Bryan wrote, >That sounds OK to me, though I wonder how it will sit with >non-kernel-developers. I think some downstream people are actually >pulling kernel code from Subversion at the moment, which strikes me as >probably a mistaken thing to do.
> <b In some cases, SVN is the only place where the code is, SDP, iSer, and what if I need a bug fix, I have to wait for Linus to open up his tree for the next 2.6.x release and for him to pull from Roland's git tree, then I can get it from kernel.org ? And when I want to backport the code to some earlier kernel, keeping close to the tip of SVN and syncing often makes it easier to keep my patches for backports up to date. If you dump big changes to the tip all at once, it breaks all the backport patches and basically I have to re-backport it all again, which is much harder than just keeping up with a few incremental changes each week. This is why I would like to see the development tree (either SVN or git I don't care which) in the open and used as the working data base, not just somewhere you dump periodic snapshots from your "behind closed doors" repository. woody _______________________________________________ openib-general mailing list [email protected] http://openib.org/mailman/listinfo/openib-general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
