Re: What's cooking in git.git (Jan 2016, #02; Mon, 11)

From: Junio C Hamano <gitster@pobox.com>
Date: Jan 12, 2016, 18:47 UTC
Message-ID: <xmqqtwmitz2a.fsf@gitster.mtv.corp.google.com>
In-Reply-To: <alpine.DEB.2.20.1601120939270.2964@virtualbox>

Johannes Schindelin <Johannes.Schindelin@gmx.de> writes:

>>  Rerolled.
>>  Needs review.
>
> Actually, it waits for a re-roll ;-)

True.

I think strbuf_getline() that handles the payload as "text" without having _crlf() suffix is an ideal endgame in the longer term, but I do not think it is a good idea to do that as a flag-day change. So I am inclined not to change the function names around that feature in this series. Others can do the wholesale rename as a separate follow-up topic when the tree is quiescent.

The other point "We prepared to be able to use line termination other than LF and NUL, but in practice nobody uses other values, so tighten the interface to clarify that we allow only these two values" you raised is something I think should be done as a preparatory step for this series, as the correctness of the series depends heavily on that assumption that nobody ever uses any other value than LF or NUL. I am looking at it but haven't published any progress yet.

Thanks.