Re: [PATCH 0/11] annotating unused function parameters

From: Junio C Hamano <gitster@pobox.com>
Date: Aug 26, 2022, 16:37 UTC
Message-ID: <xmqqpmgnyn9e.fsf@gitster.g>
In-Reply-To: <Ywhx6LLe6YcS/2xf@coredump.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 6 quoted lines

> No, but the farther away you go from the edit-compile-run cycle, the
> more painful warnings become. Catching them immediately and fully has
> real value, as it means the cost of correcting them is lower. So all
> things being equal, I think we should prefer universal solutions when
> they're available (and for example compiler errors over say, coccinelle
> or other analysis tools).

Thanks for saying it so succinctly. Making compiler errors less useful only to please Coccinelle is putting our priority wrong.

> Ugh. Yeah, that is really unfortunate. I much prefer the parenthesized
> syntax, but if we can't find a way to unconfuse third-party parsing,
> then switching is probably the least-bad solution.

Or have third-party fix the parsing ;-) Until then, perhaps we have to live with a suboptimal syntax.

Thanks.