tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
#!/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
|
2022-09-19 18:54:53 -04:00
|
|
|
#
|
|
|
|
# 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.
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
|
|
|
|
import sys, os, time, re, select
|
|
|
|
import argparse
|
2020-11-03 01:41:58 -05:00
|
|
|
import itertools
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
import subprocess
|
2020-09-14 04:11:19 -04:00
|
|
|
import tempfile
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
|
2022-09-19 18:54:58 -04:00
|
|
|
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")
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
import pyboard
|
|
|
|
|
|
|
|
if os.name == "nt":
|
|
|
|
CPYTHON3 = os.getenv("MICROPY_CPYTHON3", "python3.exe")
|
2022-11-29 10:58:21 -05:00
|
|
|
MICROPYTHON = os.getenv(
|
|
|
|
"MICROPY_MICROPYTHON", test_dir + "/../ports/windows/build-standard/micropython.exe"
|
|
|
|
)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
else:
|
|
|
|
CPYTHON3 = os.getenv("MICROPY_CPYTHON3", "python3")
|
2022-09-19 18:54:58 -04:00
|
|
|
MICROPYTHON = os.getenv(
|
|
|
|
"MICROPY_MICROPYTHON", test_dir + "/../ports/unix/build-standard/micropython"
|
|
|
|
)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
|
2020-09-14 04:11:19 -04:00
|
|
|
# For diff'ing test output
|
|
|
|
DIFF = os.getenv("MICROPY_DIFF", "diff -u")
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
PYTHON_TRUTH = CPYTHON3
|
|
|
|
|
|
|
|
INSTANCE_READ_TIMEOUT_S = 10
|
|
|
|
|
|
|
|
APPEND_CODE_TEMPLATE = """
|
|
|
|
import sys
|
|
|
|
class multitest:
|
|
|
|
@staticmethod
|
|
|
|
def flush():
|
2021-04-25 10:37:11 -04:00
|
|
|
try:
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
sys.stdout.flush()
|
2021-04-25 10:37:11 -04:00
|
|
|
except AttributeError:
|
|
|
|
pass
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
@staticmethod
|
|
|
|
def skip():
|
|
|
|
print("SKIP")
|
|
|
|
multitest.flush()
|
|
|
|
raise SystemExit
|
|
|
|
@staticmethod
|
|
|
|
def next():
|
|
|
|
print("NEXT")
|
|
|
|
multitest.flush()
|
|
|
|
@staticmethod
|
2021-08-12 01:49:27 -04:00
|
|
|
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
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
def globals(**gs):
|
|
|
|
for g in gs:
|
|
|
|
print("SET {{}} = {{!r}}".format(g, gs[g]))
|
|
|
|
multitest.flush()
|
|
|
|
@staticmethod
|
|
|
|
def get_network_ip():
|
|
|
|
try:
|
2022-04-29 08:47:58 -04:00
|
|
|
ip = nic.ifconfig()[0]
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
except:
|
2022-04-29 08:47:58 -04:00
|
|
|
try:
|
|
|
|
import network
|
|
|
|
if hasattr(network, "WLAN"):
|
|
|
|
ip = network.WLAN().ifconfig()[0]
|
|
|
|
else:
|
|
|
|
ip = network.LAN().ifconfig()[0]
|
|
|
|
except:
|
|
|
|
ip = HOST_IP
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
return ip
|
|
|
|
|
|
|
|
{}
|
|
|
|
|
|
|
|
instance{}()
|
|
|
|
multitest.flush()
|
|
|
|
"""
|
|
|
|
|
2020-04-07 01:03:31 -04:00
|
|
|
# The btstack implementation on Unix generates some spurious output that we
|
2022-01-16 22:23:24 -05:00
|
|
|
# can't control. Also other platforms may output certain warnings/errors that
|
|
|
|
# can be safely ignored.
|
2020-04-07 01:03:31 -04:00
|
|
|
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.
|
2022-01-16 22:23:24 -05:00
|
|
|
"lld_pdu_get_tx_flush_nb HCI packet count mismatch (", # From ESP-IDF, see https://github.com/espressif/esp-idf/issues/5105
|
2020-04-07 01:03:31 -04:00
|
|
|
)
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
|
2022-01-17 01:15:31 -05:00
|
|
|
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]
|
|
|
|
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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):
|
2020-06-05 01:41:40 -04:00
|
|
|
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
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
self.popen = None
|
|
|
|
self.finished = True
|
|
|
|
|
|
|
|
def __str__(self):
|
2020-06-05 01:41:40 -04:00
|
|
|
return self.argv[0].rsplit("/")[-1]
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
|
|
|
|
def run_script(self, script):
|
|
|
|
output = b""
|
|
|
|
err = None
|
|
|
|
try:
|
|
|
|
p = subprocess.run(
|
2020-06-05 01:41:40 -04:00
|
|
|
self.argv,
|
|
|
|
stdout=subprocess.PIPE,
|
|
|
|
stderr=subprocess.STDOUT,
|
|
|
|
input=script,
|
|
|
|
env=self.env,
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
)
|
|
|
|
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(
|
2021-08-12 01:49:27 -04:00
|
|
|
self.argv + ["-c", script],
|
2020-06-05 01:41:40 -04:00
|
|
|
stdin=subprocess.PIPE,
|
|
|
|
stdout=subprocess.PIPE,
|
|
|
|
stderr=subprocess.STDOUT,
|
|
|
|
env=self.env,
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
)
|
|
|
|
self.finished = False
|
|
|
|
|
|
|
|
def stop(self):
|
|
|
|
if self.popen and self.popen.poll() is None:
|
|
|
|
self.popen.terminate()
|
|
|
|
|
|
|
|
def readline(self):
|
2021-08-12 01:49:27 -04:00
|
|
|
sel = select.select([self.popen.stdout.raw], [], [], 0.001)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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
|
|
|
|
|
2021-08-12 01:49:27 -04:00
|
|
|
def write(self, data):
|
|
|
|
self.popen.stdin.write(data)
|
|
|
|
self.popen.stdin.flush()
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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):
|
2021-04-29 20:12:04 -04:00
|
|
|
@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
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
def __init__(self, device):
|
2021-04-29 20:12:04 -04:00
|
|
|
device = self.map_device_shortcut(device)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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
|
|
|
|
|
2021-08-12 01:49:27 -04:00
|
|
|
def write(self, data):
|
|
|
|
self.pyb.serial.write(data)
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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)
|
2021-05-13 02:26:07 -04:00
|
|
|
sys.stdout.flush()
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
|
|
|
|
|
|
|
|
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)]
|
|
|
|
|
2022-01-17 01:15:31 -05:00
|
|
|
# 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"
|
|
|
|
|
2020-05-02 01:58:41 -04:00
|
|
|
if cmd_args.trace_output:
|
|
|
|
print("TRACE {}:".format("|".join(str(i) for i in instances)))
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
# 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()
|
2020-04-07 01:03:31 -04:00
|
|
|
if out is not None and not any(m in out for m in IGNORE_OUTPUT_MATCHES):
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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()
|
2020-05-13 00:35:32 -04:00
|
|
|
if out is not None and not any(m in out for m in IGNORE_OUTPUT_MATCHES):
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
trace_instance_output(idx, out)
|
2021-08-12 01:49:27 -04:00
|
|
|
if out.startswith("BROADCAST "):
|
|
|
|
for instance2 in instances:
|
|
|
|
if instance2 is not instance:
|
|
|
|
instance2.write(bytes(out, "ascii") + b"\r\n")
|
|
|
|
else:
|
|
|
|
output[idx].append(out)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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
|
|
|
|
|
|
|
|
|
2020-09-14 04:11:19 -04:00
|
|
|
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)
|
|
|
|
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
def run_tests(test_files, instances_truth, instances_test):
|
2020-03-18 06:33:17 -04:00
|
|
|
skipped_tests = []
|
|
|
|
passed_tests = []
|
|
|
|
failed_tests = []
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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 = 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
|
|
|
|
)
|
2020-05-02 01:58:41 -04:00
|
|
|
|
|
|
|
if cmd_args.show_output:
|
|
|
|
print("### TEST ###")
|
|
|
|
print(output_test, end="")
|
|
|
|
if not skip:
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
print("### TRUTH ###")
|
|
|
|
print(output_truth, end="")
|
|
|
|
|
|
|
|
# Print result of test
|
|
|
|
if skip:
|
|
|
|
print("skip")
|
2020-03-18 06:33:17 -04:00
|
|
|
skipped_tests.append(test_file)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
elif output_test == output_truth:
|
|
|
|
print("pass")
|
2020-03-18 06:33:17 -04:00
|
|
|
passed_tests.append(test_file)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
else:
|
|
|
|
print("FAIL")
|
2020-03-18 06:33:17 -04:00
|
|
|
failed_tests.append(test_file)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
if not cmd_args.show_output:
|
|
|
|
print("### TEST ###")
|
|
|
|
print(output_test, end="")
|
|
|
|
print("### TRUTH ###")
|
|
|
|
print(output_truth, end="")
|
2020-09-14 04:11:19 -04:00
|
|
|
print("### DIFF ###")
|
2020-11-03 01:38:09 -05:00
|
|
|
print_diff(output_truth, output_test)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
|
|
|
|
if cmd_args.show_output:
|
|
|
|
print()
|
|
|
|
|
2020-03-18 06:33:17 -04:00
|
|
|
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
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
|
|
|
|
def main():
|
|
|
|
global cmd_args
|
|
|
|
|
2022-09-19 18:54:53 -04:00
|
|
|
cmd_parser = argparse.ArgumentParser(
|
|
|
|
description="Run network tests for MicroPython",
|
|
|
|
formatter_class=argparse.RawTextHelpFormatter,
|
|
|
|
)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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"
|
|
|
|
)
|
2020-11-03 01:41:58 -05:00
|
|
|
cmd_parser.add_argument(
|
|
|
|
"-p",
|
|
|
|
"--permutations",
|
|
|
|
type=int,
|
|
|
|
default=1,
|
|
|
|
help="repeat the test with this many permutations of the instance order",
|
|
|
|
)
|
2022-09-19 18:54:53 -04:00
|
|
|
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"
|
|
|
|
)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
cmd_parser.add_argument("files", nargs="+", help="input test files")
|
|
|
|
cmd_args = cmd_parser.parse_args()
|
|
|
|
|
2020-03-09 11:58:47 -04:00
|
|
|
# clear search path to make sure tests use only builtin modules and those in extmod
|
2022-09-19 18:54:58 -04:00
|
|
|
os.environ["MICROPYPATH"] = os.pathsep.join(("", ".frozen", "../extmod"))
|
2020-03-09 11:58:47 -04:00
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
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:
|
2020-06-05 01:41:40 -04:00
|
|
|
# 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:") :]))
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
else:
|
2020-06-05 01:41:40 -04:00
|
|
|
print("unknown instance string: {}".format(cmd), file=sys.stderr)
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
sys.exit(1)
|
|
|
|
|
|
|
|
for _ in range(max_instances - len(instances_test)):
|
|
|
|
instances_test.append(PyInstanceSubProcess([MICROPYTHON]))
|
|
|
|
|
2020-11-03 01:41:58 -05:00
|
|
|
all_pass = True
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
try:
|
2020-11-03 01:41:58 -05:00
|
|
|
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)
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
finally:
|
|
|
|
for i in instances_truth:
|
|
|
|
i.close()
|
|
|
|
for i in instances_test:
|
|
|
|
i.close()
|
|
|
|
|
2020-03-18 06:33:17 -04:00
|
|
|
if not all_pass:
|
|
|
|
sys.exit(1)
|
|
|
|
|
tests/run-multitests.py: Add new test runner for multiple Py instances.
This commit adds a test runner and initial test scripts which run multiple
Python/MicroPython instances (eg executables, target boards) in parallel.
This is useful for testing, eg, network and Bluetooth functionality.
Each test file has a set of functions called instanceX(), where X ranges
from 0 up to the maximum number of instances that are needed, N-1. Then
run-multitests.py will execute this script on N separate instances (eg
micropython executables, or attached boards via pyboard.py) at the same
time, synchronising their start in the right order, possibly passing IP
address (or other address like bluetooth MAC) from the "server" instance to
the "client" instances so they can connect to each other. It then runs
them to completion, collects the output, and then tests against what
CPython gives (or what's in a provided .py.exp file).
The tests will be run using the standard unix executable for all instances
by default, eg:
$ ./run-multitests.py multi_net/*.py
Or they can be run with a board and unix executable via:
$ ./run-multitests.py --instance pyb:/dev/ttyACM0 --instance exec:micropython multi_net/*.py
2020-03-03 22:25:29 -05:00
|
|
|
|
|
|
|
if __name__ == "__main__":
|
|
|
|
main()
|