From: Jonathan Nieder <jrnieder@gmail.com>
Date: Thu, 15 May 2014 00:27:38 GMT
Subject: Re: [PATCH v6 11/42] tag.c: use ref transactions when doing updates
Message-ID: <20140515002738.GH9218@google.com>
In-Reply-To: <1398976662-6962-12-git-send-email-sahlberg@google.com>

Ronnie Sahlberg wrote:

> --- a/builtin/tag.c
> +++ b/builtin/tag.c
> @@ -701,11 +702,12 @@ int cmd_tag(int argc, const char **argv, const char *prefix)
>  	if (annotate)
>  		create_tag(object, tag, &buf, &opt, prev, object);
>  
> -	lock = lock_any_ref_for_update(ref.buf, prev, 0, NULL);
> -	if (!lock)
> -		die(_("%s: cannot lock the ref"), ref.buf);
> -	if (write_ref_sha1(lock, object, NULL) < 0)
> -		die(_("%s: cannot update the ref"), ref.buf);
> +	transaction = ref_transaction_begin();
> +	if (!transaction ||
> +	    ref_transaction_update(transaction, ref.buf, object, prev,
> +				   0, !is_null_sha1(prev)) ||
> +	    ref_transaction_commit(transaction, NULL, &err))
> +		die(_("%s: cannot update the ref: %s"), ref.buf, err.buf);

Makes sense for the _update and _commit case.  (BTW, why is have_old
a separate boolean instead of a bit in flags?)

For the _begin() case, can ref_transaction_begin() ever fail?  xcalloc
die()s on allocation failure.  So I think it's fine to assume
transaction is non-null (i.e., drop the !transaction condition), or if
you want to be defensive, then label it as a bug --- e.g.:

	if (!transaction)
		die("BUG: ref_transaction_begin() returned NULL?");

Otherwise if ref_transaction_begin regresses in the future and this
case is tripped then the message would be

	fatal: refs/tags/v1.0: cannot update the ref:

which is not as obvious an indicator that the user should contact
the mailing list.

Thanks,
Jonathan