On 10/13/11 11:20, Marc Kleine-Budde wrote: > On 10/13/2011 11:10 AM, Wolfgang Grandegger wrote:
>> From that description it also seems that we can push only to one git >> repository per project. This again rises the question how to handle and >> name the can-utils and can-kmods. I see the following possibilities: >> >> 1. *One* project named "socketcan" or "linux-can" for both. In the git >> repo we could then store all socketcan related code in: >> >> can-utils/... The CAN Utils >> can-lib/... Eventually A Socketcan library, might be >> integrated into can-utils >> kernel/... Out-of-tree kernel modules for 2.6 and 3.x >> kernel-2.4/... Old kernel modules for 2.4 >> >> 2. *Two* different projects named "can-utils" and "can-kmod". > > At least two repos, one for the userspace one for the kernel. Maybe even > more then one userspace repo. Yes. Having separated kernel-space and user-space repos looks good. If we have the possibility to have more than one repos as Martin pointed out here http://git.savannah.gnu.org/cgit/freetype/ we could name them e.g. can/user.git can/kernel.git while user.git contains directories 'utils', 'lib', 'samples', 'apps?' and kernel.git contains directories 'linux-2.4' and 'linux' I'm not really sure about the naming though. Regards, Oliver _______________________________________________ Socketcan-core mailing list Socketcan-core@lists.berlios.de https://lists.berlios.de/mailman/listinfo/socketcan-core
