I'm having the same issue and implemented my own readline() function which I copied and modified from the serialutil.py file found in the pyserial package.
The serial connection is part of the class this function belongs to and is saved in attribute 'self.ser'
def _readline(self):
eol = b'
'
leneol = len(eol)
line = bytearray()
while True:
c = self.ser.read(1)
if c:
line += c
if line[-leneol:] == eol:
break
else:
break
return bytes(line)
This is a safer, nicer and faster option than waiting for the timeout.
EDIT:
I came across this post when trying to get the io.TextIOWrapper method to work (thanks zmo).
So instead of using the custom readline function as mentioned above you could use:
self.ser = serial.Serial(port=self.port,
baudrate=9600,
bytesize=serial.EIGHTBITS,
parity=serial.PARITY_NONE,
stopbits=serial.STOPBITS_ONE,
timeout=1)
self.ser_io = io.TextIOWrapper(io.BufferedRWPair(self.ser, self.ser, 1),
newline = '
',
line_buffering = True)
self.ser_io.write("ID
")
self_id = self.ser_io.readline()
Make sure to pass the argument 1
to the BufferedRWPair
, otherwise it will not pass the data to the TextIOWrapper after every byte causing the serial connection to timeout again.
When setting line_buffering
to True
you no longer have to call the flush
function after every write (if the write is terminated with a newline character).
EDIT:
The TextIOWrapper method works in practice for small command strings, but its behavior is undefined and can lead to errors when transmitting more than a couple bytes. The safest thing to do really is to implement your own version of readline
.
与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…