[jboss-jira] [JBoss JIRA] Commented: (JASSIST-125) Compilation fails with "javassist.bytecode.BadBytecode: jsr at 406" WebLogic 10.3 JTA data source class
[
https://jira.jboss.org/browse/JASSIST-125?page=com.atlassian.jira.plugin....
]
Shigeru Chiba commented on JASSIST-125:
---------------------------------------
You are right. But weblogic.jdbc.jta.DataSource.class is in Java 6 (version 50). This
confuses me.
My understanding is a class file can use either jsr or StackMapTable. So Javassist can
generate a
StackMapTable only if a class file does not contain jsr. That's why Javassist fails
to run your test code
(DataSource.class uses both jsr and StackMapTable).
Anyway, what I would like to discuss is how to fix this problem. One idea is not to
generate a StackMapTable
if a class file contains jsr. Another idea is to try to generate a StackMapTable as long
as it is possible (this
means Javassist may fail but normally should succeed)
Any comments are welcome.
Compilation fails with "javassist.bytecode.BadBytecode: jsr at
406" WebLogic 10.3 JTA data source class
-------------------------------------------------------------------------------------------------------
Key: JASSIST-125
URL: https://jira.jboss.org/browse/JASSIST-125
Project: Javassist
Issue Type: Bug
Affects Versions: 3.12.0.GA
Environment: Reproduced on Linux x86 and Windows XP (x86).
Reporter: Adam Honen
Assignee: Shigeru Chiba
Trying to add code at the beginning of the method getConnection in the class
weblogic.jdbc.jta.DataSource I got the following error:
Exception in thread "main" javassist.CannotCompileException: by
javassist.bytecode.BadBytecode: jsr at 399
at javassist.CtBehavior.insertBefore(CtBehavior.java:728)
at javassist.CtBehavior.insertBefore(CtBehavior.java:685)
at Test.main(Test.java:26)
Caused by: javassist.bytecode.BadBytecode: jsr at 399
at javassist.bytecode.stackmap.Tracer.throwBadBytecode(Tracer.java:144)
at javassist.bytecode.stackmap.Tracer.visitJSR(Tracer.java:133)
at javassist.bytecode.stackmap.Tracer.doOpcode148_201(Tracer.java:563)
at javassist.bytecode.stackmap.Tracer.doOpcode(Tracer.java:101)
at javassist.bytecode.stackmap.MapMaker.make(MapMaker.java:182)
at javassist.bytecode.stackmap.MapMaker.traceException(MapMaker.java:213)
at javassist.bytecode.stackmap.MapMaker.make(MapMaker.java:175)
at javassist.bytecode.stackmap.MapMaker.make(MapMaker.java:192)
at javassist.bytecode.stackmap.MapMaker.make(MapMaker.java:141)
at javassist.bytecode.stackmap.MapMaker.make(MapMaker.java:96)
at javassist.bytecode.MethodInfo.rebuildStackMap(MethodInfo.java:416)
at javassist.bytecode.MethodInfo.rebuildStackMapIf6(MethodInfo.java:398)
at javassist.CtBehavior.insertBefore(CtBehavior.java:719)
... 2 more
I've created a small program that simulates this problem outside the Weblogic
container.
The class can be found inside <Weblogic
home>\modules\com.bea.core.datasource6_1.4.0.0.jar.
And additional dependency which is required in order to load this class (although it is
not needed in order to modify the byte codes) is <Weblogic
home>\modules\com.bea.core.transaction_2.5.0.0.jar.
Both of these files are available when installing Weblogic 10.3.
I suspect this might be some sort of collision with WebLogic's Diagnostic framework.
I'm not sure how calls to WLDF are added, but it is possible they were added by some
other tool for manipulating byte codes (at compile time), in which case this issue is a
bit more interesting (could indicate collisions with certain other AOP tools).
Just for the sake of pin pointing the source of the problem, I also used AspectJ to add
code to the method. It worked fine with no errors.
--
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators:
https://jira.jboss.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira