Hi Junio, On Mon, 20 Jun 2016, Junio C Hamano wrote: > You can probably sell "when giving out put to file, we will never color > the output" as an improved new world order, but if that is what this > change wants to do, it probably deserves a separate patch. > > I however think you can avoid breaking expectations by people who are > not so sensible by overriding only when use_color is set to > GIT_COLOR_AUTO, perhaps? That is a very convincing argument. So convincing that I wanted to change the patch to guard behind `diff_use_color_default == GIT_COLOR_AUTO`. But that is the wrong variable: the variable that *has* that default value is git_use_color_default, and is private to color.c. But then I dug further to determine under which circumstances that variable can be reset to any different value. It turns out that in format-patch's case, it cannot: 787570c (format-patch: ignore ui.color, 2011-09-13) I hope you agree that it will be enough to augment the commit message with this analysis and keep the patch as-is (as per v2, that is)? Ciao, Dscho -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html