testutils.github: Only first result of parametrized test is reported

[miri64]

Task 1.3 is parametrized to run the unittests seperately and give a nicely parsable result:

Release-Specs/01-ci/test_spec01.py

Lines 35 to 39 in 676d417

	@pytest.mark.parametrize('nodes, test_suite',
	[pytest.param(['native'], t) for t in TESTS_SUITES],
	indirect=['nodes'])
	def test_task03(nodes, log_nodes, riotbase, test_suite):
	run_unittests(test_suite, nodes, log_nodes, os.path.join(riotbase, APP))

However, this has the disadvantage, that with the current implementation of the testutils.github modules, only the first test is reported: For each test report, the bot checks the list of tasks in OP and only makes a comment if the test is not marked as done (i.e. if the tick box is empty):

Release-Specs/testutils/github.py

Lines 182 to 192 in 676d417

	try:
	task_line, task = find_task_text(issue.body, tested_task)
	except GithubException as e:
	logger.error("Unable to get issue text of {}: {}"
	.format(issue, e))
	return
	if not task_line or not task:
	logger.warning("Unable to find task {spec}.{task} in the "
	"tracking issue".format(**tested_task))
	elif not task["done"]:
	comment = make_comment(pytest_report, issue, task)

However, it then marks the tasks as done, when the test passed:

Release-Specs/testutils/github.py

Lines 192 to 199 in 676d417

	comment = make_comment(pytest_report, issue, task)
	if comment:
	comment_url = comment.html_url
	else:
	comment_url = None
	if pytest_report.outcome == "passed":
	mark_task_done(github.get_user().login, comment_url, issue,
	task_line, tested_task)

This could lead to a misleading report, as all subsequent runs of the test (with different parameters) are not shown.