Re: [PATCH 3/3] format-patch: preserve subject newlines with -k

From: Junio C Hamano <gitster@pobox.com>
Date: May 26, 2011, 21:18 UTC
Message-ID: <7vboypw47q.fsf@alter.siamese.dyndns.org>
In-Reply-To: <20110526205504.GC31340@sigill.intra.peff.net>

Jeff King <peff@peff.net> writes:

> We have always treated multi-line subjects as second-class citizens, so
> this is not a must-have patch. But I think it makes sense to do,
> considering how simple it is, and the fact that it makes "format-patch
> -k | am -k" always a no-op, even with multi-line subjects.

I think this is a good thing to have, as the reason why we treated multi-line subjects as second-class citizens is exactly because we didn't try to stuff multiple lines on "Subject:" like you did with this patch.

Show 10 quoted lines

> diff --git a/builtin/log.c b/builtin/log.c
> index d8c6c28..3fdf488 100644
> --- a/builtin/log.c
> +++ b/builtin/log.c
> @@ -768,7 +768,7 @@ static void make_cover_letter(struct rev_info *rev, int use_stdout,
>  	pp_user_info(NULL, CMIT_FMT_EMAIL, &sb, committer, DATE_RFC2822,
>  		     encoding);
>  	pp_title_line(CMIT_FMT_EMAIL, &msg, &sb, subject_start, extra_headers,
> -		      encoding, need_8bit_cte);
> +		      encoding, need_8bit_cte, 0);

I do not appreciate a single-bit tweak as separate parameter to a function. Back when pp_title_line() had only "do we need 8-bit cte", it was Ok, but now that you are adding another bit, could we make it an "unsigned flag"?