On Mon, Dec 14, 2015 at 04:32:39PM -0500, Jeff King wrote:
> The intent here makes sense to me, and with the exception of the
> test_line_count thing that Torsten mentioned, the code looks good.
> 
> I briefly wondered if the option should simply be "--diffable" or
> something like that, and trigger this new behavior as well as implying
> --no-signature. Along with any other relevant options (if any; I don't
> recall if --stat-width is terminal-dependent for format-patch, for
> example).
> 
> But that is probably overkill. People can flip those switches
> individually if they want to (and even if somebody did want
> "--diffable", it may make sense to build it on top, so they can flip the
> zero-commit thing individually if they want).

That does sound like a potentially worthwhile thing to build on top at
some point.

I'll reroll with the other suggested changes and a slight tweak to make
the tests less dependent on the history in both cases.
-- 
brian m. carlson / brian with sandals: Houston, Texas, US
+1 832 623 2791 | https://www.crustytoothpaste.net/~bmc | My opinion only
OpenPGP: RSA v4 4096b: 88AC E9B2 9196 305B A994 7552 F1BA 225C 0223 B187