No Wayman <iarchivedmywholel...@gmail.com> writes:
> The logic for saving the archive buffer in org-archive-subtree
> does not consider the (default) value of 'from-org for
> org-archive-subtree-save-file-p.
Hmm, indeed, the change from 3d0282ef8 (New option
`org-archive-subtree-save-file-p', 2020-01-31) looks incomplete.
> While patching this, I realized I'm not sure I understand the
> intended logic of each value for org-archive-subtree-save-file-p.
>From digging a bit, I think the history went like this:
* All the way back in 4be4c5623 (version 4.12a, 2008-01-31),
org-archive-subtree saved the archive buffer, except when it was the
current buffer.
[ The next two aren't relevant for the save part, but I'm noting
because we should cleaned them up. ]
* As of bf09955fe (Bugfix for `org-archive-subtree', 2008-03-01), the
buffer was killed after being saved.
* 343f3c478 (Keep archive buffer after archiving something to it,
2009-10-28) reverted the killing, though it left behind a stale
comment saying the buffer will be killed.
* 63f6e851b (Do not save target buffer after archiving subtree,
2017-11-25) stopped saving the buffer for performance reasons.
* Some users preferred the older behavior. An associated Debian bug
was opened.
https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=887332
https://yhetil.org/orgmode/f74fce15-c2ea-048e-b2ef-7ad40e717...@arfer.net/
In response, Bastien committed 3d0282ef8 to offer more control over
when saving happens.
> When setting it to 't', the defcustom :tag string claims "Always
> save the archive buffer".
> This is not the case if archiving from within the current buffer.
> Perhaps a clearer :tag string?
>
> #+begin_src emacs-lisp
> (defcustom org-archive-subtree-save-file-p 'from-org
> ;;...
> (const :tag "Save the archive buffer unless it is the current
> buffer" t)
> ;;...
> ))
> #+end_src
Yeah, I'd say that's an improvement, though the same "unless it is the
current buffer" would be true for from-org as well, were it wired up.
> The value 'from-org also still saves the archive buffer when
> archiving from a buffer that is not in Org mode.
I'm confused by this statement. To me it looks like save-buffer will
_never_ be called when org-archive-subtree-save-file-p is from-org:
(unless (eq this-buffer buffer)
(when (or (eq org-archive-subtree-save-file-p t)
(and (boundp 'org-archive-from-agenda)
(eq org-archive-subtree-save-file-p 'from-agenda)))
(save-buffer)))
> I'm not entirely sure of its purpose. If the intent is to allow an
> option that prevents saving only when archiving from the agenda,
> I suggest a single option excluding that case and saving for
> other, non-nil values:
Based on the history above, I believe the main purpose is to give users
a way to reverse the "no saving" behavior made in 63f6e851b (Do not save
target buffer after archiving subtree, 2017-11-25). I'm _guessing_
that, on top of that, the idea adding a from-agenda value was that some
users may want to save only when archiving from the agenda, because in
that case they're a bit removed from the buffer and might not think to
save it, or something along those lines.
> #+begin_src emacs-lisp
> (defcustom org-archive-subtree-save-file-p 'unless-agenda
> "Conditionally save the archive file after archiving a subtree.
> The value 'unless-agenda prevents saving from the agenda-view.
> Other non-nil values save the archive buffer unless it is the
> current buffer."
> :group 'org-archive
> :package-version '(Org . "9.4")
> :type '(choice
> (const :tag "Do not save archive buffer when archiving
> from an agenda view" unless-agenda)
> (const :tag "Save the archive buffer unless it is the
> current buffer" t)
> (const :tag "Do not save the archive buffer")))
> #+end_src
>
> Then the saving logic in org-archive-subtree becomes:
>
> #+begin_src emacs-lisp
> (when org-archive-subtree-save-file-p
> (unless (or (eq buffer this-buffer)
> (and (eq org-archive-subtree-save-file-p
> 'unless-agenda)
> ;;bound when called from org-agenda.el
> (boundp 'org-archive-from-agenda)))
> (save-buffer)))
> #+end_src
Assuming what I said above is true, I think what you propose here loses
the ability to save only when archiving from the agenda. And more
importantly, users would not be able to give a blanket "don't save" in
order to retain the behavior introduced by 63f6e851b (Do not save target
buffer after archiving subtree, 2017-11-25).
