Having said that, if you say it has been working for you, I think there must be a method, one way or another to get it working as it did before.
The led pixel I/O is probably not the cause, what is the cause is that I am rewriting the whole network stuff at the moment (as said before). I already found an interesting effect in the TCP code that made all TCP communications very slow. I solved that now. Are you using TCP or UDP for the bridge? This may also be the cause for the slow UART I've been reporting earlier.
If you can have a little bit of patience yet, please wait for the next release and try again. Currently I have been rewriting yet more stuff and I think it's a bit too early to release now, though.
I'm using TCP because i need the packets received in the right order and also i'm planning on having a few robots connected to the same central unit. Another approach would be using UDP with specific ports for each robot but i'm not quite sure i would reach the desired results.
Patience. I has it ! I'll wait patiently for your update . thanks again!
Anyway, the whole network code is very much "in flux" (as not to say: "a mess"), when it's settling down, I will have a particular good look at the latency, I am quite confident I must be able to get at least the same numbers as before.
Also I realised the UARTs of the ESP8266 have a "loopback" mode. It will probably help me test this. Let's see if it works like I expect.