Add notes

This commit is contained in:
Adam Porter 2018-01-06 04:21:21 -06:00
parent 949ef58c61
commit 8d096188bb

View file

@ -6,6 +6,37 @@
[2017-12-31 Sun 17:54] I wonder if, instead of parsing the whole buffer with =org-element-parse-buffer=, we could simply work on a list of heading positions, e.g. a loop would search forward to the next heading position, then call whatever predicates it needed at the heading's position, using =save-excursion= around each function call. The predicates would need to be updated to get their data from the buffer, instead of using =org-element-property=, but that wouldn't be hard.
* Examples / testing
#+BEGIN_SRC elisp
(org-agenda-ng org-agenda-files
(and (or (date :deadline '<= (org-today))
(date :scheduled '<= (org-today)))
(not (apply #'todo org-done-keywords-for-agenda)))
((group (tags "bills"))
(group (todo "SOMEDAY"))))
(org-agenda-ng org-agenda-files
(and (or (date :deadline '<= (org-today))
(date :scheduled '<= (org-today)))
(not (apply #'todo org-done-keywords-for-agenda))))
(org-agenda-ng "~/org/main.org"
(and (or (date :deadline '<= (org-today))
(date :scheduled '<= (org-today)))
(not (apply #'todo org-done-keywords-for-agenda))))
(org-ql org-agenda-files
(and (todo "SOMEDAY")
(tags "Emacs")))
(org-ql org-agenda-files
(and (todo "SOMEDAY")
(tags "Emacs")
(priority >= "B")))
(org-ql "~/org/main.org"
(or (tags "Emacs")
(priority >= "B")))
#+END_SRC
* Profiling
This basically works, as a very basic kind of agenda view, but we can already see that it's much slower (at least, for single-day views) because =org-element-parse-buffer= is slow compared to the agenda code.
@ -303,7 +334,6 @@ outline-on-heading-p 3452 0.0101
string-match 4594 0.0064121759 1.395...e-06
mapcar 30 0.0041008740 0.0001366958
#+end_example
* Profiling position-based
** Macro