* Backport the following changes from 5.0.35-4 in Raring so that
nvidia-cuda-toolkit can also be built and installed on systems with
nvidia-experimental installed and without depending on and recommending
nvidia-current.
- debian/{control,rules}: Use substvars for the names of the libcuda1 and
nvidia-libopencl1 packages.
- Build-Depend/Depend on nvidia-current | nvidia-current-updates |
nvidia-experimental-*. instead of libcuda1 and nvidia-libopencl1.
(LP: #950963)
- Override libcuda.so.1 dependencies via shlibs.local to support
installing with nvidia-experimental-* while building against
nvidia-current. (LP: #1092259)
-- Graham Inggs <email address hidden> Wed, 20 Feb 2013 13:23:19 +0200
This bug was fixed in the package nvidia-cuda-toolkit - 4.0.17-3ubuntu0.1
---------------
nvidia-cuda-toolkit (4.0.17-3ubuntu0.1) precise-proposed; urgency=low
* Backport the following changes from 5.0.35-4 in Raring so that cuda-toolkit can also be built and installed on systems with experimental installed and without depending on and recommending {control, rules}: Use substvars for the names of the libcuda1 and libopencl1 packages. current- updates | experimental- *. instead of libcuda1 and nvidia-libopencl1. experimental- * while building against current. (LP: #1092259)
nvidia-
nvidia-
nvidia-current.
- debian/
nvidia-
- Build-Depend/Depend on nvidia-current | nvidia-
nvidia-
(LP: #950963)
- Override libcuda.so.1 dependencies via shlibs.local to support
installing with nvidia-
nvidia-
-- Graham Inggs <email address hidden> Wed, 20 Feb 2013 13:23:19 +0200