analyze_oucomes: do not return Results instance passed as parameter

Signed-off-by: Valerio Setti <valerio.setti@nordicsemi.no>
This commit is contained in:
Valerio Setti 2023-10-17 12:47:35 +02:00
parent 8070dbec6b
commit 781c23416e

View file

@ -57,7 +57,7 @@ class TestCaseOutcomes:
return len(self.successes) + len(self.failures)
def execute_reference_driver_tests(results: Results, ref_component, driver_component, \
outcome_file) -> Results:
outcome_file):
"""Run the tests specified in ref_component and driver_component. Results
are stored in the output_file and they will be used for the following
coverage analysis"""
@ -66,7 +66,7 @@ def execute_reference_driver_tests(results: Results, ref_component, driver_compo
if os.path.exists(outcome_file):
results.info("Outcome file (" + outcome_file + ") already exists. " + \
"Tests will be skipped.")
return results
return
shell_command = "tests/scripts/all.sh --outcome-file " + outcome_file + \
" " + ref_component + " " + driver_component
@ -76,8 +76,6 @@ def execute_reference_driver_tests(results: Results, ref_component, driver_compo
if ret_val != 0:
results.error("failed to run reference/driver components")
return results
def analyze_coverage(results, outcomes, allow_list, full_coverage):
"""Check that all available test cases are executed at least once."""
available = check_test_cases.collect_available_test_cases()
@ -132,13 +130,10 @@ def analyze_driver_vs_reference(results: Results, outcomes,
if(reference_test_passed and not driver_test_passed):
results.error(key)
return results
def analyze_outcomes(results: Results, outcomes, args) -> Results:
def analyze_outcomes(results: Results, outcomes, args):
"""Run all analyses on the given outcome collection."""
analyze_coverage(results, outcomes, args['allow_list'],
args['full_coverage'])
return results
def read_outcome_file(outcome_file):
"""Parse an outcome file and return an outcome collection.
@ -161,30 +156,27 @@ by a semicolon.
outcomes[key].failures.append(setup)
return outcomes
def do_analyze_coverage(results: Results, outcome_file, args) -> Results:
def do_analyze_coverage(results: Results, outcome_file, args):
"""Perform coverage analysis."""
results.info("*** Analyze coverage ***")
outcomes = read_outcome_file(outcome_file)
results = analyze_outcomes(results, outcomes, args)
return results
analyze_outcomes(results, outcomes, args)
def do_analyze_driver_vs_reference(results: Results, outcome_file, args) -> Results:
def do_analyze_driver_vs_reference(results: Results, outcome_file, args):
"""Perform driver vs reference analyze."""
results.info("*** Analyze driver {} vs reference {} ***".format(
args['component_driver'], args['component_ref']))
results = execute_reference_driver_tests(results, args['component_ref'], \
args['component_driver'], outcome_file)
execute_reference_driver_tests(results, args['component_ref'], \
args['component_driver'], outcome_file)
ignored_suites = ['test_suite_' + x for x in args['ignored_suites']]
outcomes = read_outcome_file(outcome_file)
results = analyze_driver_vs_reference(results, outcomes,
args['component_ref'], args['component_driver'],
ignored_suites, args['ignored_tests'])
return results
analyze_driver_vs_reference(results, outcomes,
args['component_ref'], args['component_driver'],
ignored_suites, args['ignored_tests'])
# List of tasks with a function that can handle this task and additional arguments if required
KNOWN_TASKS = {
@ -688,7 +680,7 @@ def main():
for task in tasks_list:
test_function = KNOWN_TASKS[task]['test_function']
test_args = KNOWN_TASKS[task]['args']
main_results = test_function(main_results, options.outcomes, test_args)
test_function(main_results, options.outcomes, test_args)
main_results.info("Overall results: {} warnings and {} errors",
main_results.warning_count, main_results.error_count)