Smalltalk › Pharo › Pharo Smalltalk Users

adding 1 year to the date - bug?

_‹ Previous Topic Next Topic _›

Classic

List

Threaded

11 messages Options

Petr Fischer

adding 1 year to the date - bug?

Hello, just simple test:

(Date year: 2019 month: 2 day: 26) + 1 year.
returns: 26 February 2020
OK

(Date year: 2020 month: 2 day: 26) + 1 year.
returns: 25 February 2021
What?

(maybe I do not understand something about dates again)

Thanks! pf

Pavel Krivanek-3

Re: adding 1 year to the date - bug?

In Pharo 1 year is a duration of 365 days so it does not reflect leap years.

-- Pavel

2018-02-26 10:43 GMT+01:00 Petr Fischer <[hidden email]>:

> Hello, just simple test:
>
> (Date year: 2019 month: 2 day: 26) + 1 year.
> returns: 26 February 2020
> OK
>
> (Date year: 2020 month: 2 day: 26) + 1 year.
> returns: 25 February 2021
> What?
>
> (maybe I do not understand something about dates again)
>
> Thanks! pf
>

NorbertHartl

Re: adding 1 year to the date - bug?

In reply to this post by Petr Fischer

> Am 26.02.2018 um 10:43 schrieb Petr Fischer <[hidden email]>:
>
> Hello, just simple test:
>
> (Date year: 2019 month: 2 day: 26) + 1 year.
> returns: 26 February 2020
> OK
>
> (Date year: 2020 month: 2 day: 26) + 1 year.
> returns: 25 February 2021
> What?
>
> (maybe I do not understand something about dates again)

That is a conceptual problem. 1 year is a duration. As it does not know which year it is the normal 365 days. And 2020 is a leap year so has 366 days. That’s why the date is shifted.

Norbert

NorbertHartl

Re: adding 1 year to the date - bug?

Forgot to say that this is something you can experience quite often. To me the problem is that 1 year gives a duration immediately. It should work as all units that 1 year is a unit of magnitude 1 and unit year. If you apply this to a date you could make it properly to give the date one year later.

I noticed that problem when I started to use the Units package of Marcus. It works for all units except durations because these are already taken. So for me there would be two benefits changing that.

Norbert

Am 26.02.2018 um 10:48 schrieb Norbert Hartl <[hidden email]>:

Am 26.02.2018 um 10:43 schrieb Petr Fischer <[hidden email]>:

Hello, just simple test:

(Date year: 2019 month: 2 day: 26) + 1 year.
returns: 26 February 2020
OK

(Date year: 2020 month: 2 day: 26) + 1 year.
returns: 25 February 2021
What?

(maybe I do not understand something about dates again)

That is a conceptual problem. 1 year is a duration. As it does not know which year it is the normal 365 days. And 2020 is a leap year so has 366 days. That’s why the date is shifted.

Norbert

Tudor Girba-2

Re: adding 1 year to the date - bug?

+1

Doru

> On Feb 26, 2018, at 10:54 AM, Norbert Hartl <[hidden email]> wrote:
>
> Forgot to say that this is something you can experience quite often. To me the problem is that 1 year gives a duration immediately. It should work as all units that 1 year is a unit of magnitude 1 and unit year. If you apply this to a date you could make it properly to give the date one year later.
> I noticed that problem when I started to use the Units package of Marcus. It works for all units except durations because these are already taken. So for me there would be two benefits changing that.
>
> Norbert
>
>
>> Am 26.02.2018 um 10:48 schrieb Norbert Hartl <[hidden email]>:
>>
>>
>>
>>> Am 26.02.2018 um 10:43 schrieb Petr Fischer <[hidden email]>:
>>>
>>> Hello, just simple test:
>>>
>>> (Date year: 2019 month: 2 day: 26) + 1 year.
>>> returns: 26 February 2020
>>> OK
>>>
>>> (Date year: 2020 month: 2 day: 26) + 1 year.
>>> returns: 25 February 2021
>>> What?
>>>
>>> (maybe I do not understand something about dates again)
>>
>> That is a conceptual problem. 1 year is a duration. As it does not know which year it is the normal 365 days. And 2020 is a leap year so has 366 days. That’s why the date is shifted.
>>
>> Norbert
>

--
www.tudorgirba.com
www.feenk.com

"What we can governs what we wish."

Petr Fischer

Re: adding 1 year to the date - bug?

In reply to this post by Petr Fischer

So "date addMonths: 12" behaves better for my needs. Thanks

jtuchel

Re: adding 1 year to the date - bug?

Am 26.02.18 um 11:13 schrieb Petr Fischer:
> So "date addMonths: 12" behaves better for my needs. Thanks
Hmm. We've had this discussion just the other day at Lunch. What is a
month? Should addMonths: 1 and then subtractMonths: 1 on January 31st
bring you back to January 31st? How would you implement that?

I guess this is a problem that cannot be solved by a machine without
additional context. Does 1 month later mean "on this day next month" or
does it mean "exactly as many days in the future as the current month
has days"? None of the two? Can one month later be in March when today
is January 30th?

It seems years and months are a bit incompatible with informatics ;-) In
my opinion, there is more info needed than "x months/years later" in
order to calculate the correct result.

Joachim

>
>
>> Hello, just simple test:
>>
>> (Date year: 2019 month: 2 day: 26) + 1 year.
>> returns: 26 February 2020
>> OK
>>
>> (Date year: 2020 month: 2 day: 26) + 1 year.
>> returns: 25 February 2021
>> What?
>>
>> (maybe I do not understand something about dates again)
>>
>> Thanks! pf
>>
>

--
-----------------------------------------------------------------------
Objektfabrik Joachim Tuchel mailto:[hidden email]
Fliederweg 1 http://www.objektfabrik.de
D-71640 Ludwigsburg http://joachimtuchel.wordpress.com
Telefon: +49 7141 56 10 86 0 Fax: +49 7141 56 10 86 1

Sean P. DeNigris

Re: adding 1 year to the date - bug?

Administrator

In reply to this post by NorbertHartl

NorbertHartl wrote
> Forgot to say that this is something you can experience quite often. To me
> the problem is that 1 year gives a duration immediately. It should work as
> all units that 1 year is a unit of magnitude 1 and unit year. If you apply
> this to a date you could make it properly to give the date one year later.

Yes!!!

-----
Cheers,
Sean
--
Sent from: http://forum.world.st/Pharo-Smalltalk-Users-f1310670.html

Cheers,
Sean

Esteban A. Maringolo

Re: adding 1 year to the date - bug?

In reply to this post by jtuchel

This is a discussion we already had years ago. And comes back every once in a while.

Pharo, and all dialects to my knowledge, uses incremental based calculations instead of field based calculations (like java.util.Calendar).

This is the best thread I remember where we discussed this.

http://forum.world.st/Interesting-Date-Time-Thread-on-Squeak-Dev-td4778652.html#a4778970

Regards,

Esteban.

El feb. 26, 2018 7:33 AM, "[hidden email]" <[hidden email]> escribió:

Am 26.02.18 um 11:13 schrieb Petr Fischer:

So "date addMonths: 12" behaves better for my needs. Thanks

Hmm. We've had this discussion just the other day at Lunch. What is a month? Should addMonths: 1 and then subtractMonths: 1 on January 31st bring you back to January 31st? How would you implement that?

I guess this is a problem that cannot be solved by a machine without additional context. Does 1 month later mean "on this day next month" or does it mean "exactly as many days in the future as the current month has days"? None of the two? Can one month later be in March when today is January 30th?

It seems years and months are a bit incompatible with informatics ;-) In my opinion, there is more info needed than "x months/years later" in order to calculate the correct result.

Joachim

Hello, just simple test:

(Date year: 2019 month: 2 day: 26) + 1 year.
returns: 26 February 2020
OK

(Date year: 2020 month: 2 day: 26) + 1 year.
returns: 25 February 2021
What?

(maybe I do not understand something about dates again)

Thanks! pf

--
-----------------------------------------------------------------------
Objektfabrik Joachim Tuchel mailto:[hidden email]
Fliederweg 1 http://www.objektfabrik.de
D-71640 Ludwigsburg http://joachimtuchel.wordpress.com
Telefon: <a href="tel:%2B49%207141%2056%2010%2086%200" value="+4971415610860" target="_blank">+49 7141 56 10 86 0 Fax: <a href="tel:%2B49%207141%2056%2010%2086%201" value="+4971415610861" target="_blank">+49 7141 56 10 86 1

Petr Fischer

Re: adding 1 year to the date - bug?

Thanks for discussion thread link. And yes, I need to implement it myself for customer needs, especially things like:

January 31 + 1 month = February 28/29
February 29 + 1 year = February 28
etc. :)

pf

> This is a discussion we already had years ago. And comes back every once in
> a while.
>
> Pharo, and all dialects to my knowledge, uses incremental based
> calculations instead of field based calculations (like java.util.Calendar).
>
> This is the best thread I remember where we discussed this.
> http://forum.world.st/Interesting-Date-Time-Thread-
> on-Squeak-Dev-td4778652.html#a4778970
>
>
> Regards,
>
> --
> Esteban.
>
> El feb. 26, 2018 7:33 AM, "[hidden email]" <[hidden email]>
> escribió:
>
> Am 26.02.18 um 11:13 schrieb Petr Fischer:
>
> So "date addMonths: 12" behaves better for my needs. Thanks
> >
> Hmm. We've had this discussion just the other day at Lunch. What is a
> month? Should addMonths: 1 and then subtractMonths: 1 on January 31st bring
> you back to January 31st? How would you implement that?
>
> I guess this is a problem that cannot be solved by a machine without
> additional context. Does 1 month later mean "on this day next month" or
> does it mean "exactly as many days in the future as the current month has
> days"? None of the two? Can one month later be in March when today is
> January 30th?
>
> It seems years and months are a bit incompatible with informatics ;-) In my
> opinion, there is more info needed than "x months/years later" in order to
> calculate the correct result.
>
> Joachim
>
>
>
>
>
>
> >
> > Hello, just simple test:
> >>
> >> (Date year: 2019 month: 2 day: 26) + 1 year.
> >> returns: 26 February 2020
> >> OK
> >>
> >> (Date year: 2020 month: 2 day: 26) + 1 year.
> >> returns: 25 February 2021
> >> What?
> >>
> >> (maybe I do not understand something about dates again)
> >>
> >> Thanks! pf
> >>
> >>
> >
> --
> -----------------------------------------------------------------------
> Objektfabrik Joachim Tuchel mailto:[hidden email]
> Fliederweg 1 http://www.objektfabrik.de
> D-71640 Ludwigsburg http://joachimtuchel.wordpress.com
> Telefon: +49 7141 56 10 86 0 Fax: +49 7141 56 10 86 1

Richard O'Keefe

Re: adding 1 year to the date - bug?

In reply to this post by Petr Fischer

It makes sense to add a Duration to a DateAndTime.

It does not make sense to add a Duration to a Date.

I can't check Pharo because I haven't managed to

get it working on this laptop (x86-64 + ubuntu 17),

but Squeak has Date>>addMonths: (edited for clarity):

addMonths: monthCount 
    |m year month maxDaysInMonth day|

    m := monthCount + self monthIndex - 1.

    year := m // 12 + self year.
    month := m \\ 12 + 1.
    maxDaysInMonth := Month daysInMonth: month forYear: year.
    day := self dayOfMonth min: maxDaysInMonth.
    ^ Date
        newDay: day
        month: month
        year: year

This may or may not do what your customer wants.

On 26 February 2018 at 22:43, Petr Fischer <[hidden email]> wrote:

Hello, just simple test:

(Date year: 2019 month: 2 day: 26) + 1 year.
returns: 26 February 2020
OK

(Date year: 2020 month: 2 day: 26) + 1 year.
returns: 25 February 2021
What?

(maybe I do not understand something about dates again)

Thanks! pf