On Tue, 5 May 2020 at 11:09, Jan Kiszka <jan.kis...@web.de> wrote: > > On 05.05.20 11:59, Philippe Mathieu-Daudé wrote: > I don't recall details anymore either (more than 10 year ago now...), > but this looks reasonable.
My guess is that it dates back to when the serial code would crash if passed a NULL pointer for the backend rather than treating it as "same as if connected to a /dev/null backend". thanks -- PMM