Re: Bug in git rev-parse @{push}?

From: Jeff King <peff@peff.net>
Date: Nov 3, 2016, 16:07 UTC
Message-ID: <20161103160716.6ygtf5iosudo2w3g@sigill.intra.peff.net>
In-Reply-To: <CAJUoZVY0-OA7Tu=cHDPH+NwmQ8MUuWc37_5aqeGQbwKpNWVzqg@mail.gmail.com>

On Thu, Nov 03, 2016 at 03:59:13PM +0000, Chris Purcell wrote:

Show 16 quoted lines

> >> Thanks, Jeff! If I remove the explicit configuration of remote.pushdefault
> >> = "origin", I get the same error message as you, so I suspect that's _not_
> >> the default.
> >
> > That's really bizarre, because I get the same behavior with or without
> > it set. Not only that, but it shouldn't even come into play, as
> > branch.foo.remote should take precedence anyway.
> >
> > So now I'm really puzzled.
> 
> That's because I lied. Sorry! Too many branches configured. I've actually got
> 
> [branch "foo"]
>     remote = .
> 
> If I change that to 'origin' like I claimed it was, it starts erroring.

OK, now I can reproduce. And changing remote.pushdefault _does_ matter because builtin/push.c:is_workflow_triangular() uses it as a key to "we are in a triangular workflow".

That was added by ed2b18292 (push: change `simple` to accommodate triangular workflows, 2013-06-19). TBH, I think the right solution is "stop using 'simple' in a triangular setup". But since that workflow exists for git-push, we probably need to support it via @{push}, too.

I think the solution would be something like:

  - move is_workflow_triangular() into remote.c so it is accessible in
    both places

  - when branch_get_push_1 sees that is_workflow_triangular() is true,
    treat "simple" as "current"

  - new tests should go into t/t1514-rev-parse-push.sh to cover this
    case

Want to take a stab at it?

-Peff