Bugzilla – Full Text Bug Listing
| Summary: | Consume ECF 3.5 | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Product: | [Eclipse Project] Equinox | Reporter: | Pascal Rapicault <pascal> | ||||||||
| Component: | p2 | Assignee: | DJ Houghton <dj.houghton> | ||||||||
| Status: | RESOLVED FIXED | QA Contact: | |||||||||
| Severity: | normal | ||||||||||
| Priority: | P3 | CC: | irbull, jeffmcaffer, kim.moir, slewis, tjwatson | ||||||||
| Version: | 3.7 | Flags: | tjwatson:
pmc_approved+
|
||||||||
| Target Milestone: | 3.7 RC1 | ||||||||||
| Hardware: | PC | ||||||||||
| OS: | Mac OS X - Carbon (unsup.) | ||||||||||
| Whiteboard: | |||||||||||
| Bug Depends on: | 324248 | ||||||||||
| Bug Blocks: | |||||||||||
| Attachments: |
|
||||||||||
|
Description
Pascal Rapicault
We may need to look into https://bugs.eclipse.org/bugs/show_bug.cgi?id=324248 Jeff, You asked about TODOs left for 3.7 at the meeting today. We should have mentioned this one too. FWIW: The ECF community is currently trying to decide when to have our next maintenance release (3.5.1). As per the discussion on http://dev.eclipse.org/mhonarc/lists/ecf-dev/msg05127.html it would be desirable for us to have Indigo's p2 use ECF 3.5.1 once it's released. This is particularly true because there has been one very minor (and not related to p2 at all) fix to ECF core bundles since 3.5. Just for reference, it's bug 341820. It would be nice for our community to have this fix included in Indigo's copy of ECF...so that they can use the fix without patching ECF. The current proposal for 3.5.1 release is for sometime around the end of May, 2011. If the timing of this doesn't sit well with people this bug (i.e. earlier is needed or something), then please let me know. Consuming 3.5.1 should be fine. In the meantime I will consume 3.5.0 Pascal, the build.properties of the platform feature will have to be updated to reflect the new source bundle versions of ECF. Please coordinate this change with me so I can update this feature and avoid build breakage. Where are we at with this bug. M7 is getting very close and we have not moved up to 3.5 ECF AFAICT. (In reply to comment #3) > > The current proposal for 3.5.1 release is for sometime around the end of May, > 2011. If the timing of this doesn't sit well with people this bug (i.e. > earlier is needed or something), then please let me know. Unfortunately this date doesn't sit well for me. End of May means we will not be able to consume it until our very last scheduled build (RC4) for Indigo. This seems quite risky to me. (In reply to comment #6) <stuff deleted> > > The current proposal for 3.5.1 release is for sometime around the end of May, > > 2011. If the timing of this doesn't sit well with people this bug (i.e. > > earlier is needed or something), then please let me know. > > Unfortunately this date doesn't sit well for me. End of May means we will not > be able to consume it until our very last scheduled build (RC4) for Indigo. > This seems quite risky to me. There have been no changes to any of the code consumed by p2 for the 3.5.1 stream (from 3.5.0), and there are none anticipated...except for one minor manifest cleanup for bug 342745. In other words, wrt the ECF filetransfer and core bundles, the ECF code will not change for 3.5.1. So that doesn't seem risky to me...as it's just another build from the equinox/p2 perspective. We need time to use the newly arrived TCK (for remote services/remote services admin), and to fix small bugs prior to ECF 3.5.1. (In reply to comment #7) > > There have been no changes to any of the code consumed by p2 for the 3.5.1 > stream (from 3.5.0), and there are none anticipated...except for one minor > manifest cleanup for bug 342745. In other words, wrt the ECF filetransfer and > core bundles, the ECF code will not change for 3.5.1. So that doesn't seem > risky to me...as it's just another build from the equinox/p2 perspective. > Any chance we can consume the bundles p2 uses from 3.5.1, right now? It sounds like you are done with development of these bits. > We need time to use the newly arrived TCK (for remote services/remote services > admin), and to fix small bugs prior to ECF 3.5.1. makes sense, I assume theses changes are outside the bundles p2 consumes from ECF? (In reply to comment #8) > (In reply to comment #7) > > > > There have been no changes to any of the code consumed by p2 for the 3.5.1 > > stream (from 3.5.0), and there are none anticipated...except for one minor > > manifest cleanup for bug 342745. In other words, wrt the ECF filetransfer and > > core bundles, the ECF code will not change for 3.5.1. So that doesn't seem > > risky to me...as it's just another build from the equinox/p2 perspective. > > > > Any chance we can consume the bundles p2 uses from 3.5.1, right now? It sounds > like you are done with development of these bits. As far was we know, we are done with the development of these bits (modulo the bug 342745 as mentioned above). However, it's a real hassle and time cost for us to separate out the build of the filetransfer and core bundles from the build of the rest of ECF...in fact such a separation is what we have been trying to move away from...as it greatly complicates the build of the non-core parts of ECF for us...and we don't have resources to waste on releng. > > > > We need time to use the newly arrived TCK (for remote services/remote services > > admin), and to fix small bugs prior to ECF 3.5.1. > > makes sense, I assume theses changes are outside the bundles p2 consumes from > ECF? Yes. Any changes are generally going to be in the RSA implementation...i.e. org.eclipse.ecf.osgi.services.remoteserviceadmin. With the exception of the two ECF core bundles, this bundle nor any of its dependencies are consumed by p2. Further, I think it's very unlikely that the ECF core bundles will be implicated in any of the RSA TCK work...as they didn't change at all with the RSA implementation. If the ECF core bundles *are* somehow implicated in the RSA changes then we would likely call off 3.5.1 completely (and Indigo would just use ECF core 3.5.0). (In reply to comment #9) > > As far was we know, we are done with the development of these bits (modulo the > bug 342745 as mentioned above). However, it's a real hassle and time cost for > us to separate out the build of the filetransfer and core bundles from the > build of the rest of ECF...in fact such a separation is what we have been > trying to move away from...as it greatly complicates the build of the non-core > parts of ECF for us...and we don't have resources to waste on releng. I was not suggesting you change anything on your releng process. What I was hoping for was a p2 repository that contained a current build for your 3.5.1 development which we could consume the ECF bits out of until 3.5.1 goes final. I assume you already have regular builds containing your 3.5.1 development. If that is not the case then we would have no choice but to wait. > If the ECF core bundles *are* somehow implicated in the RSA changes then we > would likely call off 3.5.1 completely (and Indigo would just use ECF core > 3.5.0). OK (In reply to comment #10) <stuff deleted> > I assume you already have regular builds containing your 3.5.1 development. Yes, see here https://build.ecf-project.org/jenkins/job/C-HEAD-sdk.feature/ Pascal, have we started consuming 3.5.0 yet? Do you want to do this now or in RC1? regarding comment #11 Scott, is there a location where this repo is available on an eclipse.org server? I notice that https://build.ecf-project.org/jenkins/job/C-HEAD-sdk.feature/ isn't an eclipse.org server. (In reply to comment #13) > regarding comment #11 > > Scott, is there a location where this repo is available on an eclipse.org > server? No, not yet. There will be when we do a our 3.5.1 maintenance build in a couple of weeks. >I notice that > https://build.ecf-project.org/jenkins/job/C-HEAD-sdk.feature/ isn't an > eclipse.org server. Right...it's our own server. Created attachment 194608 [details]
patch for map file
Created attachment 194622 [details]
patch for feature
Kim and I have released the patches. I released changes for the failing p2 tests on the new version. I kicked off a test build on hudson to verify the changes are OK. Created attachment 194641 [details]
patch for javadoc
The test build completed successfully. Once ECF 3.5.1 is released we will open a new bug and consider it. Closing. (In reply to comment #19) > The test build completed successfully. > Once ECF 3.5.1 is released we will open a new bug and consider it. > Closing. This timing is going to be difficult given I think we will be into RC4, our very last build. (In reply to comment #20) > (In reply to comment #19) > > The test build completed successfully. > > Once ECF 3.5.1 is released we will open a new bug and consider it. > > Closing. > > This timing is going to be difficult given I think we will be into RC4, our > very last build. When would work for a final 3.5.1 build? (In reply to comment #21) > (In reply to comment #20) > > (In reply to comment #19) > > > The test build completed successfully. > > > Once ECF 3.5.1 is released we will open a new bug and consider it. > > > Closing. > > > > This timing is going to be difficult given I think we will be into RC4, our > > very last build. > > When would work for a final 3.5.1 build? Now would be the best. But I realize that probably unreasonable ;-) I would not want it to slip past RC2 which starts building on May 19th. http://www.eclipse.org/eclipse/development/plans/freeze_plan_3_7.php (In reply to comment #22) <stuff deleted> > > > > When would work for a final 3.5.1 build? > > Now would be the best. But I realize that probably unreasonable ;-) :) Yes...unless someone would like to contribute to the effort to complete the Remote Service Admin TCK compliance work...or the ECF releng. > I would not want it to slip past RC2 which starts building on May 19th. > http://www.eclipse.org/eclipse/development/plans/freeze_plan_3_7.php Ok, I'll schedule ECF 3.5.1 final for May 19th. |