Re: Building glibc 2.5 on Hurd

On Sun, Feb 11, 2007 at 01:31:35AM +0100, Samuel Thibault wrote: > Hi, > > Aurelien Jarno, le Thu 08 Feb 2007 20:52:33 +0100, a écrit : > > sigsuspend_not_cancel is not defined for Hurd in > > sysdeps/mach/hurd/sigsuspend.c. This function is called from > > sysdeps/posix/pause.c. > > > > Any volu

Re: Building glibc 2.5 on Hurd

Hi, Aurelien Jarno, le Thu 08 Feb 2007 20:52:33 +0100, a écrit : > sigsuspend_not_cancel is not defined for Hurd in > sysdeps/mach/hurd/sigsuspend.c. This function is called from > sysdeps/posix/pause.c. > > Any volunteer to implement it? Mmm, it looks to me like the current __sigsuspend() doesn

Re: Building glibc 2.5 on Hurd: ``MSG_NOSIGNAL''

Thomas Schwinge a écrit : > Hello! > > This is not a proper fix, but a workaround that is needed on HEAD. (Have > to check glibc-2_5-branch.) Reported on > . > This workaround is also necessary for the 2.5 branch. Applied in our SVN as it seems harmless for

Re: Building glibc 2.5 on Hurd: GCC

Thomas Schwinge a écrit : > Hello! > > Needed because of GCC changes. Not everything is ready for upstream > inclusion. > > > This is no proper fix, but a workaround to not make argp-enabled programs > segfault on ``PROGRAM --help'' when glibc is build with GCC 4.1. > Reported on >

Re: Building glibc 2.5 on Hurd

Aurelien Jarno a écrit : > Hi all, > > I am currently trying to port the glibc 2.5 on Hurd. I am facing the > problem described here: > I am facing another problem now... sigsuspend_not_cancel is not defined for Hurd in sysdeps/mach/hurd/sigsuspend.c. This function is called from sysdeps/posix/

Re: Building glibc 2.5 on Hurd: ``__libc_once''

Thomas Schwinge a écrit : > Hello! > > Needed on glibc-2_5-branch and HEAD. Reported at > and finally on > , where it was > rejected by Ulrich Drepper. Roland wanted to have a look. > > 2006-12-18 Thomas

Re: Building glibc 2.5 on Hurd: trivia

Thomas Schwinge a écrit : > Needed for HEAD and glibc-2_5-branch. > > 2007-02-07 Thomas Schwinge <[EMAIL PROTECTED]> > > * stdlib/fmtmsg.c: Include . > > Index: stdlib/fmtmsg.c > === > RCS file: /cvs/glibc/libc/stdlib/fmtmsg

Re: Building glibc 2.5 on Hurd: tls

Barry deFreese a écrit : > Aurelien Jarno wrote: >> Thomas Schwinge a écrit : >> >>> Hello! >>> >>> Support for tls for Hurd systems is still not fixed properly. Details >>> are at . To get a functional >>> glibc-2_5-branch, this either needs to be fixed or

Re: Building glibc 2.5 on Hurd: ``-z relro''

Thomas Schwinge a écrit : > Hello! > > See for details. This should be > fairly easy to fix, but has to be tracked down. We are still using libc_cv_z_relro=no in the Debian package. The change to configure.in has been done after the 2.5 release, therefore I

Re: Building glibc 2.5 on Hurd: tls

Aurelien Jarno wrote: Thomas Schwinge a écrit : Hello! Support for tls for Hurd systems is still not fixed properly. Details are at . To get a functional glibc-2_5-branch, this either needs to be fixed or the tls-requiring bits of glibc-2_5-branch have

Re: Building glibc 2.5 on Hurd: tls

Thomas Schwinge a écrit : > Hello! > > Support for tls for Hurd systems is still not fixed properly. Details > are at . To get a functional > glibc-2_5-branch, this either needs to be fixed or the tls-requiring bits > of glibc-2_5-branch have to be bend over

Re: Building glibc 2.5 on Hurd: tls

Thomas Schwinge wrote: Hello! Support for tls for Hurd systems is still not fixed properly. Details are at . To get a functional glibc-2_5-branch, this either needs to be fixed or the tls-requiring bits of glibc-2_5-branch have to be bend over to not requir

Re: Building glibc 2.5 on Hurd

Hello! On Wed, Feb 07, 2007 at 12:07:14PM +0100, Aurelien Jarno wrote: > I am currently trying to port the glibc 2.5 on Hurd. Thank you very much for that! And yes, glibc-2_5-branch is the right choice for now. HEAD will need much more work. > I am facing the problem described here: > > http