TPTP-PMC-20080116

Logistics

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

4.4.1 Release Update

According to the schedule testing will be starting on January 28

Items still unresolved

• JSSE implementation:
  • Jon and Igor are reviewing the patch.
• Joanna's deferral defect

4.4.Release and IBM JVM

Eugene sent note with table of problems with IBM Java 1.5 and JVMTI:

• This relates to this defect
  • Problem described recommendation 1.5 SR6 (which does not solve everything)
• Eugene has also opened a few other related bugzillas
• In bug Alexander/Asaf have put in some defect notes
• IBM considers this quite important for 4.4.1
• Readme is not sufficient.

Why is this not resolved? We have had assorted mail threads about possible scenarios.

• Have not been treating as a stop ship defect
• Have not addressed class redefinition

What do we want here for 4.4.1.

• Good enough would be to provide all that we can w/ limitations
• Eugene will make the call from IBM side and Alex will interface from Intel side
  • With help from Asaf
• Do not know yet exact all the details
  • Come back next week with details
  • Eugene sizing for 4 defects he created

= 4.4.1 Test Pass

Any concerns with overlapped test passes

• > 200 fixes have gone in

• Joanna would rather move 4.5 test pass up than 4.4.1 test pass
  • Doesn't leave much development time for 4.5
  • but would recapture some of that time I6
• Paul -- no issues. Have new intern to help
• AlexN -- since didn't have many fixes don't believe need full test pass
  • AC and TI have made major changes
• Paul made alternate suggestion
  • 4.5 test pass 1 move up taking 1 wk and eliminates test pass 2
  • 4.4.1 in 1 wk

Missing report on the web site for upcoming test pass.

• Paul looking at it

Closure on 4.4.1 and 4.5 test pass

4.5 schedule dates

• TP1 for I5 starting on Monday

Oliver asks who consumes and tests with 4.5 I5.

• If not really consumed then why push testing heavily
• Harm says the Ganymede drop is downloaded by 1000s
• Joanna would be satisfied with some smoke tests for this iteration
  • Probably would cut time in ~half.
    • Paul ok
    • AlexN would test 4.5 more than 4.4.1 because that is where more work happens for his project
    • AlexA will need support from Eugene on Trace tests (the ones historically done by IBM that weren't done in I4)

Harm's suggestion (ratified by group)

• Skip test pass 1 for I5
• Leave 4.4.1 test pass alone as it exists in schedule
• When 4.4.1 test pass finishes do some reduced testing on 4.5
• Test pass 2 for I5 becomes full test pass

AI: Everyone should update their test plans to reflect this.

4.5 Discussion

Oliver asks if all weekly schedules up to date?

• No comments

Oliver asks what else needs discussion for 4.5

Regarding OpenSSL, Joanna says we have the ability to file an IPZilla to SHIP it with 4.5

• It turns out that the effort to get this to go thru is no worse than getting an external dependence approved
• There was general agreement that this sounds like a good idea
• The only question related to how this would impact our packaging size
  • I believe Joanna was going to have someone estimate this.

MISC

AlexN wanted to discuss some patch requests that came from a consumer of 4.4.0.3 that does not want to move to 4.4.1

• Request came for some one-off (backported) patches to 4.4.0.3
  • How/whether TPTP project would deliver these this type of patches
  • A patch would need to be in response to a specific bugzilla
  • Consumer is actually requesting an alternate build of a component.
    • Supporting such an alternate build is not scoped/planned by TPTP project
    • If a consumer wants variant not supported by project they are free to build themselves.

Paul ready to pull trigger on committer removal

• He has Chris' latest updates from PMC mailing list (accumulating requests from several Intel folks for removal)
• Harm asked if Chris' list was consistent with what is coming from internal process
• AI to everyone to go check the list.

There have been build issues with 4.4.1

• 4.5 is building
• NOTE, (as of 11a on 1/18 -- offline discussion w/ Kiryl indicates that some fixes have been checked in so next build from IBM should proceed.