From: Junio C Hamano <gitster@pobox.com>
Date: Sat, 14 Feb 2026 00:42:46 GMT
Subject: Re: [PATCH v2 04/13] config: format int64s gently
Message-ID: <xmqq8qcwi29l.fsf@gitster.g>
In-Reply-To: <2bca4d231686e33ea9d4d85b10fcffd60a63ad46.1771026918.git.gitgitgadget@gmail.com>

"Derrick Stolee via GitGitGadget" <gitgitgadget@gmail.com> writes:

> +static int format_config_int64(struct strbuf *buf,
> +			       const char *key_,
> +			       const char *value_,
> +			       const struct key_value_info *kvi,
> +			       int gently)
> +{
> +	int64_t v = 0;
> +	if (gently) {
> +		if (git_parse_int64(value_, &v))
> +			return -1;
> +	} else {
> +		/* may die() */
> +		v = git_config_int64(key_, value_ ? value_ : "", kvi);
> +	}
> +
> +	strbuf_addf(buf, "%"PRId64, v);
> +	return 0;
> +}

This establishes the pattern the next handful of patches follow.  We
already have in parse.c helpers that we can use for the gentler
parsing, and otherwise we'd use git_config_*() that the caller of
these new helpers were using originally.

I'd have preferred to have the blank line moved to the gap between
the decl and the first statement, i.e.,

> +{
> +	int64_t v = 0;
> +
> +	if (gently) {
> +		if (git_parse_int64(value_, &v))
> +			return -1;
> +	} else {
> +		/* may die() */
> +		v = git_config_int64(key_, value_ ? value_ : "", kvi);
> +	}
> +	strbuf_addf(buf, "%"PRId64, v);
> +	return 0;
> +}

These "format X gently" steps look very good.