René Scharfe <rene.scharfe@xxxxxxxxxxxxxx> writes: >> Symbolic constants for the magic values 0, 1, and 2 would make the >> code more readable. > > Yes, that was a bit too complicated. I shuffled the code around a bit, > so the patch is now a bit smaller and avoids introducing value 2 for > show_hunk_mark. Better? I guess. I still cannot shake off this nagging feeling that the complication mainly comes from the way threaded support was written as if the control flow is the same as that of the single-threaded implementation (e.g. grep_object() and friends are called from a one-path-at-a-time loop and pretend as if they have returned some meaningful results indicating they have found a match), and the logic may become more clear if the flow was restructured. But I haven't tried, so I dunno. Your patch does not make the situation worse, and looks like the right thing to do at the moment. -- 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