From: Junio C Hamano <gitster@pobox.com>
Date: Thu, 01 Apr 2010 04:54:34 GMT
Subject: Re: [PATCH 5/7] remote-helpers: add support for an export command
Message-ID: <7v634bsrg5.fsf@alter.siamese.dyndns.org>
In-Reply-To: <1269881309-19690-6-git-send-email-srabbelier@gmail.com>

Sverre Rabbelier <srabbelier@gmail.com> writes:

> ---
>
> 	This complements the 'import' capability, making it possible
> 	to use remote-helpers with only fast-import/fast-exports
> 	streams.

I've queued this in 'pu' only to say "I've seen it", but I suspect that
this has to wait until after 1.7.1.  I've been trying to shoot for 6-8
week cycle as originally envisioned, and if things go as planned, post
1.7.1 cycle opens around mid April.

> +	for(ref = remote_refs; ref; ref = ref->next) {
> +		char* private;
> +		unsigned char sha1[20];
> +
> +
> +		if(data->refspecs)
> +			private = apply_refspecs(data->refspecs, data->refspec_nr, ref->name);
> +
> +		if(private && !get_sha1(private, sha1))
> +		{
> +			strbuf_addf(&buf, "^%s", private);
> +			string_list_append(strbuf_detach(&buf, NULL), &revlist_args);
> +		}
> +
> +		string_list_append(ref->name, &revlist_args);
> +
> +	}

What's happening here?  Style violations aside, it appears that "private"
is accessed uninitialized when data->refspecs is NULL.