lglaf
lglaf copied to clipboard
Lekensteyn
Windows 10 file read failed
Running on Windows 10, I had trouble with file operations. Made a few changes and it worked:
import os
class FileCommunication(Communication):
def __init__(self, file_path):
super(FileCommunication, self).__init__()
if sys.version_info[0] >= 3:
self.f = os.open(file_path, os.O_RDWR | os.O_BINARY)
else:
self.f = os.open(file_path, os.O_RDWR | os.O_BINARY)
def _read(self, n, timeout=None):
try:
return os.read(self.f,n)
except IOError as e:
_logger.warning("failed to read")
raise e
def write(self, data):
os.write(self.f,data)
def close(self):
os.close(self.f)
The Python 2 vs 3 case can be folded:
self.f = os.open(file_path, os.O_RDWR | os.O_BINARY)
Why do you catch the IOError for _read? The caller should take care of IO errors and do logging as needed.
Can you also post any exceptions/errors that prompted for these patches? Thanks :-) (A PR would also be nice)
I guess there was no good reason for trapping IOError for _read.
I thought about folding Python 2 vs 3 case, but thought I'd leave the cases split up like they were before, just in case they'd need to be handled differently since I only tried it on Python 2.
The errors I got were:
Traceback (most recent call last):
File "lglaf.py", line 404, in <module>
main()
File "lglaf.py", line 386, in main
try_hello(comm)
File "lglaf.py", line 290, in try_hello
data = comm.read(0x20, timeout=HELLO_READ_TIMEOUT)
File "lglaf.py", line 148, in read
buff = self._read(need, timeout=timeout)
File "lglaf.py", line 189, in _read
return self.f.read(n)
IOError: [Errno 0] Error
I doubt it would work with os.O_RDONLY | os.O_BINARY since writing is required.
Oh, nevermind, thinko (r+ = O_RDWR).
Since you mentioned Python 2, I think that buffering could be a problem. From pydoc open:
The optional buffering argument specifies the file’s desired buffer size: 0 means unbuffered, 1 means line buffered, any other positive value means use a buffer of (approximately) that size (in bytes). A negative buffering means to use the system default, which is usually line buffered for tty devices and fully buffered for other files. If omitted, the system default is used. [2]
What if you just remove the if sys.version_info[0] >= 3: and else: ... lines completely, enabling the buffering parameter by default?
(Note to self: when adding such checks, add a comment why you exactly do it. I vaguely remember that buffering=0 gave some error on Windows, but do not know for sure.)
I think I tried different buffering options without success. When I get a chance I'll check it out. It probably was a python bug, and by now maybe the bug is fixed.
