On Tue, 20 Apr 2010, Eric Blake wrote:
> I'm assuming that later in configure we reject -lgnugetopt's getopt_long
> as broken (unless libgnugetopt has somehow picked up this month's glibc
> fix), meaning that we still want to declare rpl_getopt and friends;
Well, there's this:
configure:8265: ch
On 04/20/2010 04:37 PM, Joel E. Denny wrote:
> On Tue, 20 Apr 2010, Eric Blake wrote:
>
> if you use autoconf 2.64 (prior to that, if the existence and
> compilation checks disagree, autoconf went with the existence check).
We are using autoconf 2.65.
>>
>> Then I'm confused how
On Tue, 20 Apr 2010, Eric Blake wrote:
> >>> if you use autoconf 2.64 (prior to that, if the existence and
> >>> compilation checks disagree, autoconf went with the existence check).
> >>
> >> We are using autoconf 2.65.
>
> Then I'm confused how /opt/csw/include/getopt.h is getting included at
>
On 04/20/2010 02:16 AM, Dagobert Michelsen wrote:
> Here is a rationaly of why OpenCSW has an additional getopt.h:
> The /opt/csw/include/getopt.h is actually from libgnugetopt. It is needed
> for some packages without getopt_long detection during autoconfigure to
> compile on Solaris, as Solaris d
Hi Simon,
Am 31.03.2009 um 12:23 schrieb Simon Josefsson:
Dagobert Michelsen writes:
I am getting a test failure on Solaris 8 Sparc w/Sun Studio 11:
Hi! Thanks for the report. The problem is actually in a self-test
from
gnulib. Your GNU SASL library build should still be fine, if this i
Hi Joel,
Am 20.04.2010 um 06:46 schrieb Joel E. Denny:
On Thu, 15 Apr 2010, Eric Blake wrote:
On 04/15/2010 03:39 PM, Joel E. Denny wrote:
% cat > tmp.cc
#define __EXTENSIONS__ 1
#include
#include
% CC -I/opt/csw/include tmp.cc
"/opt/csw/include/getopt.h", line 122: Error: Only one of a set
This in the gnulib stdint.h:
/* Get those types that are already defined in other system include
files, so that we can "#define int8_t signed char" below without
worrying about a later system include file containing a "typedef
signed char int8_t;" that will get messed up by our macro. Ou
Hi Bruno,
Thanks, I'll make a new release with an updated gnulib and see how
this goes :)
- Sylvain
On Tue, Apr 20, 2010 at 12:29:29AM +0200, Bruno Haible wrote:
> Hi Sylvain,
>
> > https://buildd.debian.org/status/package.php?p=freedink
> >
> > What happens, apparently, is that 'install-reloc
Noticed during a libvirt test run with trap '' SIGPIPE.
* tests/test-select-in.sh: Consume all output.
* tests/test-lseek.sh: Check correct exit status, while avoiding
EPIPE.
Signed-off-by: Eric Blake
---
ChangeLog |7 +++
tests/test-lseek.sh |3 ++-
tests/test-sel
Bruno Haible writes:
> It is bizarre, though, that the warning only appears at the first function
> definition. Seen on Cygwin 1.7.2:
...
> AC_TRY_COMPILE does add a function for 'main', but this deserves a comment at
> least.
Yes, the entire semantics of this flag on Windows appears strange.
>
Bruno Haible writes:
> Simon Josefsson wrote:
>> Right now I create the directory like this:
>>
>> gnulib-tool --create-testdir --with-tests --dir build
>>
>> Is there a way to avoid all the unportable tests? For example, a
>> --without-unportable-tests?
>
> I'm currently using --avoid=havelib
Hi Simon,
Sorry, this thread was in a dark corner of my mailbox where I didn't see your
ping requests.
> >> Bruno, how about this patch?
> >>
> >> /Simon
> >>
> >> diff --git a/m4/visibility.m4 b/m4/visibility.m4
> >> index 35a6dc0..06537d6 100644
> >> --- a/m4/visibility.m4
> >> +++ b/m4/visibil
Simon Josefsson wrote:
> Right now I create the directory like this:
>
> gnulib-tool --create-testdir --with-tests --dir build
>
> Is there a way to avoid all the unportable tests? For example, a
> --without-unportable-tests?
I'm currently using --avoid=havelib-tests when I want to test anythin
13 matches
Mail list logo
