Issue1684
Created on 2010-12-08.13:05:04 by doublep, last changed 2020-05-27.09:10:28 by doublep.
|
| File name |
Uploaded |
Description |
Edit |
Remove |
|
test.java
|
doublep,
2010-12-08.13:05:03
|
short example demonstrating the problem |
|
|
|
test2.java
|
doublep,
2010-12-09.14:28:18
|
propagating exception thrown in Java code: here cause is correctly set |
|
|
|
Test.java
|
jeff.allen,
2018-03-17.20:03:33
|
Demonstration that getCause returns null |
|
|
|
| msg6274 (view) |
Author: (doublep) |
Date: 2010-12-08.13:05:03 |
|
The attached example demonstrates the problem. I have code that walks exception chain to "the real cause" to find out whether that cause should be reported as a user-level error (real_cause instanceof UserLevelException) or internal application bug. With exceptions raised from Jython code I cannot do this at the moment.
|
| msg6275 (view) |
Author: (doublep) |
Date: 2010-12-09.14:28:18 |
|
Actually, when exception is thrown by Java code and is wrapped by PyException for propagation, cause is set correctly. It seems the problem is there only when a Java exception is raised _directly_ from Python code, as in original example.
|
| msg11830 (view) |
Author: Jeff Allen (jeff.allen) |
Date: 2018-03-17.20:03:33 |
|
I explored this on 2.7.2a1 with a small adaptation of the provided test.java, attached, and the embedded cause is still null. Reproduce like this:
PS bugs> javac -cp ".;C:\Jython\2.7.2a1\jython.jar;C:\Jython\2.7.2a1\javalib\*" .\iss1684\Test.java
PS bugs> java -cp ".;C:\Jython\2.7.2a1\jython.jar;C:\Jython\2.7.2a1\javalib\*" iss1684.Test
Traceback (most recent call last):
File "<string>", line 2, in <module>
at sun.reflect.NativeConstructorAccessorImpl.newInstance0(Native Method)
at sun.reflect.NativeConstructorAccessorImpl.newInstance(NativeConstructorAccessorImpl.java:62)
at sun.reflect.DelegatingConstructorAccessorImpl.newInstance(DelegatingConstructorAccessorImpl.java:45)
at java.lang.reflect.Constructor.newInstance(Constructor.java:423)
at org.python.core.PyReflectedConstructor.constructProxy(PyReflectedConstructor.java:211)
java.lang.RuntimeException: java.lang.RuntimeException: Thrown from Python code
Cause: null
So we get the cause in the stack dump, but it isn't the cause in the Java sense, which seems a reasonable expectation.
|
| msg13068 (view) |
Author: (doublep) |
Date: 2020-05-27.09:10:28 |
|
Bump. Reproducible with 2.7.2.
|
|
| Date |
User |
Action |
Args |
| 2020-05-27 09:10:28 | doublep | set | messages:
+ msg13068 |
| 2018-03-17 20:03:33 | jeff.allen | set | files:
+ Test.java
resolution: accepted
messages:
+ msg11830
nosy:
+ jeff.allen
versions:
+ Jython 2.7 |
| 2013-02-26 17:39:39 | fwierzbicki | set | priority: normal
nosy:
+ fwierzbicki |
| 2013-02-19 23:28:05 | fwierzbicki | set | versions:
+ Jython 2.5, - 2.5.1 |
| 2010-12-09 14:28:18 | doublep | set | files:
+ test2.java
messages:
+ msg6275
title: PyException raised from a Java exception doesn't have cause set -> PyException wrapper for a Java exception raised directly from Python code doesn't have cause set |
| 2010-12-08 13:05:04 | doublep | create | |
|
