Difference between revisions of "CBI/May15 2012"
(New page: =Meeting= * April 17, 2012, 10am EST * Eclipse Conference facility ==Attendees== Igor F., Thanh H, Denis R, Paul W., Carl A., Andrew R. ==Minutes== 1) Eclipse platform build (CBI) ups...)
|Line 1:||Line 1:|
* , 2012, 10am EST
* Eclipse Conference facility
* Eclipse Conference facility
Revision as of 10:31, 30 May 2012
- May 15, 2012, 10am EST
- Eclipse Conference facility
Igor F., Thanh H, Denis R, Paul W., Carl A., Andrew R.
1) Eclipse platform build (CBI) upstream
We are inching ever closer to being ready to push the work to CBI'ify the Eclipse platform upstream. Doing this has multiple significant benefits: a) we can remove the forked repositories b) we can hopefully avoid bumping all qualifiers as would be likely if we pushed upstream after SR0.
Paul: CBI has been a topic for a while now in the eclipse project & architecture meetings.
FYI, for platform:
- RC1 this week
- RC2 next week
- RC3 the week after
Accepting the pom files makes sense.
Andrew: For non-pom changes, the platforms we don't support issue (e.g. AIX, HP-UX, Solaris,etc) - certainly won't be accepted as is since PDE uses the same data. Igor: not sure of how to do this just yet. Andrew: script it?
Andrew: Anything else to worry about? Thanh: What about the empty source bundles? (he commented out 5 or 6) Igor: source bundles that don't have any java sources (are what he commented out) should be fine. Paul? Paul: (Igor's listed some examples of empty source bundles) That's a bug upstream, so a bug should be raised. Paul took an action to figure this out... because it's a change to the PDE build it needs a bit more attention. It doesn't make sense to push up a broken Tycho build because there's a bug upstream. Igor: An alternative might be to tell Igor it's OK to generate an empty source bundle. Not sure this is the right approach.
Paul: In order for Eclipse platform to fully adopt, we need to generate & consume swt and equinox launchers binaries. This could come after SR0.
Pull request - Thanh. (plan A, plan B is patches)
Igor: Qualifier bug work - parent pom level is the goal. If all platform projects use the same format, it'll be all in parent pom.
Paul: checking to see if swt can move to a timestamp based qualifier... Paul will talk to them.
Andrew: aiming for next Tuesday. Who's going to catch the pull requests?
Paul: Can't do all of them unless I use the build ID which isn't appropriate. Paul, John A., Tom Watson, Dani should cover off 95%. Andew: heads up email today, pull request assembled after we hear if we can deal with empty source bundles.
Paul: will send groups to bundle pull requests and maybe reduce down to 6 bugs or so.
2) Eclipse platform test results
Thanh has re-run the test suite after we rebased to M7. Like last time, there is a very high pass rate. He'll share the results. We could use help to debug and solve the remaining issues.
3) CBI 1.0 definition
What does the first release of CBI mean? For example, what are the versions of components and what features do we support?
Looks like Tycho version 0.16 may be the version we fix on. No strong opinions. We'll revisit this later.
4) CBI 2.0 planning
What are the tantalizing opportunities to make life better & easier for many projects & committers using CBI?
- Andrew will write up the software distribution ideas.
- Are there opportunities around building the release repository and packaging?
No strong opinions today. We'll revisit this.
5) WTP update
Work has progressed nicely on CBI'ifying WTP. We'd like to share an update, and talk about the next steps.