Re: [PATCH] pack-protocol: document newline behavior in push commands

From: Jeff King <peff@peff.net>
Date: Sep 4, 2015, 09:00 UTC
Message-ID: <20150904090044.GB32751@sigill.intra.peff.net>
In-Reply-To: <xmqq4mjbi1w1.fsf@gitster.mtv.corp.google.com>

On Thu, Sep 03, 2015 at 03:17:18PM -0700, Junio C Hamano wrote:

Show 10 quoted lines

> Jeff King <peff@peff.net> writes:
> 
> > Frankly, this feels a bit like a step backwards to me. I am tempted to
> > suggest instead that git start sending the newlines, but I'm not sure
> > it's worth any potential fallout.
> 
> I actually think we should do both in the longer term.
> 
> If we say sender "SHOULD" and we know no existing receiver violates
> the "MUST NOT reject", our sender should follow that "SHOULD".

Right, it was the second "we know..." that made me worry. It is really "we assume". :) Whether it is right according to the spec or not, the real world is sometimes more complicated. And given that there is no real advantage to changing the sending behavior now, I didn't think it worth doing.

Show 5 quoted lines

> This documentation update is good in that it makes the examples
> easier to read (by the way, the first pre-context line ends with
> '\n', which we would eventually also address) by making the reader
> understand that the convention used in this S:/C: exchange
> illustration the optional LF is not shown.

In the second patch I left them all intact, but I agree that we could drop the "\n" entirely from the example conversations, as it is implied (and GIT_PACKET_TRACE, for example, does not even show it).

-Peff