[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-readline] [PATCH] Gnuplot on macOS does not suspend by Control-

From: Rin Okuyama
Subject: Re: [Bug-readline] [PATCH] Gnuplot on macOS does not suspend by Control-z
Date: Tue, 28 Nov 2017 23:20:29 +0900
User-agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.13; rv:52.0) Gecko/20100101 Thunderbird/52.4.0

On 2017/11/28 3:03, Chet Ramey wrote:
On 11/26/17 8:43 AM, Rin Okuyama wrote:
(1) _rl_signal_handler() is used as before:

int getc_wrapper(FILE *fp) {
      fd_set fds;
      int fd = fileno(fp);
      int ierr;
      if (external_fd >= 0) {
          do {
              FD_SET(fd, &fds);
              FD_SET(external_fd, &fds);
              ierr = select(external_fd + 1, &fds, 0, 0, NULL);
              if (ierr < 0 && errno == EINTR) {
                  if (_rl_caught_signal)

... poll and read from external_fd ...

          } while (!FD_ISSET(fd, &fds));
      return rl_getc(fp);
      return getc(fp);

(2) the 1st ifdef block is simply replaced by

                  return rl_getc(fp);

Both (1) and (2) suspend by Control-z. However, variant (2) does not
accept input from external_fd (it is actually mouse driver) after
continue from suspend, unless the first any input arrives from tty.
I guess that this is inevitable with rl_getc(). How do you think?
You're right, as long as the application doesn't install any signal
handlers. It's clear the rl_getc() workaround doesn't do everything you
want in this case.

Let's back up a bit. Why is Gnuplot attempting to use the readline signal
handling framework when there isn't any readline code executing? Is the
function fragment you showed assigned to rl_getc_function? If that's the
case, I can see a reason to allow rl_getc replacements to use the readline
signal handling infrastructure.

Exactly. The function shown above is a simplified version of getc_wrapper(),
which gnuplot substitutes into rl_getc_function. For more details, please
refer to src/readline.c, as well as x11_waitforinput() in term/x11.trm,
which is called from getc_wrapper() via term->waitforinput(). I think that
the fragment captures the essence of it although... Anyway, gnuplot assigns
that function to rl_getc_function, in order to catch inputs both from tty
and mouse (``external_fd'' in the above code). Gnuplot does not install any
signal handlers by its own.


reply via email to

[Prev in Thread] Current Thread [Next in Thread]