DeviceTimeoutError

[brianmay]

Seems that the library somehow gets into a weird state where it won't talk anymore to the light, and has to be restarted for it to work again.

Traceback (most recent call last):
  File "/home/lifx/python/local/lib/python2.7/site-packages/apscheduler/executors/base.py", line 112, in run_job
    retval = job.func(*job.args, **job.kwargs)
  File "/home/lifx/python/local/lib/python2.7/site-packages/wrapt/wrappers.py", line 522, in __call__
    args, kwargs)
  File "/home/lifx/python/local/lib/python2.7/site-packages/logtool/log_wrap.py", line 104, in __call__
    rc = fn (*args, **kwargs)
  File "./lifx_schedule", line 56, in set_white_color
    for l in lights.by_label('Main'):
  File "/home/lifx/python/local/lib/python2.7/site-packages/lifx/client.py", line 221, in by_label
    return filter(lambda d: d.label == label, self.get_devices())
  File "/home/lifx/python/local/lib/python2.7/site-packages/lifx/client.py", line 221, in <lambda>
    return filter(lambda d: d.label == label, self.get_devices())
  File "/home/lifx/python/local/lib/python2.7/site-packages/lifx/device.py", line 260, in label
    response = self._block_for_response(pkt_type=protocol.TYPE_GETLABEL)
  File "/home/lifx/python/local/lib/python2.7/site-packages/lifx/device.py", line 88, in _block_for_response
    return self._block_for(False, True, *args, **kwargs)
  File "/home/lifx/python/local/lib/python2.7/site-packages/lifx/device.py", line 136, in _block_for
    raise DeviceTimeoutError(self, timeout, DEFAULT_RETRANSMITS)
DeviceTimeoutError: Device with id:'d073d502abc8' timed out after 2.0 seconds and 10 retransmissions.

I believe the light in question is online and fully operational. Yes I know, the stack trace would imply otherwise.

[brianmay]

Thinking about this, by_label() should never return DeviceTimeoutError as we are only trying to get a list of lights, not try to talk to them.

[brianmay]

Still seems to be a problem in that if DeviceTimeoutError occurs in some situations it will forget about that light even if it is turned back on again. Still investigating.

[brianmay]

Think we really should cache the label so we don't have to do a network access every time __repr__ is called on the light.

[frakman1]

I too have run into this. The code should really keep going instead of crashing like this. I shouldn't have to restart the script because of a momentary IP glitch.