Details
-
Type:
Improvement
-
Status:
Closed
-
Priority:
Major
-
Resolution: Fixed
-
Affects Version/s: 2.2.5 (EE only)
-
Fix Version/s: 2.2.8 (EE only), 3.1.1
-
Component/s: Core: (other)
-
Labels:None
-
User impact:Medium
-
Affects Docs:Yes
-
Similar Issues:None
Description
I have a custom ThreadPoolFactory which must be used in favor of the one in mule-core-ee-2.2.5.jar
The discovery mechanism picks the first implementation jar which contains com.mulesoft.mule.config.Preferred
Since this simply chooses the first one it finds, it is not possible to guarantee that it picks mine
There are two workarounds, both unacceptable in the long term:
Disallow the mule-core-ee-2.2.5 jar from the classpath
Remove com.mulesoft.mule.config.Preferred from mule-core-ee-2.2.5.jar
A valid solution would be something like:
System property allowing exact specification
mule xml configuration allowing exact specification
Weighting as mentioned in below nabble link
etc...
See http://old.nabble.com/Thread-setup-cleanup-tt28413607.html#a28438680
NOTE: Our Mule instances are embedded, NOT standalone.
Issue Links
- relates to
-
MULE-4909
Refactor SPI dicovery mechanism to be more generic
-
Fix 2.x:
http://fisheye.codehaus.org/changelog/mule/?cs=20882
http://fisheye.codehaus.org/changelog/mule/?cs=20918
Fix 3.1.x:
http://fisheye.codehaus.org/changelog/mule/?cs=20909
http://fisheye.codehaus.org/changelog/mule/?cs=20919
Fix 3.x:
http://fisheye.codehaus.org/changelog/mule/?cs=20925
http://fisheye.codehaus.org/changelog/mule/?cs=20926