>> From: vincent.belaiche@gmail.com (Vincent Belaïche) >> Cc: Vincent Belaïche , >> 20782@debbugs.gnu.org >> Date: Wed, 10 Jun 2015 23:51:27 +0200 >> >> Program received signal SIGSEGV, Segmentation fault. >> SYMBOL_NAME (sym=2) at lisp.h:1662 >> 1662 return XSYMBOL (sym)->name; >> (gdb) bt full >> #0 SYMBOL_NAME (sym=2) at lisp.h:1662 >> No locals. >> #1 print_object (obj=2, printcharfun=printcharfun@entry=21101962, >> escapeflag=escapeflag@entry=true) at print.c:1539 >> end = >> c = >> i_byte = >> confusing = >> p = >> buf = "\377\377\377\377\000\000\000\000\000\000\366\000*\271\272w \343i\t\240\345\210\000\270\032\305w8\001\366\000\234\032\305wpi\261o" >> #2 0x0111d13c in print (obj=obj@entry=2, >> printcharfun=printcharfun@entry=21101962, >> escapeflag=escapeflag@entry=true) at print.c:1139 >> No locals. >> #3 0x0111d27e in Fprin1 (object=2, printcharfun=21101962) at print.c:581 >> old = 0x95ce900 >> old_point = >> start_point = >> old_point_byte = >> start_point_byte = >> free_print_buffer = true >> multibyte = true >> original = > >This shows that Emacs crashed in 'prin1' trying to format an invalid >Lisp symbol for display in a backtrace. 'prin1' was called by the >debugger, so I think Emacs was in the process of reporting a Lisp >error by invoking the debugger. Any idea what Lisp error could that >be? > >Can you "M-x load-file RET template.el RET" before repeating the >recipe, so that the Lisp-level backtrace shows more detailed >information? > Hello, I need anyhow to load template.el as part of the recipe to cause the crash --- should I also destroy template.elc ? What is happening is that I was developping some macros to create a file from a row in a SES spreadsheet using an Emacs template. My macro had a few bugs (notably it tries to create a file in a directory that does not exists and a few other wrong things causing template-new-file in trouble). I think that this is what causes some unusual processing going through bug#20782 source of crash. Now please note that I had included some explicit (debug) statement in my macro in order to debug it. But then when I type 'j' in order to execute fully the (template-new-file ...) statement right after the (debug) statement, then it seems that, indeed, some error is met and the debugger is recursively entered and then Emacs crashes. Also I am realizing that the recipe to reproduce the bug is not so simple. I have to do things by hand with M-x load-file (or 'L' in dired mode), as if I load with -l from the command line, then the bug does not occur whereas I do the same in the sequel. Attached is the new trace, it seems to be the roughly same as the previous one, although I do load template.el (not template.elc). Maybe we could skype together and by sharing my screen you would see the recipe to make the bug happen and could give further guidance for me to provide more information. Anyway I reported the bug because I was quite annoyed by it, but now that I have a correct version of my macros (without the errors occurring during template-new-file) there is no more crash, and I am less annoyed. But well, it would be good to know whether this is something that was corrected or not. I am not sure that my trying the same recipe with the latest version of Emacs would answer this question, because as I wrote early the recipe is not so simple... Vincent. --- Ce courrier électronique ne contient aucun virus ou logiciel malveillant parce que la protection avast! Antivirus est active. https://www.avast.com/antivirus