Re: CORE in gnus during reply

-- xemacs-beta(a)xemacs.org spake thusly: After a preliminary look at this, I'm almost positive it's not my bug. I tracked it down to the following loop in callint.c (line 932): /* If we used a marker to hold point, mark, or an end of the region, temporarily, convert it to an integer now. */ for (argnum = 0; argnum < argcount; argnum++) if (!NILP (varies[argnum])) XSETINT (args[argnum], marker_position (args[argnum])); The XSETINT line is the one that crashes, with the following backtrace: #4 0x80bed0b in assert_failed (file=0x8274a20 "/usr/local/src/xemacs-21.2/src/lisp.h", line=1397, expr=0x8274ba0 "RECORD_TYPEP (obj, lrecord_type_marker)") at /usr/local/src/xemacs-21.2/src/emacs.c:3214 #5 0x818723f in marker_position (...) at /usr/local/src/xemacs-21.2/src/lisp.h:1397 #6 0x809a34e in Fcall_interactively (...) at /usr/local/src/xemacs-21.2/src/callint.c:936 The following patch fixes the crash, but don't ask me why: Index: callint.c =================================================================== RCS file: /usr/CVSroot/XEmacs/xemacs/src/callint.c,v retrieving revision 1.12.2.8 diff -u -r1.12.2.8 callint.c --- callint.c 2000/11/13 09:00:30 1.12.2.8 +++ callint.c 2001/02/21 03:08:42 ＠＠ -933,7 +933,10 ＠＠ temporarily, convert it to an integer now. */ for (argnum = 0; argnum < argcount; argnum++) if (!NILP (varies[argnum])) - XSETINT (args[argnum], marker_position (args[argnum])); + { + Bufpos tmp = marker_position (args[argnum]); + XSETINT (args[argnum], tmp); + } single_console_state (); specbind (Qcommand_debug_status, Qnil); I'm not going to commit this, since I'm not certain it's the right thing to do. Perhaps someone with a little more experience with this code than I have can offer an opinion?