Eclipse/PMC

From Eclipsepedia

Jump to: navigation, search

Contents

Documents

Some documents written and/or used by the PMC:

Meeting Schedule

The Eclipse Project PMC has a weekly phone meeting every wednesday at 10.30am EST.

Meeting Minutes

May 1, 2013 - John, Dani, Andrew, McQ

  • John: LTS now running IBM Juno maintenance builds in LTS forge. Some small details still being worked out.
  • Dani: Gerrit can be configured to deny non-fast-forward merges. We agreed to request that all Eclipse project Gerrit repositories be configured in this way. See bug 402791 for details.
  • Dani: Request for JUnit bundle to stop re-exporting Hamcrest bundle. Hamcrest was originally part of JUnit so this was done for compatibility reasons. We agreed that we will not do this because it breaks all users of JUnit bundle. See bug 403676 for details.

April 24, 2013 - John, Dani, Martin

  • Dani: Upload Freeze Plan
  • Dani: Java 8 Update
  • John: Pawel stepping down as Platform/Debug Lead
    • Leaves Pin & Clone orphaned for next year. Working sets for expression view also not in Kepler.

April 17, 2013 - John, Dani, Martin

  • Martin: New Classic Package - Will commit SDK branding, and Egit Capability def
    • CSS - Should inherit proper styling from Platform (most probably OK)
  • John: API Baselines - remind teams to set API baselines
  • Martin: bug 402509 context menu filtering - request gentle nudge
  • John: Triage of Platform UI issues to be addressed



April 3, 2013 - McQ, Dani, John, Andrew

  • John: Kepler draft plan update ready, please take a look and provide comments
  • We discussed EclipseCon. It was a very good event for Eclipse project. On desktop community is moving forward on Eclipse 4. There was lots of good feedback, and very little grumbling. Orion was very well represented, lots of talks and demos from several companies adopting it.

March 20, 2013 - McQ, Dani, Martin, Andrew

  • Dani: Java 8 - jdt.core making very good progress - almost everything compiles in the editor
  • Dani: pom files update - to be discussed in the arch call
  • Martin: bug 402509 Context Menus - no feedback, ping again
  • Martin: bug 145635 Pin & Clone
    • Dani - the feature itself is good and makes sense, would also like it in JDT
    • The UI metaphor is new and would need some testing in the Field - McQ would like the patch in ASAP in a way that doesn't cause more swirl to start experimenting
    • CDT would need to adopt it, upstream changes needed
    • Current implementation showed quite a lot of ripples and bugs
      • Current impl breaks existing workflows, that's the big problem
    • John: only 1 milestone left, if it appears in M7 and breaks any adopters they have little time to react;
      • Can loop-in JDT and CDT, but there might be others ...
      • Was PTP involved in any way ? They could also be broken, they have very special requirements
    • McQ: Do we have ownership of this work ? - If we have a clear need and a clear owner working on it we could give it a try
    • AI Martin (1) Clarify with CDT why they didn't like the feature and what are their plans; all the rest comes after that eg PTP, Pawel, can it be done in a non-breaking way
  • John: bug 402912 Significant Memory Leak in M6 fixed this week: noticable if eclipse running for a long time
    • Workaround for M6: Turn off Workbench Autosave feature by setting it to 0
    • AI Post to cross-project - people doing demos at Eclipsecon might want to use the workaround
  • Martin: bug 397896 New Classic Package - initial package done but unusable while the SDK branding bundle's contents is not migrated
    • Martin will continue pushing for this to happen - fallback would be not
    • Not much risk in judging whether it's good or not, only binary diff needed
  • John: bug 380325 Java 7 filesystem fragment
    • McQ do we know performance implications, could we get rid of the native bundles ?
    • Martin would like to have the fragment in all builds, to make it easier to test performance etc implications
  • Martin: bug 403589 eclipse.inf
    • Different binary bits depending on distribution channel is bad ... Tycho removed
    • Defect in post-processing (why does it add eclipse.inf) or defect in bundle preparation (why isn't eclipse.inf in there) ?
    • Maybe related to pack200 conditioning - Martin thinks that eclipse.inf was used in the past as a marker to know whether pack200 conditioning has happened ... now downstream aggregator conditions and adds eclipse.inf
    • AI John investigate

March 13, 2013 - McQ, John, Martin, Dani

  • Andrew: bug 397896 final throughts on the EPP package
    • Move forward with [1]
    • Testing - could find someone if it's just "use during the test pass"
    • Q: Can the EPP build be done frequently so it's available 1 week before the milestone for testing ?
  • Martin: bug 402509 context menu filtering
    • Would like to get some advice from Platform on the approach; some trajectory of a solution moving forward
    • Not a must-have for Kepler, but would like to see where this is heading
    • AI McQ find Platform folk to discuss this
  • John: Eclipse 4 API Status
    • Publishing a "usable subset" of the former provisional API as final: model, core services; consume injection
    • Allows building a real Eclipse 4 App now on API
      • Mostly clean-up, documentation; removing some bad stuff
      • Writing an App on Kepler E4 API might actually be able to run on Juno
    • Some of the advanced stuff remains provisional for now, eg define injection
  • John: Plan Update - No permission to update in the PMI
  • McQ: CBI / Bad Launcher
    • Have an idea that doing some little extra work after the CBI build might be good enough for now
    • HP-UX CBI Build doesn't start out of the box (will likely stay a known issue for M6)
  • Martin: bug 380325 Java7 EFS Plugin: Build ?

March 6, 2013 - McQ, John, Martin, Andrew

  • Martin: bug 402509 context menu filtering
    • AI Martin ask engineers document what we're doing today
  • Martin: bug 397896 EPP vs Classic - how to proceed
    • Martin's recent findings on Marketplace
      • There's p2, Marketplace, Packages in the Open but they are only parts of the solution
      • There's commercial installers from EclipseSource (Yoxos), Genuitec, IBM
      • John: p2 simplified UI (eg Webtools Server Adaptors; Mylyn)
    • What's missing ? David too booked ?
  • John: CBI Builds in Kepler - SDK in good shape but repo and Equinox downloadpage not quite right yet, some might take until M6
  • John: Delete old Platform/UI Branches - any PMC Policy ? No, convert branches to tags
  • John: API Freeze next week - some (modest) new Eclipse 4 API's
    • McQ recommends going for a "complete, but minimal subset of what people will want to do", rather than too many convenience methods
    • John: Some API has been provisional for 3 years - can't make wholesale changes now for the fun of it
      • May phase in some changes by adding @noreference rather than deleting stuff
      • In some cases there's pressure to use interfaces rather than classes, in order to be able and leverage Java 8 constructs
  • Orion - anybody going to JSConf ?

February 27, 2013 - John, Andrew, Dani

  • Nothing noteworthy

February 20, 2013 - McQ, John, Dani, Martin, Andrew

  • John: Orion Release Review, please give feedback

February 13, 2013


February 6, 2013 - John, Dani, Martin

  • John: M5a and SR2 RC4
  • John: Pawel and Platform / Debug work (Pin & Clone)
    • Only adopt when it's really ready and satisfies all criteria, otherwise not into Kepler
    • Maybe pull out Debug plugins into a feature of their own ?

January 30, 2013 - McQ, John, Dani, Martin, Andrew

  • John: SR2 RC3 looking good - want to know about any remaining issues
  • Dani: Java 8 Update - Markus Keller will move from JDT UI to JDT Core, taking over JSR-308
  • John: NODE Dependency in Orion (exempt pre-req)
    • Not shipping NODE ourselves, pushing to NODE package manager instead (or clone from git)
    • Andrew: Is it a workswith ?
    • Martin: Is there any overlap with the Java portion ?
      • The server code is completely distinct from Java, but the same client can interact with both

January 24, 2013 -

  • Martin: 4.3 and 3.8 dual-stream - Webtools project plan
    • egit, mylyn, dtp have been doing both n + n-1 for a long time
    • Technically, running on both should be possible if Platform API is used (except for the changes in Theming)
    • The Eclipse Community needs to advance the latest (4.3)
    • IBM product teams have been told not to forward at all if they can't forward to 4.x - shipping on 4.2
    • John: Having wider ranges of tolerance is something we should be able to do (otherwise our plugin system is failing)
    • John: Java8 will be the first killer feature in 4.x
  • Dani: Java 8
    • Will provide feature patches as soon as it makes sense - early draft, participating in expert group
    • Want to have the compiler and editor ready when Java 8 goes GA - initial refactoring etc in SR1 - full feature set in the next release
  • John: CBI update
    • Consider switch after M5

January 17, 2013 - McQ, John, Martin

  • John: bug 397896 EPP vs Classic
    • Could volunteer to own the EPP package from Platform POV
    • Then Wayne would own the decision of RCP/RAP vs Classic - what's the difference, keep both or unify
  • John: Test Failures on I-Builds
    • Move away tests that have been failing for 9 months
    • Dani: Most tests are actually down to 0 failures
  • John: CBI / Maven Builds
    • Appears to be fragile at the moment - Maven .micro releases are breaking the build!

January 10, 2013 - McQ, John, Dani, Martin

  • McQ: CBI - Switching Platform build to CBI for SR2 is not doable
    • Want to make CBI work for Juno since it's the LTS stream
    • CBI SDK builds are good at this point, it's only a couple of minor issues (shape of repos, download page, ...
    • When to switch to CBI in the Kepler Stream ? - Proposal do do it immediately
  • CBI patches waiting for Platform to pick up

January 3, 2013 - McQ, Andrew, John, Dani

  • Andrew will recommend good conferences to promote Orion

Archive