jiri/org-ql
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Update commentary

Browse source 
We aren't using org-element-parse-buffer anymore.
This commit is contained in:
Adam Porter 2018-05-10 16:46:06 -05:00
parent 2aa3ba008e
commit 89ae2ecf4a
1 changed files with 12 additions and 15 deletions
Download patch file Download diff file Expand all files Collapse all files

27
org-agenda-ng.el
View file

@ -7,28 +7,25 @@
;; written in a more functional way, to avoid making multiple passes ;; written in a more functional way, to avoid making multiple passes
;; through each file when building a multi-day agenda. ;; through each file when building a multi-day agenda.
;; Unfortunately, however, it's more like a proof-of-nope, because ;; It also functions as a kind of "query language" for Org buffers.
;; it's significantly slower than the existing agenda code. Profiling
;; shows that =org-element-parse-buffer= is just not a fast function,
;; so most of the time spent is in that function, parsing the whole
;; buffer into a lisp tree. The existing agenda code is not written
;; in a functional style, and it makes multiple passes through each
;; file when preparing a multi-day agenda view, but it avoids parsing
;; the /entire/ buffer the way =org-element-parse-buffer= does.
;; So I guess this approach won't work, unless someone can come up
;; with a version of =org-element-parse-buffer= that's about 10x
;; faster.
;;;; Principles ;;;; Principles
;;;;; Work on elements ;;;;; Try to imitate traditional Org agenda code's results and methods
;; We want to work on elements as returned by =org-element-parse-buffer=. We only convert to strings at the very end, when we send the list of those to =org-agenda-finalize-entries=. ;; The traditional agenda code is complicated, but well-optimized. We should imitate it where
;; possible. We should also attempt to return the same results as the traditional agenda code
;; (ultimately, that is; but whether we reach that level of complexity depends on, e.g. performance
;; achieved, whether it looks like a feasible alternative, etc). At the least, we should return
;; results in the same format, so they can be used by other code that uses agenda output
;; (e.g. org-agenda-finalize-entries, org-agenda-finalize, org-super-agenda, etc). Of course, this
;; goal does not necessarily apply to the "query language"-like features, and ideally the
;; agenda-like features should build upon those.
;;;;; Preserve the call to =org-agenda-finalize-entries= ;;;;; Preserve the call to =org-agenda-finalize-entries=
;; We want to preserve the final call to =org-agenda-finalize-entries= to preserve compatibility with other packages and functions. ;; We want to preserve the final call to org-agenda-finalize-entries to preserve compatibility
;; with other packages and functions.
;;; Code: ;;; Code: