- Notifications
You must be signed in to change notification settings - Fork28
Commit489e431
committed
Remove obsolete nbtree insertion comment.
Remove a Berkeley-era comment above _bt_insertonpg() that admonishes thereader to grok Lehman and Yao's paper before making any changes. Thismade a certain amount of sense back when _bt_insertonpg() wasresponsible for most of the things that are now spread across_bt_insertonpg(), _bt_findinsertloc(), _bt_insert_parent(), and_bt_split(), but it doesn't work like that anymore.I believe that this comment alludes to the need to "couple" or "crab"buffer locks as we ascend the tree as page splits cascade upwards. Thenbtree README already explains this in detail, which seems sufficient.Besides, the changes to page splits made by commit40dae7e alteredthe exact details of how buffer locks are retained during splits; Lehmanand Yao's original algorithm seems to release the lock on the left childpage/buffer slightly earlier than _bt_insertonpg()/_bt_insert_parent()can.1 parent8a0f0ad commit489e431
1 file changed
+0
-3
lines changedLines changed: 0 additions & 3 deletions
| Original file line number | Diff line number | Diff line change | |
|---|---|---|---|
| |||
926 | 926 | | |
927 | 927 | | |
928 | 928 | | |
929 | | - | |
930 | | - | |
931 | | - | |
932 | 929 | | |
933 | 930 | | |
934 | 931 | | |
| |||
0 commit comments
Comments
(0)