TPTP-PMC-20090603

Logistics

Attending: Oliver, Kathy, Paul, Chris, Eugene

Any issues with previous summaries?

• no

Anyone need to leave early?

• No

4.5.2.1 milestone discussion

none

4.6 rampdown

Kathy has several topics

• release review foils (full and mini)
  • AI to Chris to provide any addl feedback he sees fit but Kathy is not waiting because they are due

Test pass 1 currently showing 93% success. A new report needs to be generated

• For a few of the missing test results, they are done and just need result checkin (technical difficulties)
• Platform project is ready for test pass 2
• Paul notes that Test project needs a minor patch to be submitted.
  • Patch is quite minor; Paul has reviewed and Eugene is looking at it too
  • Expect PMC request quite soon. 278832 is the bug
  • Paul will require a new candidate driver are this is checked in

Profile on server is noticing a bit of strange behavior (a regression) that might require some code fixes. Intel team is looking at it.

• Kathy is hoping this will be quick if it turns out to be a problem
• Since it is still being looked at we might not wait for it to build the next candidate driver

Kathy also noted that review noticed a few tweaks about plugin versioning that are needed

• Will be asking for PMC approval

Ther eare also a few minor documentation items.

Oliver notes that our final drop is on June 16th and asks about confidence levels

• Test success now quite good (~3% failure)
• Paul confident on test project assuming that test pass 2 is successful.
• Discussion about just extending test pass 2 and skipping test pass 3.
  • General agreement to merge test pass 2 and test pass 3

Oliver asks if we might want to standardize the two test pass model

• General thought that it might be good. We have noticed in the past that having back to back test passes exaggerating test overhead with very little code change. Single longer test period seems a good idea

Kathy notes that project log still needs some updates to describe contributions from Intel and IBM non-committers even though they are from member companies.

• Kathy confirms that for 4.6 notes contributions of non-committers are reflected correctly

The question was raised about the value of using automatic IP logging tool.

• Team agreed to get eugene's comments

IBM discussed latest Agent controller port patch. Because of size and timing, IBM was unsure if they were going to try to promote prior to 4.6

A new issue with anomolous crash under certain profiling situations has arised.

• Team is hoping new JVM from IBM resolves it.
• It might not actually be a regression because it is a scenario never used before.

So to summarize, the worst case risk is that there are 2-3 more patches to promote before final release

• In addition, there are 11 release engineering issues/steps to be done for the release

Pauls list

We discussing pauls issue list related to 4.6 release and next steps

1) Update the TPTP Development Plan (move TPTP 4.6 from Release Under Development to Current Releases/Previous Release Queries

• • web page sections and add TPTP 4.6.1/4.5.2.1 to Current Releases section):
  • related link
  • This step is not done yet. It must wait for the actual relase.

2) Update the Project Plan Summary:

• summary
• This is not done yet; Kathy will look at by next week

3) Update the Project Log:

• project log
• This is done

4) Create and submit the TPTP 4.6 Release Review slides:

• release review slides
• Here is the TPTP 4.5 release collateral:
• Last years material is at
  • old review material
• This is done

5) Defer all TPTP 4.6 defects to TPTP 4.6.1:

• 4.6 bug report
• We have deferred all that we _know_ must be deferred but there are a few hanging issues being worked that we are hoping to _not_ defer.

6) Announce the TPTP 4.6 release on the web site (news), mailing lists, and news group.

• We cannot do this until actual release and will do it with #1

7) Please for TPTP 4.6.1 (plan and schedule) and TPTP 4.5.2.1 M3+.

• Will do tihs after the actual release. Too much going on right now

4.6.1

Can we branch yet?

• We note that we generally wait till after we make the final drop of the last release
• We need to ask release engineering guys about setting up a 4.6.1 infrastructure.
  • Also will ask a question about branching overhead if we want to do it next week instead of waiting for final relase
• Historically we are busy that last week as we watch out for dependences changing underneath us

Misc

We had a discussion about future release naming/numbering... 4.6.1? 4.7?, etc

• 4.7 might confuse folks thinking it was an enhancement release. Its really about maintenance at this time
• Paul believes we should do 4.6.1, 4.6.2, 4.6.3, 4.6.4 until we hit a point where we must move to 4.7 because of a dependence.
  • General agreement on this point.