Mailing-List: contact commits-help@harmony.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@harmony.apache.org
Message-ID: <18491669.1171537326473.JavaMail.jira@brutus>
Date: Thu, 15 Feb 2007 03:02:06 -0800 (PST)
From: "Vladimir Beliaev (JIRA)" <jira@apache.org>
To: commits@harmony.apache.org
Subject: [jira] Commented: (HARMONY-2692) [drlvm] throwing exception in
 mixed mode leaks memory
In-Reply-To: <25359430.1166101461145.JavaMail.jira@brutus>
MIME-Version: 1.0
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: 7bit


    [ https://issues.apache.org/jira/browse/HARMONY-2692?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12473346 ] 

Vladimir Beliaev commented on HARMONY-2692:
-------------------------------------------

Regarding regression test - I believe there should be no reggression test because:

1. test runs ~7min (about 8'500'000 iterations passed by that time)  - too long for "acceptance" test. 

2. test behaviour depends on current DRLVM "state" in terms of "C heap usage":
    - if DRLVM is optimized to use less c-memory, then number of iterations in test may be not enough (test execution time is critical too)
    - if we re-develop test to check if the C-Heap-bound is achieved then such a test may fail due to DRLVM use more C Heap (say, because of new features / optimization are done)... So such the test would be unstable.

So, I vote for comitting the patch w/o regression test.

Thanks
Vladimir Beliaev

> [drlvm] throwing exception in mixed mode leaks memory
> -----------------------------------------------------
>
>                 Key: HARMONY-2692
>                 URL: https://issues.apache.org/jira/browse/HARMONY-2692
>             Project: Harmony
>          Issue Type: Bug
>          Components: DRLVM
>         Environment: Windows XP i686
>            Reporter: Salikh Zakirov
>         Assigned To: Gregory Shimansky
>         Attachments: 0001-Fix-of-memory-leak-from-rth_throw_lazy-runtime-helper.patch, BufferWrap.java, ExceptionLeakTest.java
>
>
> The function wrap() of nio buffers leaks memory when invoked with illegal arguments.
> When the attached test BufferWrap.java is run on DRLVM, it steadily increases used memory until crash.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.