TPTP-PMC-20090429

Logistics

Attending: Paul, Oliver, Kathy, Chris, Eugene

• regrets from Ernest

Any issues with last week's summary?

• no issues

4.5.2.1

• Ready for a light test pass
• Test will be mostly contained at IBM
• Kathy will probably ask for Intel to do a bit of agent and tracing test

We had some discussion around whether there were planned to be more milesones for 4.5.2.1 or not

• Plan to have a few bugs backported per quarter.

Josh has updated the build reporting. now comparing 4.5.2 to 4.6 and 4.5.2.1 to 4.6.

• Build download page for 4.5.2.1 now has more options to point to milestone driver or latest build.

4.6 discussion

PRC folks are finishing up attach/reattach scenarios. IBM originally hoped this could be backported to the current 4.5.2.1 milestone but

• Team Still working on finalizing the patch but won't have it to look at finally till about Thu
  • It sounds like a pretty significant change
  • Plan to promote to 4.6 and then after some testing there possibly backport to next milestone of 4.5.2.1

We discussed builds a bit. We had issues last week.

• Builds are okay now. worked thru compiler update at IBM and a disk system crash at Intel
• After this test pass will test out new build structures
• After this test pass we will be able to let us tell PRC team to avoid x86/em64t builds

Test pass is currently at 83% attempted, ~5% fail

• A few new bugs getting filed based on failures
• Paul noted again BVTs are included in report ~400... being run on every build
  • There is reporting glitch; if you click on the subreport, it will not appear (appears to be BIRT issue)
  • It is showing a few failures; seems to be issues with running the test not a real number (thus ~5% fail rate is inflated)
  • There was assorted discussion about the long term value of test automation.
    • Paul suggests to have for test pass 1 of 4.6.1 fall refresh to have all of the automatable tests integrated/going

There should now be a 4.6.1 target. It is time to start thinking of retargeting things that will not be contained on 4.6

• This is good time to retarget P2s/P3s that can't be contained to help focus developers.
  • Paul notes that he has some defects to defer; most P2/P3;
• There are a few P1s that will be delayed
  • some from paul
  • Some from Eugene who is going on vacation and probably won't be able to do too many P1s

Eugene has 5 P1s; 2 probably will need deferral unless we are lucky.

• Remaining 3 are release related rather than code related and will clean up later but before release.
• Eugene currently working on 115325 (requested by adopter); no fix yet.
• Segfault on IBM JVM; JVM team asked for new stack trace (previous one had issues) from Eugene who is working on it.

• Need update from Beijing team about Documentation update bugzilla.

If we think that any of the P1s we want to defer will be noticed by consumers, should bring them up as early as possible.

• Leads to create list of P1s at risk for slipping
• Discuss next week at PMC
• and then have Kathy or Oliver draft a blanket message about the list and send to interested parties.

Misc

We discussed our Toronto/Beijing communication.

• Weekly email exchanges are pretty good.
• After 4.6 release, want to get some concentration on ramping some Toronto folks on some of the agent areas
• Coarse grained communication is okay but if rapid turnaround is required, it is more difficult.
  • There was a question about how we are measuring thruput of the team
  • Kathy is monitoring bugzilla thruputs and bugfixes are coming thru