From: Junio C Hamano <gitster@pobox.com>
Date: Fri, 27 Feb 2026 23:28:09 GMT
Subject: Re: [PATCH 2/2] parseopt: check for duplicate long names and numerical options
Message-ID: <xmqqpl5pojg6.fsf@gitster.g>
In-Reply-To: <20260227230822.GA2965111@coredump.intra.peff.net>

Jeff King <peff@peff.net> writes:

> Doing this:
>
> diff --git a/parse-options.c b/parse-options.c
> index 51b72eee11..f056a4471e 100644
> --- a/parse-options.c
> +++ b/parse-options.c
> @@ -659,9 +659,8 @@ static void parse_options_check(const struct option *opts)
>  				optbug(opts, "short name already used");
>  		}
>  		if (opts->long_name) {
> -			if (strset_contains(&long_names, opts->long_name))
> +			if (!strset_add(&long_names, opts->long_name))
>  				optbug(opts, "long name already used");
> -			strset_add(&long_names, opts->long_name);
>  		}
>  		if (opts->type == OPTION_NUMBER) {
>  			if (saw_number_option)
>
> seems to shave off ~1% of my benchmark. Not that exciting, but hey, it's
> one line shorter to boot.

Yeah, it is the right thing not to hash the same thing twice which
is totally unnecessary.