The Read Until With Encoding Errors On Strict test is failing with UnicodeDecodeError: 'utf-8' codec can't decode byte 0xfa in position 88: invalid start byte (the "position" can differ). Turns out that the corrupted output of the file is leaking to the Write keyword call. It makes Write fail depending on timing, if the output following Write is corrupted. Issue introduced in #431
def _read_until(self, matcher, expected, timeout=None):
timeout = TimeEntry(timeout) if timeout else self.config.get('timeout')
max_time = time.time() + timeout.value
while time.time() < max_time:
undecoded = self._single_complete_read_to_buffer(max_time) # <<<<< Reading into corrupted text
if undecoded:
self._receive_buffer += undecoded.decode( # <<<<< Exception
self.config.encoding, "ignore"
)
match = matcher(self._receive_buffer)
if match:
if hasattr(match, "end"):
end = match.end()
else:
end = self._receive_buffer.index(expected) + len(expected)
output = self._receive_buffer[0:end]
self._receive_buffer = self._receive_buffer[end:]
return output
output = self._receive_buffer
self._receive_buffer = ""
raise SSHClientException(f"No match found for '{expected}' in {timeout}\nOutput:\n{output}.")
If the read after Write reads beyond the written text into the corrupted part, the
will read until timeout and the undecoded.decode will fail. The fix might be to try and decode as much of the beginning of the undecoded part as possible and then match.
The Read Until With Encoding Errors On Strict test is failing with
UnicodeDecodeError: 'utf-8' codec can't decode byte 0xfa in position 88: invalid start byte
(the "position" can differ). Turns out that the corrupted output of the file is leaking to the Write keyword call. It makes Write fail depending on timing, if the output following Write is corrupted. Issue introduced in #431If the read after Write reads beyond the written text into the corrupted part, the
undecoded = self._single_complete_read_to_buffer(max_time)
will read until timeout and the
undecoded.decode
will fail. The fix might be to try and decode as much of the beginning of the undecoded part as possible and then match.