Re: [PATCH v2 03/12] delete_ref(): handle special case more explicitly

[Junio C Hamano <gitster@pobox.com>]

Michael Haggerty <mhagger@alum.mit.edu> writes:

> delete_ref() uses a different convention for its old_sha1 parameter
> than, say, ref_transaction_delete(): NULL_SHA1 means not to check the
> old value. Make this fact a little bit clearer in the code by handling
> it in explicit, commented code rather than burying it in a conditional
> expression.
>
> Signed-off-by: Michael Haggerty <mhagger@alum.mit.edu>
> ---

I think this is a very sensible first step, but I wonder if
delete_ref() should be taught the difference between "we do not
expect anything" and "we expect it not to exist".

>  refs.c | 11 +++++++++--
>  1 file changed, 9 insertions(+), 2 deletions(-)
>
> diff --git a/refs.c b/refs.c
> index b575bb8..f0b6cec 100644
> --- a/refs.c
> +++ b/refs.c
> @@ -2795,10 +2795,17 @@ int delete_ref(const char *refname, const unsigned char *old_sha1,
>  	struct ref_transaction *transaction;
>  	struct strbuf err = STRBUF_INIT;
>  
> +	/*
> +	 * Treat NULL_SHA1 and NULL alike, to mean "we don't care what
> +	 * the old value of the reference was (or even if it didn't
> +	 * exist)":
> +	 */
> +	if (old_sha1 && is_null_sha1(old_sha1))
> +		old_sha1 = NULL;
> +
>  	transaction = ref_transaction_begin(&err);
>  	if (!transaction ||
> -	    ref_transaction_delete(transaction, refname,
> -				   (old_sha1 && !is_null_sha1(old_sha1)) ? old_sha1 : NULL,
> +	    ref_transaction_delete(transaction, refname, old_sha1,
>  				   flags, NULL, &err) ||
>  	    ref_transaction_commit(transaction, &err)) {
>  		error("%s", err.buf);