[eclipse-dev] Planning Meeting Notes - Dec 21, 2005

[Mike Wilson <Mike_Wilson@xxxxxxxxxx>]

[ It must be getting close to Christmas, I didn't
get much input this week. Note: There won't be a status message next week.
]

-------- 
 Status
--------

JDT UI:

- bug fixing:
  - 120579 [quick fix] NullPointerException on
Ctrl+1 quickfix suggestion
  - 120528 [clean up] inside editor behaves unexpected
  - 120585 [clean up] needs two invocations until
all clean ups are
    performed
  - 113333 [api][content assist]API for setMessage
in ContentAssistant
    without cycling mode enabled
  - 120539 [build path] Build path context menu
should contain
    'Add source folder'
  - 99675 [build path] New source folders should
always be inserted on to
  - 100818 [quick fix] Wrong label for nested
enum
    LinkedCorrectionProposal.LinkedModeProposal
  - 102142 [quick fix] Change method signature
quick fix renames parameter
  - 121546 Cannot validate Javadoc URL starting
with https
  - 121222 Sample code in code formatter settings
in wrong
- bug verification, 3.2 M4 testing
- build path: refactoring of build path code (simplify,
code sharing)
- refactoring:
  - refactoring history is now displayed in project
property page and
    can be exported as script
  - added Refactor->Apply Refactorings...
action which executes
    refactorings from a script
  - minor improvements in the JAR Import Wizard
(filtering of internal
    changes to the stubbed types)
- model support:
  - implemented rendering of incoming refactorings

Rel. Eng.:
- 3.2M4 performance testing
- 3.2M4 testing in org.eclipse.releng.basebuilder
- investigating impact of new version scheme on NL
fragment generation
- vacation

Platform/JDT Text:
- shipped 3.2 M4
- started to investigate open text/java editor performance
degradation, 
  - see http://download.eclipse.org/downloads/drops/S-3.2M3-200511021600/performance/org.eclipse.jdt.text.php?
  - findings so far:
    - after running the tests on three different
machines (2 WindowsXP
      and 1 Linux) we are not yet able
to reproduce the huge differences as
      reported, e.g. by 'Activate 30
Java editors' (could be caused
      by next item)
    - forced GC when idle (IDEIdleHelper)
can interfere here (we saw it
      happen on our machines) which
increases measured CPU and elapsed time
      ==> we need a way to disable
this during performance testing in order
          to compare the
results with 3.1 (bug 121562)
    - new EFS slows down a bit
      ==> we need to see whether
this can be improved (bug 121571) since
          we currently look
bad/red
- 3.2 M5 planning
- bug fixing
- bug inbox tracking
- vacation