Dear Eli, Sorry for not running an unoptimized version... Le 06/06/2016 à 16:49, Eli Zaretskii a écrit : >> From: vincent.belaiche@gmail.com (Vincent Belaïche) >> Date: Mon, 06 Jun 2016 07:25:49 +0200 >> Cc: Vincent Belaïche >> >> This is a very difficult to reproduce crash. It happens quite scarcely, >> but fortunately, knowing its existence I was running under the >> debugger. Please note that it was w/o the -Q option, and my init file is >> quite a few lines ;-). > > Thanks. > > Unfortunately, backtraces from optimized builds, even those with -Og > switch, are not reliable enough. So when you know about an existing > problem that is hard to reproduce, please run an unoptimized version > if you can. > > Nevertheless, let's see if we can glean anything useful from this > crash. Please show the output of these GDB commands: > > (gdb) p PT_ADDR > (gdb) p GPT_ADDR > (gdb) p GPT Here you are: --8<----8<----8<----8<----8<-- begin -->8---->8---->8---->8---->8---- (gdb) p PT_ADDR $1 = (unsigned char *) 0x521073e " long as we don't have any longer the \\NEWLINE the problem is solved. So it could be:\n\n\n\n--8<----8<----8<----8<----8<-- begin -->8---->8---->8---->8---->8----\n in_input_funnies=`echo \"$in_input\" \\\n "... (gdb) p GPT_ADDR $2 = (unsigned char *) 0x5210677 "" (gdb) p GPT $3 = 1651 --8<----8<----8<----8<----8<-- end -->8---->8---->8---->8---->8---- The PT_ADDR seems to be pointing at the content of an email that I was editing when the crash occurred --- fortunately nothing confidential -)), some discussion in Texinfo bug mailing list, concerning support for spaces in filenames. If you are interested the thread is titled: REVIEW REQUEST - user manual generation I have also attached a screen capture of crashed Emacs, that may help you see what the content of the edited buffer was when the crash occurred. It seems that the crash occurred while I was editing the sentence « So it could be: ... ». VBR, Vincent.