From: Jeff King <peff@peff.net>
Date: Tue, 04 Apr 2017 06:48:45 GMT
Subject: Re: [PATCH] Fix 'git am' in-body header continuations
Message-ID: <20170404064844.argmnrdb7tuk2wkj@sigill.intra.peff.net>
In-Reply-To: <f5763dab-b829-f216-a377-8a71fc4f0c1e@google.com>

On Mon, Apr 03, 2017 at 11:00:09AM -0700, Jonathan Tan wrote:

> > diff --git a/t/t4150-am.sh b/t/t4150-am.sh
> > index 89a5bacac..44807e218 100755
> > --- a/t/t4150-am.sh
> > +++ b/t/t4150-am.sh
> > @@ -983,7 +983,9 @@ test_expect_success 'am works with multi-line in-body headers' '
> >  	rm -fr .git/rebase-apply &&
> >  	git checkout -f first &&
> >  	echo one >> file &&
> > -	git commit -am "$LONG" --author="$LONG <long@example.com>" &&
> > +	git commit -am "$LONG
> > +
> > +    Body test" --author="$LONG <long@example.com>" &&
> 
> Instead of "Body test", I would write something more descriptive like "Not a
> continuation line because of blank line above", but I'm fine with either.

Yeah. I also wonder if we can make the indentation more obvious. I
thought at first that the patch was whitespace mangled. :-/

Maybe:

  SP=" " &&
  cat >msg <<-EOF &&
  $LONG

  $SP This line is indented but not a header continuation.
  EOF
  git commit -F msg ...

or something.

It might also be easier to understand what's going on if this gets its
own test. This is really just testing mailinfo. I wonder if it would
make more sense in t5100, where we would not have to deal with all the
commit/format-patch cruft.

-Peff