Re: [PATCH v3 5/5] Documentation: add caveats about I/O buffering for check-{attr,ignore}

From: Adam Spiers <git@adamspiers.org>
Date: Apr 12, 2013, 11:00 UTC
Message-ID: <20130412110045.GC26949@pacific.linksys.moosehall>
In-Reply-To: <7vzjx4fqex.fsf@alter.siamese.dyndns.org>

On Thu, Apr 11, 2013 at 07:12:22PM -0700, Junio C Hamano wrote:

Show 5 quoted lines

> It is usually OK to re-flow the text in the paragraph you are
> touching. After all, for the purpose of reviewing, people can just
> blindly apply and then ask "diff --color-words".  In this case,
> however, there was some changes that conflict in the vicinity, and
> reflowing made the resolution unnecessarily more cumbersome.

I see.  Thanks for the tip; I was only dimly aware of --color-words.

> I have briefly looked at this series, but it severely conflicts with
> a few topics in flight that touch the infrastructure you are using,
> so I haven't merged it to 'pu'. Perhaps after things calm down, we
> may want to ask you to reroll on top of updated codebase.

Sure, no problem.  I'll try a quick rebase now to see how ugly it is.

By the way, I've replaced my test for streaming --stdin which was based on stdbuf(1) and sleep(1) with Peff's clever hack based on mkfifo. I'll hold off from sending a reroll until pathspec activity cools down, but in the meantime it's available here:

    https://github.com/aspiers/git/compare/master...git-annex-streaming

It requires my "t: make PIPE a standard test prerequisite" patch, but I notice that's already in master which will make things easier later on.

Thanks! Adam