git/list[1] front-page[2] threads[3] people[4] search[5] about
 

Re: [PATCH] clone: send ref-prefixes when using protocol v2

From
Junio C Hamano <gitster@pobox.com>
Date
Jul 20, 2018, 19:53 UTC
Message-ID
<xmqqwotpadlh.fsf@gitster-ct.c.googlers.com>
In-Reply-To
<20180720192749.224284-1-bmwill@google.com>
Brandon Williams <bmwill@google.com> writes:

Is there an end-user visible effect, caused by the lack of "prefix" being fixed with this patch, that is worth describing here? "The server ended up showing refs that are irrelevant to the normal clone request which is only for heads and tags, wasting time and bandwidth", for example?

Show 5 quoted lines
> Signed-off-by: Brandon Williams <bmwill@google.com>
> ---
>
> Noticed we miss out on server side filtering of refs when cloning using
> protocol v2, this will enable that.
Show 59 quoted lines
>
>  builtin/clone.c | 22 +++++++++++++++++-----
>  1 file changed, 17 insertions(+), 5 deletions(-)
>
> diff --git a/builtin/clone.c b/builtin/clone.c
> index 99e73dae85..55cc10e93a 100644
> --- a/builtin/clone.c
> +++ b/builtin/clone.c
> @@ -895,7 +895,8 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
>  	int err = 0, complete_refs_before_fetch = 1;
>  	int submodule_progress;
>  
> -	struct refspec_item refspec;
> +	struct refspec rs = REFSPEC_INIT_FETCH;
> +	struct argv_array ref_prefixes = ARGV_ARRAY_INIT;
>  
>  	fetch_if_missing = 0;
>  
> @@ -1077,7 +1078,7 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
>  	if (option_required_reference.nr || option_optional_reference.nr)
>  		setup_reference();
>  
> -	refspec_item_init(&refspec, value.buf, REFSPEC_FETCH);
> +	refspec_append(&rs, value.buf);
>  
>  	strbuf_reset(&value);
>  
> @@ -1134,10 +1135,20 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
>  	if (transport->smart_options && !deepen && !filter_options.choice)
>  		transport->smart_options->check_self_contained_and_connected = 1;
>  
> -	refs = transport_get_remote_refs(transport, NULL);
> +
> +	argv_array_push(&ref_prefixes, "HEAD");
> +	refspec_ref_prefixes(&rs, &ref_prefixes);
> +	if (option_branch) {
> +		expand_ref_prefix(&ref_prefixes, option_branch);
> +	}
> +	if (!option_no_tags) {
> +		argv_array_push(&ref_prefixes, "refs/tags/");
> +	}
> +
> +	refs = transport_get_remote_refs(transport, &ref_prefixes);
>  
>  	if (refs) {
> -		mapped_refs = wanted_peer_refs(refs, &refspec);
> +		mapped_refs = wanted_peer_refs(refs, &rs.items[0]);
>  		/*
>  		 * transport_get_remote_refs() may return refs with null sha-1
>  		 * in mapped_refs (see struct transport->get_refs_list
> @@ -1231,6 +1242,7 @@ int cmd_clone(int argc, const char **argv, const char *prefix)
>  	strbuf_release(&value);
>  	junk_mode = JUNK_LEAVE_ALL;
>  
> -	refspec_item_clear(&refspec);
> +	refspec_clear(&rs);
> +	argv_array_clear(&ref_prefixes);
>  	return err;
>  }
Previous: Brandon WilliamsNext: Jonathan Nieder
Message 2 of 4 in “clone: send ref-prefixes when using protocol v2”
  1. clone: send ref-prefixes when using protocol v2Brandon Williams, Jul 20, 2018
  2. Junio C HamanoJul 20, 2018
  3. Jonathan NiederJul 20, 2018
  4. clone: send ref-prefixes when using protocol v2Brandon Williams, Jul 20, 2018

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.