Commit3c7d784

committed

Update outdated comment on WAL-logged locks with invalid XID

1 parent1a43de5 commit3c7d784Copy full SHA for 3c7d784

File tree

-7

lines changed

-7

lines changed

Lines changed: 3 additions & 7 deletions

Original file line number	Diff line number	Diff line change
`@@ -1121,6 +1121,9 @@ StandbyReleaseAllLocks(void)`
`1121`	`1121`	`* StandbyReleaseOldLocks`
`1122`	`1122`	`*Release standby locks held by top-level XIDs that aren't running,`
`1123`	`1123`	`*as long as they're not prepared transactions.`
	`1124`	`+ *`
	`1125`	`+ * This is needed to prune the locks of crashed transactions, which didn't`
	`1126`	`+ * write an ABORT/COMMIT record.`
`1124`	`1127`	`*/`
`1125`	`1128`	`void`
`1126`	`1129`	`StandbyReleaseOldLocks(TransactionIdoldxid)`
`@@ -1266,13 +1269,6 @@ standby_redo(XLogReaderState *record)`
`1266`	`1269`	`* transactions already committed, since those commits raced ahead when`
`1267`	`1270`	`* making WAL entries.`
`1268`	`1271`	`*`
`1269`		`- * The loose timing also means that locks may be recorded that have a`
`1270`		`- * zero xid, since xids are removed from procs before locks are removed.`
`1271`		`- * So we must prune the lock list down to ensure we hold locks only for`
`1272`		`- * currently running xids, performed by StandbyReleaseOldLocks().`
`1273`		`- * Zero xids should no longer be possible, but we may be replaying WAL`
`1274`		`- * from a time when they were possible.`
`1275`		`- *`
`1276`	`1272`	`* For logical decoding only the running xacts information is needed;`
`1277`	`1273`	`* there's no need to look at the locking information, but it's logged anyway,`
`1278`	`1274`	`* as there's no independent knob to just enable logical decoding. For`

Comments

(0)