This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Fixes for 2.15.1
- From: Roland McGrath <roland at hack dot frob dot com>
- To: "Joseph S. Myers" <joseph at codesourcery dot com>
- Cc: libc-alpha at sourceware dot org, carlos at systemhalted dot org
- Date: Thu, 23 Feb 2012 11:15:46 -0800 (PST)
- Subject: Re: Fixes for 2.15.1
- References: <Pine.LNX.4.64.1202231805150.31497@digraph.polyomino.org.uk>
> commit fb59b3a4f54777652dc877a1df0fcc009b741d87
> Author: David S. Miller <davem@davemloft.net>
> Date: Thu Feb 16 14:56:54 2012 -0800
>
> Add O_FSYNC define to sparc just like other platforms.
I'd call this harmless but not really necessary. Strictly speaking, it's a
feature addition. OTOH, given that there hasn't really been a 2.15 release
at all yet, we can be less strict than usual minor-release rules.
> commit 7c35ffedf144417ba2787322c7b75b4db5c3cb7a
> Author: Thomas Schwinge <thomas@codesourcery.com>
> Date: Fri Feb 10 21:05:54 2012 +0100
>
> Fix x86 PLT slot usage for feraiseexcept.
>
> Then we're elf/check-localplt.out-clean again.
I'd say yes.
> commit af850b1c978bdca648ef9fb141e785d75f74d9bf
> Author: Richard Henderson <rth@twiddle.net>
> Date: Thu Feb 9 11:21:47 2012 -0800
>
> Use <> for include of kernel-features.h.
This might be necessary for ports to have a working release (not sure).
Certainly ought to be harmless. I'd say yes.
> Also, I didn't look at whether any Hurd changes should be backported;
I'll leave that at tschwinge's discretion. The only Hurd changes have been
bug fixes, but I don't think anybody is close to using an unmodified glibc
on Hurd anyway.
> didn't look at whether any documentation changes should be backported;
Doesn't really matter either way.
> didn't look at what might be backport-appropriate in the ports repository;
Should be up to each port maintainer.
> and didn't look at the question of backporting libc changes needed for the
> Tile ports.
If the tile ports went in using GLIBC_2.15 then it might be appropriate.
But we need to consider each case for risk on its own.
> Ryan suggested backporting stdc-predef.h because he wanted it
> for libdfp use, although that change would probably be outside the normal
> scope of backporting (it's not a bug fix).
I'd tend toward no, but could be convinced otherwise given the "hasn't
really been a release yet" logic. But it is a very recent change that has
not yet had any time to shake out any unanticipated issues.
> Of course some people may have suggestions for other changes that should
> go on 2.15 branch.
I'd like to have us all reminded of the bugzilla conventions for requesting
those and then use them for each candidate.
Thanks,
Roland