This is the mail archive of the
libstdc++@gcc.gnu.org
mailing list for the libstdc++ project.
Re: uclibc and libstdc++ and tmpnam
- From: Steve Ellcey <sellcey at mips dot com>
- To: Jonathan Wakely <jwakely dot gcc at gmail dot com>, <aldot at gcc dot gnu dot org>
- Cc: libstdc++ <libstdc++ at gcc dot gnu dot org>
- Date: Thu, 23 Jan 2014 11:32:19 -0800
- Subject: Re: uclibc and libstdc++ and tmpnam
- Authentication-results: sourceware.org; auth=none
- References: <ffb3d516-8f05-4429-8f6c-d3b38656c261 at BAMAIL02 dot ba dot imgtec dot org> <CAH6eHdT4mEA28kkTQJqYW2M1EYr+1GjTdEK7s+G8MQV_xxs7Gw at mail dot gmail dot com>
On Wed, 2013-12-04 at 21:41 +0000, Jonathan Wakely wrote:
> On 4 December 2013 21:34, Steve Ellcey wrote:
> >
> > I was trying to build a GCC toolchain with uclibc and ran into this
> > problem:
> >
> > /local/home/sellcey/gcc/uclibc/obj-mips-mti-linux-uclibc/gcc/final/mips-mti-linux-uclibc/libstdc++-v3/include/cstdio:140:11: error: '::tmpnam' has not been declared
> > using ::tmpnam;
> >
> > Some searching led me to this email string from earlier this year:
> >
> > http://gcc.gnu.org/ml/libstdc++/2013-04/msg00013.html
> >
> > But it doesn't look like anything has been checked in. Does anyone know
> > what the state of this issue is? Should I go off and bug the uclibc
> > people?
>
> The thread was resurrected last month:
> http://gcc.gnu.org/ml/libstdc++/2013-11/msg00034.html
> I approved it here:
> http://gcc.gnu.org/ml/libstdc++/2013-11/msg00064.html
Jonathan,
I tried pinging Bernhard about this patch but didn't hear back, would
it be OK for me to check this in on his behalf? The only change I
made to his patch was to use tmpnam(NULL) in GLIBCXX_CHECK_TMPNAM
instead of tmpnam("NULL"). I think that makes more sense since tmpnam
is going to try and write to the buffer that you pass it as an argument
(if it is not null) and so passing a constant string does not make
sense. I tested it with my mips-mti-linux-gnu target and made sure that
tmpnam was found when using glibc and not found with uclibc and that
both builds worked.
Steve Ellcey
sellcey@mips.com