Re: Lexbind (was: Emacs 23.3 released)

[Eli Zaretskii]

> From: Juanma Barranquero <address@hidden>
> Date: Thu, 17 Mar 2011 02:59:38 +0100
> Cc: address@hidden
> 
> I've had several hangups while reading from the minibuffer. No CPU
> usage, but no response whatsoever so I had to kill Emacs.
> 
> I just had one while running under GDB, so I Ctrl-C and I got the
> attached backtrace.
> [...]
> (gdb) bt
> #0  0x0112c987 in next_element_from_buffer (it=0x88c250) at xdisp.c:6744

Is this with today's sources from the lexbind-new branch?  Because
line 6744 in xdisp.c is the closing brace of next_element_from_buffer
with today's bzr repo.  How can Emacs get stuck in such a line?

However, since this seems to be an optimized build, the backtrace is
not guaranteed to make sense.  I suggest to recompile without
optimizations and post a backtrace if and when that hangs.

> #1  0x0112171f in get_next_display_element (it=0x88c250) at xdisp.c:5630
> #2  0x0112292b in move_it_in_display_line_to (it=<value optimized
> out>, to_charpos=-1, to_x=-1, op=0) at xdisp.c:6956
> #3  0x0112620f in move_it_to (it=0x88c250, to_charpos=-1, to_x=-1,
> to_y=-1, to_vpos=1, op=4) at xdisp.c:7333
> #4  0x0112e87f in move_it_by_lines (it=0x88c250, dvpos=1, need_y_p=1)
> at xdisp.c:7757
> #5  0x01136a6e in try_scrolling (window=<value optimized out>,
> just_this_one_p=0, arg_scroll_conservatively=143165576, scroll_step=0,
>     temp_scroll_step=0, last_line_misfit=0) at xdisp.c:13128
> #6  0x0114e0cd in redisplay_window (window=<value optimized out>,
> just_this_one_p=0) at xdisp.c:14168
> #7  0x011502b3 in redisplay_window_0 (window=56712709) at xdisp.c:12362
> #8  0x01012961 in internal_condition_case_1 (bfun=0x1150290
> <redisplay_window_0>, arg=56712709, handlers=53689030,
>     hfun=0x1109770 <redisplay_window_error>) at eval.c:1536
> #9  0x0110f137 in redisplay_windows (window=<value optimized out>) at
> xdisp.c:12342

This all looks like a normal redisplay cycle to me, and I don't see
how it could possibly hang, especially without using CPU.  If you can
reproduce this in a non-optimized build, try stepping after
interrupting Emacs, and see where it loops or in what function call it
gets stuck.

Also, perhaps try bt in the other threads, maybe they are stuck, not
the Lisp thread.