From: Torsten Bögershausen <tboegi@web.de>
Date: Mon, 15 Aug 2022 06:34:41 GMT
Subject: Re: [PATCH/RFC 1/1] diff.c: When appropriate, use utf8_strwidth()
Message-ID: <20220815063441.uxrtdqsggmrqxxl2@tb-raspi4>
In-Reply-To: <xmqqfshy773n.fsf@gitster.g>

On Sun, Aug 14, 2022 at 04:12:12PM -0700, Junio C Hamano wrote:
> tboegi@web.de writes:
>
> > The choosen solution is to split code in diff.c like this
> >
> > strbuf_addf(&out, "%-*s", len, name);
> >
> > into 2 calls, like this:
> >
> > strbuf_addf(&out, "%s", name);
> > if (len > utf8_strwidth(name))
> >     strbuf_addchars(&out, ' ', len - utf8_strwidth(name));
>
> Makes sense.  Is utf8_strwidth(name) cheap enough that we can call
> it twice in a row on the same string casually, or should we avoid it
> with a new variable?
>
> It might be worth doing a helper function, even?
>
> 	static inline strbuf_pad(struct strbuf *out, const char *s, size_t width)
> 	{
> 		size_t w = utf8_strwidth(s);
>
> 		strbuf_addstr(out, s);
> 		if (w < width)
> 			strbuf_addchars(out, ' ', width - w);
> 	}
>
> Other than that, sounds very sensible.
>

Thanks for the review.

Actually, the commit message is wrong - after writing it, the code
was changed into

if (len > utf8_strwidth(name))
        num_padding_spaces = len - utf8_strwidth(name);

and later

if (num_padding_spaces)
	strbuf_addchars(&out, ' ', num_padding_spaces);

(And having written this, there is probably room for test cases,
IOW: a V2 will come the next days)