Crash using gnuclient -nw

Hi, I started to use gnuclient -nw to read my mail from home, and sometimes XEmacs will be gone when I get back to work. It seems to happen when I close the text mode frame. So I decided to find my core file and look into it. The Lisp backtrace doesn't look interesting: # (condition-case ... . error) # (catch top-level ...) The gdb 'where' command gave more interesting results: #0 0x4034cd21 in __kill () from /lib/libc.so.6 #1 0x80bb4a1 in fatal_error_signal (sig=6) at emacs.c:536 #2 <signal handler called> #3 0x4034cd21 in __kill () from /lib/libc.so.6 #4 0x4034c996 in raise (sig=6) at ../sysdeps/posix/raise.c:27 #5 0x4034e0b8 in abort () at ../sysdeps/generic/abort.c:88 #6 0x80be7db in assert_failed (file=0x824d5c9 "events.h", line=505, expr=0x824d620 "RECORD_TYPEP (obj, lrecord_type_command_builder)") at emacs.c:3320 #7 0x810736c in execute_command_event (command_builder=0x9200358, event={ gu = {type = Lisp_Type_Record, val = 39211188}, s = {bits = 0, val = 78422376}, u = {bits = 0, val = 78422376}, ui = 156844752, i = 156844752, v = 0x95942d0, cv = 0x95942d0}) at events.h:505 #8 0x8107cac in Fdispatch_event (event={gu = {type = Lisp_Type_Record, val = 39211188}, s = {bits = 0, val = 78422376}, u = {bits = 0, val = 78422376}, ui = 156844752, i = 156844752, v = 0x95942d0, cv = 0x95942d0}) at event-stream.c:4186 #9 0x80a0ae2 in Fcommand_loop_1 () at cmdloop.c:585 #10 0x80a0dcd in command_loop_1 (dummy={gu = {type = Lisp_Type_Record, val = 270084330}, s = {bits = 0, val = 540168660}, u = {bits = 0, val = 540168660}, ui = 1080337320, i = 1080337320, v = 0x4064a3a8, cv = 0x4064a3a8}) at cmdloop.c:496 #11 0x80c1168 in condition_case_1 (handlers={gu = {type = Lisp_Type_Record, val = 270083754}, s = {bits = 0, val = 540167508}, u = {bits = 0, val = 540167508}, ui = 1080335016, i = 1080335016, v = 0x40649aa8, cv = 0x40649aa8}, bfun=0x80a0d84 <command_loop_1>, barg={gu = { type = Lisp_Type_Record, val = 270084330}, s = {bits = 0, val = 540168660}, u = {bits = 0, val = 540168660}, ui = 1080337320, i = 1080337320, v = 0x4064a3a8, cv = 0x4064a3a8}, hfun=0x80a0e60 <cmd_error>, harg={gu = {type = Lisp_Type_Record, val = 270084330}, s = {bits = 0, val = 540168660}, u = {bits = 0, val = 540168660}, ui = 1080337320, i = 1080337320, v = 0x4064a3a8, cv = 0x4064a3a8}) at eval.c:1692 #12 0x80a0fa3 in command_loop_2 (dummy={gu = {type = Lisp_Type_Record, val = 270084330}, s = {bits = 0, val = 540168660}, u = {bits = 0, val = 540168660}, ui = 1080337320, i = 1080337320, v = 0x4064a3a8, cv = 0x4064a3a8}) at cmdloop.c:258 #13 0x80caee3 in internal_catch (tag={gu = {type = Lisp_Type_Record, val = 270035886}, s = {bits = 0, val = 540071772}, u = {bits = 0, val = 540071772}, ui = 1080143544, i = 1080143544, v = 0x4061aeb8, cv = 0x4061aeb8}, func=0x80a0f5c <command_loop_2>, arg={gu = { type = Lisp_Type_Record, val = 270084330}, s = {bits = 0, val = 540168660}, u = {bits = 0, val = 540168660}, ui = 1080337320, i = 1080337320, v = 0x4064a3a8, cv = 0x4064a3a8}, threw=0x0) at eval.c:1317 #14 0x80a029d in initial_command_loop (load_me={gu = {type = Lisp_Type_Record, val = 270084330}, s = {bits = 0, val = 540168660}, u = {bits = 0, val = 540168660}, ui = 1080337320, i = 1080337320, v = 0x4064a3a8, cv = 0x4064a3a8}) at cmdloop.c:307 #15 0x80bc6ff in xemacs_21_5_b4_i686_pc_linux () at emacs.c:2358 #16 0x80bea9b in main (argc=4, argv=0xbffff424, envp=0xbffff438) at emacs.c:2725 #17 0x403469cb in __libc_start_main (main=0x80bea24 <main>, argc=4, argv=0xbffff424, init=0x807e068 <_init>, fini=0x8220a14 <_fini>, rtld_fini=0x4000aea0 <_dl_fini>, stack_end=0xbffff41c) at ../sysdeps/generic/libc-start.c:92 Coincidentally, Robert Pluim sent out a message about gnuclient -nw and how it led to a crash here: #4 0x080b659b in assert_failed ( file=0x82180e0 "/home/rpluim/xemacs-21.5/src/events.h", line=505, expr=0x8218160 "RECORD_TYPEP (obj, lrecord_type_command_builder)") at /home/rpluim/xemacs-21.5/src/emacs.c:3320 This looks a lot like what I had: #6 0x80be7db in assert_failed (file=0x824d5c9 "events.h", line=505, expr=0x824d620 "RECORD_TYPEP (obj, lrecord_type_command_builder)") at emacs.c:3320 So I'm going to try Robert's patch. Thanks Robert! - Amit P.S. I compiled XEmacs a few days ago from CVS. It reports itself as 21.5 (beta 4) "bamboo" XEmacs Lucid -- Amit J Patel, Computer Science Department, Stanford University http://www-cs-students.stanford.edu/~amitp/ ``Parkinson's Other Law: Perfection is achieved only at the point of collapse.''