From: Junio C Hamano <gitster@pobox.com>
Date: Mon, 07 May 2012 18:10:23 GMT
Subject: Re: [PATCH] correct git-status Porcelain Format documentation
Message-ID: <7vobpz6ewg.fsf@alter.siamese.dyndns.org>
In-Reply-To: <20120506135141.GA17176@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

> What should "git status --porcelain -z -b" look like? With my patch, it
> prints the branch-line with only a newline, not respecting the NUL
> termination. Which sounds like a bug, except that's what happens _now_
> with "git status --short -z -b".  Which seems like a bug to me, but
> maybe somebody is relying on that. It seems kind of broken to me.

It is broken; and it is not a problem if somebody relied on a broken
output without giving --porcelain in a script and such a script needs to
be updated.  If we are updating the code to give "## <branch>" in the
output under the "--porcelain", let's do that right from the beginning.

> Also, while looking at the documentation, we say of "-z": "This implies
> the --porcelain output format if no other format is given". But the only
> other format you could give is "--short", since there is no way to ask
> for the long output (and nor do we handle NUL-termination in that code,
> anyway).

... nor do we define long porcelain format to begin with.

> Should this be simplified to just "this implies porcelain"?
>
> That would technically break somebody who wanted their
> status.relativePaths config option respected. But it kind of seems crazy
> to me.

I do not think it is worth changing it.  It is not too much touble to
spell "status --porcelain -z <whatever else>" in a Porcelain script only
once and forget about it.