Dear Ivan, Followup to yor unanswered questions of yesterday :
[ Snip... ] > The best would be if you could give a minimal recipe starting from `emacs >> -Q`. >> > > Won't do any good to you : my "(require sage "sage") is in my .emacs, > which won't get searchde when starting emacs with -Q flag. > > > What I meant was something like > > $ emacs -Q > > ;; Evaluate in scratch: > (add-to-list 'load-path "/usr/local/sage/local/share/emacs") > (require 'sage "sage") > (setq sage-command "/usr/local/bin/sage") > (require 'sage-view "sage-view") > (add-hook 'sage-startup-after-prompt-hook 'sage-view) > > M-x sage RET > > C-x C-c > > Repeat. > > Is that enough to reproduce it? That eliminates other problems in your > .emacs file, though I don't think that's very likely the problem. I think > it has to be something in .sage, but I've been wrong before. :-) > > Same result as before : first time (after removing ~/.sage), wiorks like a charm, second time : dead emacs. > The simplest way : start emacs from command line or from a gnome icon, > type "M-x sage" in the startum screen, and voilà : first time it works like > a charm, second time (= second emacs invocation) : a dead emacs. > > > What happens if you turn on debugging (M-x toggle-debug-on-quit RET) and > then run M-x sage and press C-g when it's frozen? > Aha ! When, I press ^G, the frame gets divided in two windows : *SAGE-newerr* and *Backtrace*. Content off *SAGE-newerr* : Debugger entered--Lisp error: (quit) python-send-string("%colors NoColor") sage-send-command("%colors NoColor" nil) sage-send-startup-before-prompt-command() run-hooks(sage-startup-before-prompt-hook) sage(nil) call-interactively(sage record nil) command-execute(sage record) execute-extended-command(nil "sage") call-interactively(execute-extended-command nil nil) Content of *Backtrace* : ---------------------------------------------------------------------- | Sage Version 5.9, Release Date: 2013-04-30 | | Type "notebook()" for the browser-based notebook interface. | | Type "help()" for help. | ---------------------------------------------------------------------- RET KeyboardInterrupt sage: Quite interestingly, sage is still active, and answers to command, but output is *not* typeset : sage: 2+2 4 sage: integrate(tan(x), x) log(sec(x)) sage: quitting sage then C-x C-c does a "normal" emacs exit (no hangup). Repeating leads to the same result. I tried to disable online typesetting and online plot in the .emacs file : I still get a dead emacs when requesting a sage session... > Now for the weirder part : I observed this behaviour on two machines : a > "normal" desktop and a large desktop used as a server. On a third machine > (a small notebook I bring with me almost everywhere), things work as > advertised. I can't for the life of me state what differs between these > machines re: sage and/or emacs. All of them also have texlive, R and maxima > installations, all of them run Debian testing (updated often). > > > Is there any way I could get a temporary account on the server to see if I > can debug it there? Email me privately with username and password if so. > OK : what is your mail address ? Mail me at emanuel period charpentier hat gmail.com > Do they all have the same versions of Emacs, Sage, and same configuration > files? Perhaps .bashrc could come into play as well by setting the > environment variable RANDOM_BUGS=1 :-) > Currently : "big desktop" : sage 5.10, "small desktop" and "notebook" : sage 5.9. In all three cases, I updated R to 3.0.1 (see TRAC#14706) and installed database_gap_4.5.7, dot2tex and sage_mode_0.9.1. All of them are running Debian jessie, updatete about daily. The relevant .bashrc are about pristine. "Big_desktop" has a sage-5.10 directory belonging to user "sage", symlinked to /usr/local/sage. "Small desktop" has a /usr/local/sage-5.9 directory belonging to me, symlinked as /usr/local/sage. "notebook" has a sage5-9 directory in my home directory, symlinked as /usr/local/sage. In all cases, /usr/local/bn/sage is a symlink to /usr/local/sage/sage. > > Heisenbug ? > > > Well, I've certainly never seen it, but there's probably a good reason for > it. > Schrödinbug ?? Again, thank you for your time and your help ! Emmanuel Charpentier -- You received this message because you are subscribed to the Google Groups "sage-support" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-support+unsubscr...@googlegroups.com. To post to this group, send email to sage-support@googlegroups.com. Visit this group at http://groups.google.com/group/sage-support. For more options, visit https://groups.google.com/groups/opt_out.