Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
RE: [cdt-dev] Problem with executing dsf-gdb JUnit tests

I've fixed this last remaining floating point math issue.

So, as for the intermittent failures, we'll have to look into them. Can you please open a bugzilla report and indicate your target environment (OS and toolchain/gdb vendor and versions) you are testing?

Thanks,
John

At 12:22 AM 3/11/2010, sneha_apte wrote:
Content-Language: en-US
Content-Type: multipart/alternative;
         boundary="_000_14B8F27D7CE40C4C9E481B2B845C2E0D03E06E8EA4EXCHANGEpersi_"

Hi,
 
I tested the updated plugin.
I can still observe the failure for same test case.
Following is the details of the failed test case:
java.lang.AssertionError: "Failed to correctly evalutate '3.1415 + 1.1111': expected '4.2526', got '4.2525'",
"Failed to correctly evalutate '3.1415 + 1.1111': expected '4.2526', got '4.2525'"
 
As you have mentioned, the output of the binaries is same for both cases.
Also after executing more rounds of tests, I could observe that the errors are intermittent.
The test cases for which errors are obtained are not same for every execution.         
 
But the failure mentioned above is consistent for all executions.
 
Thanks,
Sneha
 
 
 
From: cdt-dev-bounces@xxxxxxxxxxx [ mailto:cdt-dev-bounces@xxxxxxxxxxx] On Behalf Of John Cortell
Sent: Thursday, March 11, 2010 3:12 AM
To: CDT General developers list.; CDT General developers list.
Subject: RE: [cdt-dev] Problem with executing dsf-gdb JUnit tests
 
As for the first issue, I've loosened expectations for tests involving floating point math. I committed the changes to MIExpressionsTest.java. Please check out and verify

As for the other failures you're seeing when altering the order of the gcc switches, that doesn't make much sense to me. Any chance these are just intermittent failures that seem to be tied to the switch ordering but really aren't? The first thing I'd check is whether the output of the build is acutally different. Do a binary compare of the executables built each way. If there's no difference, then it's some intermittent failure you're seeing not having anything to do with the switches.

John


At 02:26 AM 3/10/2010, sneha_apte wrote:

Hi,

Following test case fails while executing JUnit Plugin Tests for Suite 6.8 after adding -m32 in GCC flags.

In the class MIExpressionsTest.java, following test case is failing:
public void testLiteralFloatingPointExpressions() throws Throwable

After debugging the code we observed that an Assertion error with following message is logged:
"Failed to correctly evaluate '100.0 / 3.0': expected '33.333333333333336', got '33.333333333333329'"

In addition to above findings, I observed one more strange behavior.
If the tests are executed with GCCFLAGS = -g -m32 -pthread then only above mentioned error is observed.
However if tests are executed with GCCFLAGS = -g -pthread -m32 then following problems seem to occur:

In the current /data/launch/src/Makefile, -m32 seems to be added to GCC flags at the end.
On execution with this change, we observed errors and failures in following test cases:
1 Failure:  This is same as mentioned above (In the class MIExpressionsTest.java, testLiteralFloatingPointExpressions()).
2 Errors:
1) Null Pointer Exception from test case writeRegisterHEXFormat() in class MIRegistersTest.java
2) NullPointerException from test case fillInvalidCount() in class MIMemoryTest.java

 I am not sure if the sequence of adding -m32 is affecting that error or not.
But above mentioned results are obtained by making only this change in the code.

I will continue checking this. If anybody has any suggestions / comments, please let me know.

-Thanks and Regards,
Sneha

-----Original Message-----
From: cdt-dev-bounces@xxxxxxxxxxx [ mailto:cdt-dev-bounces@xxxxxxxxxxx] On Behalf Of Marc Khouzam
Sent: Tuesday, March 09, 2010 8:06 PM
To: CDT General developers list.
Subject: RE: [cdt-dev] Problem with executing dsf-gdb JUnit tests


[...]

> To resolve this issue I had to add -m32 in GCC_FLAGS of
> data/launch/src/Makefile.
> With this change, I faced no issues in test execution.

Cool.  I have added this flag to the Makefile.
John, can you confirm that it does not cause a problem on Windows?
 
> Results of tests for gdb 6.8:
> On running 6.8 Test Suite we can observe 138 test cases with
> 1 failure.

Which one fails?
I'm getting everything passing for 6.8, on Linux 32bit.


 
_______________________________________________
cdt-dev mailing list
cdt-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cdt-dev

DISCLAIMER ========== This e-mail may contain privileged and confidential information which is the property of Persistent Systems Ltd. It is intended only for the use of the individual or entity to which it is addressed. If you are not the intended recipient, you are not authorized to read, retain, copy, print, distribute or use this message. If you have received this communication in error, please notify the sender and delete all copies of this message. Persistent Systems Ltd. does not accept any liability for virus infected mails.
_______________________________________________
cdt-dev mailing list
cdt-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cdt-dev

DISCLAIMER ========== This e-mail may contain privileged and confidential information which is the property of Persistent Systems Ltd. It is intended only for the use of the individual or entity to which it is addressed. If you are not the intended recipient, you are not authorized to read, retain, copy, print, distribute or use this message. If you have received this communication in error, please notify the sender and delete all copies of this message. Persistent Systems Ltd. does not accept any liability for virus infected mails.
_______________________________________________
cdt-dev mailing list
cdt-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/cdt-dev

Back to the top