On Thu, 2011-03-17 at 09:33 +0100, Thomas Egerer wrote: > This patch fixes that the pretty-formats tformat and format ignore > git's color option. It is my understanding that this is intentional, the logic being: If you normally don't want color, but have specified it directly on the command-line, you probably want color. It is arguable that this logic does not hold for format aliases, of course. (and personally I've never agreed with the above logic anyway) iirc, there are a couple of other places beyond log-tree.c which need to propagate COLOR_DIFF into the pretty context if you want to respect the colour option in user-specified formats. Skimming my own diffs: rev-list.c and shortlog.c Of course, explicitly propagating "diff_opts" settings like this just shows off how poorly named these things are if we're going to use them for this sort of thing. That's probably a place for another patch, though. I've got a patch to add support for conditional formats, including such things as: %(opt-color: %Cred%h%Creset, %h). I'm currently in the process of re-rolling just the "long options" part of that into something more-manageable, but honestly that's the bulk of the change, so if that gets accepted then adding %(opt-color...) would be pretty trivial. --Will -- 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