🚀 go-pugleaf

Mike Russell at IBM Richland (?) sent a few bug reports to the list a
while back, but he's not a registered Internet gateway user within IBM
(which has some inordinately stringent rules for Internet access) so I
can't get email direct to him...... hopefully he'll see this.

Message follows:

------------------------------------------------------------------------------
Thanks for the bug reports!  They're greatly appreciated...
    
    In working with Epoch 4.0 patchlevel 2, I have found a number of 
    possible bugs:
    
    1)  In keyboard.c, function quit_throw_to_read_command_char:
    
     quit_throw_to_read_command_char ()
    {
    /*  quit_error_check (); */
      sigfree ();
      /* Prevent another signal from doing this before we finish.  */
      waiting_for_input = 0;
      input_pending = 0;
      unread_command_char = -1;
    #ifdef POLL_FOR_INPUT
      if (poll_suppress_count != 0)
        abort ();
    #endif
      _longjmp (getcjmp, 1);
    }
    
    The check to see if poll_suppress_count is not 1 is an error.  It should
    really check to see if poll_suppress_count is zero.  This problem arose
    from using the elisp function read-char in an event handler, and
    pressing C-g.  At this point, poll_suppress_count was 2, and C-g caused
    the editor to abort.
    
But your version has a check (poll_suppress_count != 0)... and that
means that if poll_suppress_count is 2 the abort would happen
anyway... which doesn't mesh with your comments... ???

    3)  Slow redraw:  This may or may not be a bug, but it is a change in
    behavior from Epoch 3.2  When another x-window partially or completely
    overlaps with an Epoch screen, redraws on that screen are very slow--you
    can see individual lines as they are redrawn.  When there are no
    overlapping windows, the redraws are quite fast.

I've been noticing this on RS/6000's, but it doesn't hit me on some
other platforms (notably SGI machines).  I assume you're using
RS/6000's?

Thanks,

Marc
------------------------------------------------------------------------------