[RFC] warn time for appointments

[RFC] warn time for appointments

[Ivan Kanis]

Hi Bastien,

In bzr emacs there is a new variable called appt-warning-time-regexp. It
contains a string for how many minutes we want to be reminded of an
appointment. 

For example, to be warned 30 minutes in advance of an appointment you
would put the following in the diary file:
   2011/06/01 12:00 Do something ## warntime 30

I would very much like this feature in org. How about we use the same
regexp? We could use this when we call org-agenda-to-appt. A typical
entry would look like this:

* Do something
  <2012-02-04 Sat> warntime 30

What do you think of the specifications? If it's OK with you I will find
the time to do it.

Take care,
-- 
Ivan Kanis
http://kanis.fr

Nothing in life is to be feared. It is only to be understood.
    -- Marie Curie

Re: [RFC] warn time for appointments

[Ivan Kanis]

> In bzr emacs there is a new variable called appt-warning-time-regexp. It
> contains a string for how many minutes we want to be reminded of an
> appointment...

I got no reply to my RFC and to my tentative patches. Does that mean
everybody is happy with the default 12 minutes warning time?
-- 
Ivan Kanis
http://kanis.fr

We make a living by what we get, we make a life by what we give.
    -- Winston Churchill

I am listening to "Supersystem - The Love Story".

Re: [RFC] warn time for appointments

[Peter Münster]

On Fri, Mar 23 2012, Ivan Kanis wrote:

> I got no reply to my RFC and to my tentative patches. Does that mean
> everybody is happy with the default 12 minutes warning time?

No, I'm not happy with 12 minutes. But there is
https://github.com/p-m/org-notify where you can have arbitrary warning
times (seconds, weeks, whatever...).
Does it fit your needs?

(I'll ask Bastien for inclusion in contrib...)

-- 
           Peter

Re: [RFC] warn time for appointments

[Bernt Hansen]

Peter Münster <pmlists@free.fr> writes:

> On Fri, Mar 23 2012, Ivan Kanis wrote:
>
>> I got no reply to my RFC and to my tentative patches. Does that mean
>> everybody is happy with the default 12 minutes warning time?
>
> No, I'm not happy with 12 minutes. But there is
> https://github.com/p-m/org-notify where you can have arbitrary warning
> times (seconds, weeks, whatever...).
> Does it fit your needs?
>
> (I'll ask Bastien for inclusion in contrib...)

I'm happy with the 12 minute warning time.  If I have an appointment at
2PM I stick that on my calendar... if it takes 2 hours to drive there I
put second task (Drive to wherever) at noon so I get a 12 minute warning
before I have to jump in my car and go.

Regards,
Bernt

Re: [RFC] warn time for appointments

[Nick Dokos]

Peter Münster <pmlists@free.fr> wrote:

> On Fri, Mar 23 2012, Ivan Kanis wrote:
> 
> > I got no reply to my RFC and to my tentative patches. Does that mean
> > everybody is happy with the default 12 minutes warning time?
> 
> No, I'm not happy with 12 minutes. But there is
> https://github.com/p-m/org-notify where you can have arbitrary warning
> times (seconds, weeks, whatever...).
> Does it fit your needs?
> 
> (I'll ask Bastien for inclusion in contrib...)
> 

BTW, I had done some simple performance measurements and hit some hiccups
previously. I just repeated them and I believe things are much improved:

org-notify-process                8           0.521915      0.065239375
org-notify-todo-list              8           0.512575      0.064071875
org-notify-make-todo              194         0.0485780000  0.0002504020

Another time

org-notify-process                11          1.6243200000  0.1476654545
org-notify-todo-list              11          1.516659      0.1378780909
org-notify-make-todo              596         0.2396870000  0.0004021593

and I've seen the avg time (last column) of org-notify-process go up to
0.22s, whereas before it used to be 5s or so, and even after the first
round of optimization that Peter did, it went down but still was over a
second (1.6s iirc). That caused significant hesitation when typing.

At this point,I do not notice any such hesitation - otoh, I'm sure I'm
not stressing it very much: that would require constructing a suitable
load generator, which I have not done.

But previously, things were bad enough that I had to turn off the notify
process. Now, I'm leaving it on and (mostly) not noticing that
it's there, so that's definitely progress.

Thanks Peter!
Nick

Re: [RFC] warn time for appointments

[Ivan Kanis]

Peter Münster <pmlists@free.fr> wrote:

> On Fri, Mar 23 2012, Ivan Kanis wrote:
>
>> I got no reply to my RFC and to my tentative patches. Does that mean
>> everybody is happy with the default 12 minutes warning time?
>
> No, I'm not happy with 12 minutes. But there is
> https://github.com/p-m/org-notify where you can have arbitrary warning
> times (seconds, weeks, whatever...).
> Does it fit your needs?

Thanks for the link, I will try it out! In the meantime I am using a
hack that I posted on the list.
-- 
Ivan Kanis
http://ivan.kanis.fr

Men of lofty genius when they are doing the least work are the most
active.
    -- Leonardo da Vinci

Re: [RFC] warn time for appointments

[Bastien]

Hi Ivan,

Ivan Kanis <ivan.kanis@googlemail.com> writes:

> Peter Münster <pmlists@free.fr> wrote:
>
>> On Fri, Mar 23 2012, Ivan Kanis wrote:
>>
>>> I got no reply to my RFC and to my tentative patches. Does that mean
>>> everybody is happy with the default 12 minutes warning time?
>>
>> No, I'm not happy with 12 minutes. But there is
>> https://github.com/p-m/org-notify where you can have arbitrary warning
>> times (seconds, weeks, whatever...).
>> Does it fit your needs?
>
> Thanks for the link, I will try it out! In the meantime I am using a
> hack that I posted on the list.

Note that you can use it from latest master HEAD in the git repo.

-- 
 Bastien

Re: [RFC] warn time for appointments

[Peter Münster]

On Fri, Mar 23 2012, Nick Dokos wrote:

> But previously, things were bad enough that I had to turn off the notify
> process. Now, I'm leaving it on and (mostly) not noticing that
> it's there, so that's definitely progress.

There is still one thing, that we can do: call org-notify-process
periodically, but only when the user is idle for at least some seconds.
This logic exists already in gnus-demon-add-handler but the gnus-demon
runs only when gnus is running. So, what we can do is:
- patch gnus-demon.el to allow running the daemon independently of gnus
- copy parts of gnus-demon.el
- perhaps something else...

Some day I'll take a look.

-- 
           Peter