[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [O] [PATCH] Timestamps: Handle sub-10-min ranges when updating times

From: Trevor Murphy
Subject: Re: [O] [PATCH] Timestamps: Handle sub-10-min ranges when updating timestamps
Date: Wed, 07 Aug 2013 23:31:18 -0700
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/24.3 (gnu/linux)

Wow, thanks for parsing that, Nicolas.  I didn't realize until just now
that my git skills were so poor - I neglected the --cover-letter option,
hence the utter lack of explanatory material.

To answer your points:

Nicolas Goaziou <address@hidden> writes:

> Thanks for your patch. Would you mind providing a test-case for it? I'm
> not sure about the use of `org-get-compact-tod'.

Schedule an event for today with a five-minute duration.  E.g:

* TODO test out bug in `org-schedule'
SCHEDULED: <2013-08-07 Wed 17:00-17:05>

Then hit C-c C-s (or however you have `org-schedule' bound).  With the
default setup, you'd expect to see the following prompt in the

Date+time [2013-08-07]: 17:00+0:05

however what you'll get instead is:

Date+time [2013-08-07]: 17:00+0:5

The latter is not a valid time spec.  If you simply accept it, then at
least on my install org reschedules the event to:

SCHEDULED: <2013-08-07 Wed-17:00>

Which is not what I intended.  I'll add that you can get the same buggy
behavior from any command that calls `org-time-stamp' on an
already-timestamped event with <10 minute duration.

>> -    (if (< dm 0) (setq dm (+ dm 60) dh (1- dh)))
>> +    (when (< dm 0) (setq dm (+ dm 60) dh (1- dh)))
> Although I agree with this change, this is not strictly necessary here.

Agree.  I just couldn't resist.  Since I'll likely be rewriting this
patch anyways, I'll revert this back.

>>      (concat t1 "+" (number-to-string dh)
>> -            (if (/= 0 dm) (concat ":" (number-to-string dm))))))))
>> +            (when (/= 0 dm) (concat ":"
>> +                                        (if (< dm 10)
>> +                                            (concat "0" (number-to-string 
>> dm))
>> +                                          (number-to-string dm)))))))))
> It would be better to use a 0-padded format string, e.g.,
>   (and (/= 0 dm) (format ":%02d" dm))

I tested that and it felt noticeably slower when I called
`org-reschedule'.  The extra `if' and `concat' did not feel slower.  I
didn't do explicit timings because of the subjective feel (also because
I'm not really sure how to do those tests yet).  That being said, I
agree with you.

If you prefer, I'll resubmit the patch without the if => when and using
the format string.  Let me know if you'd prefer I do some timing tests
on format vs if/concat.

Trevor Murphy
GnuPG Key: 0xCB06EAAF

reply via email to

[Prev in Thread] Current Thread [Next in Thread]