On Mar 31, 2006, at 1:10 AM, Don Brown wrote:
Ok, as I understand the current system, we'd need a GA release of
all the subprojects, including Action 1, in order to declare a GA
1.3 release.
How much work does those involved think this will be? If a bit,
I'm thinking it might be easier in the long run to go ahead,
consolidate the subprojects, and even merge a few like taglibs and
EL. Then, we can push for a single release instead of having to
roll and evaluate the 6 subprojects, each might take a few
releases to get GA.
I'm curious about the relationship between a 'release' and a
'artifact'. Does a single artifact (e.g. struts-taglib-x.y.z.jar)
require it's own separate release? Is this a requirement?
I also want to ask how everyone feels about our current build
process. Are you using Ant, Maven 1 or Maven 2?
I'm not against merging subprojects, especially if it means
simplifying the build and dependencies. When we initially setup the
Maven 1 build, most subproject's build configuration simply inherited
from the master (which get's svn:external'd into each). This
approach reduces a lot of confusion between project interdependencies.
At this point we should probably take a vote and finalize a plan to
merge (as suggested below) what we can.
I'm willing to defer to those who have been doing all the hard work
to get this far, and I'd like to help do whatever it takes to get a
stable release out there. I'm just wondering if one larger release
will be quicker and easier than 6 releases. In fact, here is an
idea at how to getting a release out ASAP and make Action easier to
manage:
- Put all subprojects under Action 1
What do you mean exactly? Since there isn't a directory called
'action1', I can only assume you mean to put them all under the same
named release, the 'Action 1' release and not actually moving (svn
mv) them.
If so, I'd be +1
- Merge Faces into Extras
By this, I assume you mean physically 'svn mv' the source and modify/
merge the dependencies for the build. Also +1
- Merge EL into taglibs
+1 again (Does that affect our minimum supported servlet spec?)
- Get strict with tickets - only let bugs hold us up, and then,
only ones we agree should hold us back. Most projects ship with
bugs, just we'd make sure to document them.
Agreed.
Bottom line, the fewer subprojects and even "modules", the quicker
a GA release will happen and the easier the project will be to
manage. We are _way_ overdue on this release and I'm now hearing
some literally don't think it will _ever_ be finished, and worse,
that we don't care.
It's funny how 'being excited about something new' somehow translates
into 'not caring about something else'. I find this to be pretty
characteristic of nay-sayers and negative people in general (e.g. our
trolls).
I agree 100% that we are overdue for a release and I'd like to help
make that happen.
Don
--
James Mitchell
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
