On Tue, 2010-12-21 at 16:32 -0500, Colin McCabe wrote: > On Tue, Dec 21, 2010 at 9:44 AM, Laszlo Boszormenyi <gcs@xxxxxxxxx> wrote: > > I've a question. gceph needs everything noted in debian/ceph.install > > (cephfs, cconf, ... the rados classes and manpages), right? The best > > would be to make a ceph-common package, depending on ceph | gceph and > > the latter packages would contain only the binaries ceph and gceph > > respectively. Am I right? > > Yeah. gceph should be a separate, optional package. If you install > gceph, you definitely want the rest of the ceph system, but not vice > versa. Do you mean that installing gceph, the CLI ceph is also needed? > I'm not sure if it makes sense to split off the command-line ceph tool > from the rest of the ceph binaries. It's not a lot of code, and it is > a handy little binary. I guess potentially we could do it if it makes > the packaging cleaner. OK, play with letters. 'A' will mean cephfs via rados classes to manpages; with the package name ceph-common. 'B' will mean the ceph binary and its manpage. 'C' is for gceph and its manpage (if any). (A+B) <-> (A+C) definitely a no go, as they would need to conflict with each other. A - B - C can be a choice, even if B and C will be small in size. (A+B) - C is an other option, this time C will depend on (A+B). Either of the two latter is fine with me. I'd go A-B-C if gceph is fine without the ceph binary. Otherwise (A+B)-C is the only option as I see. It's up to you, upstream which one would be the best. Reading your last reply, you would like to have (A+B)-C as C depending on (A+B), right? Laszlo/GCS -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
