myfaces-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "simon.kitching@chello.at" <simon.kitch...@chello.at>
Subject Re: [myfaces core] Is there any task left for release myfaces core 1.1.6?
Date Fri, 01 Aug 2008 13:07:36 GMT
simon.kitching@chello.at schrieb:
> Leonardo Uribe schrieb:
>>
>>
>> On Fri, Aug 1, 2008 at 7:33 AM, simon.kitching@chello.at 
>> <mailto:simon.kitching@chello.at> <simon.kitching@chello.at 
>> <mailto:simon.kitching@chello.at>> wrote:
>>
>>     Matthias Wessendorf schrieb:
>>
>>         On Fri, Aug 1, 2008 at 2:20 PM, Leonardo Uribe
>>         <lu4242@gmail.com <mailto:lu4242@gmail.com>> wrote:
>>         
>>             Hi
>>
>>             As planned to release tomahawk, It could be good
>>             (optional) to release
>>             myfaces core 1.1.6. <http://1.1.6.>
>>
>>             Is there any task left to do this? If no I'll start the
>>             release procedure.
>>               
>>
>>         sounds good to me
>>         
>>     Sounds good to me too.
>>
>>     One thing I have meant to do is to check that all the info from
>>     the (long obsolete) .xml files that sit next to the component
>>     .java files has been migrated to an appropriate place, then remove
>>     them. But that's not critical for a release.
>>
>>     I would suggest giving the release candidate for this a reasonable
>>     time (>1 week) for the user community to test. There have been
>>     some radical changes since the last release, and our unit tests
>>     are not great so getting real-world testing for this would be very
>>     useful.
>>
>>     But I would also suggest stating in the RC announcement that only
>>     *regressions* from the 1.1.5 release will be looked at during the
>>     rc cycle. We need to get the release cycles going again, even with
>>     known issues - as long as they are not regressions.
>>
>>
>> Ok, sounds good taking into account the latest changes, but I have 
>> never seen how a release candidate procedure works. I suppose it is 
>> the same as normal release but there is no vote, just an announcement 
>> about it and where to find the artifacts, and those artifacts are not 
>> published on main maven repo, right?
> I think that passing around something that has the final version 
> number in it is too dangerous. So instead how about creating a tag 
> dir, and updating the version within that dir to "1.1.7-rc1", then 
> just doing a build and putting the artifacts up on people.apache.org?
>
> Then if testing goes ok, we can either generate the final release from 
> the rc tag dir, or just do a normal release again from trunk 
> (presuming not too much has changed since the rc was tagged).
>
> Ideally we would also push the 1.1.7-rc1 artifacts to the apache 
> snapshot repo, so that maven users can test this rc really easily. I'm 
> not sure how to do that, but it shouldn't be difficult; we can get the 
> manually downloadable artifacts there first, and figure out how to 
> push to the snapshot repo later...

(sorry, please read 1.1.6 instead of 1.1.7 above; got confused between 
core and tomahawk versions :-)

Hmm..actually, what if its version is named 1.1.6-rc1-SNAPSHOT? That's 
more appropriate for pushing to the snapshot repo (and "mvn deploy will 
do so automatically). Does that version# come before or after 
1.1.6-SNAPSHOT? Probably doesn't matter, as people will be pointing 
directly at one or the other.

We could presumably do a tomahawk release candidate in the same way, and 
send it out for testing at the same time (ie tomahawk-1.1.7-rc1 can be 
sent out after core-1.1.6-rc1 but before core-1.1.6 has been released).

Cheers,
Simon


Mime
View raw message