Re: [PATCH] log --oneline: put decoration at the end of the line

From: Jeff King <peff@peff.net>
Date: Sep 19, 2012, 20:05 UTC
Message-ID: <20120919200526.GA22650@sigill.intra.peff.net>
In-Reply-To: <7vr4pxg507.fsf@alter.siamese.dyndns.org>

On Wed, Sep 19, 2012 at 12:57:28PM -0700, Junio C Hamano wrote:

Show 13 quoted lines

> Jeff King <peff@peff.net> writes:
> 
> > We should also consider briefly whether anybody is relying on --oneline
> > for machine parsing. I think "log --oneline" is fair game, but I wonder
> > if people calling "rev-list --decorate --oneline" should be considered.
> > It seems kind of unlikely to me, considering that the decorate output is
> > ambiguous to parse anyway (if you see parentheses, you cannot tell if it
> > is decorate output or part of the commit subject).
> 
> Yeah, I do not think it is likely.  Among the in-tree scripts,
> git-stash does use rev-list --oneline but the purpose of the call
> exactly is to grab a human readable one line summary, and it will be
> happy with any change to make --oneline more human readble.

Yeah, that makes sense.

> t4202 has many invocations of "log --oneline --decorate", though;
> these things do get tested.

I think it is just trying to compare the "log.decorate" variable to the "--decorate" command-line option. Notice that it generates the expected output by running the latter. So I think it would be OK (and if not, I think it would make sense to update the test, as it does not care about the specific format).

Google Code Search doesn't show anything interesting, though I suppose its results are getting continually out of date (they claim to have shut it down, though you can still query it if you use the right URL).

-Peff