Ihor Radchenko <yanta...@posteo.net> writes:

> David Masterson <dsmaster...@gmail.com> writes:
>>> AFAIK, closed date intervals with time are currently not supported by
>>> org-agenda. You can instead use diary sexp timestamps, a bunch of active
>>> timestamps in the body, or M-x org-clone-subtree-with-time-shift
>> "a bunch of active timestamps"?!?  Ooh!  I hadn't gotten the impression
>> that that was allowed from the "Dates and Times" section of the manual.
>> It always talked about "a timestamp" and it made sense to me that a task
>> would only have one timestamp.  Can a statement be added to the manual
>> to make this explicit?
> How about the attached patch?

I think that's good, but...

>> By closed date interval, I assume you mean "<time1>--<time2>".  That is
>> mentioned in 8.1 of the manual. It's header is "Time/Date range", but it
>> only talks about dates (no times) including in the example.
> Handling of date intervals when TIME1/2 have time specifications is
> undefined behaviour for now. And I am not in favour of making it defined -
> we may consider using this semantics to define repeater intervals with
> end date in future. That would allow what you attempted to do in the
> original email, but using <2023-06-05 Mon 10:30-12:15>--<2023-06-08 Thu>
> semantics. See https://orgmode.org/list/877cxp1fbx.fsf@localhost
>> If times are allowed, then the paragraph could include a time example.
>> If not, then the header should be changed and "timestamp" should be
>> changed to "datestamp".
> No. The same section talks about time range: <2006-11-02 Thu 10:00-12:00>
> "Timestamp" is the general term we use. We make it explicit in the
> parent section that timestamps may or may not have time specification:
>     A timestamp is a specification of a date (possibly with a time) in a
>     special format, either =<2003-09-16 Tue>= or
>     =<2003-09-16 Tue 09:39>=

Maybe I'm not explicit enough.  In section 8.1 of the Org 9.6 manual is
a subsection "Time/Date Range" that *implies* times are supported in
ranges by the use of words "time" and "timestamp" when, above, you're
saying they are undefined (unsupported?) for now.  I'm merely saying
adjust the manual to remove the implication.

David Masterson

Reply via email to