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

[PATCH v5 4/4] sha1_name: minimize OID comparisons during disambiguation

From
Derrick Stolee <dstolee@microsoft.com>
Date
Oct 12, 2017, 12:02 UTC
Message-ID
<20171012120220.226427-5-dstolee@microsoft.com>
In-Reply-To
<61168095-d392-39d2-ba65-823525239b5c@gmail.com>
Minimize OID comparisons during disambiguation of packfile OIDs.

Teach git to use binary search with the full OID to find the object's position (or insertion position, if not present) in the pack-index. The object before and immediately after (or the one at the insertion position) give the maximum common prefix. No subsequent linear search is required.

Take care of which two to inspect, in case the object id exists in the packfile.

If the input to find_unique_abbrev_r() is a partial prefix, then the OID used for the binary search is padded with zeroes so the object will not exist in the repo (with high probability) and the same logic applies.

This commit completes a series of three changes to OID abbreviation code, and the overall change can be seen using standard commands for large repos. Below we report performance statistics for perf test 4211.6 from p4211-line-log.sh using three copies of the Linux repo:

| Packs | Loose  | HEAD~3   | HEAD     | Rel%  |
|-------|--------|----------|----------|-------|
|  1    |      0 |  41.27 s |  38.93 s | -4.8% |
| 24    |      0 |  98.04 s |  91.35 s | -5.7% |
| 23    | 323952 | 117.78 s | 112.18 s | -4.8% |
Signed-off-by: Derrick Stolee <dstolee@microsoft.com>
---
 sha1_name.c | 76 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++----
 1 file changed, 71 insertions(+), 5 deletions(-)
diff --git a/sha1_name.c b/sha1_name.c
index fdd2711a6..7fd5f5f71 100644
--- a/sha1_name.c
+++ b/sha1_name.c
@@ -153,7 +153,9 @@ static void unique_in_pack(struct packed_git *p,
 	uint32_t num, last, i, first = 0;
 	const struct object_id *current = NULL;
 
-	open_pack_index(p);
+	if (open_pack_index(p) || !p->num_objects)
+		return;
+
 	num = p->num_objects;
 	last = num;
 	while (first < last) {
@@ -478,6 +480,7 @@ struct min_abbrev_data {
 	unsigned int init_len;
 	unsigned int cur_len;
 	char *hex;
+	const unsigned char *hash;
 };
 
 static inline char get_hex_char_from_oid(const struct object_id *oid,
@@ -505,6 +508,67 @@ static int extend_abbrev_len(const struct object_id *oid, void *cb_data)
 	return 0;
 }
 
+static void find_abbrev_len_for_pack(struct packed_git *p,
+				     struct min_abbrev_data *mad)
+{
+	int match = 0;
+	uint32_t num, last, first = 0;
+	struct object_id oid;
+
+	if (open_pack_index(p) || !p->num_objects)
+		return;
+
+	num = p->num_objects;
+	last = num;
+	while (first < last) {
+		uint32_t mid = first + (last - first) / 2;
+		const unsigned char *current;
+		int cmp;
+
+		current = nth_packed_object_sha1(p, mid);
+		cmp = hashcmp(mad->hash, current);
+		if (!cmp) {
+			match = 1;
+			first = mid;
+			break;
+		}
+		if (cmp > 0) {
+			first = mid + 1;
+			continue;
+		}
+		last = mid;
+	}
+
+	/*
+	 * first is now the position in the packfile where we would insert
+	 * mad->hash if it does not exist (or the position of mad->hash if
+	 * it does exist). Hence, we consider a maximum of three objects
+	 * nearby for the abbreviation length.
+	 */
+	mad->init_len = 0;
+	if (!match) {
+		nth_packed_object_oid(&oid, p, first);
+		extend_abbrev_len(&oid, mad);
+	} else if (first < num - 1) {
+		nth_packed_object_oid(&oid, p, first + 1);
+		extend_abbrev_len(&oid, mad);
+	}
+	if (first > 0) {
+		nth_packed_object_oid(&oid, p, first - 1);
+		extend_abbrev_len(&oid, mad);
+	}
+	mad->init_len = mad->cur_len;
+}
+
+static void find_abbrev_len_packed(struct min_abbrev_data *mad)
+{
+	struct packed_git *p;
+
+	prepare_packed_git();
+	for (p = packed_git; p; p = p->next)
+		find_abbrev_len_for_pack(p, mad);
+}
+
 int find_unique_abbrev_r(char *hex, const unsigned char *sha1, int len)
 {
 	struct disambiguate_state ds;
@@ -536,19 +600,21 @@ int find_unique_abbrev_r(char *hex, const unsigned char *sha1, int len)
 	if (len == GIT_SHA1_HEXSZ || !len)
 		return GIT_SHA1_HEXSZ;
 
-	if (init_object_disambiguation(hex, len, &ds) < 0)
-		return -1;
-
 	mad.init_len = len;
 	mad.cur_len = len;
 	mad.hex = hex;
+	mad.hash = sha1;
+
+	find_abbrev_len_packed(&mad);
+
+	if (init_object_disambiguation(hex, mad.cur_len, &ds) < 0)
+		return -1;
 
 	ds.fn = extend_abbrev_len;
 	ds.always_call_fn = 1;
 	ds.cb_data = (void*)&mad;
 
 	find_short_object_filename(&ds);
-	find_short_packed_object(&ds);
 	(void)finish_object_disambiguation(&ds, &oid_ret);
 
 	hex[mad.cur_len] = 0;
-- 
2.14.1.538.g56ec8fc98.dirty
Previous: Derrick StoleeNext: Derrick Stolee
Message 21 of 22 in “Improve abbreviation disambiguation”
  1. 0/4 Improve abbreviation disambiguationDerrick Stolee, Oct 8, 2017
  2. 1/4 p4211-line-log.sh: add log --online --raw --parents perf testDerrick Stolee, Oct 8, 2017
  3. 3/4 sha1_name: parse less while finding common prefixDerrick Stolee, Oct 8, 2017
  4. Jeff KingOct 9, 2017
  5. 4/4 sha1_name: minimize OID comparisons during disambiguationDerrick Stolee, Oct 8, 2017
  6. Jeff KingOct 9, 2017
  7. Derrick StoleeOct 10, 2017
  8. Jeff KingOct 10, 2017
  9. Junio C HamanoOct 10, 2017
  10. Jeff KingOct 10, 2017
  11. Derrick StoleeOct 10, 2017
  12. Jeff KingOct 10, 2017
  13. Derrick StoleeOct 11, 2017
  14. 0/4 Improve abbreviation disambiguationDerrick Stolee, Oct 12, 2017
  15. Derrick StoleeOct 12, 2017
  16. Junio C HamanoOct 12, 2017
  17. Jeff KingOct 12, 2017
  18. 3/4 sha1_name: parse less while finding common prefixDerrick Stolee, Oct 12, 2017
  19. 2/4 sha1_name: unroll len loop in find_unique_abbrev_rDerrick Stolee, Oct 12, 2017
  20. 1/4 p4211-line-log.sh: add log --online --raw --parents perf testDerrick Stolee, Oct 12, 2017
  21. 4/4 sha1_name: minimize OID comparisons during disambiguationDerrick Stolee, Oct 12, 2017
  22. 2/4 sha1_name: unroll len loop in find_unique_abbrev_rDerrick Stolee, Oct 8, 2017

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.