This ticket holds the test results for OSAL 4.2.1 (4.2.0 with #189 changes applied).
Results of interest:
Ran the functional tests and unit tests twice, once for the default mqueue implementation, and once for the socket implementation.
The only difference between them was the expected queue depth semantics problem with sockets, resulting in one expected unit test failure (see #191 discussion).
Execution notes of interest:
Executed on x86_64 machine, -m32 edit in osal-config.mak required.
Encountered #150, and required a manual edit before testing.
There was a minor build problem with the osloader unit test (x86_64 test files were built, instead of the requested 32-bit item under test). A re-build using a makefile in the osloader-test was required. (#200) Without this the osloader unit tests failed on this x86_64 platform.
There were differences (improvements) in these unit test results compared to the OSAL 4.2.0 unit test results in the tarball. The released results had more failures, whereas these current tests had none. Could be a packaging problem? Details:
ut_oscore: all tests now passing (was 3 failures), only one is now not-applicable (was 7), overall several formerly-failing and N/A tests were not executed (quick look: not found in unit test source code either, so looks intentional?).
ut_osfile: all tests now pass (was 1 test setup failure)
ut_osfilesys: all tests pass
ut_osloader & ut_osnetwork: all tests pass, identical results as 4.2.0 files
ut_ostimer: all tests now pass (was 1 failure)
Attached Files:
delivered OSAL 4.2.0 test results from the tarball (for comparison)
mqueue test results for this version
socket test results for this version
test notes
Script used to execute the tests filed under #194.
(In response to cFS CCB AI 117).
This ticket holds the test results for OSAL 4.2.1 (4.2.0 with #189 changes applied).
Results of interest:
Execution notes of interest:
There were differences (improvements) in these unit test results compared to the OSAL 4.2.0 unit test results in the tarball. The released results had more failures, whereas these current tests had none. Could be a packaging problem? Details:
Attached Files: