org-ql-find does not take you to the search result you select

[oantolin]

Or rather it does, but it does so inside a org-with-point-at which expands to a form including save-excursion so the jumping gets undone right away! (There's an easy work around: just use embark-dwim, but I would much rather just press RET.) This is with org-ql version 0.8-pre from MELPA, build 20231020.244.

[alphapapa]

Hi Omar,

Thanks for reporting that. I hadn't noticed since I use org-tree-to-indirect-buffer in the org-ql-find-goto-hook.

I think this commit should fix it: 631cd80 Please let me know if it does for you.

[oantolin]

Yes, that works perfectly. Thank you!

[telenieko]

THANK YOU, @oantolin !!! This has been driving me crazy for weeks now, and I had no idea where, even on which package, the problem was!!

[alphapapa]

@telenieko Glad that it's fixed for you. Could you tell me a bit more about how you're using org-ql? I'm surprised it wasn't more obvious where the problem was.

[oantolin]

Yeah, I'm curious about that too, @telenieko.

[telenieko]

@telenieko Glad that it's fixed for you. Could you tell me a bit more about how you're using org-ql? I'm surprised it wasn't more obvious where the problem was.

There are two places that were broken in my system,

In a notmuch-show buffer I can call mf-email-get-links-to-current-thread and, using org-ql I will find any Org entries that contain links to any e-mails of the e-mail thread being visited. (#279 has some background on the org-backlinks predicate):

(defun mf-email-get-links-to-current-thread()
  "Find org items linking to this email thread."
  (interactive)
  (let* ((message-ids (notmuch-show-get-messages-ids))
         (targets (mapconcat (lambda (tgt) (format "\"notmuch:%s\"" tgt))
			     message-ids))
         (query (format "org-backlinks:%s" targets))
         (title (format "Tasks linked to: %s"
                        (notmuch-show-get-subject)))
         )
    (org-ql-find (org-agenda-files t t)
		 :query-prefix query
		 :prompt "Tareas relacionadas: ")))

I also have a very simple search function bound to a key for "full text" searching:

(defun mf-pkm-todo-search (arg)
  "Search for TODO. If ARG search archives too."
  (interactive "P")
  (let ((include-archives (not (eq nil arg))))
    (org-ql-find (org-agenda-files nil include-archives))))

You see both functions use org-ql-find to allow me to refine the search in the minibuffer (specially on the e-mail searching function).

Back to the obviousness: Now that @oantolin mentioned it... I feel ashamed of missing this behaviour of org-with-point-at (well, not ashamed, I'm a newbie still!). But before then my brain just did not make the connection.

[alphapapa]

In a notmuch-show buffer I can call mf-email-get-links-to-current-thread and, using org-ql I will find any Org entries that contain links to any e-mails of the e-mail thread being visited. (#279 has some background on the org-backlinks predicate):

Cool, thanks. Quick thing I noticed that you might want to be aware of: you seem to be concatting targets without commas between tokens, i.e. IIUC it will turn into something like org-backlinks:"notmuch:FOO""notmuch:BAR". Instead the query should look something like org-backlinks:notmuch:FOO,notmuch:BAR (assuming that the org-backlinks predicate handles multiple arguments correctly).

Back to the obviousness: Now that @oantolin mentioned it... I feel ashamed of missing this behaviour of org-with-point-at (well, not ashamed, I'm a newbie still!). But before then my brain just did not make the connection.

Well, you shouldn't feel bad, after all, I misused it myself. :)

Also, friendly tip: Instead of (not (eq nil arg)) you can use the classic trick to "booleanize" it, (not (not arg)).

[telenieko]

Thanks for the tips @alphapapa! will polish my code a little bit! :D

For anyone coming here in the future: I fixed the targets concatenation by passing a SEPARATOR parameter to mapconcat.

[bram85]

Commit 8e9a8fe reintroduces org-with-point-at, essentially reintroducing this bug again in release 0.8.1.

[alphapapa]

Commit 8e9a8fe reintroduces org-with-point-at, essentially reintroducing this bug again in release 0.8.1.

Argh. Thanks for reporting.

[alphapapa]

This should be fixed now, on master and in v0.8.2. Thanks, @bram85.

[oantolin]

I'm still seeing the bug after that commit. You did replace org-with-point-at with without-restriction, but you forgot to actually move point. The code in commit 631cd80 included a (goto-char marker) which you don't have anymore.

[alphapapa]

@oantolin Thanks. I've just pushed fixes as v0.8.3 and to master. I've tested it more carefully, and it seems to work correctly in various circumstances. Please let me know if you find any more problems.

[oantolin]

Thank you! I can confirm that fix works for me.