You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@harmony.apache.org by "Gregory Shimansky (JIRA)" <ji...@apache.org> on 2007/04/24 17:58:15 UTC
[jira] Commented: (HARMONY-2081) [drlvm][interpreter] Synthetic
test with monexit without monenter before passed on RI but failed on DRLVM
[ https://issues.apache.org/jira/browse/HARMONY-2081?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12491342 ]
Gregory Shimansky commented on HARMONY-2081:
--------------------------------------------
In fact there are 2 different but related bugs in interpreter.
1. Monitor exit bytecode doesn't check which monitor it removes from the frame.locked_monitors list. It always removes the last added, so it makes it look like a stack. But this is not correct because bytecode may enter and exit monitors in different order.
2. When exiting from a synchronized method interpreter doesn't do monitorexit when there are no locked monitors left.
> [drlvm][interpreter] Synthetic test with monexit without monenter before passed on RI but failed on DRLVM
> ---------------------------------------------------------------------------------------------------------
>
> Key: HARMONY-2081
> URL: https://issues.apache.org/jira/browse/HARMONY-2081
> Project: Harmony
> Issue Type: Bug
> Components: DRLVM
> Reporter: Mikhail Fursov
> Assigned To: Gregory Shimansky
> Priority: Minor
> Attachments: interpreter-2081.diff, monexit.zip, TestME.zip
>
>
> Here is synthetic test in jasmin to demonstrate the problem: we do not handle the situation when monexit is called without monenter.
> The test passes on RI (BEA/SUN) and failed with -Xem:jet, -Xem:opt, -Xem:interpreter
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.