Kyle Meyer writes: > Thanks for the patch. > > stardiviner writes: > >> On Wed, Dec 2, 2020 at 5:30 PM stardiviner wrote: >> >>> The default [C-c C-q] completing tags only retrieve tags from current >>> buffer locally. >>> >>> By this patch, will merge both buffer-local tags and user defined global >>> `org-tags-alist`. > > It does a bit more than that. It uses org-global-tags-completion-table, > which considers tags in all agenda files by default and takes into > account org-tag-alist (as well as org-tag-persistent-alist) via the use > of the org-current-tag-alist variable. > That's what I want. Why obviously user pre-defined tags can't be used globally. Right? It should be. >>> This is more reasonable. > > I'd guess that depends on the user. I personally wouldn't like to see > tags from all of my agenda files, and I'm fine not seeing > org-tag{-persistent}-alist ones that aren't in the current buffer given > that they have fast selection keys. Currently tags selection is now slow, even with 100 pre-defined tags. Should be fine. Tags sometimes is abundant. So if it is slow, then it means the command =org-set-tags-command= should be optimized. > >> Subject: [PATCH] org.el: Complete tags from both global and buffer local >> >> * lisp/org.el: (org-fast-tag-selection): merge buffer local tags with >> global alist of tags. > > Convention/consistency nits: spurious ":" after ".el" and > s/merge/Merge/. Updated. Thanks > >> --- >> lisp/org.el | 11 +++++++---- >> 1 file changed, 7 insertions(+), 4 deletions(-) >> >> diff --git a/lisp/org.el b/lisp/org.el >> index 0e12e4b15..287b8c407 100644 >> --- a/lisp/org.el >> +++ b/lisp/org.el >> @@ -12256,10 +12256,13 @@ (defun org-fast-tag-selection (current inherited table &optional todo-table) >> (condition-case nil >> (setq tg (completing-read >> "Tag: " >> - (or buffer-tags >> - (with-current-buffer buf >> - (setq buffer-tags >> - (org-get-buffer-tags)))))) >> + (delq nil >> + (delete-dups >> + (append (or buffer-tags >> + (with-current-buffer buf >> + (setq buffer-tags >> + (org-get-buffer-tags)))) >> + (org-global-tags-completion-table)))))) > > This change in behavior should come with a NEWS entry and a > documentation update. What the manual currently says is now stale: > > - {{{kbd(TAB)}}} :: > > #+kindex: TAB > Enter a tag in the minibuffer, even if the tag is not in the > predefined list. You can complete on all tags present in the > buffer. You can also add several tags: just separate them with > a comma. > > As I mentioned above, though, I'm not sure always adding agenda tags is > desirable. However, I think it'd probably be safe to look at > org-complete-tags-always-offer-all-agenda-tags as an indication of > whether the user wants this behavior. org-set-tags-command already > considers that option when it generates the table that it passes to > org-fast-tag-selection. So perhaps we could just consider the table > when calling completing-read for the tab key (something along the lines > of the patch at the end of the email). > Indeed, I add condition on ~org-complete-tags-always-offer-all-agenda-tags~ now. > Conceptually that's been discussed/tried before, but it was then backed > out of: > > * https://orgmode.org/list/F753E612-2D5D-4BA7-AF0C-D49C7A8DDA24@pobox.com/T/#u > * 647396464 (org.el: Include tags from `org-tag-alist' when completing > with the TAB key, 2012-03-27) > * d4ddcbb8b (Revert "org.el: Include tags from `org-tag-alist' when > completing with the TAB key.", 2012-04-10) > * acc7a0b2b (org.el: Include `org-tag-alist' in the list for tag > completions, 2012-03-27) > I checked this mailing list thread, the original commit use ~(mapcar 'car table)~. It's more simple. I updated my code to this. My patch only changed the ~completing-read~ part. And I found ido support option is been removed in Org. I think this is right decision. > At a quick glance, I think the patch below avoids the problems that led > to 647396464 being reverted, but that'd need to be checked more > carefully. > > > diff --git a/lisp/org.el b/lisp/org.el > index 5b0ae389c..9383719e3 100644 > --- a/lisp/org.el > +++ b/lisp/org.el > @@ -12139,7 +12139,7 @@ (defun org-fast-tag-selection (current inherited table &optional todo-table) > fulltable)))) > (buf (current-buffer)) > (expert (eq org-fast-tag-selection-single-key 'expert)) > - (buffer-tags nil) > + (tab-tags nil) > (fwidth (+ maxlen 3 1 3)) > (ncol (/ (- (window-width) 4) fwidth)) > (i-face 'org-done) > @@ -12274,16 +12274,22 @@ (defun org-fast-tag-selection (current inherited table &optional todo-table) > (setq current nil) > (when exit-after-next (setq exit-after-next 'now))) > ((= c ?\t) > + (unless tab-tags > + (setq tab-tags > + (delq nil > + (mapcar (lambda (x) > + (let ((item (car-safe x))) > + (and (stringp item) > + (list item)))) > + (org--tag-add-to-alist > + (with-current-buffer buf > + (org-get-buffer-tags)) > + table))))) > (condition-case nil > - (setq tg (completing-read > - "Tag: " > - (or buffer-tags > - (with-current-buffer buf > - (setq buffer-tags > - (org-get-buffer-tags)))))) > + (setq tg (completing-read "Tag: " tab-tags)) > (quit (setq tg ""))) > (when (string-match "\\S-" tg) > - (cl-pushnew (list tg) buffer-tags :test #'equal) > + (cl-pushnew (list tg) tab-tags :test #'equal) > (if (member tg current) > (setq current (delete tg current)) > (push tg current))) Hmm, I tested your upper patch, works same. But looks safer. So I applied your patch in my patch. Actually it's totally your code.... Hahaha Anyway, I re-generated patch. Bother you to review it. :smile: -- [ stardiviner ] I try to make every word tell the meaning that I want to express. Blog: https://stardiviner.github.io/ IRC(freenode): stardiviner, Matrix: stardiviner GPG: F09F650D7D674819892591401B5DF1C95AE89AC3