[PATCH 1/5] diff_tree_sha1: skip diff_tree if old == new

From: Dan McGee <dpmcgee@gmail.com>
Date: Mar 31, 2011, 01:37 UTC
Message-ID: <1301535481-1085-1-git-send-email-dpmcgee@gmail.com>

This was seen to happen in some invocations of git-log with a filtered path. Only do it if we are not recursively descending, as otherwise we mess with copy and rename detection in full tree moves.

We need to exclude times when FIND_COPIES_HARDER is enabled as the diff machinery requires traversal of all files in this case to find copies from files that did not change at all in the given revision.

Signed-off-by: Dan McGee <dpmcgee@gmail.com>
---

These next few patches are all derived from trying to make operations on this
repository a bit faster:
    http://projects.archlinux.org/svntogit/packages.git/

As far as a different shape of repository, this one qualifies. Some stats from after running `git gc --prune` and looking at things from count-objects among others:

                    arch-packages   linux-2.6
in-pack:                   496718     1979274
size-pack:                 112582      513977
ls | wc -l:                  2401          34
ls-tree -r | wc -l          14039       36706
ls-tree -r -t | grep ' tree ' | wc -l
                            11211        2256
time git log -- zzzzz_not_exist >/dev/null
                          35.558s      0.976s

I didn't find some golden switch to turn that made things instantly fast, but these patches did at least give some speedups. Suggestions/feedback/etc. welcome. Most profiling was done with `valgrind --tool=callgrind` and then using kcachegrind to chase down slow spots.

 tree-diff.c |    3 +++
 1 files changed, 3 insertions(+), 0 deletions(-)

diff --git a/tree-diff.c b/tree-diff.c
index 76f83fc..ab90f1a 100644
--- a/tree-diff.c
+++ b/tree-diff.c
@@ -286,6 +286,9 @@ int diff_tree_sha1(const unsigned char *old, const unsigned char *new, const cha
 	unsigned long size1, size2;
 	int retval;
 
+	if (!DIFF_OPT_TST(opt, FIND_COPIES_HARDER) && !hashcmp(old, new))
+		return 0;
+
 	tree1 = read_object_with_reference(old, tree_type, &size1, NULL);
 	if (!tree1)
 		die("unable to read source tree (%s)", sha1_to_hex(old));

-- 
1.7.4.2