diff options
| author | Ying Wang <wangying@google.com> | 2015-12-22 12:21:17 -0800 |
|---|---|---|
| committer | Ying Wang <wangying@google.com> | 2015-12-22 14:08:15 -0800 |
| commit | 9d99b19da7ab70bfff55327c5e4719addb1955f7 (patch) | |
| tree | 82ea8780254ef7fddadb5c570143c10e62bc9815 /lib/retrace.jar | |
| parent | 72a7fde89fd0fa930baf7b47d1bd19c73317de1a (diff) | |
| download | proguard-9d99b19da7ab70bfff55327c5e4719addb1955f7.tar.gz | |
Fix 'Exception in thread "main" java.lang.AbstractMethodError:'
When running with OpenJDK 8 (openjdk version "1.8.0_45-internal") on
module messaging, sometimes an odd exception is caught in main:
Exception in thread "main" java.lang.AbstractMethodError:
java.lang.Exception.getMessage()Ljava/lang/String;
at proguard.ProGuard.main(ProGuard.java:519)
It's not 100% reproducible. But it's certain to happen if you run the
same ProGuard tasks in parallel (e.g. x40).
The exception runtime type is actually
java.lang.ArrayIndexOutOfBoundsException, which I tracked down to
raise from MappingPrinter.visitLineNumberTableAttribute().
It's strange that the exception object is in a weird state that
e.printStatckTrace() throws java.lang.AbstractMethodError.
Other people have seen similar things too:
http://stackoverflow.com/questions/6777710/abstractmethoderror-on-calling-exception-printstacktrace
Using temparary variables low/high fixes the crash.
It seems to be an OpenJDK 8 bug.
Rebuilt the jar files by running build/build.sh.
Bug: 26274804
Change-Id: I773023b2dc57bd048c6781aded84e0c0df923a90
Diffstat (limited to 'lib/retrace.jar')
| -rw-r--r-- | lib/retrace.jar | bin | 6973 -> 7556 bytes |
1 files changed, 0 insertions, 0 deletions
diff --git a/lib/retrace.jar b/lib/retrace.jar Binary files differindex c00fbe7..94c8c6b 100644 --- a/lib/retrace.jar +++ b/lib/retrace.jar |