Re: [RFCv2 09/16] remote.h: add get_remote_capabilities, request_capabilities

From: Junio C Hamano <gitster@pobox.com>
Date: Jun 2, 2015, 21:24 UTC
Message-ID: <xmqqd21d248d.fsf@gitster.dls.corp.google.com>
In-Reply-To: <1433203338-27493-10-git-send-email-sbeller@google.com>

Stefan Beller <sbeller@google.com> writes:

Show 23 quoted lines

> Instead of calling get_remote_heads as a first command during the
> protocol exchange, we need to have fine grained control over the
> capability negotiation in version 2 of the protocol.
>
> Introduce get_remote_capabilities, which will just listen to
> capabilities of the remote and request_capabilities which will
> tell the selection of capabilities to the remote.
>
> Signed-off-by: Stefan Beller <sbeller@google.com>
> ---
>  connect.c | 37 +++++++++++++++++++++++++++++++++++++
>  remote.h  |  3 +++
>  2 files changed, 40 insertions(+)
>
> diff --git a/connect.c b/connect.c
> index a2c777e..4ebe1dc 100644
> --- a/connect.c
> +++ b/connect.c
> @@ -105,6 +105,43 @@ static void annotate_refs_with_symref_info(struct ref *ref)
>  	string_list_clear(&symref, 0);
>  }
>  
> +void get_remote_capabilities(int in, char *src_buf, size_t src_len)

We need to clarify that this is for v2 only in its name, no?

Show 15 quoted lines

> +{
> +	string_list_clear(&server_capabilities, 1);
> +	for (;;) {
> +		int len;
> +		char *line = packet_buffer;
> +
> +		len = packet_read(in, &src_buf, &src_len,
> +				  packet_buffer, sizeof(packet_buffer),
> +				  PACKET_READ_GENTLE_ON_EOF |
> +				  PACKET_READ_CHOMP_NEWLINE);
> +		if (len < 0)
> +			die_initial_contact(0);
> +
> +		if (!len)
> +			break;

OK, so we get sequence of capabilities, one per packet, until packet_flush(). I wonder if we want to hint that with:

		if (!len)
			break; /* flush */

or something like that.

> +		string_list_append(&server_capabilities, line);
> +	}
> +}

> +int request_capabilities(int out, struct transport_options *options)
> +{
> +	if (options->multi_ack == 2)    packet_write(out, "multi_ack_detailed");

Spell these just like you would do any other functions, i.e.

	if (options->multi_ack == 2)
        	packet_write(out, "multi_ack_detailed");

I think this step is sensible, and anticipate that a new method that corresponds to this will be introduced to the transport layer.