Hello,
I have a `org-init.el' file in my home directory, with common settings used
when in batch mode.
That works well when that file is a regular file:
--8<---------------cut here---------------start------------->8---
sva@MEDIACENTER$ make html
Exporting file /cygdrive/c/home/sva/file.txt to HTML...
emacs --batch -Q --eval "(setq debug-on-error t) (add-to-list 'load-path
\""~/src/org-mode/lisp"\") (add-to-list 'load-path
\""~/src/org-mode/contrib/lisp"\")" -l "~/org-init.el" file.txt -f
org-export-as-html
Emacs 24.0.93.1
Org-mode version 7.8.03
OVERVIEW
Loading vc-svn...
Loading vc-git...
Exporting...
Exporting...
Saving file c:/home/sva/file.html...
Wrote c:/home/sva/file.html
HTML export done, pushed to kill ring and clipboard
--8<---------------cut here---------------end--------------->8---
However, if I put it somewhere, and make a symlink from my home directory to
its real place, Emacs fails loading it:
--8<---------------cut here---------------start------------->8---
sva@MEDIACENTER$ make html
Exporting file /cygdrive/c/home/sva/file.txt to HTML...
emacs --batch -Q --eval "(setq debug-on-error t) (add-to-list 'load-path
\""~/src/org-mode/lisp"\") (add-to-list 'load-path
\""~/src/org-mode/contrib/lisp"\")" -l "~/org-init.el" file.txt -f
org-export-as-html
Debugger entered--Lisp error: (void-variable !<symlink>ÿþ/)
eval-buffer(#<buffer *load*> nil "c:/home/sva/org-init.el" nil t) ; Reading
at buffer position 14
load-with-code-conversion("c:/home/sva/org-init.el" "c:/home/sva/org-init.el"
nil t)
load("c:/home/sva/org-init.el" nil t)
command-line-1(("--eval" "(setq debug-on-error t) (add-to-list 'load-path
\"~/src/org-mode/lisp\") (add-to-list 'load-path
\"~/src/org-mode/contrib/lisp\")" "-l" "~/org-init.el" "file.txt" "-f"
"org-export-as-html"))
command-line()
normal-top-level()
Makefile:33: recipe for target `file.html' failed
make: *** [file.html] Error 127
--8<---------------cut here---------------end--------------->8---
I find this very strange, as I thought the symlink feature would be
transparent for upper applications. This does not seem to be the case, though.
Any idea on how to circumvent this, if possible?
Best regards,
Seb
--
Sebastien Vauban
