Re: [PATCH 2/2] git-gui: support for diff3 conflict style

From: Pratyush Yadav <me@yadavpratyush.com>
Date: Oct 3, 2019, 20:54 UTC
Message-ID: <20191003205407.uaguw7qyabfy3aym@yadavpratyush.com>
In-Reply-To: <dc210bf0-eef3-ed58-f596-0999fd9d9a98@iee.email>

On 03/10/19 09:02PM, Philip Oakley wrote:

Show 46 quoted lines

> On 30/09/2019 13:17, Bert Wesarg wrote:
> > Pratyush,
> > 
> > On Sun, Sep 29, 2019 at 5:04 PM Pratyush Yadav <me@yadavpratyush.com> wrote:
> > > Hi Philip, Bert,
> > > 
> > > Is there any way I can test this change? Philip, I ran the rebase you
> > > mention in the GitHub issue [0], and I get that '9c8cba6862abe5ac821' is
> > > an unknown revision.
> > > 
> > > Is there any quick way I can reproduce this (maybe on a sample repo)?
> > The easiest way to produce a merge conflict, is to change the same
> > line differently in two branches and try to merge them. I added a
> > fast-import file to demonstrate this for you.
> > 
> > $ git init foo
> > $ cd foo
> > $ git fast-import <../conflict-merge.fi
> > $ git reset --hard master
> > $ git merge branch
> > 
> > this gets you into the conflict, probably the usual style. Which looks
> > in liek this on the terminal:
> > 
> > @@@ -2,7 -2,7 +2,11 @@@ Lorem ipsum dolor sit amet, consectetu
> >    Sed feugiat nisl eget efficitur ultrices.
> >    Nunc cursus metus rutrum, mollis lorem vitae, hendrerit mi.
> >    Aenean vestibulum ante ac libero venenatis, non hendrerit orci pharetra.
> > ++<<<<<<< HEAD
> >   +Proin bibendum purus ut est tristique, non pharetra dui consectetur.
> > ++=======
> > + Proin placerat leo malesuada lacinia lobortis.
> > ++>>>>>>> branch
> >    Pellentesque aliquam libero et nisi scelerisque commodo.
> >    Quisque id velit sed magna molestie porttitor.
> >    Vivamus sed urna in risus molestie ultricies.
> > 
> > and this in git gui: https://kgab.selfhost.eu/s/gHHaQqowGp7mXEb
> 
> The snapshot of the Gui looks just the thing! (I've been away).
> 
> I'm sure this would solve my immediate issue.
> 
> My only remaining bikeshed question it prompted was to check which parts
> would be committed as part of committing the whole "hunk". But haven't had
> time to look at all!

I'm not sure what you mean by "committing the whole hunk". In a merge conflict state, you don't get the usual "Stage hunk" and "Stage lines" options, but instead get 3 options:

  Use Remote Version
  Use Local Version
  Revert To Base

You can use these to choose how you want to resolve the conflict.

These 3 options seem to work fine on my quick testing.

-- 
Regards,
Pratyush Yadav