Re: [PATCH] send-email: don't cc *-by lines with '-' prefix

From: Jeff King <peff@peff.net>
Date: Apr 4, 2019, 12:14 UTC
Message-ID: <20190404121416.GC22324@sigill.intra.peff.net>
In-Reply-To: <dd8160f8-0e5e-1024-53c1-1a9f23423af5@rasmusvillemoes.dk>

On Thu, Apr 04, 2019 at 11:42:23AM +0200, Rasmus Villemoes wrote:

Show 10 quoted lines

> My ack for Baruch's original patch, which AFAICT is identical with
> Junio's suggestion, still stands. FWIW, I'm against Joe's suggestion of
> stopping at a line matching /^---/, since it's not unlikely somebody
> does something like
> 
> ---- dmesg output ----
> bla bla
> ----
> 
> in the commit message.

Keep in mind that on the receiving end, we are going to stop reading the commit message at a triple-dash, too, which is done as (from mailinfo.c's patchbreak()):

  /^---( [^\s]|\s*$)/

So it might make sense to use the same rule here. That said:

> Since all lines (except for some of the diff header lines) in the patch
> part begin with space, - or +, insisting on a the line starting with a
> letter should be sufficient for excluding any random Foo-by lines that
> may appear in the patch part.

Yeah, I think this mostly makes it a non-issue, unless we care about efficiency (and I doubt it is even measurable).

Technically you could have other cruft after the diff, too. But I think putting "signed-off-by: somebody" in your email sig is a case of "if it hurts, don't do it".

-Peff