Bugzilla – Full Text Bug Listing
| Summary: | Migrate technology.emft.teneo.releng to modeling.emft.teneo.releng | ||
|---|---|---|---|
| Product: | [Modeling] EMF | Reporter: | Nick Boldt <nboldt> |
| Component: | Teneo | Assignee: | Nick Boldt <nboldt> |
| Status: | VERIFIED FIXED | QA Contact: | |
| Severity: | normal | ||
| Priority: | P3 | CC: | Ed.Merks, janet.campbell, mtaal, sharon.corbett, webmaster |
| Version: | unspecified | ||
| Target Milestone: | --- | ||
| Hardware: | PC | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Bug Depends on: | |||
| Bug Blocks: | 194409 | ||
|
Description
Nick Boldt
The catch here is that mtaal will need to file a new ECF to cover the EMF project Can this move wait until after we receive this paperwork, or should I remove him from the groups and do the move anyway and we can re-add him at a later date? -M. Matt: don't do anything until I've finished the steps assigned to me! Why does he need an ECF to move from being an EMFT committer (in tech) to being an EMFT committer (in modeling)? He's not moving to EMF per se, just changing CVS paths. Essentially the problem is that if someone has the ability to put code in the tree somewhere we have to have paperwork covering the code to protect the community. (In reply to comment #3) > Essentially the problem is that if someone has the ability to put code in the > tree somewhere we have to have paperwork covering the code to protect the > community. Fair enough. We should make sure then that the other EMFT folks have gone thru this paperwork too. Everyone with a "emft-*" group id will eventually move into /cvsroot/modeling/org.eclipse.emf/* -- the last projects to move are Eike Stepper's Net4j and CDO components. Did Christian have to do this for Query, Transaction and Validation? What about the EMF Search, EMF Compare, JCR Management and MWE components? I assume so, but if not, then I guess Ed has a bunch of paperwork to do. (Sorry, Ed!) ;-) adding block relationship -- can't complete bug 194409 without doing this bug first Nick, I want to be sure that if we put components into EMF's module that they can still be incubating, and new components can still use parallel IP. It will sure save a lot of hassle if we don't need to move things between a special EMFT module to an EMF module. (In reply to comment #6) > I want to be sure that if we put components into EMF's module that they can > still be incubating, and new components can still use parallel IP. It will > sure save a lot of hassle if we don't need to move things between a special > EMFT module to an EMF module. I agree completely -- see my email to Matt on this subject. Less downstream churn would be great, but only if it's kosher with the EMO and legal folks. Can't see that it's much different from the Equinox incubator being part of the platform project... though they may have gone through a lot of refactoring to move from incubation to production, and we of course want to avoid that, if at all possible. (In reply to comment #2) > Why does he need an ECF to move from being an EMFT committer (in tech) to being > an EMFT committer (in modeling)? Are there two EMFT projects, one in technology and one in modeling? I'm confused. Please see Bjorn's note from today titled 'Fw: Committers who aren't committers". I hope that clarifies everything... (In reply to comment #9) > Please see Bjorn's note from today titled 'Fw: Committers who aren't > committers". I hope that clarifies everything... And if you weren't on the cc: list, the answer is... there's only one EMFT project, but it's currently spread across two places (pending the completion of the move from tech to modeling). DONE: * [Nick] copy cvs files from /cvsroot/technology/org.eclipse.emft/teneo to /cvsroot/modeling/org.eclipse.emf/org.eclipse.emf.teneo * [Nick] copy cvs files from /cvsroot/technology/org.eclipse.emft/releng/teneo to /cvsroot/modeling/org.eclipse.emf/org.eclipse.emf.teneo.releng * [Nick / Martin] mark old CVS files 'deprecated' via a README.txt,v stating that these files are deprecated, where they moved to, and when they will be deleted (I've just said "late 2007 after 0.8.0 is released".) * [Nick] change teneo.releng module to work with the new paths and builder * [Nick] run some test builds to verify it's working [1] * [Nick] update promote script/properties to use new rss/update/downloads/etc. location: should be done but requires verification with first build when Martin's back from vacation * [Nick] update relupdate.php to use new path * [Nick] update website [1]http://emft.eclipse.org/modeling/emft/downloads/?project=teneo&sortBy=date&hlbuild=0#latest TODO: + [Nick] fix Search CVS database (purge old records in filetags, cvsfiles, commits tables; modify releases table) + [Nick/Martin] verify new build and promote tools work properly Once that's done, we'll need to complete the user/group reassignment stuff. Is everyone (EMO, Webmaster, PMC) cool with having emft projects in /cvsroot/modeling/org.eclipse.emf/* ? FYI, changes are listed here: http://www.eclipse.org/modeling/emf/searchcvs.php?q=days%3A14+file%3A%25teneo%25 I'll hold off on the database purge until after Martin confirms a good build/promote. Also, will a maintenance branch be required? If so, please backport my changes to that branch too, as appropriate. Martin (via email) said: > I have moved to the new cvs location (and updated cvs). However I get a build error: http://emft.eclipse.org/modeling/emft/teneo/downloads/drops/0.8.0/I200709031434/buildlog.txt > Do you have any idea what I am doing wrong? Probably something which I should have read in the wiki..... And I replied (via email): > http://wiki.eclipse.org/Modeling_Project_Releng/Building_Zips_And_Jars#Build_Problems_.26_Solutions And he replied (via email): > Hmm the feature.xml and map file and customtargets.xml seem to be okay. > So I am not sure if this is the problem. My guess is that it's your map file. Compare my latest N build [1] to your latest I build [2]: [1]http://emft.eclipse.org/modeling/emft/teneo/downloads/drops/0.8.0/N200708272242/directory.txt [2]http://emft.eclipse.org/modeling/emft/teneo/downloads/drops/0.8.0/I200709031434/directory.txt And you'll notice that I'm building from HEAD whereas you're building from an old tag which doesn't include my fixes [3], [4]. [3]http://www.eclipse.org/modeling/emf/searchcvs.php?q=file%3Ateneo%25feature.xml+author%3Anickb&project=0 [4]http://dev.eclipse.org/viewcvs/index.cgi/org.eclipse.emf/org.eclipse.emf.teneo/tests/org.eclipse.emf.teneo.tests-feature/feature.xml?root=Modeling_Project&r1=1.30&r2=1.31 If you want to tag your files and use a different tag than the old one or HEAD, please feel free. Note that the new build page has options for "use static map in CVS" and "generate map using branch tag (HEAD or Rx_y_maintenance)" rather than "tagged build" vs. "untagged build". This change is for security reasons, so that an anonymous shared ID does not change your cvs repository; instead, if you want to tag your files you need to now do so yourself. Slightly more effort, but also much more control over what goes into your builds. (In reply to comment #13) Building from HEAD passed: http://emft.eclipse.org/modeling/emft/downloads/?project=teneo&sortBy=date&hlbuild=0#latest Please verify build contents (zips), then respin as an I build. You can tag your sources, update your map file (replace string "HEAD" with your tag(s)), then build; or, you can build from HEAD (which is what's in your mapfile now -- I updated the file so that both 'gen map' and 'use map' modes will currently do the same thing). It works fine afaics. So for me this issue can be closed. Thanks! gr. Martin (In reply to comment #15) > It works fine afaics. So for me this issue can be closed. > Thanks! Can you do a promote too, to ensure that part works? Then we can sort out the ECF stuff and fix the group permissions... (In reply to comment #16) > Can you do a promote too, to ensure that part works? Promote works. > Then we can sort out the ECF stuff and fix the group permissions... Group changes can wait until bug 198541 is sorted out. Closing bug. Move to verified as per bug 206558. |