Stefan Johnson tigerphoenixdragon@yahoo.com wrote:
By removing the test for -fstack-protector from the configure script, this compiles and runs just fine.
Yes; Alex Barton already reported that the test doesn't work for gcc 4.X; this was fixed in CVS. I _hope_ that the detection now works for your platform as well.
The diff is here: --- ngircd/configure.in 2006/08/02 10:29:11 1.118.2.2 +++ ngircd/configure.in 2006/08/13 18:21:31 1.118.2.3 @@ -8,7 +8,7 @@ # (at your option) any later version. # Please read the file COPYING, README and AUTHORS for more information. # -# $Id: configure.in,v 1.118.2.2 2006/08/02 10:29:11 alex Exp $ +# $Id: configure.in,v 1.118.2.3 2006/08/13 18:21:31 alex Exp $ #
# -- Initialisation -- @@ -68,7 +68,7 @@ AC_MSG_CHECKING([whether ${CC} accepts -fstack-protector]) ssp_old_cflags="$CFLAGS" CFLAGS="$CFLAGS -fstack-protector" - AC_TRY_COMPILE(,,, ssp_cc=no) + AC_TRY_LINK(,,, ssp_cc=no) echo $ssp_cc if test "X$ssp_cc" = "Xno"; then CFLAGS="$ssp_old_cflags"
"' conftest.c >&5 ld: -fstack-protector: Unknown flag ld: Usage: ld [options] file [...] configure:7020: $? = 1
Hmph, so compiling worked; thats why the test (wrongly) assumed that your cc support -fstack-protector. Using AC_TRY_LINK should work for your platform, then.
Thanks, Florian