Summary: | StackOverflowError: infinite recursion in org.aspectj.weaver.ReferenceType.isAssignableFrom() | ||||||||
---|---|---|---|---|---|---|---|---|---|
Product: | [Tools] AspectJ | Reporter: | Troy Anderson <tanderson> | ||||||
Component: | Compiler | Assignee: | aspectj inbox <aspectj-inbox> | ||||||
Status: | RESOLVED FIXED | QA Contact: | |||||||
Severity: | normal | ||||||||
Priority: | P2 | CC: | aclement | ||||||
Version: | unspecified | ||||||||
Target Milestone: | 1.6.8 | ||||||||
Hardware: | PC | ||||||||
OS: | Windows Vista | ||||||||
Whiteboard: | |||||||||
Attachments: |
|
Description
Troy Anderson
2009-09-22 11:10:22 EDT
Created attachment 147786 [details]
Java file that breaks iajc
unfortunately I cannot reproduce the problem. I probably need to know what aspect you are using against that java class - are you able to include it? Compiling that code standalone just works (with the necessary annotations stubbed out), or compiling it with an aspect containing: before(): within(*) {} just works. It must be something about the pointcuts in use that cause the type analysis that leads to the recursion. Created attachment 147837 [details]
Aspects applied
Here are the asspects applied
hmm, unfortunately even with the aspects compiled against that type, I can't trigger the failure. It is possible due to the other types that are missing - I'm just stubbing them out when their inter relationships may be important to drive the failure case. Believed fixed! Reported as occurring more frequently against 1.6.7 and knowing what I had changed in that release led me to find the cause. Discussed fully in https://bugs.eclipse.org/bugs/show_bug.cgi?id=298908 fix is in 1.6.8 |