I understand and agree to the change up to this part, but I do
not necessarily agree with what follows:
To me, if the user explicitly says --diff-filter or -S, it seems
more natural to interpret that the user wanted _some_ sort of
diff. Now, there are people who say raw format is anti-human,
which I consider is a valid view, but I think it is better than
NO_OUTPUT in that case.
I wonder if doing something like this instead makes more sense
perhaps?
-- >8 --
diff --git a/builtin-log.c b/builtin-log.c
index 69f2911..e68bfad 100644
--- a/builtin-log.c
+++ b/builtin-log.c
@@ -23,6 +23,35 @@ static int cmd_log_wc(int argc, const ch
if (argc > 1)
die("unrecognized argument: %s", argv[1]);
+ if (rev->always_show_header) {
+ /* Log command is primarily about the message for human
+ * consumption, but if the user asks for any diff, it
+ * is human unfriendly to give the raw diff.
+ *
+ * Show command is the same way, but there the default is
+ * always give patch output, so this does not trigger.
+ */
+ if (rev->diffopt.output_format == DIFF_FORMAT_RAW) {
+ if (rev->diffopt.pickaxe)
+ rev->diffopt.output_format = DIFF_FORMAT_PATCH;
+ else {
+ rev->diffopt.output_format = DIFF_FORMAT_DIFFSTAT;
+ rev->diffopt.summary = 1;
+ }
+ }
+
+ /* If the user is limiting the commits to the ones
+ * that have particular classes of diff, we should not
+ * show the log message for irrelevant ones.
+ *
+ * git show --diff-filter=R -M --all can be used to view
+ * the branch tips that renames something. I do not know
+ * how useful that is, though.
+ */
+ if (rev->diffopt.pickaxe || rev->diffopt.filter)
+ rev->always_show_header = 0;
+ }
+
prepare_revision_walk(rev);
setup_pager();
while ((commit = get_revision(rev)) != NULL) {
-
To unsubscribe from this list: send the line "unsubscribe git" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html