PR#1355: Do not build package module-build-macros for tests in test_build.py - fm-orchestrator

fm-orchestrator

#1355 Do not build package module-build-macros for tests in test_build.py

Closed 4 years ago by cqi. Opened 4 years ago by cqi.

cqi/fm-orchestrator dont-build-module-build-macros-for-test_build into master

Do not build package module-build-macros for tests in test_build.py

Chenxiong Qi • 4 years ago

4ba6954

tests/test_build/test_build.py

file modified

+22

		`@@ -357,6 +357,28 @@`

		`class BaseTestBuild:`

		`+ @classmethod`
		`+ def setup_class(cls):`
		`+ # Do not really build module-build-macros SRPM`
		+ # get_disttag_rpm builds package module-build-macros by `rpmbuild -bs'
		`+ # and return the path of result SRPM.`
		`+`
		`+ cls.p_Popen = patch("subprocess.Popen")`
		`+ cls.mock_Popen = cls.p_Popen.start()`
		`+ cls.mock_Popen.return_value.returncode = 0`
		`+ cls.mock_Popen.return_value.communicate.return_value = ('', '')`
		`+`
		`+ # Let get_disttag_rpm return a fake SRPM path`
		`+ cls.p_glob = patch(`
		`+ "glob.glob",`
		`+ return_value=["/tmp/module-build/module-build-macros.src.rpm"])`
		`+ cls.mock_glob = cls.p_glob.start()`
		`+`
		`+ @classmethod`
		`+ def teardown_class(cls):`
		`+ cls.p_glob.stop()`
		`+ cls.p_Popen.stop()`
		`+`
		`def run_scheduler(self, db_session, msgs=None, stop_condition=None):`
		`module_build_service.scheduler.main(`
		`msgs or [],`

cqi commented 4 years ago

Signed-off-by: Chenxiong Qi cqi@redhat.com

jkaluza commented on line 16 of tests/test_build/test_build.py 4 years ago

get_disttag_rpm

jkaluza commented 4 years ago

What's the main reason to do it here? To remove the rpmbuild dependency?

rebased onto 4ba6954

4 years ago

cqi commented 4 years ago

@jkaluza Fixed that typo. Regarding the removal, I found out the built module-build-macros SRPM is not used in those tests, and this change could make tests run faster.

jkaluza commented 4 years ago

@cqi, I've tested this and for me it makes the tests 10 seconds slower, not sure why :(. Can you try yourself with and without your patch and check the results?

cqi commented 4 years ago

@jkaluza Interesting. Not sure why as well. However, I suspect it could probably because of the virtual machine I'm using and run tests inside. I test it again with and without this patch, there is no significant difference this time. rpmbuild -bs is fast enough. Seems this patch is useless. I'm going to close it.

Pull-Request has been closed by cqi

4 years ago