Re: [PATCH v6 2/2] config: allow giving separate author and committer idents

From: Junio C Hamano <gitster@pobox.com>
Date: Feb 6, 2019, 22:43 UTC
Message-ID: <xmqqo97opmg8.fsf@gitster-ct.c.googlers.com>
In-Reply-To: <20190206182612.GA10231@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

Show 28 quoted lines

> An alternative view is that anybody who calls git_author_info() to
> create a commit _should_ be checking author_ident_sufficiently_given(),
> and it's a bug that they're not.
>
> I.e., should we be doing something like this (and probably some other
> spots, too):
>
> diff --git a/commit.c b/commit.c
> index a5333c7ac6..c99b311a48 100644
> --- a/commit.c
> +++ b/commit.c
> @@ -1419,8 +1419,11 @@ int commit_tree_extended(const char *msg, size_t msg_len,
>  	}
>  
>  	/* Person/date information */
> -	if (!author)
> +	if (!author) {
>  		author = git_author_info(IDENT_STRICT);
> +		if (!author_ident_sufficiently_given())
> +			warning("your author ident was auto-detected, etc...");
> +	}
>  	strbuf_addf(&buffer, "author %s\n", author);
>  	strbuf_addf(&buffer, "committer %s\n", git_committer_info(IDENT_STRICT));
>  	if (!encoding_is_utf8)
>
> I dunno. It seems pretty low priority, and nobody has even noticed after
> all these years. So I'm not sure if it's worth spending too much time on
> it.

That's quite tempting. But I agree that this is something we can leave for a later clean-up, as the topic to add the config variables is pretty much orthogonal to it, and we are not making things that much worse than the status quo.

Thanks.