I think I addressed the raised issues. Just merged org-fold upstream.
Kévin Le Gouguec <kevin.legoug...@gmail.com> writes: > My recipe (based on commit f9dd109bc, Emacs 29.0.50 commit 864c8013fd): > > $ git switch feature/org-fold-universal-core-tidy > $ make autoloads > $ emacs -Q -L lisp -eval "(setq org-startup-folded t)" repro.org > > Restarting Emacs with the above between each step: > > (1) C-s abc ⇒ no logbook is unfolded, > (2) C-s def ⇒ no logbook is unfolded, > (3) C-s ghi ⇒ some logbooks are unfolded. > > Assuming you can reproduce: is it expected that logbooks are expanded in > case (3)? I don't see what's "conceptually" different in situation (3) > vs. (1) and (2), so I'm puzzled to get different results. I had a hard time to reproduce your recipe because it was related to font-locking. The only way was increasing the font size, so that line widths were larger than the frame width. I believe I fixed the issue now. > Also, a bit of idle curiosity: > >> (defun org-fold--isearch-reveal (&rest _) >> "Reveal text at POS found by isearch." >> (org-fold-show-set-visibility 'isearch)) > > org-fold-show-set-visibility calls either > org-fold-show-set-visibility--overlays, or > org-fold-show-set-visibility--text-properties, and AFAICT neither of > these handle 'isearch as an argument… Is there a (cdr (assq 'isearch > org-fold-show-context-detail)) missing? > > (This comes from a very cursory reading of the code; apologies if I've > missed something) You are right. I supposed to use org-fold-show-context. Fixed now. Best, Ihor