summaryrefslogtreecommitdiff
path: root/sha1-lookup.c
diff options
context:
space:
mode:
authorJunio C Hamano <gitster@pobox.com>2007-12-30 11:13:27 (GMT)
committerJunio C Hamano <gitster@pobox.com>2008-04-09 08:30:18 (GMT)
commit12ecb01107c4e77d3bccb5be5a0230c4546dafaf (patch)
tree022b8c212b7bd7cbd5bd7636f9d8ab8f04646960 /sha1-lookup.c
parent628522ec1439f414dcb1e71e300eb84a37ad1af9 (diff)
downloadgit-12ecb01107c4e77d3bccb5be5a0230c4546dafaf.zip
git-12ecb01107c4e77d3bccb5be5a0230c4546dafaf.tar.gz
git-12ecb01107c4e77d3bccb5be5a0230c4546dafaf.tar.bz2
sha1-lookup: make selection of 'middle' less aggressive
If we pick 'mi' between 'lo' and 'hi' at 50%, which was what the simple binary search did, we are halving the search space whether the entry at 'mi' is lower or higher than the target. The previous patch was about picking not the middle but closer to 'hi', when we know the target is a lot closer to 'hi' than it is to 'lo'. However, if it turns out that the entry at 'mi' is higher than the target, we would end up reducing the search space only by the difference between 'mi' and 'hi' (which by definition is less than 50% --- that was the whole point of not using the simple binary search), which made the search less efficient. And the risk of overshooting becomes very high, if we try to be too precise. This tweaks the selection of 'mi' to be a bit closer to the middle than we would otherwise pick to avoid the problem. Signed-off-by: Junio C Hamano <gitster@pobox.com>
Diffstat (limited to 'sha1-lookup.c')
-rw-r--r--sha1-lookup.c33
1 files changed, 26 insertions, 7 deletions
diff --git a/sha1-lookup.c b/sha1-lookup.c
index 4faa638..da35747 100644
--- a/sha1-lookup.c
+++ b/sha1-lookup.c
@@ -50,6 +50,12 @@
* the midway of the table. It can reasonably be expected to be near
* 87% (222/256) from the top of the table.
*
+ * However, we do not want to pick "mi" too precisely. If the entry at
+ * the 87% in the above example turns out to be higher than the target
+ * we are looking for, we would end up narrowing the search space down
+ * only by 13%, instead of 50% we would get if we did a simple binary
+ * search. So we would want to hedge our bets by being less aggressive.
+ *
* The table at "table" holds at least "nr" entries of "elem_size"
* bytes each. Each entry has the SHA-1 key at "key_offset". The
* table is sorted by the SHA-1 key of the entries. The caller wants
@@ -119,11 +125,25 @@ int sha1_entry_pos(const void *table,
if (hiv < kyv)
return -1 - hi;
- if (kyv == lov && lov < hiv - 1)
- kyv++;
- else if (kyv == hiv - 1 && lov < kyv)
- kyv--;
-
+ /*
+ * Even if we know the target is much closer to 'hi'
+ * than 'lo', if we pick too precisely and overshoot
+ * (e.g. when we know 'mi' is closer to 'hi' than to
+ * 'lo', pick 'mi' that is higher than the target), we
+ * end up narrowing the search space by a smaller
+ * amount (i.e. the distance between 'mi' and 'hi')
+ * than what we would have (i.e. about half of 'lo'
+ * and 'hi'). Hedge our bets to pick 'mi' less
+ * aggressively, i.e. make 'mi' a bit closer to the
+ * middle than we would otherwise pick.
+ */
+ kyv = (kyv * 6 + lov + hiv) / 8;
+ if (lov < hiv - 1) {
+ if (kyv == lov)
+ kyv++;
+ else if (kyv == hiv)
+ kyv--;
+ }
mi = (range - 1) * (kyv - lov) / (hiv - lov) + lo;
if (debug_lookup) {
@@ -142,8 +162,7 @@ int sha1_entry_pos(const void *table,
if (cmp > 0) {
hi = mi;
hi_key = mi_key;
- }
- else {
+ } else {
lo = mi + 1;
lo_key = mi_key + elem_size;
}