export are now pretty large, but
>> they accurately (at least from what I have checked out) the current
>> situation.  I did grab a chunk of one and mail separately to Jeff
>> McAffer to sanity check the defaults we are using now.  Seems to me we
>> could be ok with o.e.j.server.* without specifically stating each
>> subpackage but hopefully he can clue is in if that finer grained
>> resolution is valuable.
>>
>>> I definitely broke annotations when I gutted them from the early
>>> 3.0 version, but I'm not sure if we want to hold up an M0 for that.
>>
>> not a big deal imo
>>
>>> I've built org.cometd against a 7.0.0.M0 snapshot using the
>>> new continuation API and that is working fine (albeit with only
>>> a little testing other than the test harnesses).
>>
>>
>>> As part of that, I switched the 7.0.0.M0 to use the javax.servlet
>>> servlet-api rather than the mortbay servlet-api.  It felt cleaner
>>> than having excludes to resolve conflicts.  Do we need to do a new
>>> CQ for that?
>>
>> Already added that after you mentioned wanting to do this last week :)
>>
>> https://dev.eclipse.org/ipzilla/show_bug.cgi?id=3225
>>
>>> Other little issues are:
>>>
>>> The distribution is not pulling in the new jetty-continuation jar
>>
>> fixed
>>
>>> The distribution is not creating an empty logs directory
>>
>> fixed
>>
>> cheers,
>> jesse
>>
>