Sven,
I'm triaging old Emacs bugs.
Thanks for doing this! I'm sure with a package like Emacs it must be quite a task.
Your e-mail prompted me to do some binary searching on my .emacs and finally make some progress on this very old annoyance.
1. When spell-checking using ispell. After finding a misspelled word, Emacs highlights it and brings up a buffer offering some replacements, each one tagged with a digit. What is supposed to happen is that I type the digit, and that replacement is substituted; what actually happens is that Emacs complains with the message above. There are several other command that one can give ispell at this point; those work, but typing a letter which is not a valid command gives the same message.
OK, here's a specific recipe to reproduce. This is with emacs21 version 21.4a+1-5.6.
1. Create a file "dotemacs-test" with the contents (one line): (setq help-char "\M-h") 2. Run emacs with "emacs -q -l dotemacs-test".3. Type the characters "glox" into the *scratch* buffer. (I assume any misspelled word will work. I use the American dictionary.)
4. Type M-$ (meta-shift-4). 5. Emacs opens a small buffer "*Choices*" with the following content: (0) glob (1) glop (2) glow (3) g lox (4) g-lox (5) lox 6. Type the character "1". Expected behavior: - In the scratch buffer, "glox" is replaced with "glop". Actual behavior: - No replacement. Emacs complains (in the minibuffer): Wrong type argument: number-or-marker-p, "\350"
If you are still seeing this, please send a Lisp backtrace for the error (use M-x toggle-debug-on-error to enable the backtrace) and also check if this appears in "emacs21 -q".
It does happen with emacs -q (on my system emacs -> emacs21). It also does not happen if I comment out the line noted in step #1 above from my regular .emacs.
Here you go on the backtrace:
Debugger entered--Lisp error: (wrong-type-argument number-or-marker-p "è") ispell-command-loop(("glob" "glop" "glow" "g lox" "g-lox" "lox") nil "glox" #<marker at 1 in *scratch*> #<marker at 5 in *scratch*>) byte-code("Ä^...@^haaa@^H@ \n%
For case #2 (file changed on disk), I have attached the backtrace as it seems to contain some funny characters which prevent me from pasting it.
Also, do you see this with emacs22 as well?
I haven't tried emacs22. I tend to resist Emacs version upgrades because the changelogs are so massive and it takes a while to read them and figure out what's going to affect me. However, I can install it and try if needed.
Please let me know what additional help I can provide. Thanks, Reid -- http://blog.reidster.net
case2-backtrace
Description: Binary data