Fix oversight in coding of _bt_start_vacuum: we can't assume that the LWLock
authorTom Lane <[email protected]>
Fri, 30 Mar 2007 00:13:05 +0000 (00:13 +0000)
committerTom Lane <[email protected]>
Fri, 30 Mar 2007 00:13:05 +0000 (00:13 +0000)
commit93e34979dfcc7468b27aba670636d4703927794d
tree6c974199fcc5b691576bc9cace0b08becb0ddc69
parent0a01c08d6c39bdb55f038a9c0c9a9f1ebc189af3
Fix oversight in coding of _bt_start_vacuum: we can't assume that the LWLock
will be released by transaction abort before _bt_end_vacuum gets called.
If either of these "can't happen" errors actually happened, we'd freeze up
trying to acquire an already-held lock.  Latest word is that this does
not explain Martin Pitt's trouble report, but it still looks like a bug.
src/backend/access/nbtree/nbtutils.c