TPTP-PMC-20080206

Logistics

Attending: AlexN, Chris, Paul, Joanna, Harm, Oliver, AlexA

Are there any items that need to be moved to front of agenda?

• AlexN needs to step out early.

Any comments on notes from last week?

• no

4.4.1 Discussion

Oliver asks if the only remaining items are 4.4.1 profiler

• Joanna comments that there are some build issues related to SSL solution that required versioning
  • Waiting for code review
  • Will kick off build in next few minutes

4.4.1 Profiler

Oliver asks if we need to resolve any 4.4.1 profiler issues here.

• Regarding IBM VM issues, AlexA has proposed a solution for 4.4.1
  • Still trying to get sufficient feedback in 4.4.1
  • Agrees that detecting IBM JVM and special casing it may be ideal
• One other defect awaiting defect closure from IBM

Oliver asks if profiler stuff be resolved next week?

• The intent is to resolve them next week
• Oliver will be following up with Eugene and asks who all should be involved in discussion?
  • First Eugene and Alex
  • If have problems closing add Harm, Chris, Paul
• Harm noted that after resolution happens, some testing will be needed.

Reminder to everyone who may have forgotton... TPTP 4.4.1 is bundled with Eclipse spring release of Europa (3.3.2)

• Absolute Final due date Feb 29
• Does anyone see a flaw that will cause a delay past the 29th
  • Not at this time

4.4.1 Test Pass

Harm asks about 4.4.1 test pass report. It is far from complete right now.

• Is it because test lists are bad or that people have not checked in results?
  • Joanna suggests that it is a bit of both.
  • Trace -- Alex asks about manual generation

• • Paul will try to do 2x a day until automation works again
• Paul notes BIRT autogeneration issue makes him manually generate (at 58% this morning)
  • Will regenerate this PM and expects better results
  • AlexA asked for a bit of clarification on the intent to get BIRT autogenerating again.
    • Paul states that right now it is manual because of a bug. Supposed to be automatic and twice a day.

Harm wants folks to be aware that if we are only at 58% completion that after the build comes thru on Monday that there is still about half of the test pass remaining!

• Oliver notes that we need to do a little more due diligence
• Joanna notes that some testing is blocked by a few key defects
• Harm asks if folks will finish test pass by Friday...
  • Platform pretty much agreed (with exception of items directly in path with known defects)
  • Monitoring will need to rerun a few scenarios but pretty much will be done Friday
  • Test notes that have 3-4 test suites blocked on AC but will finish remainder by Friday
  • Trace does not see problems finishing by Friday

4.5 Discussion

Don't currently have a candidate driver

• Failure to build for both 441 and 45
• Should have a driver midafternoon (Joanna was about to hit the switch to build)
• Reminder: Ganymede delivery of driver is on 15th. If we get candiate today, we are already a few days behind.
  • Paul notes that we need the new build driver.
    • Not sure if starting from the build of the 4th would be appropriate since a bunch of fixes have gone in
• Is there anything we could have done to avoid this build issue to avoid it in the future?
  • no comments.

Supposed to be totally finished with I5 on 13th

• Should be starting test pass at this point
• We are NOT going to be on the memory stick for EclipseCon on the 15th
• For Ganymede release, could potentially be in EPP for WTP kit
  • Do we want this as a checkpoint of how far we've come with POG
    • Harm, Oliver, AlexA think good
• Harm notes that we would need to make a commitment (i.e., follow thru) instead of just a goal

4.5 Weekly Schedule Discussion

Monitoring

• Has requested to move remaining I5 defects to I6. Has posted request for approval to PMC.
• Should he record this in schedule?
  • Oliver responds not yet
  • P1s that could not be contained should appear in table as "slipped" (e.g., even the ones moving to I6)

Test

• Slipped a bunch forward.
• Paul asks when does I6 schedule plans to incorporate them into I6 weekly schedule plan?
  • Oliver notes that when we start I6, schedules should have I6 details.

Platform

• Good detail

Trace

• It just says testing at this point.

Oliver notes that there is a bunch moved to I6. Does this mean that we will miss some stuff for GA or do we have sufficient flexibility in the schedule to accomodate?

• It is better to know sooner rather than later.

Platform

• Most worried about enhancements
• Looks managable in all but a few cases
  • Oliver asks if we want to look at master list and discuss key items that are at risk of missing GA
    • Joanna suggests talking next week.
    • Paul suggests we discuss when finish testing for I5 and start development for I6
    • Oliver agrees with this approach

Harm notes that we are not using standard keywords in weekly schedule making searching harder.

• e.g., finish, finished, done
• e.g., slipped items that don't appear with their final date
• e.g., pretty blank starting from a few weeks from now

When we start I6, oliver would like everything that we want to get done between then and GA to be annotated in the table.

• Does anyone have comments about this?

Test

• No major problems
• Wonders if it will cause extra work later in project
• Harm suggests that at least P1s targeted should be placed
• Paul questions if doing stuff past I6 (i.e., all the way to release) will be too much work

Monitoring

• I7/I8 are listed as being only for critical/blocking defects
• Can lead place noncritical items to those iterations
  • Notes that have normal defects that can't be contained by I6 and they would love to schedule them into I7
  • Harm suggests that if nothing else is going on in I7, can work to some extent on clearing backlog of bugs

Harm suggests using schedule to note P1 and critical (i.e., not necessarily ALL items)

• If something that we are currently committed to is going to slip off of the release need to know soonest so that we can go to community and state what is being lost.

Chris raised the question of whether it may be possible to use bugzilla fields and change history and then automatically generate weekly schedules

• General agreement that this is something to think about
• Unsure of exact technology that could accomplish it (can't create weekly targets)
  • Oliver may be willing to think of tracking at iteration level if iterations are stay pretty short (i.e., close to 3-4 weeks development)

MISC

A few words on BVT

Paul gave a BVT update(build validate test)

• System up and running some core junits
• Joel to take bits that work and hook it into the end of build process

Oliver asks if they will be plugged in by next week?

• Paul is optimistic
• Joanna thinks end of week sounds doable

Oliver would like to be able to change "close to working" to "working" in his report to get more bang for the buck.

• Paul suggests oliver wait till friday to hopefully get BVT done

A few words on Oliver Report

Harm does have statistics about community finding defects and concept that community is actively filing bugs is good to mention.

• Harm was surprised to see 50% of bug reports came from outside of core team. and pleased.

Joanna notes that bug fix trend is downward

• Oliver notes that in addition to this, resourcing levels are down as well
• Question whether it may be possible to normalize the data relative to overall resources
  • Want to show (if possible) that productivity level is up even though total count is down
  • Statement about correlation of that fix bug rate to resourcing

Joel and harm tried to get download details.

• Mirrors don't keep metrics for mirror downloads except for core platform
• A weakness to remind board about again.

A meeting at EclipseCon

Paul reminded everyone that we wanted to discuss meeting at EclipseCon.

• Discussion centered around meeting on Sunday Afternoon
• AI Oliver to will have his folks work on scheduling

POG Update

Eugene and AlexA have agreed on most things Harm has not tested latest because of build problems.

Do we expect Harms use case to work from end to end in I5

• Most issues resolved but a few key issues still exist
• Hope Harm will be satisfied with our progress at end of I5
  • Harm notes that we have about a dozen POG issues
    • We always knew we were going to get only some done in this iteration
• Alex asked Harm to look at bugzilla and respond to igor question http://bugs.eclipse.org/214098
  • Harm took this AI

Will we be in WTP packaging?

• If we commit, need to be in their weekly builds and have to manage/track to it.
• Oliver says this is a mechanical detail
• Oliver wonders if our profiler is good enough to suggest that it be there.
  • Is it quality enough that we believe it should be there?
  • Harm notes that we have already been invited. if stay on POG it will be good.
    • If want to do this, need formal test cases not just manual user Harm.
  • We need to contact David Williams to get more details on logistics
  • Do we want test tools in addition to profiler?
    • Paul suggests he sees no reason why not.
• AI Oliver to find out more about the overhead involved
  • Harm comments:
  • EPP just has a feature file that pulls
  • We just would have to commit to testing the feature we want to stress

AI: Chris to get new telecon numbers for meeting