You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@avalon.apache.org by Leo Simons <le...@apache.org> on 2003/03/31 10:10:49 UTC
excalibur-compatibility deadlock problem
http://lsd.student.utwente.nl/gump/excalibur-compatibility.html:
[junit] Testcase: testDeadLock took 0.328 sec
[junit] FAILED
[junit] null
[junit] junit.framework.AssertionFailedError
[junit] at
org.apache.avalon.excalibur.concurrent.test.ReadWriteLockTestCase.testDeadLock(ReadWriteLockTestCase.java:406)
[junit] at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
[junit] at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
[junit] at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
[junit] Testcase: testDeadLock BUILD FAILED
file:///data/gump/avalon-excalibur/compatibility/build.xml:169: Test
org.apache.avalon.excalibur.concurrent.test.ReadWriteLockTestCase failed
Total time: 47 seconds
---
this is on Red Hat 8.0 with Sun JDK 1.4.1. I'll argue that's an
important platform to support!
is this a known issue?
- LSD
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@avalon.apache.org
For additional commands, e-mail: dev-help@avalon.apache.org
Re: excalibur-compatibility deadlock problem
Posted by Berin Loritsch <bl...@apache.org>.
Leo Simons wrote:
> http://lsd.student.utwente.nl/gump/excalibur-compatibility.html:
>
> [junit] Testcase: testDeadLock took 0.328 sec
> [junit] FAILED
> [junit] null
> [junit] junit.framework.AssertionFailedError
> [junit] at
org.apache.avalon.excalibur.concurrent.test.ReadWriteLockTestCase.testDeadLock(ReadWriteLockTestCase.java:406)
[junit] at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
[junit] at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
[junit] at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
[junit] Testcase: testDeadLock BUILD FAILED
>
> file:///data/gump/avalon-excalibur/compatibility/build.xml:169: Test
>
> org.apache.avalon.excalibur.concurrent.test.ReadWriteLockTestCase
failed Total time: 47 seconds
>
> ---
>
> this is on Red Hat 8.0 with Sun JDK 1.4.1. I'll argue that's an
important platform to support!
>
> is this a known issue?
It's a race condition leading to a deadlock--it passes on my machine.
But that is why we are deprecating the use of all the stuff in
compatibility. I don't want to spend time and energy trying to
bug fix code we aren't supporting.
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@avalon.apache.org
For additional commands, e-mail: dev-help@avalon.apache.org
Re: excalibur-compatibility deadlock problem
Posted by Sam Ruby <ru...@apache.org>.
Leo Simons wrote:
>
> nope! I've tried it on a couple of machines, and a normal 'ant test'
> doesn't result in failure, whereas running GUMP (either completely or
> using '/data/gump/bin/build excalibur-compatibility', which just
> rebuilds the excalibur-compatibility project and nothing else) does
> result in the failure. Which is real weird. Maybe it has to do with
> using the cvs version of ant or junit.
There are a number of ways to debug this.
One is to simply:
cd /data/gump/ant
cvs update -D "2003/01/01"
build ant
build excalibur-compatibility
Another is to edit /data/gump/build.sh. Search for
"excalibur_compatibility;" (note underscore and semicolon). Modify the
classpath immediately below this line and rerun.
The second approach is great for isolating which jar caused the problem.
The first approach can then be used to isolate the individual commit.
As always, any changes you make will be cleaned up with the next run.
- Sam Ruby
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@avalon.apache.org
For additional commands, e-mail: dev-help@avalon.apache.org
Re: excalibur-compatibility deadlock problem
Posted by Berin Loritsch <bl...@apache.org>.
Leo Simons wrote:
> http://lsd.student.utwente.nl/gump/excalibur-compatibility.html:
>
> [junit] Testcase: testDeadLock took 0.328 sec
> [junit] FAILED
> [junit] null
> [junit] junit.framework.AssertionFailedError
> [junit] at
org.apache.avalon.excalibur.concurrent.test.ReadWriteLockTestCase.testDeadLock(ReadWriteLockTestCase.java:406)
[junit] at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
[junit] at
sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:39)
[junit] at
sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:25)
[junit] Testcase: testDeadLock BUILD FAILED
>
> file:///data/gump/avalon-excalibur/compatibility/build.xml:169: Test
>
> org.apache.avalon.excalibur.concurrent.test.ReadWriteLockTestCase
failed Total time: 47 seconds
>
> ---
>
> this is on Red Hat 8.0 with Sun JDK 1.4.1. I'll argue that's an
important platform to support!
>
> is this a known issue?
It's a race condition leading to a deadlock--it passes on my machine.
But that is why we are deprecating the use of all the stuff in
compatibility. I don't want to spend time and energy trying to
bug fix code we aren't supporting.
RE: excalibur-compatibility deadlock problem
Posted by Leo Sutic <le...@inspireinfrastructure.com>.
> From: news [mailto:news@main.gmane.org] On Behalf Of Leo Simons
>
> Leo Sutic wrote:
> > If you run "ant test" on the compatibility project (and not GUMP,
> > which you seem to be running), do you still get a failure there?
>
> nope! I've tried it on a couple of machines, and a normal 'ant test'
> doesn't result in failure, whereas running GUMP (either completely or
> using '/data/gump/bin/build excalibur-compatibility', which just
> rebuilds the excalibur-compatibility project and nothing else) does
> result in the failure. Which is real weird. Maybe it has to do with
> using the cvs version of ant or junit.
I have inserted debugging code of the most crap version
(System.out.println)
So we'll see just why it fails. I'll track this bug and remove the code
before a release.
/LS
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@avalon.apache.org
For additional commands, e-mail: dev-help@avalon.apache.org
Re: excalibur-compatibility deadlock problem
Posted by Leo Simons <le...@apache.org>.
Leo Sutic wrote:
>>From: news [mailto:news@main.gmane.org] On Behalf Of Leo Simons
>>
>>is this a known issue?
>
> No, it's not. I am unable to duplicate that error.
>
> You seem to have problems with many testcases that involve multiple
> threads.
the failure with the monitor thing was on a different machine, different
ant version, different junit version, different jdk, and a windows os.
Somehow I doubt these issues are related :D
> If you run "ant test" on the compatibility project (and not GUMP, which
> you seem to be running), do you still get a failure there?
nope! I've tried it on a couple of machines, and a normal 'ant test'
doesn't result in failure, whereas running GUMP (either completely or
using '/data/gump/bin/build excalibur-compatibility', which just
rebuilds the excalibur-compatibility project and nothing else) does
result in the failure. Which is real weird. Maybe it has to do with
using the cvs version of ant or junit.
gotta go for now,
- Leo
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@avalon.apache.org
For additional commands, e-mail: dev-help@avalon.apache.org
Re: excalibur-compatibility deadlock problem
Posted by Leo Simons <le...@apache.org>.
Leo Sutic wrote:
>>From: news [mailto:news@main.gmane.org] On Behalf Of Leo Simons
>>
>>is this a known issue?
>
> No, it's not. I am unable to duplicate that error.
>
> You seem to have problems with many testcases that involve multiple
> threads.
the failure with the monitor thing was on a different machine, different
ant version, different junit version, different jdk, and a windows os.
Somehow I doubt these issues are related :D
> If you run "ant test" on the compatibility project (and not GUMP, which
> you seem to be running), do you still get a failure there?
nope! I've tried it on a couple of machines, and a normal 'ant test'
doesn't result in failure, whereas running GUMP (either completely or
using '/data/gump/bin/build excalibur-compatibility', which just
rebuilds the excalibur-compatibility project and nothing else) does
result in the failure. Which is real weird. Maybe it has to do with
using the cvs version of ant or junit.
gotta go for now,
- Leo
RE: excalibur-compatibility deadlock problem
Posted by Leo Sutic <le...@inspireinfrastructure.com>.
> From: news [mailto:news@main.gmane.org] On Behalf Of Leo Simons
>
> is this a known issue?
No, it's not. I am unable to duplicate that error.
You seem to have problems with many testcases that involve multiple
threads.
If you run "ant test" on the compatibility project (and not GUMP, which
you seem to be running), do you still get a failure there?
/LS
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@avalon.apache.org
For additional commands, e-mail: dev-help@avalon.apache.org