jt/clone-server-option (was What's cooking in git.git (Apr 2019, #03; Tue, 16))

[Jonathan Tan <jonathantanmy@google.com>]

> * jt/clone-server-option (2019-04-15) 4 commits
>  - SQUASH???
>  - clone: send server options when using protocol v2
>  - SQUASH???
>  - transport: die if server options are unsupported
> 
>  "git clone" learned a new --server-option option when talking over
>  the protocol version 2.
>
>  Getting there.

With the two "SQUASH???" commits, I think that all outstanding comments
have been addressed, and Jonathan Nieder has given his Reviewed-by [1]
[2].

[1] https://public-inbox.org/git/20190409204525.GA92879@google.com/
[2] https://public-inbox.org/git/20190409204616.GB92879@google.com/

I understand that the following references are not exhaustive (in
particular, Szeder made some comments too), but I'll address them:

>  cf. <xmqqk1g342xi.fsf@gitster-ct.c.googlers.com>

This was about (1) what happens if the given option is not understood by
the protocol-v2-supporting other side, and (2) what happens if protocol
v2 is not used.

For (2), I have made the client die upon Jonathan Nieder's suggestion,
similar to what "git push" does when the other side does not support
push options. For (1), I have documented (in Documentation/ and in the
commit message) that handling of server options are server-specific,
just like push options (which are just sent directly to hooks).

>  cf. <20190406115728.GB219876@google.com>

This was about (1) giving an example of what a server option can be used
for, (2) using NODUP instead of DUP, and (3) a long line. (2) and (3)
have been changed. As for (1), I replied in [3] that since it is
server-specific, I don't think I can give a meaningful example.

[3] https://public-inbox.org/git/20190408171149.125872-1-jonathantanmy@google.com/