Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
[wtp-dev] Minutes of WTP Status Telecon, 2005-06-30


Raghu Srinivasan, Rob Frost, Igor Shabalov, Paul Meijer, Ted Bashr, Hoi Lam, Karl Lum, Thomas Yip, Gorkem Ercan, Naci Dai, Amy Wu, Arthur Ryman, Chuck Bridgham, Craig Salter, David Williams, Jeffrey Liu, John Lanuti, Keith Chong, Larry Dunnell, Lawrence Mandel, Nitin Dahyabhai, Peter Moogk, Phillip Avery, Susan Yeshin, Timothy Deboer


1. Approval of Minutes

Are there any correctives to last week's minutes? [1]



2. Review of Open Action Items

1.1 2005-06-16 [ACTION] Jeffrey to fix Bugzilla 82185 [1] which tags releng component build scripts and maps.
                Jeffrey - I submitted a patch to Bugzilla. I am waiting for David and Naci to review it before I commit it.

                David - I'll review it. Should we tag each build? Certainly each I-build. Why not tag them all?

                Arthur - Even is we don't recreate the builds, it makes comparing source code easier. You can at least find the source used for a build.

                Jeffrey - Does Eclipse tag I-builds?

                David - Yes.

                Jeffrey - Naci was concerned about tagging I-builds.

                David - Perhaps this is a UI problem. Too many tags make some views hard to read. Jeffrey should tag all builds except Nightly builds

                [RESOLVED] We will tag all builds except Nighlty builds.  We will not retag Milestone builds since they will already have valid tags.

                Jeffrey - We holding off until doing this until M5 is released. Naci is also looking at it.

                Naci - We are splitting releng into the maps and scripts. We will tag the maps.

                Arthur - Are we tagging the scripts?

                Naci - No. It might be hard o reproduce builds.

                Jeffrey - We are getting the maps from HEAD.

                Arthur - I'm leaving this open until the fix is implemented.


3. JavaOne Update - Tim Wagner (if available)

Igor - I attended. Exadel had Eclipse 3.1 in its booth. Very good reaction. We had a product based on WTP.  IBM held a good reception.

4. M5 Status - All

David - The M5 candidate is pretty good quality. We'll do one more build. Ready in 30 minutes. Testing looks OK. It has some problems but no blockers. My overall assessment is Go.

Naci - I was testing with this morning. Saw some stack traces. Had to restart Eclipse. I will assess the next build.

Ted - Can David pls send out a note and give us a window for testing.

David - You can download now. The zips are there. The tests are completing. We need more testing. Please report status by 5pm.

Peter - Go for Web service.

Criag - Go for XML.

Tim - Go for Server, modulo Gorkem.

Gorkem - (phone line dropped)

Chuck - Go for J2EE after we verify this driver has our fixes.

Larry - Go for Data.

David - Go for SSE.

Arthur - Only Chuck has approval to commit fixes to resolve problems in the J2EE fixes. If these problems are resolved, then we'll declare M5.

5. Bug Status - Jeffrey Liu

Component leads should be prepared to summarise the status of their components with respect to the following metrics:

5.1 Backlog of Resolved but Unverified

Jeffrey  - over 700 remain - look at top 10 reporters. and Verified but Unclosed Bugs - please close these. Nitin wins!

[ACTION] Nitin will verify his fixes by next week.

[ACTION] Arthur to add personalized Bug List queries to people pages (by reporter, by owner).

[ACTION] Component owners to follow up with reporters of fixed bugs and ask them to verify the fixes.

Chuck - I choose the blockers and sent out note this morning.

Jeffrey - I volunteer to help follow up.

Arthur - The reporter doesn't have to verify.

Jeffrey - I asked them to put a note in the bug and let me verify it if they don't have time to verify on M5.

Lawrence - In general we should get reporters to verify.

Jeffrey - Note that there are some differences in the queries dues to Bugzilla limitations.

5.2 Backlog of Bugs in Inboxes

Jeffrey  - last week we agreed to get this to zero, only a few components

Larry will follow up for Data.

Naci will followup for releng.

5.3 Review of Major, Critical and Blocker Bugs not Targetted to 0.7

Jeffrey - Per last week the suggestion was to add a rationale comment and then set the priority to P4. None of deferred bugs have justification.

[ACTION] All component owners will add rationale and set priority to P4 for serious defered bug by next week.

5.4 Backlog of Major, Critical and Blocker Bugs Targetted to 0.7

Jeffrey has posted a Web page that contains the Bugzilla queries we need to review the status. [1]

Ted - Thx to Jeffrey.


6. Stability of I-Builds During Shutdown - Tim deBoer

Tim started a thread on this topic. [1] Do we have a concensus?

Tim -  We are planning RCs on 13, 22, and end of month. This is too much time delay. We need more frequent stable builds.

Naci - I was planning to continue to create I-builds.

Ted - What about continuous map-builds?

Naci - The warm up I builds are currently visible. They are not really I-builds until they pass manual testing. They should be hidden until we declare they are good.

Jeff - Why not just link them from a different page?

Naci - Yes, or a mailing list, but not main download. We still need manual testing because we don't have enough automatic testing.

Arthur - Are we seeing different test failures in HEAD versus map builds?

Naci - There is a consistent pattern dues to test quality but this has improved.

David - Shouldn't we be skipping nightly builds?

Tim - What about experimental fixes?

David - Developers should be able to resolve problems in PDE.

Ted - N-builds are useful for us.

David - Developers should have clean local builds before committing to HEAD.

Craig - What's the benefit of skipping HEAD builds?

Naci - Machine resources.

Jeffrey - I agree with David. There should only be high quality fixes at this point.

David - The Eclipse Platform only does HEAD builds on request. We need to discourage the behavioir of reliance of HEAD builds to find problems.

Tim - What about collecting multiple fixes from team members?

Nitin - Only bugs with well understood fixes should be committed.

David - I've been developing with all WTP plug-ins loaded and have no performance problem. Full builds from CVS take 30 min. The only problem local builds can't find is map file errors. I don't run all JUnits.

Arthur - We are not going to achieve a concensus in the remaining time. Pls continue the discussion in the wtp-dev list. However, it does sound like we should be expecting developers to at least resolve most compile and build problems locally before they commit fixes now.


7. CVS Brownout Status - Arthur Ryman

Arthur - We experienced erratic CVS connectivity during the download frenzy following the release of Eclipse 3.1. This prevented some builds. I've escalated the issue and requested isolation of the development and Web infrastructures to prevent this in the future.

8. Help System - Susan Yeshin

Susan - User docs are componentized. How should ISV docs be laid out. Should we have one book for WST and one for JST?

Peter - Why not one for each component.

Igor - As an ISV, I would like to see more introductory information for each component.

9. Other Issues - All


Arthur Ryman,
Rational Desktop Tools Development

phone: +1-905-413-3077, TL 969-3077
assistant: +1-905-413-2411, TL 969-2411
fax: +1-905-413-4920, TL 969-4920
mobile: +1-416-939-5063, text: 4169395063@xxxxxxx

Back to the top