[Remind-Fans] Changes coming to Remind...

[Jochen Sprickerhof]

* Dianne Skoll <dianne at skoll.ca> [2020-01-13 18:35]:
>> - Add an element with the complete body, i.e. %\"bar%\" Foo today,
>> or maybe two elements one with the calendar mode entry (bar) and one
>> with the rest (Foo today)
>
>Hmm, that's tricky.  I have to think about that.

Which part would be tricky? The complete body or splitting it into two 
elements?

>> - Why is there date, time, d, m, y and eventstart? Shoudn't one
>> format be enough?
>
>No, they all mean different things:
>
>date is the trigger date
>d, m and y are present ONLY if the corresponding part is present in the
>REM command.  Have a look:
>
>$ echo "REM 13 AT 20:00 MSG test" | ./src/remind -pp -b2 -
>{"date":"2020-01-13", "time":1200, "d":13, "eventstart":"2020-01-13T20:00",
>"priority":5000, "body":"test"}
>
>There are no "y" and "m" elements because those were missing.  So the back-end
>can know that this reminder will trigger on the 13th of every month.

I'm not sure if this is really useful, at least not for python-remind. 
As far as I understand we can always have a context that limits the 
repetition (like have a global OMIT and add a SKIP). So I can't deduce 
the repetition from the absence alone but would have to evaluate the 
complete input file at which point it defeats the purpose of reading the 
Remind output.
So for python-remind I let Remind evaluate the input file and try to 
generate repetition rules based on the evaluated output. Or am I missing 
something?

>time and eventstart can differ for multi-day events, as in:
>
>$ echo "REM 12 AT 20:00 DURATION 24:00 MSG test" | ./src/remind -pp -b2 -
>{"date":"2020-01-12", "duration":1440, "time":1200, "eventduration":1440,
>"d":12, "eventstart":"2020-01-12T20:00", "priority":5000, "body":"test"}
>{"date":"2020-01-13", "duration":1200, "time":0, "eventduration":1440, "d":12,
>"eventstart":"2020-01-12T20:00", "priority":5000, "body":"test"}

Good point, so for python-remind I would condense this again to one 
icalendar entry using the eventstart as the key (is the assumption that 
I can merge it always true)?
Would it be possible to have a mode to only output this event once?

>> - Have a pure json output without the rem2ps2 lines wrapped in a
>> json array [].
>
>I was thinking of doing that, but didn't because it reduces the amount of
>memory you need for parsing JSON, especially in C code like rem2ps.  It
>is a little bit ugly, I admit, and I might be persuaded to change that. :)

I see where you are coming from, but up to now I didn't hit any memory 
barrier with my reminders. Also, as detailed above, I would probably 
read in all elements anyhow to do repetition calculation and so on 
(python-remind currently has a logic to cache all reminders and to rerun 
Remind only on file or day change). So for me the question is more if I 
can dump the Remind output directly into a json parser or have to filter 
it first.

Cheers Jochen
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://dianne.skoll.ca/pipermail/remind-fans/attachments/20200114/31188d57/attachment.sig>