Mailing-List: contact commits-help@harmony.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@harmony.apache.org
Message-ID: <12765754.1165854928210.JavaMail.jira@brutus>
Date: Mon, 11 Dec 2006 08:35:28 -0800 (PST)
From: "Vasily Zakharov (JIRA)" <jira@apache.org>
To: commits@harmony.apache.org
Subject: [jira] Updated: (HARMONY-2615) [drlvm][test] Stress test for
 massive class loading from bootclasspath
In-Reply-To: <28655536.1165854803895.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit

     [ http://issues.apache.org/jira/browse/HARMONY-2615?page=all ]

Vasily Zakharov updated HARMONY-2615:
-------------------------------------

    Description: 
This is a stress test which loads all the classes it finds in VM bootclasspath using simple Class.forName() call.
 
Not all of the classes are supposed to be loaded correctly this way, so the test ignores any class loading errors - its main purpose is to make sure VM survives loading thousands of classes without a crash. Also, investigating why particular classes fail to load may provide clues to some other bugs in VM or classlib.

I recommend including this test to some Harmony test suite that is run regularly, as it may detect many VM issues that may occur in future.

Currently this test crashes both on Windows/IA32 and Linux/IA32 in a very strange way - at some point, unhandled exception occurs inside the catch(Throwable) block:

9879: org.mmtk.plan.Plan
java.lang.ClassNotFoundException: null.Factory
        at java.lang.Class.forName(Unknown Source)
        at java.lang.Class.forName(Unknown Source)
        at org.mmtk.vm.VM.<clinit>(Unknown Source)
        at org.mmtk.utility.Constants.<clinit>(Unknown Source)
        at org.mmtk.plan.Plan.<clinit>(Unknown Source)
        at java.lang.VMClassRegistry.initializeClass(Native Method)
        at java.lang.Class.forName(Unknown Source)
        at java.lang.Class.forName(Unknown Source)
        at CrashTestBCP.load(CrashTestBCP.java:75)
        at CrashTestBCP.main(CrashTestBCP.java:98)


  was:
This is a stress test which loads all the classes it finds in VM bootclasspath using simple Class.forName() call.
 
Not all of the classes are supposed to be loaded correctly this way, so the test ignores any class loading errors - its main purpose is to make sure VM survives loading of thousands classes without crashes. Also, investigating why particular classes fail to load may provide clues to some new bugs in VM or classlib.

I recommend including this test to some Harmony test suite that is run regularly, as it may detect many VM issues that may occur in future.

Currently this test crashes both on Windows/IA32 and Linux/IA32 in a very strange way - at some point, unhandled exception occurs inside the catch(Throwable) block:

9879: org.mmtk.plan.Plan
java.lang.ClassNotFoundException: null.Factory
        at java.lang.Class.forName(Unknown Source)
        at java.lang.Class.forName(Unknown Source)
        at org.mmtk.vm.VM.<clinit>(Unknown Source)
        at org.mmtk.utility.Constants.<clinit>(Unknown Source)
        at org.mmtk.plan.Plan.<clinit>(Unknown Source)
        at java.lang.VMClassRegistry.initializeClass(Native Method)
        at java.lang.Class.forName(Unknown Source)
        at java.lang.Class.forName(Unknown Source)
        at CrashTestBCP.load(CrashTestBCP.java:75)
        at CrashTestBCP.main(CrashTestBCP.java:98)


> [drlvm][test] Stress test for massive class loading from bootclasspath
> ----------------------------------------------------------------------
>
>                 Key: HARMONY-2615
>                 URL: http://issues.apache.org/jira/browse/HARMONY-2615
>             Project: Harmony
>          Issue Type: Test
>          Components: DRLVM
>            Reporter: Vasily Zakharov
>         Attachments: CrashTestBCP.java
>
>
> This is a stress test which loads all the classes it finds in VM bootclasspath using simple Class.forName() call.
>  
> Not all of the classes are supposed to be loaded correctly this way, so the test ignores any class loading errors - its main purpose is to make sure VM survives loading thousands of classes without a crash. Also, investigating why particular classes fail to load may provide clues to some other bugs in VM or classlib.
> I recommend including this test to some Harmony test suite that is run regularly, as it may detect many VM issues that may occur in future.
> Currently this test crashes both on Windows/IA32 and Linux/IA32 in a very strange way - at some point, unhandled exception occurs inside the catch(Throwable) block:
> 9879: org.mmtk.plan.Plan
> java.lang.ClassNotFoundException: null.Factory
>         at java.lang.Class.forName(Unknown Source)
>         at java.lang.Class.forName(Unknown Source)
>         at org.mmtk.vm.VM.<clinit>(Unknown Source)
>         at org.mmtk.utility.Constants.<clinit>(Unknown Source)
>         at org.mmtk.plan.Plan.<clinit>(Unknown Source)
>         at java.lang.VMClassRegistry.initializeClass(Native Method)
>         at java.lang.Class.forName(Unknown Source)
>         at java.lang.Class.forName(Unknown Source)
>         at CrashTestBCP.load(CrashTestBCP.java:75)
>         at CrashTestBCP.main(CrashTestBCP.java:98)

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira