Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
RE: [cdt-dev] Issues with unit tests for CDT 1.2.1 RC1

	Hi, Alain!

I've made some research on the problem with org.eclipse.cdt.debug.core.tests.LocationTests.testIsEqual(). The test fails at line 172 where it is checked whether the length of the breakpoints array is 1: 

But before that first all of existing breakpoints are removed and then another one is created, so the assertion should pass successfully (and it does when the test is run step by step in debug mode). So I think that the problem is in poor synchronisation of threads - under normal conditions (not step-by-step running) the code where the deletion is actually performed runs later than the assertion.



-----Original Message-----
From: cdt-dev-admin@xxxxxxxxxxx [mailto:cdt-dev-admin@xxxxxxxxxxx]On Behalf Of Alain Magloire
Sent: Wednesday, February 18, 2004 7:09 PM
To: cdt-dev@xxxxxxxxxxx
Subject: Re: [cdt-dev] Issues with unit tests for CDT 1.2.1 RC1

> 	Hi, Alain!
> Thanks for your response! I'm running the tests on a single P4 processor =
> under Red Hat EL Workstation 3 using Sun 1.4.2 JVM.
> I've checked the tools you wrote about (addr2line/c++filt) and I found =
> that they present in my path. So it doesn't seem to be the reason. But I =
> did the following: I checked out a subsequent revision (1.3) of =
> org.eclipse.cdt.core.model.tests.BinaryTests from CVS. The revision I =
> used before was 1.2 tagged with CDT_1_2_1_RC1. The later revision =
> differs by some addition in the code of the setUp method. As I could =
> understand it the addition replaces the default binary parser used by =
> CCorePlugin by GNUElfParser, which is able to extract debug symbols from =
> binaries as it is seen from the comment. After I built the test plugin =
> with the newer revision of BinaryTests the testGetChildren method =
> started to work correctly and now all of 560 core tests pass =
> successfully. Alain, I'm not sure that the decision I applied is =
> absolutely correct - it more looks like a kind of hack. But it makes me =
> think that in the future final release of CDT 1.2.1 the problem with =
> BinaryTests.testGetChildren() is going to disappear.

Yes, the problem was fix in the branch.  I would recommand to use
the latest branch of CDT_1.2, it contains many fixes.

> Well, I'll work more on the second issue with LocationTests.testIsEquals =
> and if I find something I'll post it here.


cdt-dev mailing list

Back to the top