TPTP-PMC-20071031

Attending

Chris, Alan Haggarty, Joanna, Harm, Paul, Oliver, Alex Nan

Follow up from last week

Responding to question about last weeks meeting, Harm noted:

• Harm posted some questions to mailing list
  • Schedule from last weeks minutes
  • Doing plugin refactoring as part of M3
    • Yet test pass was supposed to start that date
    • In future should not create this confusion from minutes
  • Harm posted question to list but didn't get a response on list
• Paul and Oliver did have a discussion but didn't follow up on list
  • For future reference will try to be better following up on list
• Topic to be addressed in meeting section related to schedule

4.5 line coverage

Paul talked about rolling out line coverage use by TPTP projects. Because just smoke tested in 4.4.0.3, and minimally tests, didn't roll out.

• Want to roll out for real in 4.5.
• I1(M3) test pass for Test
• Other projects in I2(M4) (as part of test update)
  • Instructions for running are on wiki in MISC section
    • How_to_test_TPTP_with_line_level_code_coverage_enabled
• Alex got a bit of info off the web site but it didn't seem sufficient.
• Paul: There is some custom config but it is documented on wiki
  • For now does not cover RCP apps... Just the SDK

Harm question: If test project uses it does this mean we will be able to have a meaningful report

• Paul: Yes, we will have real coverage stats for TEST after I1(M3)
  • Test project is committed and will make it happen.
    • This does mean that we need RCP and other related things.

4.5 feature removal

Alex started discussion about refactoring/removing features from GA (related to plugins above)

• Scapa components are still there. Do we want to remove them?
  • Scapa stuff includes JMX agents: jboss/JoNAS; perfmon and stats viewer
• Oliver question: Should oliver ask for half a head from Scapa?
  • Its not tested, and needs some rework, small slice of engineer is unlikely to help
    • We should take it out as we discussed at end of last release.
    • Removing these particular components is fairly easy. Removing from GA stream and not putting into separate packaging.
  • Alex noted need to inform the community that this feature will no longer be there.
    • Harm -- documented in plan and visible in milestone is the informing the community

• Question -- Will we check in all these changes to these and other modules to M3 or wait for M4
• General rule -- do it as soon as possible.
  • The Scapa one is proabably possible in M3 w/o killing test pass (because is simple)
  • All projects should definately done with all cleanup by M4
• As early in M4 as possible

Question from Alex: Is one particular enhancement request approved (I didn't catch the enhancement number)? If not, needs to be removed from CVS

• Harm suggests that it should not be in CVS if not approved yet.
• Discussion moved offline to AG call

4.5 Bugzilla cleanup

Cleaning up bugs/plans keywords in CVS

• What does plan word mean for defects?
  • Joanna: Triaged for 4.5 by leads -- and 4.4.1 bugs that were not made
• In general there are inconsistencies (such as priority settings)
  • In past bumped priorities for planned items to P1. Joanna did not do this this time

Harm gave a tutorial on historical [preferred] use of terminology in bugzilla.

• Correct use of "Plan" keyword
  • Plan keyword used only for enhancements rather than defects
  • Used for driving discussion not to indicate commitment
• Correct use of Priority
  • P1 (definately there and staffed)
  • P2 (possibly there/stretch)
  • P3 (probably not to be there
• There are enhancements tagged as plan keyword with priorities that are low but are discussed as approved.
• Should be able to look at a enhancements/plan keyword and defects with priorities to see what is resourced/committed
• All items (defect/enhancements) tagged P1 should indicate fully committed and resourced
• Scrub in past has marked everything as P3/future then start reprioritizing and targeting specific items for upcoming release and setting priority.
• AI: Action item for leads [due next week]
  • Ensure that defects/enhancements targeted for 4.5 are appropriately set and defects for future are appropriate "unset".
  • Deal with misassigned defects (items assigned to folk no longer involved in project)

4.5 Test Pass I1(M3)/beyond

We are at 1.15% completion on test pass (row 4)[1]

• This is bad... :-/
• Some folks have not checked in some results yet (e.g., Monitoring)
  • We need to educate the team to check into results closer to when the tests are run.
• Did leads check in what they intended to test into test plan yet?
  • Leads need to make sure reports reflect intended tests
• Folks need to start testing
  • AI: Leads need to review test suites today (and state what will test now and for future test passes)

Everyone commits to completing test pass by:

• Monitoring on track for Friday.
• Other projects okay with next Wednesday slipped date (see below for platform discussion).
• We will have 1 week at end of test pass to import other project's changes
  • What smoke tests/full tests at that time? near full test pass.
  • Oliver noted unchanged pieces of code probably don't need much right now
    • NOTE: Leads may be checking in suites (test) for each test pass (removing them up front from early iterations)

Joanna knows that some platform test items won't be done (e.g., Kiryl is out this week) this week and will lag till next week.

• Do we need to slip schedule?
  • Harm asked as usual: Did we enter the test pass expecting that there would be no new defects?
  • The drop schedule to Ganymede is not going to move

Joanna noted IBM JVM SR6 drop was delayed to about Oct 31.

• If it drops in time, will be tested otherwise won't.

We did not enter test pass on time because of build issues.

• Positive comment: No new TI defects (critical blocking ones) from Joel's testing
• We are set up to drop 4.5 codebase for M3
  • What is date that we will have 4.5 build without OpenSSL?
  • Assuming sampson finishes code review, when next successful build happens (tommorrow)
    • No good build today (Hope to get a response from Kiryl regarding Intel side of build)

4.5 Misc

AI: All leads need to put in their weekly milestones/dependences into week to week wiki page at TPTP_45_Weekly_Schedule

• See Paul's initial draft for Test project for example

Can we rename our milestones to be the same as the milestones that we are delivering to Ganymede?

Discussion about end of year holiday gap in schedule

• Has Ganymede taken this into account?
• Note that some countries take gap at different time
• item to bring up with planning council

Profiler use case at EclipseWorld

• Have not seem update from Alex (Intel) regarding ideal users/instructions.
  • Goal, high level overview for Oliver to present
    • Wiki to point potential users to
    • Commit to respond to users
• Action item
  • wiki created for folks to go to to review whether their use matches desired use profile
    • NOTE: Fill in under TPTP_User_Experiences
• Due by next week!!!!
  • Alex be prepared to give brief update in PMC call about status