centipede wrote: > Funny you mention the bluetooth stack. By now I have disemboweled the kbdd > driver, which after all just opens a serial connection to /dev/rfcommN. Yes, there is no bluetooth magic in kbdd. And rfcomm must be reliable In case of problems it should close connection completely, not loose data. If this does not happen it is bluetooth protocol or device problem. I would blame the keyboard but who knows. Try with same kbdd recompiled on linux PC with bluetooth if it works the same. > I've played with the baud-rate - slower or faster - but not really any > other settings. No gain. this does nothing by definition, it is only for backward compatibility with real serial device > It > may fool the bluetooth stack, as if the keyboard and the driver needs to > negotiate before accepting another keystroke. "Is that an S-up?", "Yes, it > is a T-down", "???? never mind. T-up you say?". Nice theory but frcomm does not care at all about nature of the data sent (i.e. key up/down) so there is nothing like this on bluetooth stack level. Key up/down codes are resolved in kbdd. Frantisek
