| Summary: | Most test cases fail when oe-selftest comes from a tarball | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [QA/Testing] Build Testing | Reporter: | Leonardo Sandoval Gonzalez <leonardo.sandoval.gonzalez> | ||||
| Component: | general | Assignee: | Anuj Mittal <anuj.mittal> | ||||
| Status: | RESOLVED NOTABUG | QA Contact: | |||||
| Severity: | normal | ||||||
| Priority: | Medium+ | CC: | anuj.mittal, ee.peng.yeoh, randy.macleod, ross.burton | ||||
| Version: | 2.5 | ||||||
| Target Milestone: | Future | ||||||
| Hardware: | x86 | ||||||
| OS: | Multiple | ||||||
| Whiteboard: | |||||||
| OS type for building Yocto: | --- | Type of Regression: | --- | ||||
| Verified: | Documentation change: | No (bug/feature does not impact docs) | |||||
| Attachments: |
|
||||||
Created attachment 4147 [details]
Log obtained when running oe-selftest -a when the latter comes from a release tarball
They're all failing in the common setup at least... I came across this bug while searching for something and tried it out. I can't reproduce this. From the log attached, it looks like paths included were from two different setups: 2017-12-06 08:02:26,673 - oe-selftest - WARNING - meta-selftest layer not found in BBLAYERS, adding it 2017-12-06 08:02:33,575 - oe-selftest - INFO - Adding layer libraries: 2017-12-06 08:02:33,575 - oe-selftest - INFO - /home/lsandov1/poky-testing/poky-parallel-selftest/meta/lib 2017-12-06 08:02:33,575 - oe-selftest - INFO - /home/lsandov1/poky-testing/poky-parallel-selftest/meta-yocto-bsp/lib 2017-12-06 08:02:33,576 - oe-selftest - INFO - Running bitbake -p 2017-12-06 08:02:35,078 - oe-selftest - INFO - Adding: "include selftest.inc" in /home/lsandov1/tmp/poky-rocko-18.0.0/build/conf/local.conf 2017-12-06 08:02:35,078 - oe-selftest - INFO - Adding: "include bblayers.inc" in bblayers.conf right, so this means that on my setup there was some variable (BB*) polluting the environment. Perhaps I hit into another bug? setting twice the environment causes the second to fail? (In reply to comment #4) > right, so this means that on my setup there was some variable (BB*) > polluting the environment. Perhaps I hit into another bug? setting twice the > environment causes the second to fail? Yes, please see bug 12163. BBPATH wasn't exported from the tarball resulting in selftest scripts/layer to use the old BBPATH that may have been pointing to another builddir. Marking this as resolved. |
Most selftest cases fail when executed with oe-selftest **and** the latter is part of the tarball release [1] This is the exception shown which present multiple times, one for each test case: 2017-12-06 08:05:50,467 - oe-selftest - INFO - ====================================================================== 2017-12-06 08:05:50,467 - oe-selftest - INFO - ERROR: test_archiver_allows_to_filter_on_recipe_name (archiver.Archiver) 2017-12-06 08:05:50,467 - oe-selftest - INFO - ---------------------------------------------------------------------- 2017-12-06 08:05:50,467 - oe-selftest - INFO - Traceback (most recent call last): File "/home/lsandov1/tmp/poky-rocko-18.0.0/meta/lib/oeqa/core/case.py", line 41, in _oeSetUp self.setUpMethod() File "/home/lsandov1/tmp/poky-rocko-18.0.0/meta/lib/oeqa/selftest/case.py", line 131, in setUp for root, _, files in os.walk(self.testlayer_path): File "/usr/lib64/python3.4/os.py", line 371, in walk if isdir(join(top, name)): File "/usr/lib64/python3.4/posixpath.py", line 82, in join path += b TypeError: unsupported operand type(s) for +=: 'NoneType' and 'str' [1] http://downloads.yoctoproject.org/releases/yocto/yocto-2.4/poky-rocko-18.0.0.tar.bz2