Minutes Jun 15 2011

From The Linux Foundation
Jump to: navigation, search


Stew Benedict (LF), Mats Wichmann (Intel), Robert Schweikert (Suse), Jeff Licquia (LF), Alan Clark (Suse)



Jeff: FHS progress - mailing list has quieted down, going through the bugs in the rollup, making changes and pushing, with the intent to push a draft spec for comments. People are welcome to contribute. Questions?
Jeff: SDK - apparently we have a problem with the SDK wrt ppc32. We're inheriting the host machine settings from gcc, resulting in binaries that won't work on lower end (embedded) systems. Fix is easy, question is what the release schedule should be. Do a 4.1.2 with just this fix, or roll in some of the other sdk bugs?
Robert: If we have other bugs to fix, do them and include the ppc32 fix. Don't drop what we're doing for this issue.
Jeff: Sort of impossible to waive tests, when they don't run at all.
Robert: Right, so let's just say the sdk is the next thing we refresh, in the next few weeks.
Jeff: OK, so lets make the SDK the top priority for the next refresh

Some discussion of logistics of releasing sdk and ppc32 fixed packages

Jeff: Anything else on the sdk?
Mats: Still have to deal with the gcc-4.5.x+ issues. less trivial work than the ppc32 fix

Jeff: library uplift (http://www.linuxfoundation.org/en/Uplift_Target), 50-some libraries
Jeff: looking at the list, glibc is probably a hot-button item, 87 new interfaces
Robert: factory just went to 2.13, given that RHEL is at 2.12, 2.13 seems reasonable
Robert: so if we go to 2.13, will we take all 87 interfaces?
Jeff: We don't use all interfaces now, so some may be excluded (looks at list on navigator to comment on list)


Jeff: libz doesn't look like it's changed much, but I seem to remember some change they made?
Stew: yes, there was a change that impacted some of our python stuff
Robert: some entries don't have a version for the LSB version, why?
Stew: couldn't locate a version for some
Robert: should we specify a minimum version in the spec?
Mats: It's really just a reference, the spec if functional based, not version based. Our zlib is based on 1.2.2.
Jeff: ncurses, we base things off POSIX, do we really want to pull in the new stuff?
Mats: I think pam went to symbol versioning, may be a good reason to take a look at it
Jeff: Only one new symbol according to navigator. Probably just the symbol versioning.
Jeff: libstdc++ - uplift may get tied to sdk issues. 101 new symbols
Robert: Most are probably from the new draft standard
Jeff: Some may be bugfixes or inheritence tree changes. Looking at the list, some look familiar
Robert: Are we in agreement to uplift?
Jeff: Need to uplift the sdk's copy of the headers. Changes should be backwards compatible, but we have to make sure we don't pull in new symbols
Jeff: libgcc_s, 4 new interfaces, looks like all double-underscore "stuff"
Mats: good time to remember that there may be other interfaces, never imported the first time that may be of interest

End Of Call