Followup-For: Bug #319544
Package: nvidia-kernel-source
*** Please type your report below this line ***
This bug is caused by an error in debian/rules.
The script correctly determines what gcc version was used to compile the
kernel in question,with the following assignment in debian/rules:
CC = gcc-$(shell grep LINUX_COMPILER $(KSRC)/include/linux/compile.h | sed
's/.*[gcc|GCC].* \([0-9]\.[0-9]\+\).*/\1/')
However, at no stage is this $CC variable checked for validity. At a
later stage, nv/conftest.sh is called, to check for rivafb. The method
conftest.sh uses creates a tiny file called conftest$$.c, then compiles
it with $CC - if conftest$$.o exists, then rivafb isn't on the system
and continue - else report a rivafb problem and quit.
Since $CC isn't checked for existance, if it isn't on the system,
conftest$$.c
isn't compiled for reasons not related to CONFIG_FB_RIVA being defined;
the error is entirely misleading.
$CC missing is a very likely situation on any system with an amd64
kernel (where /usr/bin/gcc is gcc 3.3, and the kernels are compiled with
gcc 3.4), and also possible on any system where kernels are being
installed rather than compiled.
The correct behaviour should be a sinple check for a valid $CC binary,
and bailing out with a descriptive message if it is missing.
--Jo Shields
-- System Information:
Debian Release: 3.1
APT prefers unstable
APT policy: (500, 'unstable')
Architecture: amd64 (x86_64)
Kernel: Linux 2.6.12-1-amd64-k8
Locale: LANG=en_GB.UTF-8, LC_CTYPE=en_GB.UTF-8 (charmap=UTF-8)
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
