harmony-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Alexey Varlamov" <alexey.v.varla...@gmail.com>
Subject Re: svn commit: r520804 - in /harmony/enhanced/classlib/trunk/modules/luni: make/exclude.common make/exclude.linux.x86_64.drl make/exclude.windows.x86_64.drl src/test/java/tests/api/java/net/MulticastSocketTest.java
Date Mon, 26 Mar 2007 05:00:32 GMT
2007/3/24, Alexei Zakharov <alexei.zakharov@gmail.com>:
> Things become more interesting. With my current firewall configuration
> it passes on IBM VME and fails on DRLVM (on Linux). :-/
>
> Stacktraces:
> ---
> test_joinGroupLjava_net_SocketAddressLjava_net_NetworkInterface:
> junit.framework.AssertionFailedError: Did not get expected exception
> when joining for second time on same interface at
> tests.api.java.net.MulticastSocketTest.test_joinGroupLjava_net_SocketAddressLjava_net_NetworkInterface(MulticastSocketTest.java:557)
> at java.lang.reflect.VMReflection.invokeMethod(VMReflection.java)
>
> test_leaveGroupLjava_net_InetAddress
> junit.framework.AssertionFailedError: Failed to throw exception
> leaving non-member group at
> tests.api.java.net.MulticastSocketTest.test_leaveGroupLjava_net_InetAddress(MulticastSocketTest.java:595)
> at java.lang.reflect.VMReflection.invokeMethod(VMReflection.java)
> ---
>
> Guess this is not a firewall issue. Some problems with luni natives -
> DatagramSocketImpl doesn't throw required exception? Any ideas?

This might be JIT-related problem - you can try to constrain execution
mode to run with JET or OPT only to localize. Simply put mode args to
<jre>/bin/default/harmonyvm.properties, either -Xem:jet or -Xem:opt or
-Xem:interpreter and run the tests.
Well, theoretically there is simpler way to supply extra args to VM,
via hy.test.vmargs property - but AFAIK it works not for all modules
(HARMONY-3311).

--
Alexey

>
> Regards,
>
> 2007/3/22, Alexey Petrenko <alexey.a.petrenko@gmail.com>:
> > I've tested on Windows and Linux behind our firewalls.
> > And I believe that CC machines are behind the firewall too.
> >
> > SY, Alexey
> >
> > 2007/3/22, Alexei Zakharov <alexei.zakharov@gmail.com>:
> > > Well, I've just tried to run this test on my Windows laptop with
> > > direct connection to internet - everything works just fine. So I
> > > believe this was a firewall problem indeed. IMO it worths mentioning
> > > somewhere on Wiki and / or site.
> > >
> > > Thanks,
> > >
> > > 2007/3/22, Tony Wu <wuyuehao@gmail.com>:
> > > > Hi Alexei,
> > > > 3 Multicast IP addresses were used in this tests. so you should set
> > > > all of them as ACCPET( 224.0.0.3, 224.0.0.4 and 224.1.2.3) into your
> > > > iptables.
> > > >
> > > > On 3/22/07, Alexei Zakharov <alexei.zakharov@gmail.com> wrote:
> > > > > The same picture for 224.1.2.3 :(
> > > > >
> > > > > 2007/3/22, Tony Wu <wuyuehao@gmail.com>:
> > > > > > sorry I mean,
> > > > > > iptables -A HYBLD -d 224.1.2.3 -j ACCEPT
> > > > >
> > > > > > On 3/22/07, Tony Wu <wuyuehao@gmail.com> wrote:
> > > > > > > plus iptables -A HYBLD -d 224.0.0.4 -j ACCEPT
> > > > > > >
> > > > > > > On 3/22/07, Tony Wu <wuyuehao@gmail.com> wrote:
> > > > > > > > Hi,
> > > > > > > > It caused by my patch on harmony-3433.
> > > > > > > >
> > > > > > > > It will fail on some machines because of firewall,
the access to some
> > > > > > > > multicast addresses used in tests are not permitted
on your machines.
> > > > > > > > running following lines will help,
> > > > > > > > iptables -A HYBLD -d 224.0.0.3 -j ACCEPT
> > > > > > > > iptables -A HYBLD -d 224.0.0.4 -j ACCEPT
> > > > > > > >
> > > > > > > > In the meanwhile I was thinking about is there any
way to avoid this
> > > > > > > > operation, because asking user to set up a rule for
testing is always
> > > > > > > > not a good idea. Do you have any suggestion? Thanks.
> > > > > > > >
> > > > > > > > On 3/22/07, Alexei Zakharov <alexei.zakharov@gmail.com>
wrote:
> > > > > > > > > Hi Alexey,
> > > > > > > > >
> > > > > > > > > MulticastSocketTest fails each time I run luni
tests. At least on
> > > > > > > > > Linux. And it seems CC encounters the same issue.
> > > > > > > > >
> > > > > > > > > Regards,
> > > > > > > > >
> > > > > > > > > 2007/3/21, Alexey Petrenko <alexey.a.petrenko@gmail.com>:
> > > > > > > > > > oops...
> > > > > > > > > > That's a patch for HARMONY-3433 "[classlib][luni]move
out
> > > > > > > > > > src\test\java\tests\api\java\net\MulticastSocketTest.java"
> > > > > > > > > >
> > > > > > > > > > SY, Alexey
> > > > > > > > > >
> > > > > > > > > > 2007/3/21, apetrenko@apache.org <apetrenko@apache.org>:
> > > > > > > > > > > Author: apetrenko
> > > > > > > > > > > Date: Wed Mar 21 01:44:06 2007
> > > > > > > > > > > New Revision: 520804
> > > > > > > > > > >
> > > > > > > > > > > URL: http://svn.apache.org/viewvc?view=rev&rev=520804
> > > > > > > > > > > Log:
> > > > > > > > > > > Patch for HARMONY-3454 "[classlib][swing]
JEditorPane.createEditorKitForContentType() uses wrong classloader"
> > > > > > > > > > >
> > > > > > > > > > > Modified:
> > > > > > > > > > >    harmony/enhanced/classlib/trunk/modules/luni/make/exclude.common
> > > > > > > > > > >    harmony/enhanced/classlib/trunk/modules/luni/make/exclude.linux.x86_64.drl
> > > > > > > > > > >    harmony/enhanced/classlib/trunk/modules/luni/make/exclude.windows.x86_64.drl
>
>
>
> --
> Alexei Zakharov,
> Intel ESSD
>

Mime
View raw message