Notice: This Wiki is now read only and edits are no longer possible. Please see: https://gitlab.eclipse.org/eclipsefdn/helpdesk/-/wikis/Wiki-shutdown-plan for the plan.
CBI/Mar20 2012
Contents
Meeting
- March 20, 2012, 10am EST
- Eclipse Conference facility
Attendees
Andrew, Thanh Ha, Krzysztof Daniel (Chris), Paul Webster, David Williams, Kim Moir
Regrets:
Igor Fedorenko
Minutes
Bug 374666
Bug: Run CBI & PDE builds of platform in parallel
Discussed this bug in today's cbi-dev call. The consensus was:
1) If the qualifier(s) used for the CBI build were different, without development, the comparator tool wouldn't be useful. Conversely, with the qualifiers the same, it would work like it does today and help provide some level of confidence the builds were the same.
2) SR0 with one set of qualifiers and SR1 with another would likely trigger a re-download of all platform bundles. This isn't what we want.
M6 Rebase status
Thanh provided an update that he's rebased the code, and now rebasing the natives. He'll communicate any issues he runs into via. cbi-dev.
M5 or M6 packaging status
Markus was not present.
David questioned the emphasis on testing packaging at this point. Andrew noted it was for testing purposes and to discover any potential issues sooner rather than later. Andrew acknowledged the paradox of decreasing risk by trying packaging now but increasing the risk of the CBI dopple-ganger packages getting into the wild and causing confusion.
Agreement that it does no harm to try packaging early (i.e. now) to test, and that the emphasis should be on closing gaps to make the CBI build output the same as PDE build output. Our packages built should be kept low-profile to avoid them getting out into the wild. [dw edit: To clarify, I think the statement was that the EPP Packaging does very little actual building, it is almost entirely assembly, so the thought was it does no harm to to try and make sure that assembly process still works (and it will be obvious if it works or not) but there is no use in "making the results available" for testing (low profile or not :) since there's nothing to be gained by that, some danger of "letting into the wild", and distracts from the more important work of confirming bits are the same, the JUnits pass both cases, etc. ... all stated as just my point of view].
Bug 355430
Bug: change primary builder to 4.2
Discussed this briefly to get a sense of how it would affect the CBI build. For context, the M4 & M5 platform builds used the 3.8 build as the primary build and the 4.2 builds re-used bundles from 3.8. The plan moving forward is to decouple them into two complete builds for 3.8 and 4.2.
This is expected to land for M7.
Bug 370707
Bug: reproducible build version qualifiers
Igor can't make it today so Andrew requested we keep discussion for this one in the Bug.