Brett Wooldridge
|
7dfdf597e2
|
Because housekeeper runs immediately now, 'previous' timestamp must be initialised as if it has run before, otherwise retrograde detection triggers.
|
9 years ago |
|
Brett Wooldridge
|
e2f1974dc9
|
Add null check.
|
9 years ago |
|
Brett Wooldridge
|
69dbe20cff
|
Issue #559 Adjust illegal clock motion detection, use scheduleWithFixedDelay() for housekeeper, and enlarge closeConnectionExecutor queue.
|
9 years ago |
|
Brett Wooldridge
|
8cb3d4aedc
|
Merge pull request #564 from EddieCurtis/patch-1
Fixed typo in README.md
|
9 years ago |
Eddie Curtis
|
f21347b5d3
|
Fixed typo in README.md
|
9 years ago |
Nitin
|
b78fde9121
|
full stop :)
|
9 years ago |
|
Nitin
|
96c91d549a
|
disable leakDetectionThreshold if not valid.
|
9 years ago |
|
Nitin
|
0e3f0eb0f6
|
leakDetectionThreshold should be less than maxLifetime
|
9 years ago |
|
Brett Wooldridge
|
d4d6d759de
|
Issue #547 Decreased allowable minimum connectionTimeout and validationTimeout to 250ms.
|
9 years ago |
|
Brett Wooldridge
|
bd7e3ba6db
|
Issue #563 Do not start the house-keeping thread until after pool initialisation has completed.
|
9 years ago |
|
Nitin
|
2b0e3f7553
|
changed log level: warn instead of debug, during driver support check
(logs only once)
|
9 years ago |
|
Brett Wooldridge
|
f7c8118a86
|
Issue #559 cap the number of connections added by fillPool() if there are already pending tasks to do so.
|
9 years ago |
|
Brett Wooldridge
|
91dbe31ff2
|
Fixes #559 improve logging of time shifts, allow retrograde time adjustments
|
9 years ago |
|
Nitin
|
4f865cabb5
|
do not set maxLifetime when disabling idle timeout
|
9 years ago |
|
Nitin
|
f40a13b5be
|
added synchronized to getObjectInstance()
|
9 years ago |
|
Nitin
|
a40e4856c9
|
renamed method to checkDriverSupport() and query driver default
transaction isolation once
|
9 years ago |
|
Brett Wooldridge
|
e9eec65ba2
|
Update CHANGES
|
9 years ago |
Ruben Dijkstra
|
8adf36bc3d
|
Skip the `testMetricUsage` test on Windows
My 75th percentile is indeed 250ms, but _sometimes_ returns 249ms
|
9 years ago |
|
Brett Wooldridge
|
c0bd8bfeac
|
Fixes #555 check for default method is not longer needed because proxy generation is performed at build-time not run-time now.
|
9 years ago |
|
Nitin
|
99cacb33b2
|
changes as suggested. reverted resetBits and method.getParameterCount()
|
9 years ago |
|
Brett Wooldridge
|
73491abb31
|
Fixes #555 check for default method is not longer needed because proxy generation is performed at build-time not run-time now.
|
9 years ago |
|
Nitin
|
573660dfe1
|
future safe from statement reorder
|
9 years ago |
|
Nitin
|
64dd93fe8f
|
minor
|
9 years ago |
|
Nitin
|
85e41abf10
|
context.close is correct. 542 may have other reason/s!
|
9 years ago |
|
Nitin
|
b525e0f7ef
|
set maxPoolSize = minIdle only if undefined, consistent error messages
|
9 years ago |
|
Nitin
|
ad754d085b
|
made transactionIsolation final
|
9 years ago |
|
Nitin
|
fde88d0659
|
if user specified, set transaction isolation level
|
9 years ago |
|
Nitin
|
ce50c2cfd6
|
removed resetBits. setting transaction isolation level like readonly &
autocommit.
|
9 years ago |
|
Nitin
|
c8cd7ab1bb
|
swap test condition
|
9 years ago |
|
Nitin
|
9a63ebf876
|
replaced getParameterTypes().length with getParameterCount()
|
9 years ago |
Mikhail Mazurskiy
|
743dc09d9a
|
Slightly improve generics
|
9 years ago |
|
Mikhail Mazurskiy
|
a30dcd87b8
|
Cache volatile reads in local variables
Avoid unnecessary read memory barriers
|
9 years ago |
|
Mikhail Mazurskiy
|
c3dcb08023
|
Preserve tread interrupted status
|
9 years ago |
|
Mikhail Mazurskiy
|
877025fb87
|
Add EditorConfig
See http://EditorConfig.org
|
9 years ago |
|
Nitin
|
aa90bd62e0
|
close statement before commit/rollback
|
9 years ago |
|
Nitin
|
6769539f10
|
close ds
|
9 years ago |
|
Nitin
|
fbb31cc2bb
|
replaced DriverSapDB with DataSourceSapDB
|
9 years ago |
|
Nitin
|
7706922f26
|
removed AS400, was 'driver' for DB2 on AS400 not datasource. added
informix
|
9 years ago |
|
Brett Wooldridge
|
5291e6bbb3
|
Merge pull request #548 from mgazanayi/patch-1
Add SQLite DataSource class
|
9 years ago |
Marouane Gazanayi
|
239f97017e
|
Add SQLite DataSource class
|
9 years ago |
|
Nitin
|
23b4909155
|
sizing queue for connection closing proportionally
|
9 years ago |
|
Nitin
|
c362e8bb35
|
fixed test
|
9 years ago |
|
Nitin
|
9daa51313f
|
I've added context.close() before but not sure anymore! see 542
|
9 years ago |
|
Nitin
|
87757fcd0a
|
keeping connectionBag.remove and totalConnections.decrementAndGet()
close by
|
9 years ago |
|
Nitin
|
5f851a4cfb
|
reverted. using cachedExecutor for setting network timeout as before
|
9 years ago |
|
Nitin
|
c9ad521660
|
cleanup + rearrange statements
|
9 years ago |
|
Nitin
|
2f3a1079ed
|
merged DefaultThreadFactory with UtilityElf
|
9 years ago |
|
Nitin
|
cf1b7738c0
|
unused import
|
9 years ago |
|
Nitin
|
a56f5dcb2d
|
removed ThreadPoolExecutor for use with setNetworkTimeout
|
9 years ago |
|
Brett Wooldridge
|
1a64ab2357
|
This change in incorrect. Even the pure JDBC4 isValid() check can execute a query, and it is known that some drivers do so.
|
9 years ago |