On Fri, 8 Apr 2022, Andy Shevchenko wrote:
> On Fri, Apr 08, 2022 at 03:34:27PM +0300, Ilpo Järvinen wrote:
> > On Fri, 8 Apr 2022, Andy Shevchenko wrote:
> > > On Fri, Apr 08, 2022 at 02:39:53PM +0300, Ilpo Järvinen wrote:
>
> > What I'd like to do here is to take advantage of the function that was
> > added:
> >
> > if (!n_tty_receive_char_flow_ctrl(tty, c) &&
> > tty->flow.stopped && !tty->flow.tco_stopped && I_IXANY(tty) &&
> > c != INTR_CHAR(tty) && c != QUIT_CHAR(tty) &&
> > c != SUSP_CHAR(tty))) {
> > start_tty(tty);
> > process_echoes(tty);
> > }
> > ...but it will change STOP_CHAR vs START_CHAR precedence for the case
> > where they're the same characters. I don't know if it matters.
>
> No idea of impact of such change.
What I could do, is to create a separate change out of this outlined
change alone so it would be possible to bisect to the very change, if it
would cause a regression for somebody.
It doesn't feel very useful to have START_CHAR and STOP_CHAR match but
who knows what is out there. I tested it briefly with stty and I know it
"works" on the normal receive path (doesn't stop) but usefulness of such
a config seems rather doubtful.
One interesting difference is that here in the closing path, stop has
higher precedence and on the normal path, start takes preference. I don't
know if that's intentional or not.
--
i.
