|
From: Julian S. <js...@ac...> - 2014-05-16 12:47:41
|
Greetings. It'll soon be time for a 3.10 release. I'd like to propose the following: 1 August 2014: feature freeze 1 September 2014: release This puts it at roughly a year since 3.9, and also ties in conveniently with the tentative Fedora 21 schedule. On the whole I think the tree is in a pretty good shape. Items I'd like to see completed for 3.10 are: Implement EXIDX reading (EXIDX is an arm32 specific unwind format) Improve MacOSX 10.9 support Improve AArch64 support Merge the ppc64le port Make sure that gcc 4.9 and glibc 2.20 support is solid Fix as many bugs in docs/internals/3_9_BUGSTATUS.txt as possible Comments on the timing? Other stuff people want get in 3.10 ? J |
|
From: Philippe W. <phi...@sk...> - 2014-05-17 09:53:41
|
On Fri, 2014-05-16 at 14:47 +0200, Julian Seward wrote:
> Greetings.
>
> It'll soon be time for a 3.10 release. I'd like to propose
> the following:
>
> 1 August 2014: feature freeze
> 1 September 2014: release
>
> This puts it at roughly a year since 3.9, and also ties in
> conveniently with the tentative Fedora 21 schedule.
>
> On the whole I think the tree is in a pretty good shape. Items
> I'd like to see completed for 3.10 are:
>
> Implement EXIDX reading (EXIDX is an arm32 specific unwind format)
> Improve MacOSX 10.9 support
> Improve AArch64 support
> Merge the ppc64le port
> Make sure that gcc 4.9 and glibc 2.20 support is solid
> Fix as many bugs in docs/internals/3_9_BUGSTATUS.txt as possible
>
> Comments on the timing? Other stuff people want get in 3.10 ?
Looks all ok to me.
On my side, there are several things in the pipe line, some
are ready or should be ready soon:
- improved stack address info description (patch ready, need review)
- enable vgdb ptrace invoker for aarch64 (and re-enable ptrace for arm)
- helgrind better support for Ada tasks
(still waiting feedback from Ada developers!!!)
Otherwise, working (time permitting) on various small or bigger subjects
- some more improvements in helgrind error messages
(e.g. thread names)
- fix the helgrind false positive with TLS variable
- optimise memcheck vbits storev32/storev64 for
known at JIT time Vbits values
- inlined stacktrace printing in stacktrace
(this would be useful for properly suppressing errors in inlined
functions). Probably not a minor work (and I need to improve
my dwarf knowledge for this)
|