Re: [PATCH] transport: Catch non positive --depth option value

From: Junio C Hamano <gitster@pobox.com>
Date: Nov 18, 2013, 16:51 UTC
Message-ID: <xmqqzjp1bqm3.fsf@gitster.dls.corp.google.com>
In-Reply-To: <5283A380.9030308@gmail.com>

"Andrés G. Aragoneses" <knocte@gmail.com> writes:

Show 8 quoted lines

> Instead of simply ignoring the value passed to --depth
> option when it is zero or negative, now it is caught
> and reported.
>
> This will let people know that they were using the
> option incorrectly (as depth<0 should be simply invalid,
> and under the hood depth==0 didn't mean 'no depth' or
> 'no history' but 'full depth' instead).

My initial knee-jerk reaction was: doesn't this change break existing use to unplug a shallow repository and bring it to a repository with an unshallow one to disallow depth=0, though?

I somehow thought that the code supports unshallowing with --depth=0 even though since 4dcb167f (fetch: add --unshallow for turning shallow repo into complete one, 2013-01-11), the officially supported way to tell Git to unshallow is with that option.

But apparently that is not the case; I do not think depth==0 meant 'full depth' (i.e. "git fetch --depth=0" did not unshallow); it was simply ignored in fetch_pack.c::find_common() and friends.

So I think it should be a safe change to disallow non-positive depth like this patch does, but the proposed commit log message may need polishing.

Thanks.

Show 19 quoted lines

> Signed-off-by: Andres G. Aragoneses <knocte@gmail.com>
> ---
>  transport.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/transport.c b/transport.c
> index 7202b77..edd63eb 100644
> --- a/transport.c
> +++ b/transport.c
> @@ -483,6 +483,8 @@ static int set_git_option(struct
> git_transport_options *opts,
>  			opts->depth = strtol(value, &end, 0);
>  			if (*end)
>  				die("transport: invalid depth option '%s'", value);
> +			if (opts->depth < 1)
> +				die("transport: invalid depth option '%s' (non positive)", value);
>  		}
>  		return 0;
>  	}