circuitpython/tests/run-multitests.py
Damien George af42634866 tests/run-multitests.py: Support outputting test metrics.
If a multitest calls `multitest.output_metric(...)` then that output will
be collected separately, not considered as part of the test verification
output, and instead be printed at the end.  This is useful for tests that
want to output performance/timing metrics that may change from one run to
the next.

Signed-off-by: Damien George <damien@micropython.org>
2023-03-23 13:18:52 +11:00

633 lines
20 KiB
Python
Executable File

#!/usr/bin/env python3
# This file is part of the MicroPython project, http://micropython.org/
# The MIT License (MIT)
# Copyright (c) 2020 Damien P. George
#
# run-multitests.py
# Runs a test suite that relies on two micropython instances/devices
# interacting in some way. Typically used to test networking / bluetooth etc.
import sys, os, time, re, select
import argparse
import itertools
import subprocess
import tempfile
test_dir = os.path.abspath(os.path.dirname(__file__))
if os.path.abspath(sys.path[0]) == test_dir:
# remove the micropython/tests dir from path to avoid
# accidentally importing tests like micropython/const.py
sys.path.pop(0)
sys.path.insert(0, test_dir + "/../tools")
import pyboard
if os.name == "nt":
CPYTHON3 = os.getenv("MICROPY_CPYTHON3", "python3.exe")
MICROPYTHON = os.getenv(
"MICROPY_MICROPYTHON", test_dir + "/../ports/windows/build-standard/micropython.exe"
)
else:
CPYTHON3 = os.getenv("MICROPY_CPYTHON3", "python3")
MICROPYTHON = os.getenv(
"MICROPY_MICROPYTHON", test_dir + "/../ports/unix/build-standard/micropython"
)
# For diff'ing test output
DIFF = os.getenv("MICROPY_DIFF", "diff -u")
PYTHON_TRUTH = CPYTHON3
INSTANCE_READ_TIMEOUT_S = 10
APPEND_CODE_TEMPLATE = """
import sys
class multitest:
@staticmethod
def flush():
try:
sys.stdout.flush()
except AttributeError:
pass
@staticmethod
def skip():
print("SKIP")
multitest.flush()
raise SystemExit
@staticmethod
def next():
print("NEXT")
multitest.flush()
@staticmethod
def broadcast(msg):
print("BROADCAST", msg)
multitest.flush()
@staticmethod
def wait(msg):
msg = "BROADCAST " + msg
while True:
if sys.stdin.readline().rstrip() == msg:
return
@staticmethod
def globals(**gs):
for g in gs:
print("SET {{}} = {{!r}}".format(g, gs[g]))
multitest.flush()
@staticmethod
def get_network_ip():
try:
ip = nic.ifconfig()[0]
except:
try:
import network
if hasattr(network, "WLAN"):
ip = network.WLAN().ifconfig()[0]
else:
ip = network.LAN().ifconfig()[0]
except:
ip = HOST_IP
return ip
@staticmethod
def expect_reboot(resume, delay_ms=0):
print("WAIT_FOR_REBOOT", resume, delay_ms)
@staticmethod
def output_metric(data):
print("OUTPUT_METRIC", data)
{}
instance{}()
multitest.flush()
"""
# The btstack implementation on Unix generates some spurious output that we
# can't control. Also other platforms may output certain warnings/errors that
# can be safely ignored.
IGNORE_OUTPUT_MATCHES = (
"libusb: error ", # It tries to open devices that it doesn't have access to (libusb prints unconditionally).
"hci_transport_h2_libusb.c", # Same issue. We enable LOG_ERROR in btstack.
"USB Path: ", # Hardcoded in btstack's libusb transport.
"hci_number_completed_packet", # Warning from btstack.
"lld_pdu_get_tx_flush_nb HCI packet count mismatch (", # From ESP-IDF, see https://github.com/espressif/esp-idf/issues/5105
)
def get_host_ip(_ip_cache=[]):
if not _ip_cache:
try:
import socket
s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM)
s.connect(("8.8.8.8", 80))
_ip_cache.append(s.getsockname()[0])
s.close()
except:
_ip_cache.append("127.0.0.1")
return _ip_cache[0]
class PyInstance:
def __init__(self):
pass
def close(self):
pass
def prepare_script_from_file(self, filename, prepend, append):
with open(filename, "rb") as f:
script = f.read()
if prepend:
script = bytes(prepend, "ascii") + b"\n" + script
if append:
script += b"\n" + bytes(append, "ascii")
return script
def run_file(self, filename, prepend="", append=""):
return self.run_script(self.prepare_script_from_file(filename, prepend, append))
def start_file(self, filename, prepend="", append=""):
return self.start_script(self.prepare_script_from_file(filename, prepend, append))
class PyInstanceSubProcess(PyInstance):
def __init__(self, argv, env=None):
self.argv = argv
self.env = {n: v for n, v in (i.split("=") for i in env)} if env else None
self.popen = None
self.finished = True
def __str__(self):
return self.argv[0].rsplit("/")[-1]
def run_script(self, script):
output = b""
err = None
try:
p = subprocess.run(
self.argv,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT,
input=script,
env=self.env,
)
output = p.stdout
except subprocess.CalledProcessError as er:
err = er
return str(output.strip(), "ascii"), err
def start_script(self, script):
self.popen = subprocess.Popen(
self.argv + ["-c", script],
stdin=subprocess.PIPE,
stdout=subprocess.PIPE,
stderr=subprocess.STDOUT,
env=self.env,
)
self.finished = False
def stop(self):
if self.popen and self.popen.poll() is None:
self.popen.terminate()
def readline(self):
sel = select.select([self.popen.stdout.raw], [], [], 0.001)
if not sel[0]:
self.finished = self.popen.poll() is not None
return None, None
out = self.popen.stdout.raw.readline()
if out == b"":
self.finished = self.popen.poll() is not None
return None, None
else:
return str(out.rstrip(), "ascii"), None
def write(self, data):
self.popen.stdin.write(data)
self.popen.stdin.flush()
def is_finished(self):
return self.finished
def wait_finished(self):
self.popen.wait()
out = self.popen.stdout.read()
return str(out, "ascii"), ""
class PyInstancePyboard(PyInstance):
@staticmethod
def map_device_shortcut(device):
if device[0] == "a" and device[1:].isdigit():
return "/dev/ttyACM" + device[1:]
elif device[0] == "u" and device[1:].isdigit():
return "/dev/ttyUSB" + device[1:]
else:
return device
def __init__(self, device):
device = self.map_device_shortcut(device)
self.device = device
self.pyb = pyboard.Pyboard(device)
self.pyb.enter_raw_repl()
self.finished = True
def __str__(self):
return self.device.rsplit("/")[-1]
def close(self):
self.pyb.exit_raw_repl()
self.pyb.close()
def run_script(self, script):
output = b""
err = None
try:
self.pyb.enter_raw_repl()
output = self.pyb.exec_(script)
except pyboard.PyboardError as er:
err = er
return str(output.strip(), "ascii"), err
def start_script(self, script):
self.pyb.enter_raw_repl()
self.pyb.exec_raw_no_follow(script)
self.finished = False
def stop(self):
self.pyb.serial.write(b"\r\x03")
def readline(self):
if self.finished:
return None, None
if self.pyb.serial.inWaiting() == 0:
return None, None
out = self.pyb.read_until(1, (b"\r\n", b"\x04"))
if out.endswith(b"\x04"):
self.finished = True
out = out[:-1]
err = str(self.pyb.read_until(1, b"\x04"), "ascii")
err = err[:-1]
if not out and not err:
return None, None
else:
err = None
return str(out.rstrip(), "ascii"), err
def write(self, data):
self.pyb.serial.write(data)
def is_finished(self):
return self.finished
def wait_finished(self):
out, err = self.pyb.follow(10, None)
return str(out, "ascii"), str(err, "ascii")
def prepare_test_file_list(test_files):
test_files2 = []
for test_file in sorted(test_files):
num_instances = 0
with open(test_file) as f:
for line in f:
m = re.match(r"def instance([0-9]+)\(\):", line)
if m:
num_instances = max(num_instances, int(m.group(1)) + 1)
test_files2.append((test_file, num_instances))
return test_files2
def trace_instance_output(instance_idx, line):
if cmd_args.trace_output:
t_ms = round((time.time() - trace_t0) * 1000)
print("{:6} i{} :".format(t_ms, instance_idx), line)
sys.stdout.flush()
def run_test_on_instances(test_file, num_instances, instances):
global trace_t0
trace_t0 = time.time()
error = False
skip = False
injected_globals = ""
output = [[] for _ in range(num_instances)]
output_metrics = []
# If the test calls get_network_ip() then inject HOST_IP so that devices can know
# the IP address of the host. Do this lazily to not require a TCP/IP connection
# on the host if it's not needed.
with open(test_file, "rb") as f:
if b"get_network_ip" in f.read():
injected_globals += "HOST_IP = '" + get_host_ip() + "'\n"
if cmd_args.trace_output:
print("TRACE {}:".format("|".join(str(i) for i in instances)))
# Start all instances running, in order, waiting until they signal they are ready
for idx in range(num_instances):
append_code = APPEND_CODE_TEMPLATE.format(injected_globals, idx)
instance = instances[idx]
instance.start_file(test_file, append=append_code)
last_read_time = time.time()
while True:
if instance.is_finished():
break
out, err = instance.readline()
if out is None and err is None:
if time.time() > last_read_time + INSTANCE_READ_TIMEOUT_S:
output[idx].append("TIMEOUT")
error = True
break
time.sleep(0.1)
continue
last_read_time = time.time()
if out is not None and not any(m in out for m in IGNORE_OUTPUT_MATCHES):
trace_instance_output(idx, out)
if out.startswith("SET "):
injected_globals += out[4:] + "\n"
elif out == "SKIP":
skip = True
break
elif out == "NEXT":
break
else:
output[idx].append(out)
if err is not None:
trace_instance_output(idx, err)
output[idx].append(err)
error = True
if error or skip:
break
if not error and not skip:
# Capture output and wait for all instances to finish running
last_read_time = [time.time() for _ in range(num_instances)]
while True:
num_running = 0
num_output = 0
for idx in range(num_instances):
instance = instances[idx]
if instance.is_finished():
continue
num_running += 1
out, err = instance.readline()
if out is None and err is None:
if time.time() > last_read_time[idx] + INSTANCE_READ_TIMEOUT_S:
output[idx].append("TIMEOUT")
error = True
continue
num_output += 1
last_read_time[idx] = time.time()
if out is not None and not any(m in out for m in IGNORE_OUTPUT_MATCHES):
trace_instance_output(idx, out)
if out.startswith("WAIT_FOR_REBOOT"):
_, resume, delay_ms = out.split(" ")
if wait_for_reboot(instance, delay_ms):
# Restart the test code, resuming from requested instance block
if not resume.startswith("instance{}".format(idx)):
raise SystemExit(
'ERROR: resume function must start with "instance{}"'.format(
idx
)
)
append_code = APPEND_CODE_TEMPLATE.format(injected_globals, resume[8:])
instance.start_file(test_file, append=append_code)
last_read_time[idx] = time.time()
if out.startswith("BROADCAST "):
for instance2 in instances:
if instance2 is not instance:
instance2.write(bytes(out, "ascii") + b"\r\n")
elif out.startswith("OUTPUT_METRIC "):
output_metrics.append(out.split(" ", 1)[1])
else:
output[idx].append(out)
if err is not None:
trace_instance_output(idx, err)
output[idx].append(err)
error = True
if not num_output:
time.sleep(0.1)
if not num_running or error:
break
# Stop all instances
for idx in range(num_instances):
instances[idx].stop()
output_str = ""
for idx, lines in enumerate(output):
output_str += "--- instance{} ---\n".format(idx)
output_str += "\n".join(lines) + "\n"
return error, skip, output_str, output_metrics
def wait_for_reboot(instance, extra_timeout_ms=0):
# Monitor device responses for reboot banner, waiting for idle.
extra_timeout = float(extra_timeout_ms) * 1000
INITIAL_TIMEOUT = 1 + extra_timeout
FULL_TIMEOUT = 5 + extra_timeout
t_start = t_last_activity = time.monotonic()
while True:
t = time.monotonic()
out, err = instance.readline()
if err is not None:
print("Reboot: communication error", err)
return False
if out:
t_last_activity = t
# Check for reboot banner, see py/pyexec.c "reset friendly REPL"
if re.match(r"^MicroPython v\d+\.\d+\.\d+.* on .*; .* with .*$", out):
time.sleep(0.1)
break
if t_last_activity == t_start:
if t - t_start > INITIAL_TIMEOUT:
print("Reboot: missed initial Timeout")
return False
else:
if t - t_start > FULL_TIMEOUT:
print("Reboot: Timeout")
return False
instance.pyb.enter_raw_repl()
return True
def print_diff(a, b):
a_fd, a_path = tempfile.mkstemp(text=True)
b_fd, b_path = tempfile.mkstemp(text=True)
os.write(a_fd, a.encode())
os.write(b_fd, b.encode())
os.close(a_fd)
os.close(b_fd)
subprocess.run(DIFF.split(" ") + [a_path, b_path])
os.unlink(a_path)
os.unlink(b_path)
def run_tests(test_files, instances_truth, instances_test):
skipped_tests = []
passed_tests = []
failed_tests = []
for test_file, num_instances in test_files:
instances_str = "|".join(str(instances_test[i]) for i in range(num_instances))
print("{} on {}: ".format(test_file, instances_str), end="")
if cmd_args.show_output or cmd_args.trace_output:
print()
sys.stdout.flush()
# Run test on test instances
error, skip, output_test, output_metrics = run_test_on_instances(
test_file, num_instances, instances_test
)
if not skip:
# Check if truth exists in a file, and read it in
test_file_expected = test_file + ".exp"
if os.path.isfile(test_file_expected):
with open(test_file_expected) as f:
output_truth = f.read()
else:
# Run test on truth instances to get expected output
_, _, output_truth, _ = run_test_on_instances(
test_file, num_instances, instances_truth
)
if cmd_args.show_output:
print("### TEST ###")
print(output_test, end="")
if not skip:
print("### TRUTH ###")
print(output_truth, end="")
# Print result of test
if skip:
print("skip")
skipped_tests.append(test_file)
elif output_test == output_truth:
print("pass")
passed_tests.append(test_file)
else:
print("FAIL")
failed_tests.append(test_file)
if not cmd_args.show_output:
print("### TEST ###")
print(output_test, end="")
print("### TRUTH ###")
print(output_truth, end="")
print("### DIFF ###")
print_diff(output_truth, output_test)
# Print test output metrics, if there are any.
if output_metrics:
for metric in output_metrics:
print(test_file, ": ", metric, sep="")
if cmd_args.show_output:
print()
print("{} tests performed".format(len(skipped_tests) + len(passed_tests) + len(failed_tests)))
print("{} tests passed".format(len(passed_tests)))
if skipped_tests:
print("{} tests skipped: {}".format(len(skipped_tests), " ".join(skipped_tests)))
if failed_tests:
print("{} tests failed: {}".format(len(failed_tests), " ".join(failed_tests)))
return not failed_tests
def main():
global cmd_args
cmd_parser = argparse.ArgumentParser(
description="Run network tests for MicroPython",
formatter_class=argparse.RawTextHelpFormatter,
)
cmd_parser.add_argument(
"-s", "--show-output", action="store_true", help="show test output after running"
)
cmd_parser.add_argument(
"-t", "--trace-output", action="store_true", help="trace test output while running"
)
cmd_parser.add_argument(
"-i", "--instance", action="append", default=[], help="instance(s) to run the tests on"
)
cmd_parser.add_argument(
"-p",
"--permutations",
type=int,
default=1,
help="repeat the test with this many permutations of the instance order",
)
cmd_parser.epilog = (
"Supported instance types:\r\n"
" -i pyb:<port> physical device (eg. pyboard) on provided repl port.\n"
" -i micropython unix micropython instance, path customised with MICROPY_MICROPYTHON env.\n"
" -i cpython desktop python3 instance, path customised with MICROPY_CPYTHON3 env.\n"
" -i exec:<path> custom program run on provided path.\n"
"Each instance arg can optionally have custom env provided, eg. <cmd>,ENV=VAR,ENV=VAR...\n"
)
cmd_parser.add_argument("files", nargs="+", help="input test files")
cmd_args = cmd_parser.parse_args()
# clear search path to make sure tests use only builtin modules and those in extmod
os.environ["MICROPYPATH"] = os.pathsep.join(("", ".frozen", "../extmod"))
test_files = prepare_test_file_list(cmd_args.files)
max_instances = max(t[1] for t in test_files)
instances_truth = [PyInstanceSubProcess([PYTHON_TRUTH]) for _ in range(max_instances)]
instances_test = []
for i in cmd_args.instance:
# Each instance arg is <cmd>,ENV=VAR,ENV=VAR...
i = i.split(",")
cmd = i[0]
env = i[1:]
if cmd.startswith("exec:"):
instances_test.append(PyInstanceSubProcess([cmd[len("exec:") :]], env))
elif cmd == "micropython":
instances_test.append(PyInstanceSubProcess([MICROPYTHON], env))
elif cmd == "cpython":
instances_test.append(PyInstanceSubProcess([CPYTHON3], env))
elif cmd.startswith("pyb:"):
instances_test.append(PyInstancePyboard(cmd[len("pyb:") :]))
else:
print("unknown instance string: {}".format(cmd), file=sys.stderr)
sys.exit(1)
for _ in range(max_instances - len(instances_test)):
instances_test.append(PyInstanceSubProcess([MICROPYTHON]))
all_pass = True
try:
for i, instances_test_permutation in enumerate(itertools.permutations(instances_test)):
if i >= cmd_args.permutations:
break
all_pass &= run_tests(test_files, instances_truth, instances_test_permutation)
finally:
for i in instances_truth:
i.close()
for i in instances_test:
i.close()
if not all_pass:
sys.exit(1)
if __name__ == "__main__":
main()