On 2010-01-30 3:48 PM, Luigi Mantellini wrote: > On Sat, Jan 30, 2010 at 11:25 AM, Weedy <weedy2...@gmail.com> wrote: >> >> https://dev.openwrt.org/wiki/GetSource --> git://nbd.name/openwrt.git >> -- >> A: Because it messes up the order in which people normally read text. >> Q: Why is top-posting such a bad thing? > > I know this link... but using git like just a "SVN" site doesn't > permit to have a complete git-like workflow (like u-boot and linux > kernel projects). > For example, using a full featured git repo, you can have a > fork/branch for each "topic": a branch/fork for the > "toolchain"-related patches, a fork for the "internal infrastructure", > a fork for some class of platform (arm, x86, mips, ...) and provide a > development cycle that merge periodically (during a merge window) the > various flows in order to have a release. Actually a simplified version of the distributed development model would work even with the current repository structure. Anybody can clone the git repo, make a branch, publish it, and merge changes from the git origin. The main difference is that when integrating changes back into the core repo, they have to be rebased before they can be committed. Not a big problem IMHO.
Right now, I have no intentions of moving my development work to branches instead of the central SVN repository, because I think for the core development, the centralized development model works better for us. Also, we don't have to choose between those workflows for the whole project - it can be done on a case by case basis. If one of the OpenWrt developers wants to maintain a specific subsystem in a branch, there's not much in the way of that - it's just that nobody's doing it yet ;) So, aside from simply repeating what some of the other successful projects have done, what specific advantages do you see in changing the development model? - Felix _______________________________________________ openwrt-devel mailing list openwrt-devel@lists.openwrt.org https://lists.openwrt.org/mailman/listinfo/openwrt-devel
