Hi, On Thu, Aug 25, 2005, Blars Blarson wrote: > > Please point me at the full buildd log, and I'd really be interested in > > the config.log. > http://buildd.debian.org/fetch.php?&pkg=gstreamer0.8&ver=0.8.10-2&arch=sparc&stamp=1122583775&file=log&as=raw > http://packages.qa.debian.org/g/gstreamer0.8.html has a "buildd logs" > link if you wish to compare other buildd logs.
Yes, I'm aware of such resources. I asked for yours because the one on vore is older than the latest toolchain packages. > My pbuilder system is down right now (failed hardware upgrade) so I > don't have access to those logs at the moment. For config.log I'd > need to manually redo the build. That's too bad, the config.log is really needed, look at configure's output: ... checking for working mmap... yes checking for usable SVR4/SUSv2 makecontext(2)/swapcontext(2)... no ... Please note that on all other arches, this last check on makecontext succeeds, this is why it looks like a toolchain issue to me, and libc6 had precisely issues with ucontext.h on sparc, hence my doubts. The configure script tries to build and run a test program as defined in the following macro: AC_DEFUN([GST_CHECK_MAKECONTEXT], [ AC_MSG_CHECKING([whether we have makecontext]) AC_TRY_RUN([ #include <ucontext.h> #include <stdlib.h> void test(void) { exit(0); } int main(int argc, char *argv[]) { ucontext_t ucp; int ret; ret = getcontext(&ucp); if(ret<0)exit(1); ucp.uc_stack.ss_sp = malloc(65536); ucp.uc_stack.ss_size = 65536; makecontext(&ucp,test,0); setcontext(&ucp); exit(1); } ], HAVE_MAKECONTEXT="yes", HAVE_MAKECONTEXT="no") AC_MSG_RESULT($HAVE_MAKECONTEXT) if test "$HAVE_MAKECONTEXT" = "yes"; then AC_DEFINE_UNQUOTED(HAVE_MAKECONTEXT, $HAVE_MAKECONTEXT, [defined if we have makecontext ()]) fi ]) I would be interested in knowing how the above behaves. Maybe I should bring this to debian-sparc? Bye, -- Loïc Minier <[EMAIL PROTECTED]> Come, your destiny awaits!