This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: v2: The GNU C Library 2.16 release plan
- From: Allan McRae <allan at archlinux dot org>
- To: Mike Frysinger <vapier at gentoo dot org>
- Cc: libc-alpha at sourceware dot org, "Joseph S. Myers" <joseph at codesourcery dot com>,Carlos O'Donell <carlos_odonell at mentor dot com>, Andreas Jaegar <aj at suse dot com>,Paul Eggert <eggert at cs dot ucla dot edu>, "Ryan S. Arnold" <ryan dot arnold at gmail dot com>,"H.J. Lu" <hjl dot tools at gmail dot com>
- Date: Sat, 02 Jun 2012 10:36:54 +1000
- Subject: Re: v2: The GNU C Library 2.16 release plan
- References: <4FAC3FCE.1020402@mentor.com> <Pine.LNX.4.64.1205302354240.8161@digraph.polyomino.org.uk> <Pine.LNX.4.64.1205311847020.15944@digraph.polyomino.org.uk> <201205311551.54950.vapier@gentoo.org>
On 01/06/12 05:51, Mike Frysinger wrote:
> On Thursday 31 May 2012 14:48:52 Joseph S. Myers wrote:
>> Oh, and one other point about release testing:
>>
>> GCC routinely gets tested before release by full GNU/Linux distribution
>> rebuilds - is anyone able to test distribution rebuilds with new glibc
>> after we freeze for 4.6? This may be less likely to show up problems than
>> rebuilds with new GCC are, but it could still be useful.
>
> in testing glibc-2.16 and building up a new chroot, i've noticed two issues:
>
> - a lot of gnulib packages fail with gets() errors ... the gnulib project has
> fixed this, but it'll take a while for that to trickle down into all the
> respective projects. i've just added patches like this to fix it:
> http://sources.gentoo.org/app-arch/gzip/files/gzip-1.4-no-gets.patch
>
> - the fortify warning change wrt optimization is too noisy. i glanced
> through gcc to see if the fortify-by-default could be made to work in a
> different way, but couldn't see a clean way of doing it. instead, i'll just
> carry this in Gentoo:
>
> --- a/include/features.h
> +++ b/include/features.h
> @@ -325,10 +325,11 @@
> # define __USE_REENTRANT 1
> #endif
>
> +#if !defined __OPTIMIZE__ || __OPTIMIZE__ <= 0
> +# undef _FORTIFY_SOURCE
> +#endif
> #if defined _FORTIFY_SOURCE && _FORTIFY_SOURCE > 0
> -# if !defined __OPTIMIZE__ || __OPTIMIZE__ <= 0
> -# warning _FORTIFY_SOURCE requires compiling with optimization (-O)
> -# elif !__GNUC_PREREQ (4, 1)
> +# if !__GNUC_PREREQ (4, 1)
> # warning _FORTIFY_SOURCE requires GCC 4.1 or later
> # elif _FORTIFY_SOURCE > 1
> # define __USE_FORTIFY_LEVEL 2
> -mike
I have rebuilt the Arch Linux [core] repo (basically software needed for
boot-up and building more software...) with glibc-2.15-1124-gebc64a1 for
x86_64. I get the following issues:
A bunch of the afore mentioned gnulib/gets issues (bison, diffutils,
gettext, gzip, inetutils, libpipeline, m4, man-db, tar, wget)
This is a genuine bug in glibc:
pciutils-3.1.9:
/usr/include/sys/io.h: In function 'outsw':
/usr/include/sys/io.h:168:51: error: '____addr' undeclared (first use in
this function)
Patch on the way...
I assume these two are missing header issues but have not investigated yet:
busybox-1.19.4:
shell/shell_common.c: In function 'printlim':
shell/shell_common.c:369:2: error: unknown type name 'rlim_t'
shell/shell_common.c:371:13: error: dereferencing pointer to incomplete type
shell/shell_common.c:373:14: error: dereferencing pointer to incomplete type
shell/shell_common.c:375:13: error: 'RLIM_INFINITY' undeclared (first
use in this function)
pam-1.1.5:
pam_unix_acct.c: In function '_unix_run_verify_binary':
pam_unix_acct.c:97:19: error: storage size of 'rlim' isn't known
pam_unix_acct.c:106:19: error: 'RLIMIT_NOFILE' undeclared (first use in
this function)
I saw a couple of other build failures, but I am not sure they are
related at the moment.
Allan