19:01 <_whitenotifier-f> [libfx2] sidd-kishan opened pull request #16: Please consider this update for making the repo more appealing - https://github.com/whitequark/libfx2/pull/16

22:09 <chuckmcm[m]> I haven't gotten to the end of this discussion yet, but the mis-match here is that the FTDI USB chip, on a development board, has some additional I/O pins called "modem control" (part of the original serial port design), there are two "inputs" (DSR, CTS), and two "outputs" (DTR, RTS). DSR stands for "Dataset ready" and it was sent from the modem to the computer to say it was good to go, and CTS stood for "Clear to send" which indicated

22:09 <chuckmcm[m]> CTS), you could send control messages to the serial port to make RTS or DTR go "high" or "low". And with that, serial device programming was born.

22:09 <chuckmcm[m]> the modem had space in its input buffer to accept more characters from the computer. The outputs were RTS which stood for "Ready to send", and DTR which stood for "Data Terminal Ready". Folks have "repurposed" those pins and attached the outputs to "reset" and "mode" on chips (my first encounter with this was Arduino), and since serial drivers on computers had functions to set the state of RTS and DTR (and read the state of DSR and

22:11 <chuckmcm[m]> It is entirely possible to add four GPIOs to the Glasgow applet to represent these signals, and update the USB driver so when "`SET_RTS`" or "`SET_DTR`" is sent in the control channel of the USB device the pins reflect the appropriate state. But That doesn't let you use it as an ESP programmer unless you actually hook those pins up to something on the ESP.

22:56 <stary2001[m]> "Glasgow applet [...] sent in the control channel of the usb device"

22:56 <stary2001[m]> afaik the uart applet uses a pty, not the actual usb device, and the pty doesn't expose those

23:02 <omnitechnomancer> Yea the difficulty here is that there isn't a nice uniform across OSes manner to provide a software serial port. The FTDI device gets to have a kernel driver that knows it's a serial port

23:03 <omnitechnomancer> And connects all the relevant APIs up to userland

23:05 <stary2001[m]> yes, let's not redo that whole conversation

23:05 <stary2001[m]> s/let's/lets/

23:16 <omnitechnomancer> Indeed

23:38 <Wanda[cis]> clearly we need to make a TUSE layer

23:49 <josHua[m]> Catherine and I vaguely investigated this sort of thing at this point in history. we did not like it