Nicolas Goaziou <n.goaz...@gmail.com> writes: > Ryo TAKAISHI <ryo.takaish...@gmail.com> writes: > >> I did'nt come up with to use it. >> But "%(func %:description)" or "%(func (plist-get >> org-store-link-plist :description))", I think the former is readble >> template than the latter. > > Probably, but it's also more error-prone. > > For example, your code operates only at top-level, i.e. it won't handle > something like: > > %(fun arg1 (sub-fun %:description)) > > It will also error out if any atom isn't a symbol, i.e.: > > %(format "%d" 1) > > It's all about pro and cons. > > > Regards,
I fix these problem. A new patch expand keyword recursively, and only symbol. Regards, --- lisp/org-capture.el | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/lisp/org-capture.el b/lisp/org-capture.el index 1dfffc6..b8c9844 100644 --- a/lisp/org-capture.el +++ b/lisp/org-capture.el @@ -1621,10 +1621,23 @@ The template may still contain \"%?\" for cursor positioning." (goto-char (match-beginning 0)) (let ((template-start (point))) (forward-char 1) - (let ((result (org-eval (read (current-buffer))))) + (let* ((sexp (mapcar 'org-capture-expand-keyword-in-embedded-elisp + (read (current-buffer)))) + (result (org-eval sexp))) (delete-region template-start (point)) (insert result)))))) +(defun org-capture-expand-keyword-in-embedded-elisp (attr) + (cond ((consp attr) + (mapcar 'org-capture-expand-keyword-in-embedded-elisp attr)) + ((symbolp attr) + (let* ((attr-symbol (symbol-name attr)) + (key (and (string-match "%\\(:.*\\)" attr-symbol) + (intern (match-string 1 attr-symbol))))) + (or (plist-get org-store-link-plist key) + attr))) + (t attr))) + (defun org-capture-inside-embedded-elisp-p () "Return non-nil if point is inside of embedded elisp %(sexp)." (let (beg end) -- 1.7.9.5