Re: [Jamvm-general] Crash running OpenJDK 8 build.tools.generatenimbus.Generator, a regression sinc
Brought to you by:
rlougher
From: Robert L. <rob...@gm...> - 2013-04-26 02:06:14
|
Hi Xerxes, On 22 April 2013 12:38, Xerxes Rånby <xe...@za...> wrote: > > The jamvm+openjdk 8 bootstrap compiler crash while running the Nimbus LAF > build.tools.generatenimbus.Generator. > /home/xranby/icedtea-8-b80/bootstrap/jdk1.7.0/bin/java -version > openjdk version "1.8.0-fcs" > OpenJDK Runtime Environment (build 1.8.0-fcs-xranby_2013_04_22_12_10-b00) > JamVM (build 1.6.0-devel, inline-threaded interpreter with stack-caching) > based on jamvm source git commit: > http://git.berlios.de/cgi-bin/cgit.cgi/jamvm/commit/?id=da518b74fed8d3fe986543d222a45c54efdad5ef > > I can run the same generator correctly without observing any of the below > mentioned three crash variants using JamVM + OpenJDK 7 based on JamVM source > git commit: > http://git.berlios.de/cgi-bin/cgit.cgi/jamvm/commit/?id=0972452d441544f7dd29c55d64f1ce3a5db90d82 > > All logs below was produced on a 32bit x86 Ubuntu 12.04 system. > > Testcase: > # inside a openjdk 8 source tree run: > cd /home/xranby/icedtea-8-b80/openjdk/jdk/makefiles > /home/xranby/icedtea-8-b80/bootstrap/jdk1.7.0/bin/java -Xms64M -Xmx1100M > -XX:PermSize=32m -XX:MaxPermSize=160m -XX:ThreadStackSize=1536 > -XX:-PrintVMOptions -XX:-UnlockDiagnosticVMOptions -XX:-LogVMOutput -cp > /home/xranby/icedtea-8-b80/openjdk.build/jdk/btclasses > build.tools.generatenimbus.Generator -skinFile > /home/xranby/icedtea-8-b80/openjdk/jdk/src/share/classes/javax/swing/plaf/nimbus/skin.laf > -buildDir /home/xranby/icedtea-8-b80/openjdk.build/jdk/gensrc > -packagePrefix javax.swing.plaf.nimbus -lafName Nimbus > With the testcase I was able to reproduce the problem. Unfortunately I don't see exactly the same behaviour but I'm using OpenJDK 8 rather than IcedTea and AMD64. So it may be a different issue but hopefully not... The issue I'm seeing is related to JSR335, specifically an invokespecial on a default interface method (isAnnotationPresent),. This should now be fixed: http://git.berlios.de/cgi-bin/cgit.cgi/jamvm/commit/?id=b0651cde57c7a48651c99d7f4fb44ba7b2fae9b3 Thanks, Rob. |