Lack of timestamp makes the lib unreliable

[AlexisTM]

We wanted to use the libmotioncapture python interface to use the Optitrack positions, but the lack of timestamps made it impossible.
We do not receive the data at the expected interval (sometimes we got no message then receive two at the same time) so a post synchronization (time.monotonic_ns()) is also invalid.

Below is our branch forwarding both the Optitrack (open source) and VRPN timestamps.

main...AlexisTM:libmotioncapture:feat/add_stamps

[whoenig]

Thanks. I think having timestamps is valuable in general, although some more work is needed to get this across the other platforms, too.

I am wondering what the reason for the weird behavior you are seeing is, though. This is certainly not normal. Are you on a wired (Ethernet) connection? Do you see the same issue with the C++ examples, or is that something that might be caused by the Python bindings?

[AlexisTM]

I used the Python bindings. I have that behaviour both on a Linux machine connected to the host through Wifi as well as on the Windows host itself (the one running Motive).

It is indeed a bit of work to add the timestamps, that is why I didn't do a PR as it requires to update all interfaces.