Re: [PATCH] use C99 declaration of variable in for() loop

From: brian m. carlson <sandals@crustytoothpaste.net>
Date: Feb 15, 2024, 23:12 UTC
Message-ID: <Zc6abO6RV9RwpABR@tapette.crustytoothpaste.net>
In-Reply-To: <xmqqcysxskd9.fsf@gitster.g>

On 2024-02-15 at 17:33:22, Junio C Hamano wrote:

Show 14 quoted lines

> Elia Pinto <gitter.spiros@gmail.com> writes:
> 
> > With the exception of cbtree.c, which would need initial
> > reworking to remove the usage of goto, it expands the
> > use of variable scope reduction in for loops as
> > permitted by the C99 standard, which was first introduced
> > in the git codebase with commit 44ba10d6.
> 
> Thanks, but ...
> 
> Our test balloon may have proven that nobody will be inconvenienced,
> and it does mean we can be liberal using it when we add new code or
> update existing loops "while at it", but I personally do not think
> such a code churn is very welcome.

I will also say that sending one giant patch for this may be a bit hard to review. While I will defer to Junio's opinion as the maintainer, I would be more inclined to review this kind of series if it came in in smaller patches, a few at a time, in which case I would find it a welcome improvement.

Since my time to work on Git is relatively limited, having, say, a five-patch series where we each update a single file would let me review these changes in a relatively short amount of time, which I would be more likely to be able to find time for. Looking at the large patch, I'd be worried that I wouldn't be able to get through the entire thing in one sitting.

Of course, if you bring in a nice Coccinelle patch for it, then that may make a longer (but still one-file-per-commit) series more viable, since it will help reviewers have more confidence in your change.

-- 
brian m. carlson (he/him or they/them)
Toronto, Ontario, CA