Re: git tag -s: TAG_EDITMSG should not be deleted upon failures

From: Jeff King <peff@peff.net>
Date: Dec 6, 2008, 21:54 UTC
Message-ID: <20081206215400.GA29440@coredump.intra.peff.net>
In-Reply-To: <7v8wqtvvql.fsf@gitster.siamese.dyndns.org>

On Sat, Dec 06, 2008 at 01:28:50PM -0800, Junio C Hamano wrote:

> Thanks.  I love patches that addresses bugs during -rc period.

Well, I'm not sure this was a bug fix versus an improvement, but at least wasn't an all new feature. And it was short enough to look at in one sitting.

Of course, I did still manage to introduce a bug in my 4-line change...;)

Show 5 quoted lines

>     - the "path" variable is uninitialized if we do not start editor at
>       all, so unlink(path) and free(path) have a very high chance of
>       failing.
> 
>       I think you need [Update #1] below squashed in to fix this.

Oops. Yes, that is definitely a problem.

> [Update #1]
> [...]
> -	char *path;
> +	char *path = NULL;

Right, that fix looks good.

Show 6 quoted lines

> +	if (build_tag_object(buf, sign, result) < 0) {
> +		if (path)
> +			fprintf(stderr, "What you edited in your editor is left in %s",
> +				path);
> +		exit(128);
> +	}

Much better, though the message is a bit awkward. How about

  "The tag message has been left in %s"

Do you want me to resend, or do you want to fix up locally?

-Peff