Difference between revisions of "Eclipse/Status meetings/Minutes"

From Eclipsepedia

Jump to: navigation, search
(JDT UI (Markus))
Line 38: Line 38:
 
==== JDT UI (Markus) ====
 
==== JDT UI (Markus) ====
 
* cleaning up website for transfer to Git
 
* cleaning up website for transfer to Git
 +
* planning to remove org.junit 3.8.2 from the SDK builds; JUnit 3 tests can also be run with JUnit 4 now
  
 
==== PDE (Curtis) ====
 
==== PDE (Curtis) ====

Revision as of 08:47, 10 October 2012


October 10, 2012

Discussion/Reminders

Status

Platform Workspace (Szymon)

Platform UI (Paul)

Platform Text (Dani)

  • bug fixing / triage / investigation
  • removed the obsolete "File > Convert Line Delimiters To > MacOS 9" (Mac OS X uses Unix' \n)

Platform Debug (Mike / Pawel)

Platform Ant (Mike)

Platform SWT (Silenio)

Platform Releng (David)

  • Mac 64 bit machine (VM and Eclipse) is now used instead of 32 bit. (bug 354774) It is now quickest run of unit tests (though, admittedly we do run a few fewer tests there). Some labeling will remain wrong until bug 390986 is fixed. But it is 64 bit, using a Java 6 VM.
- I plan to remove the "old" Mac 32 bit Hudson jobs unless someone thinks we might need them for some special case?
  • Windows. Sigh. After our performance "tests of the tests" we have had trouble getting ordinary JUnit's running again in a reasonable period of time due to "virtualization tweaking" that's needed. Join in or watch the fun at bug 389857. I think progress is being made, but seems clear it will never "be the same" as running plain windows.

JDT Core (Srikanth)

  • Good progress on JSR308 & JSR335.

JDT Debug (Mike)

  • fixed bugs in 'Open from Clipboard' action

JDT UI (Markus)

  • cleaning up website for transfer to Git
  • planning to remove org.junit 3.8.2 from the SDK builds; JUnit 3 tests can also be run with JUnit 4 now

PDE (Curtis)

Equinox (Tom Watson)

Equinox p2 (Ian Bull)

October 03, 2012

Discussion/Reminders

Status

Platform Workspace (Szymon)

  • back from vacation

Platform UI (Paul)

Platform Text (Dani)

  • vacation
  • bug fixing

Platform Debug (Mike / Pawel)

  • investigating UI issues with Debug actions / tool bars / views
  • started converting all deprecated actions, etc to commands / handlers (bug https://bugs.eclipse.org/390869)
  • bug fixing / triage

Platform Ant (Mike)

  • bug fixing / triage

Platform SWT (Silenio)

  • converted SWT source to 64 bit by default
    • this does *not* affect end users that consume the compiled 32- or 64-bit SWT bundles
    • developers who use SWT from source on 32-bit machines can install the "64to32" charset to avoid tons of changes, see bug 377107

Platform Releng (David)

  • tests of performance tests
  • nearly done with "windows tests" and will resume windows unit tests in day or so

JDT Core (Srikanth)

  • Good progress on JSR 335 and JSR 308.

JDT Debug (Mike)

  • investigating UI issues with Debug actions / tool bars
  • started converting all deprecated actions, etc to commands / handlers (bug https://bugs.eclipse.org/390869)
  • bug fixing / triage

JDT UI (Markus)

  • bug fixing / triage / investigation

PDE (Curtis)

  • Moved core/ui bundles to 1.5 BREE (starting with tonight's N build)
  • Investigating moving API Tools to use ASM 4.0
  • Bug fixing

Equinox (Tom Watson)

  • Juno SR1 updates were halted because of Mac issues. Determined that a buckminster fix is needed bug 390984

Equinox p2 (Ian Bull)

September 26, 2012

Discussion/Reminders

Status

Platform Workspace (Szymon)

  • vacation

Platform UI (Paul)

Platform Text (Dani)

  • vacation

Platform Debug (Mike / Pawel)

  • bug fixing / triage

Platform Ant (Mike)

  • bug fixing / triage

Platform SWT (Silenio)

Platform Releng (David)

  • prepared for Juno SR1 release on Friday
  • I'm prepared to give up trying to fix the 2 p2-related DNFs on Mac tests. I think "manual" proxies really don't work with p2!? Opened bug 390392. Also requested webmasters use transparent proxies which (I'm guessing) would make the tests work again. bug 390401. Should we disable them? DNFs add hours and hours to total test time.
  • bug 389857 - run performance tests on eclipse.org hardware - with windows job
got one run so far. Took 13 hours! May disable some and run shorter set for now?
some performance suites will likely need work, just to get them to run (just a heads up)
  • 27 of 43 (37%) bugs marked fixed in "umbrella bug" for move to eclipse.org bug 381873.
Is that good progress? Or does it need focus?
In terms of actual junit failures:
SR0: 528 failures 5 DNFs
SR1: 506 failures 2 DNSs
one suite (still) disabled: org.eclipse.ui.tests.rcp

JDT Core (Srikanth)

  • New engineer onboard ! Welcome Manoj Palat to Eclipse.
  • Progress on JSR 335 & JSR 308
  • Shipped M2.
  • Routine matters

JDT Debug (Mike)

  • bug fixing / triage

JDT UI (Markus)

  • fixed leak test failures in automated tests by closing the intro view

PDE (Curtis)

  • org.eclipse.pde.core and org.eclipse.pde.ui bundles will increase BREE to 1.5 (Bug 319744)
  • CQ approved for new plug-in image browser view
  • bug fixing

Equinox (Tom Watson)

Equinox p2 (Ian Bull)

September 19, 2012

Status

Platform Workspace (Szymon)

Platform UI (Paul)

  • Fixed button ordering that was broken in some dialogs (e.g. "Delete Resource" didn't use "Cancel | OK" on Mac and GTK).

Platform Text (Dani)

  • Juno SR1 status: done
  • working towards Kepler M2 (one day test pass, bug fixing)

Platform Debug (Mike / Pawel)

  • Welcome Markus Keller as new committer

Platform Ant (Mike)

  • bug fixing / triage

Platform SWT (Silenio)

  • committer vote for Alexander Kurtakov (Red Hat) just concluded successfully
  • investigating CTabFolder performance
  • beginning UIA investigation (UIA is new Windows accessibility framework)
  • reviewing and committing some GTK3 patches from Red Hat

Platform Releng (David)

  • Juno SR1 simrel
  • Kepler M2
  • performance tests (basics)
  • Reminder: Kepler M2 sign-offs on Thursday

JDT Core (Srikanth)

  • Good progress on JSR 335 & JSR 308.

JDT Debug (Mike)

  • content assist in Debug and Variables view now shows additional info (Javadoc)

JDT UI (Markus)

  • Fixed various coloring problems for platforms with white-on-black tooltips (e.g. Variable value hover on Ubuntu with Ambiance theme). Common flaw: Code sets bg to SWT.COLOR_INFO_BACKGROUND but doesn't set fg to COLOR_INFO_FOREGROUND.

PDE (Curtis)

  • API Tools EE Descriptions categories now distributed in a single feature
  • Bug fixing

Equinox (Tom Watson)

Equinox p2 (Ian Bull)

September 12, 2012

Status

Platform Workspace (Szymon)

Platform UI (Paul)

Platform Text (Dani)

  • Juno SR1 status: done
  • bug fixing
  • local holiday

Platform Debug (Mike / Pawel)

  • SR1 done
  • bug fixing / inbox triage

Platform Ant (Mike)

Platform SWT (Silenio)

  • Juno SR1 status: done
  • group is back at full strength
  • 4.x performance investigation
  • fixed the ld-linux crash that affects the JUnit tests of some groups
  • nominated Alex from Red Hat as a committer for GTK 3 work
  • 4.3-stream bug fixing

Platform Releng (David)

  • bug 389273 Final (again) EMF udpate for Juno SR1 RC4
  • bug 389053 For Kepler, with Curtis, simplifying o.e.pde.api.tools.ee.feature categories (and adding license)
  • bug 389369 Baby steps to running performance tests on build.eclipse.org

JDT Debug (Mike)

  • SR1 done
  • bug fixing / inbox triage

JDT UI (Dani)

  • Juno SR1 status: done
  • replaced old Linux build machine install with a new one
  • bug fixing
  • local holiday

PDE (Curtis)

  • I (Curtis) may miss the meeting due to an unexpected appointment
  • No changes planned for SR1

Equinox (Tom Watson)

Equinox p2 (Ian Bull)

September 5, 2012

Reminders

RC3 this week with a test pass on Thursday and official sign-off. All non-doc fixes need PMC approval. For details see Juno 4.2 SR1 Endgame Plan. Please make sure that all bugs are verified and marked as such in bugzilla. Currently we still have many unverified bugs.

Discussion

Sign-off details [David]

I propose a slight tightening of "sign off" procedure, so there is a concrete deadline. Typically, if "test day" is Thursday, we'd say project leads or delegates 1) must sign-off by 5 PM (Eastern) Thursday, or, 2) by 5 PM, comment in sign-off bugzilla entry when they plan to be done (Such as "by noon Friday"). If no sign-off and no word from component, then Friday morning a note sent to PMC List asking for PMC member to "sign off", or not. If no word from PMC, by 5 PM Friday, we would assume a "go", hope for the best, and sort out errors in process or delegation later. See discussions in bug 388377 for the background and motivation for this proposal.

Suggestions? Counter proposals?

Status

Platform Workspace (Szymon)

  • Juno SR1 status: no further fixes planned

Platform UI (Paul)

Platform Text (Dani)

  • Juno SR1 status: no further fixes planned
  • bug fixing

Platform Debug (Mike / Pawel)

  • Juno SR1 status: no further fixes planned
  • bug fixing / triage
  • Vacation

Platform Ant (Mike)

  • Juno SR1 status: no further fixes planned
  • Vacation

Platform SWT (Silenio)

  • Juno SR1 status: no further fixes planned
  • reviewing more GTK 3 patches
  • bug fixing

Platform Releng (David)

prep for Juno SR1, minor work with tests, test framework Orbit work

JDT Core (Srikanth)

* Progress on JSR335 - we are able to execute code with default methods now.
* Progress on JSR308 - Annotations on nested types, wildcards.
* Progress on hiring - Two candidates short listed - looks promising barring last
  minute hitches.
* Nothing more in plan for SR1.


JDT Debug (Mike)

  • Juno SR1 status: no further fixes planned
  • bug fixing / triage
  • Vacation

JDT UI & Text (Dani)

  • Juno SR1 status: no further fixes planned
  • replacing old Linux build machine install with a new one (continued)
  • bug fixing

PDE (Curtis)

  • No further fixes planned for SR1
  • bug fixing

Equinox (Tom Watson)

  • Last planned fix went into todays SR1 build bug 388299
  • No more planned fixes for SR1

Equinox p2 (Ian Bull)

August 29, 2012

Reminder

RC2 this week with a test pass on Thursday and official sign-off. All non-doc fixes after RC3 need PMC approval. For details see Juno 4.2 SR1 Endgame Plan.

Discussion

Status

Platform Workspace (Szymon)

  • one minor thing planned for SR1, but can be moved to SR2, see bug 381670
  • new critical bug against 4.2.1 to investigate, see bug 388289

Platform UI (Paul)

Platform Text (Dani)

  • Juno SR1 status: made last fix (bug 387095 ) for SR1
  • bug fixing

Platform Debug (Dani for Mike / Pawel Piech)

  • Juno SR1 status: no further fixes planned
  • vacation

Platform Ant (Dani for Mike)

  • Juno SR1 status: no further fixes planned
  • vacation

Platform SWT (Silenio)

Platform Releng (David)

investigating unit tests and (occasionally) improving test infrastructure.
Can now (again) run 32 bit linux tests (automatically)
no further bug fixes planned for Juno SR2
will be working/investigating/testing (locally) from feature branches until SR1 is done.
reminder once Juno SR1 is done (in approx. 3 weeks), only one week till Kepler M2 +0 deadline

JDT Core (Srikanth)

JDT Debug (Dani for Mike)

  • Juno SR1 status: no further fixes planned
  • vacation

JDT UI & Text (Dani)

  • Juno SR1 status: no further fixes planned
  • replacing old Linux build machine install with a new one (continued)
  • bug fixing

PDE (Dani for Curtis)

  • Juno SR1 status: no further fixes planned
  • vacation

Equinox (Tom Watson)

  • Juno SR1 - One more bug we are investigating bug 388299

Equinox p2 (Ian Bull)

August 22, 2012

Discussion

Status

Platform UI (Paul)

Platform Text (Dani)

  • Juno SR1 status: added a workaround for a Platform UI bug 375576 - one further fix under investigation (bug 387095 )
  • bug fixing

Platform Debug (Michael Rennie / Pawel Piech)

  • nothing more planned for SR1
  • vacations
  • bug fixing

Platform Ant (Michael Rennie)

  • nothing more planned for SR1
  • vacations

Platform SWT (Silenio)

Vacations
Investigating remaining issues for SR1
GTK 3 patches review

Platform Releng (davidw)

Promoted and Simrel work for Juno SR1 RC1
Improving Unit test diagnostic info (90% Markus, THANKS!): Properties button, Crash logs saved, Screen Captures?, log sizes displayed.
(still) looking at bug 387052 for SR1
Next week hope to start on 3rd party updates (ant 1.8.4, javax.annotations 1.1, IUC4J 49.1?) for M2

JDT Core (Srikanth)

JDT Debug (Michael Rennie)

  • nothing more planned for SR1
  • vacations
  • bug fixing

JDT UI & Text (Dani)

  • replacing old Linux build machine install with a new one
  • bug fixing

PDE (Curtis Windatt)

  • nothing more planned for SR1
  • Curtis is a new dad
  • vacations
  • bug fixing

Equinox (Tom Watson)

Equinox p2 (Ian Bull)

August 15, 2012

Discussion

  • (Markus) Auto-tagging from master is problematic in the milestone week
    • No possibility for a controlled rebuild
    • Not all committers seem to follow (or know about) the Eclipse process (lockdown rules; changes on or after test day need to be reviewed)
    • Team leads should decide and stick to one of two working modes:
      • Either work with integration and master branches, and only push integration in a controlled fashion
      • Or work with master only, but make sure all committers follow the rules and do not commit ongoing work to master in milestone weeks after the last scheduled test build and until the milestone is released
Decision: Team leads who keep working with master only remind their committers to not push unreviewed work in milestone weeks
  • (John) What to do in the Git repository when bundles or features are removed?
Decision: Keep a .project file and add a readme.

Status

Platform UI (Paul)

Returning from vacations
Juno SR1 status: still have open bugs for 4.2.1, will need to bump some to 4.2.2

Platform Text (Dani)

Juno SR1 status: no further bug fixes planned so far
bug fixing

Platform Debug (Michael Rennie / Pawel Piech)

Vacation

Platform Ant (Michael Rennie)

Vacation

Platform SWT (Silenio)

Vacation
Still investigating one issue for SR1 bug 321887
GTK 3 patches review
Bug fixing

Platform Releng (davidw)

Promoted Kepler M1 +0 (congratulations and thanks!)
Updated Build Schedule
Prep for Juno RC1
looking at bug 387052 for SR1
hope to simplify unit tests soon bug 377718

JDT Core (Srikanth)

Aug 15th & Aug 20th are public holidays in India.
Juno SR1 status: no further bug fixes planned.
bug fixing
Java 8 on lower gear past 15 days and next week due to M1 and SR1 related work and vacations.
We bid adieu to Ayushman Jain.


JDT Debug (Michael Rennie)

  • Vacation

JDT UI & Text (Dani)

Juno SR1 status: fixed 1 remaining issue - no further fixes planned
bug fixing

PDE (Curtis Windatt)

One PDE UI bug being backported to SR1 - bug 386866 Translations for tracing preference page
Bug fixing
Sick time

Equinox (Tom Watson)

Equinox p2 (Ian Bull)

August 08, 2012

Discussion

  • Should we collect status notes on a wiki?
http://wiki.eclipse.org/Eclipse/Status_meetings
answer was "yes" but suggested a "one page" format so always same page, easier to "watch" for changes.
  • What build schedule would we like for Juno SR1 RC1? Our +0 day is 8/17 (then a week off, then RCs three Friday's in a row).
Assume we'll be doing both 3.8.1 and 4.2.1 "in sync" (Though 4.2.1 is the official Juno SR1 deliverable).
How to fit these two "M-builds" in with other I-builds, N-builds?
M builds scheduled for Mondays? 10 AM, 12 AM Eastern, with rebuilds upon request?
Decision was to move to Wednesday's 10, 12:00; always, instead of Thursday, and then schedule nothing special for RC weeks. This aligns with the 4.2.1 SR 1 Endgame Plan.
  • Question was asked about how to "change history" to correct authorship of a commit. Basic answer was that changing history risks messing up others who have that branch checked out, so best to add a new commit explaining the connection to a bug, the other commit hash, and ?fixing? authorship if possible ... but, not sure any of us on call knew all details of how that all works with auto IP log (especially, if desire was to remove an author? ... though, not sure that's legitimate ... maybe if someone just added whitespace formatting or something?). [Its not clear, to me, how "authorship" could ever be "wrong"? Interesting case. ]

Status

Platform Text (Dani)

working towards Kepler M1 (one day test pass, bug fixing, N&N)
Juno SR1 status: no bug fixes planned so far

Platform Debug (Michael Rennie)

Holidays
Testing 4.3 M1
SR1 - fixed two bugs, no additional fixes planned

Platform Ant (Michael Rennie)

Holidays
Testing 4.3 M1
No fixes planned for SR1

Platform SWT (Silenio)

Holidays
SR1 fixes
Reviewing GTK 3 patches

Platform Releng (davidw)

Sim Release:
Completed migration of repository to git.
Converted and setup aggregation build for Kepler and Juno SR1

JDT Core (Srikanth)

Working towards Kepler M1 (one day test pass, bug fixing, N&N)
Nothing more planned for M1.
Juno SR1 status: Still a few bugs being worked on.

JDT Debug

Holidays
Testing 4.3 M1
SR1 - fixed two bugs, no additional fixes planned

JDT UI & Text (Dani)

working towards Kepler M1 (one day test pass, bug fixing, N&N)
Juno SR1 status: fixed 2 bugs - no further bug fixes planned to fix so far

PDE (Curtis Windatt)

testing for M1
no more fixes planned for SR1