Re: vma-iter: support Solaris

2017-04-18 Thread Bruno Haible
Tom G. Christensen wrote: > On Solaris 2.6 and 7 there is the additional issue of missing MAP_ANONYMOUS: > > In file included from vma-iter.c:41:0: > /usr/include/sys/procfs.h:44:2: error: #error "Cannot use procfs in the > large file compilation environment" > vma-iter.c: In function 'vma_iterat

Re: vma-iter: support Solaris

2017-04-18 Thread Bruno Haible
Tom G. Christensen wrote: > This is causing my daily gnulib builds to fail on Solaris. > > On Solaris 8 and 9 I'm seeing this error: > > In file included from vma-iter.c:41:0: > /usr/include/sys/procfs.h:44:2: error: #error "Cannot use procfs in the > large file compilation environment" > #err

Re: Merge getopt from gnulib to glibc and vice versa, eliminate __need_getopt

2017-04-18 Thread Bruno Haible
Hi Paul, > 0002-getopt-gnu-omit-some-duplicate-code.patch It took me some time to convince myself that this patch is correct. Here's a test case: $ tar xvfz hello.tar.gz (attached) $ cd hello $ $GNULIB_TOOL --import --lib=libposix --source-base=libp --m4-base=m4p --macro-prefix=gp getopt-posix $

Re: gnulib poll() returns POLLERR on Win32 platform, breaking libvirt

2017-04-18 Thread Gisle Vanem
Daniel P. Berrange wrote: that makes the recv() work, but now poll() just gets into an infinite loop calling select() over and over again, never reporting that the socket is now read()able. So something else is now broken. I guess we're now calling into gnulib's rpl_select() instead of the bare

Re: vma-iter: support Solaris

2017-04-18 Thread Tom G. Christensen
On 19/03/17 16:38, Bruno Haible wrote: I'm adding support for Solaris to the 'vma-iter' module. It's useful for GNU clisp. This is causing my daily gnulib builds to fail on Solaris. On Solaris 8 and 9 I'm seeing this error: In file included from vma-iter.c:41:0: /usr/include/sys/procfs.h:44:

gnulib poll() returns POLLERR on Win32 platform, breaking libvirt

2017-04-18 Thread Daniel P. Berrange
We're seeing a failure in libvirt on Win32 platforms whereby poll() often returns POLLERR. I traced this down to the rpl_recv() function calling FD_TO_SOCKET() and getting INVALID_SOCKET back. This is propagated back to windows_compute_revents_socket() which sets POLLERR, because WSAGetLastError()

Re: Problem with valgrind-tests: relies on bash not causing error

2017-04-18 Thread Bernhard Voelker
On 04/14/2017 10:56 PM, Marc Nieper-Wißkirchen wrote: > While your proposed solution would work for anyone who is aware of this > problem, it won't work for an unaware user who would simply run ./configure, > wondering why valgrind isn't detected. As it seems > rather involved to resolve the inco