On Wed, Feb 12, 2020 at 02:19:40PM -0800, Dan Williams wrote: > On Wed, Feb 12, 2020 at 2:07 PM Darrick J. Wong <darrick.wong@xxxxxxxxxx> wrote: > [..] > > I've really really wanted to be able to tell people to just send a pull > > request for large series and skip all the email patch review stuff, but > > I'm well aware that will start a popular revolt. But maybe we can do > > both? Is it legit to ask that if you're sending more than a simple > > quickfix, to please push a branch somewhere so that I can just yank it > > down and have a look? I try to do that with every series that I send, I > > think Allison has been doing that, Christoph does it sometimes, etc. > > This is begging me to point out that Konstantin has automated this > with his get-lore-mbox tool [1]. As long as the submitter uses "git > format-patch --base..." then the tool can automate recreating a local > branch from a mail series. > > [1]: https://lore.kernel.org/workflows/20200201030105.k6akvbjpmlpcuiky@chatter.i7.local/ Yes, I'm aware of the development of git-lore-mbox, but I'd rather just pull directly from a developer's git repo than create branches on my computer. I /already/ have my own script to extract patches and apply/paste/staple them onto a git tree, and xfsprogs already has scripts to automate porting of libxfs change from the kernel. My goal here is more to change the balance of who does what work a little bit back towards patch authors, than it is to spend less time putting together git trees. I mean, most of you already use git already anyway, right? :) --D
