Gunnar Evermann writes: | Can one of you guys check with a debugger what exactly is causing the | crash? BTW, is it only on HPUX? Good question! Anyone else hit this except me? | Maybe this is some old problem which is only now exposed because of the | additional call to beginning_of_line_p. If you cannot fix it than back | out my redisplay patch for 21.0 and I will try again for 21.2 when I have | found the right maschine here. Can you resend this patch since I am not subscribed to xemacs-patches. --Marcus

GE> On Fri, 2 Oct 1998, Marcus Thiessel wrote: GE> I just saw a message from some guy using solaris on c.e.xemacs I too just experienced the crash on Solaris 2.next However, I successfully expunged. The crash happened when I hit 'n' to go to the next message... Here's my data, enjoy: First, stdout: Lisp backtrace follows: set-window-buffer(#<window on "INBOX" 0xd646> #<buffer "INBOX">) # bind (buf norecord bufname) switch-to-buffer(#<buffer "INBOX">) # (unwind-protect ...) # (unwind-protect ...) # bind (buffer) vm-unbury-buffer(#<buffer "INBOX">) # (unwind-protect ...) # bind (vm-sbe-buffer) vm-preview-current-message() # bind (oldmp use-marks error direction count signal-errors retry count) vm-next-message(1 t nil) # bind (vm-circular-folders used-marks mlist del-count count) vm-delete-message(1) # bind (command-debug-status) call-interactively(vm-delete-message) # (condition-case ... . error) # (catch top-level ...) Now, dbx: program terminated by signal ABRT (Abort) Current function is fatal_error_signal 262 kill (getpid (), fatal_error_code); (dbx) where [1] _kill(0x0, 0x6, 0x0, 0x0, 0xfecb2100, 0x1278e4), at 0xfec969f0 =>[2] fatal_error_signal(sig = 6), line 262 in "emacs.c" ---- called from signal handler with signal 6 (SIGABRT) ------ [3] _kill(), at 0xfec969f0 [4] abort(0xfecb2100, 0x4fc828, 0x45d, 0x4fc864, 0x1659b0f, 0x4), at 0xfec392cc [5] assert_failed(file = 0x4fc828 "/import/verve/georgn/xemacs/xemacs-21.0-b57/src/insdel.c", line = 1117, expr = 0x4fc864 "(retval) >= BI_BUF_BEG (buf) && retval <= BI_BUF_Z (buf)"), line 2608 in "emacs.c" [6] bufpos_to_bytind(buf = 0xae6000, x = 12296508), line 1117 in "insdel.c" [7] beginning_of_line_p(b = 0xae6000, pt = 12296509), line 540 in "editfns.c" [8] Fset_window_buffer(window = UNION, buffer = UNION), line 3170 in "window.c" [9] funcall_recording_as(recorded_as = UNION, nargs = 2, args = 0xffbec820), line 3197 in "eval.c" [10] Ffuncall(nargs = 3, args = 0xffbec820), line 3238 in "eval.c" [11] Fbyte_code(bytestr = UNION, vector = UNION, maxdepth = UNION), line 416 in "bytecode.c" [12] funcall_lambda(fun = UNION, nargs = 1, arg_vector = 0xffbeca98), line 3599 in "eval.c" [13] apply_lambda(fun = UNION, numargs = 1, unevalled_args = UNION), line 3501 in "eval.c" [14] Feval(form = UNION), line 3067 in "eval.c" [15] Fprogn(args = UNION), line 758 in "eval.c" [16] Fbyte_code(bytestr = UNION, vector = UNION, maxdepth = UNION), line 554 in "bytecode.c" [17] funcall_lambda(fun = UNION, nargs = 1, arg_vector = 0xffbed4c4), line 3599 in "eval.c" [18] funcall_recording_as(recorded_as = UNION, nargs = 1, args = 0xffbed4c0), line 3200 in "eval.c" [19] Ffuncall(nargs = 2, args = 0xffbed4c0), line 3238 in "eval.c" [20] Fbyte_code(bytestr = UNION, vector = UNION, maxdepth = UNION), line 416 in "bytecode.c" [21] funcall_lambda(fun = UNION, nargs = 0, arg_vector = 0xffbedae4), line 3599 in "eval.c" [22] funcall_recording_as(recorded_as = UNION, nargs = 0, args = 0xffbedae0), line 3200 in "eval.c" [23] Ffuncall(nargs = 1, args = 0xffbedae0), line 3238 in "eval.c" [24] Fbyte_code(bytestr = UNION, vector = UNION, maxdepth = UNION), line 416 in "bytecode.c" [25] funcall_lambda(fun = UNION, nargs = 3, arg_vector = 0xffbee16c), line 3599 in "eval.c" [26] funcall_recording_as(recorded_as = UNION, nargs = 3, args = 0xffbee168), line 3200 in "eval.c" [27] Ffuncall(nargs = 4, args = 0xffbee168), line 3238 in "eval.c" [28] Fbyte_code(bytestr = UNION, vector = UNION, maxdepth = UNION), line 416 in "bytecode.c" [29] funcall_lambda(fun = UNION, nargs = 1, arg_vector = 0xffbee564), line 3599 in "eval.c" [30] funcall_recording_as(recorded_as = UNION, nargs = 1, args = 0xffbee560), line 3200 in "eval.c" [31] Ffuncall(nargs = 2, args = 0xffbee560), line 3238 in "eval.c" [32] Fcall_interactively(function = UNION, record_flag = UNION, keys = UNION), line 949 in "callint.c" [33] Fcommand_execute(cmd = UNION, record = UNION, keys = UNION), line 2584 in "eval.c" [34] execute_command_event(command_builder = 0x7a8880, event = UNION), line 4346 in "event-stream.c" [35] Fdispatch_event(event = UNION), line 4645 in "event-stream.c" [36] Fcommand_loop_1(), line 579 in "cmdloop.c" [37] command_loop_1(dummy = UNION), line 494 in "cmdloop.c" [38] condition_case_1(handlers = UNION, bfun = 0xd20a0 = &`xemacs`cmdloop.c`command_loop_1(Lisp_Object dummy), barg = UNION, hfun = 0xd0e00 = &`xemacs`cmdloop.c`cmd_error(Lisp_Object data, Lisp_Object dummy), harg = UNION), line 1653 in "eval.c" [39] command_loop_3(), line 256 in "cmdloop.c" [40] command_loop_2(dummy = UNION), line 267 in "cmdloop.c" [41] internal_catch(tag = UNION, func = 0xd1138 = &`xemacs`cmdloop.c`command_loop_2(Lisp_Object dummy), arg = UNION, threw = (nil)), line 1328 in "eval.c" [42] initial_command_loop(load_me = UNION), line 305 in "cmdloop.c" [43] xemacs_21_0_sparc_sun_solaris2_8(argc = 1, argv = 0xffbeede4, envp = 0xffbeedec, restart = 0), line 1705 in "emacs.c" [44] main(argc = 1, argv = 0xffbeede4, envp = 0xffbeedec), line 2122 in "emacs.c"

GE> Please try this patch: Aside from the fact that the first part had to be applied by hand it seems to work. Thanks. Incidentally, I was able to avoid the crash by saving my mailbox before expunging, then expunging and saving again.

Gunnar Evermann writes: | OK, I found the problem. I had actually intended to do it this way | but somehow forgot. Sorry for that. Applied it but I can still crash XEmacs when hitting the catchup&exit button running Gnus. It is reproducable with -q and just says Error: Object "" does not have windowed ancestor then XEmacs disappears without a trace. --Marcus

Gunnar Evermann writes: | Have you applied both hunks? I'm afraid I cannot produce a real patch | against b57 currently, as I only have a non-networked Linux machine with | hacked on beta55 sources on it available. Yep! I applied the failed one by hand (I am not sure why it failed). | As far as the above crash is concerned, I have no idea how it happens but | I will grep the source for the message. Perhaps this one is related to | Motif? If you have some time on your hand, can you take a vanilla b57 | back out the widening/redisplay patch (take it from the ml-archive) and | try again? I can try but the next two day I will be out of the office. :-( | Or you could always give me an account on your machine... :-) | I only have 5M quota around here :-( 5G would be more appropriate. The account on my machine is no problem but you will find it difficult to login because of a firewall. :-( I don't have any access to a machine outside the firewall running 11.x. | PS: what happened to xemacs-beta? I don't seem to get any messages anymore. It seems it has become a very low traffic list for a reason I don't know (vacation?). --Marcus

Kyle Jones writes: | Village gadflies too busy in Real Life to make trouble here. :) I definitley can't help with any Real Life related business. --Marcus