You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by Stefano Bagnara <ap...@bago.org> on 2008/08/29 15:30:08 UTC
[javaflow] wrong assertions in AbstractSimpleTestCase
Hi all,
i looked in depth in some test failure for javaflow and I think some of
them are bugs in the test suite and not in the code.
E.g:
1) the class ClassAccess2.java does not have a Continuation.suspend()
so it is ok that the caller does not fine a Continuation. Adding a
Continuation.suspend before returning fixed the issue.
2) AbstractSimpleTestCase expects the wrong results for the try/catch issue.
2A) testSimpleTryCatchWithoutException expect 3 continuations (note the
last assertTrue(c != null)) but this is wrong. When no exception happen
the continuations will be 2.
2B) testSimpleTryCatchWithException expect 3 continuations, expect the
wrong result on the 3rd and does not complete/test the last continuation.
Am I totally wrong or they are issues in the test suite?
Should I open a JIRA for this and attach a patch?
AFAICT the try/catch test works fine once the testsuite assertions have
been corrected, the only failure is on the synchronized-test.
About the synchronized block and the illegal monitor i'm not sure I
understand how it should work.. should javaflow keep the lock open
during the suspension or tracking the locked objects and try to lock
them again on resume?
Stefano
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [javaflow] wrong assertions in AbstractSimpleTestCase
Posted by Stefano Bagnara <ap...@bago.org>.
Torsten Curdt ha scritto:
>> Am I totally wrong or they are issues in the test suite?
>
> Will have a closer look when I am back from vacation. About to walk out
> the door. Back online in two weeks.
So you should be back, or you should do soon.. Hope you are relaxed
because I prepared some work for you ;-)
>> Should I open a JIRA for this and attach a patch?
>
> Sounds good :)
Here we are:
- https://issues.apache.org/jira/browse/SANDBOX-253
Fix pom dependencies (update asm to latest 2.x and fix jci to a
non-snapshot)
- https://issues.apache.org/jira/browse/SANDBOX-254
wrong expectations in failing junit tests (SimpleTestCase)
- https://issues.apache.org/jira/browse/SANDBOX-255
synchronized(obj) support is missing (locks are not released when
suspending resulting in IllegalMonitorExceptions)
>> AFAICT the try/catch test works fine once the testsuite assertions
>> have been corrected, the only failure is on the synchronized-test.
>
> Hm
Please, have a look at SANDBOX-254. I hope I understood enough of
javaflow and that the patch is good!
>> About the synchronized block and the illegal monitor i'm not sure I
>> understand how it should work.. should javaflow keep the lock open
>> during the suspension or tracking the locked objects and try to lock
>> them again on resume?
>
> No. The lock needs to be released.
Done in SANDBOX-255. Please review.
I had to introduce a new Frame extension "MonitoringFrame" that also
take care of tracking monitorenter/monitorexit so in any given frame it
knows what local variables are locked.
The test included in javaflow and a more advanced test I wrote for it
seems to work, but I'm sure you can think of better tests that can prove
limits for this approach.
Stefano
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org
Re: [javaflow] wrong assertions in AbstractSimpleTestCase
Posted by Torsten Curdt <tc...@apache.org>.
> Am I totally wrong or they are issues in the test suite?
Will have a closer look when I am back from vacation. About to walk
out the door. Back online in two weeks.
> Should I open a JIRA for this and attach a patch?
Sounds good :)
> AFAICT the try/catch test works fine once the testsuite assertions
> have been corrected, the only failure is on the synchronized-test.
Hm
> About the synchronized block and the illegal monitor i'm not sure I
> understand how it should work.. should javaflow keep the lock open
> during the suspension or tracking the locked objects and try to lock
> them again on resume?
No. The lock needs to be released.
Sorry ...later!
cheers
--
Torsten
---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@commons.apache.org
For additional commands, e-mail: dev-help@commons.apache.org