bug#13864: 24.3.50; emacsclient -t loops when connected to emacs server running in X11

[Eli Zaretskii <eliz@gnu.org>]

> From: ashish.is@lostca.se (Ashish SHUKLA)
> Cc: 13864@debbugs.gnu.org
> Date: Fri, 08 Mar 2013 15:38:30 +0530
> 
> >   (gdb) break dispnew.c:4509
> >   (gdb) commands
> >> p i
> >> p desired_matrix->nrows
> >> if i < desired_matrix->nrows
> >> pgrowx desired_matrix->rows+i
> >> end
> 
> I later added a 'continue' in here as Breakpoint 6 in the output.

Yes, sorry about that, I forgot.

> Not sure if the attached gdb output is any useful.

It is, thanks.  I think we are making progress.

> - emacs -Q 
> - M-x server-start
> - gdb stuff, breakpoints + loading .gdbinit
> - Started an xterm of dimensions (maybe 20-25 rows)
> - emacsclient -t
> - key presses (none of them is C-x C-f)

Emacs would begin to flicker after these, right?

> - Breakpoint 1 being hit and requiring me to type 'cont' on every hit
> - After some 'cont' inputs, I deleted it, and re-added it as Breakpoint 6 with
>  'continue' as mentioned above.
> - Breakpoint 6 being continuously hit

So you are saying that scrolling_1 is never called, is that right?

> - C-x 5 0 in emacsclient window
> - appended '====EMACSCLIENT STOPPED====' to logfile
> - emacsclient -t
> - Breakpoint 6 being hit
> - I resized window to 4-5 rows in an effort to reduce no. of redraw messages
> - Killed gdb after 2 minutes, which killed emacs.

To have a way of terminating the session in a more civilized way, I
frequently use the following trick: put a breakpoint at Fredraw_display,
before you start the debugging session.  Then, whenever you want to
change something or finish the session, just "M-x redraw-display RET"
and GDB kicks in.

> Let me know if you need more information.

Hmm...  Some observations:

  . update_frame_1 is constantly called with either the entire frame,
    starting with the menu bar, or just with the last line of the
    frame, which is the echo area.

  . I see tooltip messages being displayed in the echo area.  You have
    a mouse active (as far as Emacs is concerned) on the xterm frame,
    is that right?  Can you disable it and see if the flickering
    persists?

  . There are some instances where Emacs displayed "Quit" in the echo
    area.  Did you type C-g or some such?

Moving on in the investigation of the problem (assuming that
disabling the mouse doesn't fix it), I assume that the function
update_frame_line is being called to redraw the xterm frame, given the
evidence you gathered this far.  First, let's make sure this is indeed
so.  Use this breakpoint:

  (gdb) break update_frame_line
  (gdb) commands
   > p vpos
   > continue
   > end
  (gdb)

Please see if you see all the line numbers when you recreate the
situation with flickering.

If you indeed see all the line numbers of a frame, next thing is to
find out whether on your system the FRAME_CHAR_INS_DEL_OK macro
returns zero or non-zero.  (Depending on that, Emacs uses two separate
portions of code in update_frame_line which try to avoid redrawing the
parts of screen that didn't change.)  You could, for example, put a
breakpoint inside this block:

  if (!FRAME_CHAR_INS_DEL_OK (f))
    {
      int i, j;

      /* Find the first glyph in desired row that doesn't agree with
	 a glyph in the current row, and write the rest from there on.  */
      for (i = 0; i < nlen; i++)
	{
	  if (i >= olen || !GLYPH_EQUAL_P (nbody + i, obody + i))
	    {
	      /* Find the end of the run of different glyphs.  */
	      j = i + 1;
	      while (j < nlen
		     && (j >= olen
			 || !GLYPH_EQUAL_P (nbody + j, obody + j)
			 || CHAR_GLYPH_PADDING_P (nbody[j])))
		++j;

	      /* Output this run of non-matching chars.  */
	      cursor_to (f, vpos, i);
	      write_glyphs (f, nbody + i, j - i);
	      i = j - 1;

	      /* Now find the next non-match.  */
	    }
	}

      /* Clear the rest of the line, or the non-clear part of it.  */
      if (olen > nlen)
	{
	  cursor_to (f, vpos, nlen);
	  clear_end_of_line (f, olen);
	}

      /* Make current row = desired row.  */
      make_current (desired_matrix, current_matrix, vpos);
      return;  <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<
    }

on the marked line, and see if it ever breaks.

Thanks.