Meeting

Attendees

Andrew R., Thanh H, Paul W., Jan S., Tobias O., John A., Steve F., Szymon P., Krzysztof D.

The PMC has decided resolving performance and behaviour issues are more important than the CBI build. Paul continuing to work on CBI would risk this. Thus Juno SR2 won't be released based on a CBI build.

In terms of the remaining work, of 20 possible bugs, only the following 7 are needed for Juno to release based on CBI:

384873, 385154, 385959 have an existing fix. If the fix tests cleanly, then these are resolved. Thanh H. to test
378234 and 377190 involve some comparisons of CBI vs. PDE build software. David Williams will be swapping for Paul. Thanh H. will assist.
381057 Krzysztof offered to take on, thanks!
393922 involves multiple sub-components. Paul's out of time to work on it (as per above PMC comment). David W. will be taking over.

We'll meet again next week to discuss the schedule.

John A. - What's there for CBI may be useful for approximately 95% of the community. Perhaps this is enough for LTS?
Andrew R. - Will talk with the LTS steering committee on Thursday to explore.
Thanh H. - prior to the Holidays Kepler was fully building with CBI based on patches that were submitted to bugzilla. He'll check the current status and update if needed.
John A. - The platform team can & should be more aggressive about switching to CBI and shutting down PDE for Kepler. History has shown it is the only way people will really dig in.
Jan S. - CBI is still using a development version of Tycho. Is this desirable?
Andrew R./ Paul W. - Not quite there to be able to lock down on a released version of Tycho. We still depend on features in the development version and periodically need new features/fixes.
Paul W. - In an ideal world, we could delay Juno SR2 2.5 months to give it a good soak when built on CBI.