On 12/14/2010 09:42 AM, Dale Henrichs wrote:
> It seems to me that the hudson server is already building the needed
> release artifacts as well as running the tests ... so building the
> release artifacts manually seems redundant and error prone.
>
> The only fly in the ointment is that the hudson server (presumably)
> loads the latest code whenever it does it's build/test cycle.
>
> Which means that 5 minutes after the decision to make a release someone
> could check in a new package that wasn't intended to be part of the release.
>
> If the hudson server built a set of artifacts based on the Metacello
> configuration for Seaside30, then we'd have a stable definition for the
> product artifacts...then the process would be something like:
>
> 1. Update version numbers...
> 2. Load all latest code....
> 3. Update the Metacello config...
> 3.5 wait 24 hours
> 3.6 verify that hudson tests pass
> 4. (should be done by hudson server?)
> 5. snap off a new change log
> 6. copy hudson artifacts to website...
> etc...
>
> Dale
> _______________________________________________
> seaside-dev mailing list
>
[hidden email]
>
http://lists.squeakfoundation.org/mailman/listinfo/seaside-devBTW, I downloaded the archive of seaside3-oneclick, seaside3-tests,
seaside3-swazoo, and seaside3 and verified that (from a Grease and
Seaside30 perspective) no unwanted mcz files were included, so I've
captured the artifacts that are needed to perform step 6, but I don't
think I've got permission to actually update the website...
Dale
_______________________________________________
seaside-dev mailing list
[hidden email]
http://lists.squeakfoundation.org/mailman/listinfo/seaside-dev
Locked
