On Sat, 2018-06-02 at 13:10 -0400, Rob Clark wrote: > On Sat, Jun 2, 2018 at 12:07 PM, Jason Ekstrand <ja...@jlekstrand.net> wrote: > > 1) Forks share objects behind the user's back so it should use > > significantly less space on the GitLab server if everyone's personal mesa > > repo is an actual GitLab fork and not a bare repo that they pushed.
It'd be cool if gitlab could figure out object sharing without this, but I totally understand why that's hard. > kinda on the topic, but not directly related to mesa.. at least a few > of us have private kernel trees on p.fd.o, and kernel git trees are > *much* bigger. Will there be (or is there already) a gitlab kernel > tree that we could recreate our private kernel trees from, to share > objects? > > I guess since it is mostly stuff that feeds into drm-next, maybe > Dave's drm-next tree should move to gitlab, and then others fork off > of that? A fresh clone's linux/.git is about 2.3G at the moment. That's probably not a huge burden in terms of just storage, but might be bad in terms of I/O (since identical data competing for page cache is silly), so having a standard kernel tree mirror does sound wise. - ajax _______________________________________________ mesa-dev mailing list mesa-dev@lists.freedesktop.org https://lists.freedesktop.org/mailman/listinfo/mesa-dev