From: Alex Riesen <raa.lkml@gmail.com>
Date: Thu, 21 Feb 2008 20:35:06 GMT
Subject: Re: [PATCH] A test for commit --amend allowing changing of a very empty commit message
Message-ID: <20080221203506.GA20143@steel.home>
In-Reply-To: <7vwsoyt5xt.fsf@gitster.siamese.dyndns.org>

Junio C Hamano, Thu, Feb 21, 2008 21:23:10 +0100:
> Alex Riesen <raa.lkml@gmail.com> writes:
> >
> > Well, it fails... In current master
> 
> Then please mark it with test_expect_failure.

That's because I expect it to be fixed (in the next mail)

> > +test_expect_success 'allow amend of an empty message' '
> > +
> > +	git reset --hard &&
> > +	sha=$(:|git commit-tree HEAD^{tree} -p HEAD) &&
> 
> You do not have to pipe; </dev/null would do ;-).

Sadly: no. There are very challenged systems. Like, for instance, ones
having no idea of what "/dev/null" is. Windows/Cygwin/ActiveState Perl
in a bloody stupid corporate combination.

I'm trying to minimize differences between my tree for my system at
work and the main tree. It is honestly the only purpose for this
construct.