[ Readding tomás since this thread got more high-level again ... ]
On 2025-04-20 13:27, Ihor Radchenko wrote:
> Looks reasonable.
Thanks. More or less by chance I let Org's ERT tests run with
`org-timestamp-formats' set to `("%Y-%m-%d" . "%Y-%m-%d %H:%M")'.
(I do not intend to do anything like that in the final patch, of
course: There I'd just let-bind `org-timestamp-formats' in the new
tests.)
That inadvertent change resulted in quite some failures and me
adding the following comment to variable `org-timestamp-formats':
;; The weekday name "%a" is considered semi-optional in these formats,
;; see https://list.orgmode.org/87fricxatw.fsf@localhost/. It is
;; "optional" because the `org-timestamp-*' functions work alright on
;; weekday-less timestamps in paragraphs when one omits the "%a". But
;; it is only "semi"-optional since Org cannot properly process
;; timestamps in CLOCK, DEADLINE, and SCHEDULED lines when one omits
;; the "%a".
I haven't yet followed up on the failing tests, and whether it
would be easy to make them work without the "%a". Plus I do not
consider this a problem for me, since I am a happy deadline-free
Org mode user.
But I wonder if your opinion w.r.t. my previous patch or omitting
the "%a" from `org-timestamp-formats' in general changes by this
new fact?
Thanks.
