[PATCH v2 2/2] dim: Make sure we push the branch before creating a tag

[Maxime Ripard]

On Thu, Oct 10, 2019 at 03:03:55PM +0200, Daniel Vetter wrote:
> On Thu, Oct 10, 2019 at 11:49:03AM +0200, Maxime Ripard wrote:
> > The current dim code will create the tag using a local branch as an
> > argument, but while we make sure to fetch the branch first in the callers,
> > we don't actually make sure that any local change have been pushed.
> >
> > Make sure it's the case before creating the tag.
> >
> > Signed-off-by: Maxime Ripard <mripard at kernel.org>
> >
> > ---
> >
> > Changes from v1:
> >   - Switch to git_push
> > ---
> >  dim | 1 +
> >  1 file changed, 1 insertion(+)
> >
> > diff --git a/dim b/dim
> > index 0b1cd042345f..b581f246c5cc 100755
> > --- a/dim
> > +++ b/dim
> > @@ -1962,6 +1962,7 @@ function tag_branch
> >  	branch=$2
> >  	remote=$(branch_to_remote $branch)
> >
> > +	git_push $remote $branch
>
> From a tool safety pov, not sure anymore this is the right thing to do.
> You might end up push accidental garbage and fun stuff like that, since
> this bypasses the usual checks dim push-branch does.

Ah, right.

> I think a check using git merge-base --is-ancestor and bailing out if your
> local branch isn't contained in upstream (tagging something that's older
> than latest upstream is legit imo) would be much better. Thoughts?

It could work. Another option would be to just tag the remote branch
instead of the local one. Since the first thing we do is fetching
everything (and since you just said that tagging something older than
latest upstream is fine), it would allow to just ignore any local
state of the branch we want to create a pull request for

Maxime
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/dim-tools/attachments/20191010/7e96a6bb/attachment.sig>