Hello, In commit 0260d2fcf603f30210e2b95d37727edd832c12e9 of 2021-04-26, the regexp used in `org-agenda-prefix-format' was updated to use a non-greedy regexp, in order to allow multiple %(expression) instances in `org-agenda-prefix-format'. Unfortunately, this change breaks the ability to use nested forms in %(expression) instances. For example, I use the following setting for `org-agenda-prefix-format': (setf (alist-get 'agenda org-agenda-prefix-format) " %(format-time-string \"%a\" (org-get-deadline-time (point))) ") When running `org-agenda', I now receive the error `org-compile-prefix-format: End of file during parsing'. This is because of the updated regexp match, which stops at the first closing parenthesis: ELISP> myfmt " %(format-time-string \"%a\" (org-get-deadline-time (point))) " ELISP> (and (string-match "%\\(\\?\\)?\\([-+]?[0-9.]*\\)\\([ .;,:!?=|/<>]?\\)\\([cltseib]\\|(.+?)\\)" myfmt) (match-string 4 myfmt)) "(format-time-string \"%a\" (org-get-deadline-time (point)" With the previous regexp, it works fine: ELISP> (and (string-match "%\\(\\?\\)?\\([-+]?[0-9.]*\\)\\([ .;,:!?=|/<>]?\\)\\([cltseib]\\|(.+)\\)" myfmt) (match-string 4 myfmt)) "(format-time-string \"%a\" (org-get-deadline-time (point)))" Thanks, Radon Rosborough
Radon Rosborough <radon.neon@gmail.com> writes:
> Hello,
>
> In commit 0260d2fcf603f30210e2b95d37727edd832c12e9 of 2021-04-26, the
> regexp used in `org-agenda-prefix-format' was updated to use a
> non-greedy regexp, in order to allow multiple %(expression) instances
> in `org-agenda-prefix-format'. Unfortunately, this change breaks the
> ability to use nested forms in %(expression) instances. For example, I
> use the following setting for `org-agenda-prefix-format':
Confirmed.
Sorry, I noticed this problem after sending the patch, but I do not know
an easy way to solve it. To handle sexps properly, we need to do
something more complex than just a simple regexp. Probably, using
org-element--parse-paired-brackets or similar code.
Fixing this is (down) in my todo-list, but patches are welcome.
Best,
Ihor
[-- Attachment #1: Type: text/plain, Size: 343 bytes --] If you insert the format string into a temporary buffer, then the `read' function will read up to the closing parenthesis and update point accordingly. That would be a foolproof way to handle all possible Lisp forms: use a regexp to find the start, and then use `read' to find the end. Then proceed with regexp matching, and repeat like that. [-- Attachment #2: Type: text/html, Size: 373 bytes --]
[-- Attachment #1: Type: text/plain, Size: 466 bytes --] Radon Rosborough <radon.neon@gmail.com> writes: > If you insert the format string into a temporary buffer, then the `read' > function will read up to the closing parenthesis and update point > accordingly. That would be a foolproof way to handle all possible Lisp > forms: use a regexp to find the start, and then use `read' to find the end. > Then proceed with regexp matching, and repeat like that. Thanks for the suggestion! The patch is attached. Best, Ihor [-- Warning: decoded text below may be mangled, UTF-8 assumed --] [-- Attachment #2: 0001-Allow-sexps-with-nested-parenthesis-in-org-agenda-pr.patch --] [-- Type: text/x-diff, Size: 1862 bytes --] From ffbf7fe29913e33e3f20544be49b6cc2c0ef6409 Mon Sep 17 00:00:00 2001 Message-Id: <ffbf7fe29913e33e3f20544be49b6cc2c0ef6409.1622361756.git.yantar92@gmail.com> From: Ihor Radchenko <yantar92@gmail.com> Date: Sun, 30 May 2021 16:00:07 +0800 Subject: [PATCH] Allow sexps with nested parenthesis in org-agenda-prefix-format * lisp/org-agenda.el (org-compile-prefix-format): Use `read' instead of pure regexp matching to determine bounds of %(sexp) expressions in `org-agenda-prefix-format'. Fixes https://orgmode.org/list/87fsy8yi1e.fsf@localhost/ --- lisp/org-agenda.el | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/lisp/org-agenda.el b/lisp/org-agenda.el index f49a24f88..fcfcd872e 100644 --- a/lisp/org-agenda.el +++ b/lisp/org-agenda.el @@ -6942,7 +6942,7 @@ (defun org-compile-prefix-format (key) (and (string-match "\\.[0-9]+" x) (string-to-number (substring (match-string 0 x) 1))))))) (if (eq var 'eval) - (setq varform `(format ,f (org-eval ,(read (match-string 4 s))))) + (setq varform `(format ,f (org-eval ,(read (substring s (match-beginning 4)))))) (if opt (setq varform `(if (member ,var '("" nil)) @@ -6951,7 +6951,12 @@ (defun org-compile-prefix-format (key) (setq varform `(format ,f (if (member ,var '("" nil)) "" (concat ,var ,c (get-text-property 0 'extra-space ,var))))))) - (setq s (replace-match "%s" t nil s)) + (if (eq var 'eval) + (setf (substring s (match-beginning 0) + (+ (match-beginning 4) + (length (format "%S" (read (substring s (match-beginning 4))))))) + "%s") + (setq s (replace-match "%s" t nil s))) (push varform vars)) (setq vars (nreverse vars)) (with-current-buffer (or org-agenda-buffer (current-buffer)) -- 2.26.3
[-- Attachment #1: Type: text/plain, Size: 170 bytes --] Ah yes, you can also check the length of the form that was read. Good call, I didn't think of that option, which is clearly better than using a temporary buffer. Thanks! [-- Attachment #2: Type: text/html, Size: 195 bytes --]
Ihor Radchenko <yantar92@gmail.com> writes:
> Thanks for the suggestion! The patch is attached.
Applied, thanks.
--
Bastien