git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[RFC/PATCH 1/4] tag: speed up --contains calculation

From
Jeff King <peff@peff.net>
Date
Jul 5, 2010, 12:33 UTC
Message-ID
<20100705123335.GA25699@sigill.intra.peff.net>
In-Reply-To
<20100705122723.GB21146@sigill.intra.peff.net>

When we want to know if commit A contains commit B (or any one of a set of commits, B through Z), we generally calculate the merge bases and see if B is a merge base of A (or for a set, if any of the commits B through Z have that property).

When we are going to check a series of commits A1 through An to see whether each contains B (e.g., because we are deciding which tags to show with "git tag --contains"), we do a series of merge base calculations. This can be very expensive, as we repeat a lot of traversal work.

Instead, let's leverage the fact that we are going to use the same --contains list for each tag, and mark areas of the commit graph is definitely containing those commits, or definitely not containing those commits. Later tags can then stop traversing as soon as they see a previously calculated answer.

This sped up "git tag --contains HEAD~200" in the linux-2.6 repository from:

  real    0m15.417s
  user    0m15.197s
  sys     0m0.220s
to:
  real    0m5.329s
  user    0m5.144s
  sys     0m0.184s
Signed-off-by: Jeff King <peff@peff.net>
---
Note that this is a depth first search, whereas we generally traverse in
a more breadth-first way. So it can actually make things slightly slower
than the current merge-base code, if:
  1. You don't have any merge base calculations that involve going very
     far back in history.
  2. We depth-first down the wrong side of a merge.

However, in the usual cases, I think it will perform much better. Comparing an old tag with a recent commit means we have to look at a lot of the commit graph, anyway.

If anybody has a clever idea for making it breadth-first, I would be happy to hear it.

Other caveats:
  1. This can probably be a one-liner change to use in "git branch
     --contains", as well. I didn't measure it, as that already tends to
     be reasonably fast.
  2. It uses commit marks, which means it doesn't behave well with other
     traversals. I have no idea if I should be using alternate marks or
     not.
  3. We never clear the marks, or check them against the "want" list. So
     we just assume that repeated calls to contains will have the same
     "want" list.
 builtin/tag.c |    2 +-
 commit.c      |   42 ++++++++++++++++++++++++++++++++++++++++++
 commit.h      |    2 ++
 3 files changed, 45 insertions(+), 1 deletions(-)
diff --git a/builtin/tag.c b/builtin/tag.c
index d311491..c200e1e 100644
--- a/builtin/tag.c
+++ b/builtin/tag.c
@@ -49,7 +49,7 @@ static int show_reference(const char *refname, const unsigned char *sha1,
 			commit = lookup_commit_reference_gently(sha1, 1);
 			if (!commit)
 				return 0;
-			if (!is_descendant_of(commit, filter->with_commit))
+			if (!contains(commit, filter->with_commit))
 				return 0;
 		}
 
diff --git a/commit.c b/commit.c
index e9b0750..20354c6 100644
--- a/commit.c
+++ b/commit.c
@@ -845,3 +845,45 @@ int commit_tree(const char *msg, unsigned char *tree,
 	strbuf_release(&buffer);
 	return result;
 }
+
+static int in_commit_list(const struct commit_list *want, struct commit *c)
+{
+	for (; want; want = want->next)
+		if (!hashcmp(want->item->object.sha1, c->object.sha1))
+			return 1;
+	return 0;
+}
+
+static int contains_recurse(struct commit *candidate,
+			    const struct commit_list *want)
+{
+	struct commit_list *p;
+
+	/* was it previously marked as containing a want commit? */
+	if (candidate->object.flags & TMP_MARK)
+		return 1;
+	/* or marked as not possibly containing a want commit? */
+	if (candidate->object.flags & UNINTERESTING)
+		return 0;
+	/* or are we it? */
+	if (in_commit_list(want, candidate))
+		return 1;
+
+	if (parse_commit(candidate) < 0)
+		return 0;
+
+	/* Otherwise recurse and mark ourselves for future traversals. */
+	for (p = candidate->parents; p; p = p->next) {
+		if (contains_recurse(p->item, want)) {
+			candidate->object.flags |= TMP_MARK;
+			return 1;
+		}
+	}
+	candidate->object.flags |= UNINTERESTING;
+	return 0;
+}
+
+int contains(struct commit *candidate, const struct commit_list *want)
+{
+	return contains_recurse(candidate, want);
+}
diff --git a/commit.h b/commit.h
index eb2b8ac..1a7299e 100644
--- a/commit.h
+++ b/commit.h
@@ -153,6 +153,8 @@ extern struct commit_list *get_shallow_commits(struct object_array *heads,
 int is_descendant_of(struct commit *, struct commit_list *);
 int in_merge_bases(struct commit *, struct commit **, int);
 
+int contains(struct commit *, const struct commit_list *);
+
 extern int interactive_add(int argc, const char **argv, const char *prefix);
 extern int run_add_interactive(const char *revision, const char *patch_mode,
 			       const char **pathspec);
-- 
1.7.2.rc1.209.g2a36c
Previous: Jeff KingNext: Jonathan Nieder
Message 12 of 46 in “Why is "git tag --contains" so slow?”
  1. Theodore Ts'oJul 1, 2010
  2. Shawn O. PearceJul 1, 2010
  3. Sam VilainJul 3, 2010
  4. Avery PennarunJul 1, 2010
  5. tytso@mit.eduJul 1, 2010
  6. Jeff KingJul 1, 2010
  7. Jeff KingJul 1, 2010
  8. tytso@mit.eduJul 2, 2010
  9. Jeff KingJul 3, 2010
  10. tytso@mit.eduJul 4, 2010
  11. Jeff KingJul 5, 2010
  12. 1/4 tag: speed up --contains calculationJeff King, Jul 5, 2010
  13. Jonathan NiederOct 13, 2010
  14. Clemens BuchacherOct 13, 2010
  15. Ævar Arnfjörð BjarmasonFeb 23, 2011
  16. Jeff KingFeb 23, 2011
  17. 2/4 limit "contains" traversals based on commit timestampJeff King, Jul 5, 2010
  18. Jonathan NiederOct 13, 2010
  19. 3/4 default core.clockskew variable to one dayJeff King, Jul 5, 2010
  20. 4/4 name-rev: respect core.clockskewJeff King, Jul 5, 2010
  21. Jeff KingJul 5, 2010
  22. Jonathan NiederOct 14, 2010
  23. Clemens BuchacherOct 16, 2010
  24. Jeff KingOct 27, 2010
  25. Clemens BuchacherOct 28, 2010
  26. tytso@mit.eduJul 5, 2010
  27. Jeff KingJul 6, 2010
  28. Will PalmerJul 6, 2010
  29. tytso@mit.eduJul 6, 2010
  30. Jeff KingJul 8, 2010
  31. Will PalmerJul 8, 2010
  32. tytso@mit.eduJul 8, 2010
  33. Jeff KingJul 7, 2010
  34. Theodore TsoJul 8, 2010
  35. Jakub NarebskiJul 8, 2010
  36. Nicolas PitreJul 8, 2010
  37. Avery PennarunJul 8, 2010
  38. Nicolas PitreJul 8, 2010
  39. Jakub NarebskiJul 8, 2010
  40. Sverre RabbelierJul 8, 2010
  41. Nicolas PitreJul 8, 2010
  42. Nicolas PitreJul 8, 2010
  43. Jeff KingJul 8, 2010
  44. Johan HerlandJul 8, 2010
  45. Nicolas PitreJul 8, 2010
  46. Jeff KingJul 7, 2010

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.