Hi
I am running xemacs from the carbon2 branch on my mac. I find that
after a few hours of inactivity, xemacs takes up a lot of cpu
constantly, using about 60-80% of the cpu (and firefox the
remaining). I had one such xemacs instance under gdb, but I know
neither xemacs internals nor carbon. So any help getting to the root
of this problem will be useful. I can help debug the problem when it
happens next, if I am given clues about where to look for what.
I ran dapptrace (dtrace script for tracing function calls) and I
have put the output here:
http://mirthmakers.com/code/xemacs-dapptrace.txt
This file is really large, as I turned on all the options for
dapptrace, just to get as much data as possible (may be too much
though, now that I look at the output).
Here is the gdb backtrace, in case this is useful for someone. It
just does this all the time.
#0 get_tty_and_process_events [inlined] () at event-carbon.c:590
#1 get_events (can_block_p=<value temporarily unavailable, due to
optimizations>) at event-carbon.c:638
#2 0x00019b45 in emacs_carbon_event_pending_p (how_many=1) at event-
carbon.c:638
#3 0x000692b0 in event_stream_event_pending_p (how_many=<value
temporarily unavailable, due to optimizations>) at event-carbon.c:638
#4 0x0006965b in run_pre_idle_hook () at event-carbon.c:638
#5 0x00122a6b in redisplay () at event-carbon.c:638
#6 0x0006de90 in Fnext_event (event=21180348, prompt=2048652) at
event-carbon.c:638
#7 0x0002eff4 in Fcommand_loop_1 () at event-carbon.c:638
#8 0x0002f0e5 in command_loop_1 (unused_dummy=2048652) at event-
carbon.c:638
#9 0x0005ca2e in condition_case_1 (handlers=-1073744472,
bfun=0x2f0c0 <command_loop_1>, barg=2048652, hfun=0x2e970
<cmd_error>, harg=2048652) at event-carbon.c:638
#10 0x0002ebf8 in get_tty_and_process_events [inlined] () at event-
carbon.c:262
#11 0x0002ebf8 in command_loop_2 (unused_dummy=2048652) at event-
carbon.c:638
#12 0x00056419 in internal_catch (tag=-1073744472, func=0x2eb90
<command_loop_2>, arg=2048652, threw=0x0, thrown_tag=0x0,
backtrace_before_throw=0x0) at event-carbon.c:638
#13 0x0002ed43 in initial_command_loop (load_me=2048652) at event-
carbon.c:638
#14 0x00053e96 in xemacs_21_5_b28_i386_apple_darwin9_6_0 (argc=2,
argv=0xbffffbb0, unused_envp=0x0, restart=0) at event-carbon.c:638
#15 0x00054b50 in main (argc=-1073744472, argv=0xbffff5a8,
unused_envp=0xbffffbbc) at event-carbon.c:638
Thanks for any help
I've noticed the same thing, but have not gotten around to looking
into it. I have assumed that it is related to network handling, as it
normally occurs after using gnus or slime from XEmacs, both of which
communicate via sockets. Further, it appears that Emacs (AquaEmacs)
does not behave this way.
event-carbon.c defines a variable named "debug_carbon_events"; is
there a way to set this from inside XEmacs, or must it be done from a
debugger (or by recompiling)?
Regards,
Raymond Wiker
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org