[Spice-devel] Survey of repository preferences

[Frediano Ziglio]

> 
> On Mon, Jul 31, 2017 at 09:57:41AM +0200, Christophe de Dinechin wrote:
> > > What is a difference between a personal branch and a branch at the
> > > offical repo?
> > 
> > Visibility. Having an implicit way to share not just the diffs (the way
> > patches do)
> > but the actual branch. There is a reason kernel.org hosts git repos and not
> > just a mailing list with patches: git branches are much easier to deal
> > with.
> > That’s why Linus “lieutenants” send him pull requests for whole branches,
> > and not individual patches.
> 
> I'm not sure if you are implying that Linus "Lieutenants" push their
> work on kernel.org before sending him a pull request, but Linus pulls
> code from many different places, and each lieutenant decides which place
> is best for him. Some examples from a kernel clone I have around:
> 
> Merge branch 'fixes' of git://git.armlinux.org.uk/~rmk/linux-arm
> Merge branch 'upstream' of
> git://git.linux-mips.org/pub/scm/ralf/upstream-linus
> Merge branch 'for-linus' of
> git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
> (the last one being hosted on kernel.org, but in a personal repo)
> 
> Christophe
> 

One of the reason (as far as I remember) we setup patchwork was
a tool to easily grab the patches. Series can be applied with a
single command.

The Linux case is a bit different, mostly for size reasons. Linus
merges entire features of Linux, not single patches or series. A
feature can contain tons of different patches and is quite impossible
to ask to merge such feature through e-mail (would requite possibly
hundred of mails for a single feature) so pull requests are used.
In Linus the way to ask for a pull request is through e-mail. Kind
of "please pull the branch XXX at YYYY". At low level accepting a
pull request (in whichever way is done) require something like
"git pull GIT_REPO:BRANCH && git push". GitHub and GitLab offer
web interfaces to do it.

In our case using the ML as source of pull requests would require
people to setup a public branch in a repository and send a git URL
asking for merge. However if we would like to retain the reviews
on the ML the author have to send the patches content too so this
would end up with the suggestion (3b on my recap?) to add an URL
to the cover/patch message. But this to me don't seem to help
much tracking the status of patches merged/reviewed.

Frediano