[Gambas-user] Difference DateDiff(...) ?

[Tony Morehen]

You're right.  Print is not the problem.  Another part of Gambas could 
raise an error.  But Benoit is right.  Doing this kind of check is too 
resource intensive.  You're also right.  Since DateDiff(gb.second) will 
overflow for a ~70 year difference, over the longer term, it should 
return a long.  Since this kind of edge case is rare, it's not an urgent 
change and,in fact, may cause problems for existing code.


On 2017-08-06 02:27 PM, d4t4full at ...626... wrote:
> Toni,
>
> I don't think Print is at fault here at all. My guess is that DateDiff may need an "attitude adjustment" in these cases: It may need to either raise an overflow error or return a 64bit Long.
>
> Hope this helps,
> zxMarce.
>
>
>
> On Aug 6, 2017, 12:00, at 12:00, Tony Morehen <tmorehen at ...3602...> wrote:
>> zxMarce, you are exactly right.  DateDiff returns an integer so when
>> print adds the two datediff it adds 2 signed integers and gets an
>> signed
>> integer.  Since the max value of a signed integer is 2147483647, the
>> answer of 2208988800 overflows to -2085978496. So print's behaviour is
>> as expected.  The only question is: Should print raise an overflow
>> error?  Note that there has not been a hardware overflow so the
>> gambas's
>> math engine must do the overflow check (by matching signs).
>>
>>
>> On 2017-08-06 10:30 AM, d4t4full at ...626... wrote:
>>> Without a machine nearby, I'd risk you're in the right track.
>>>
>>> The DIM AS LONG creates 64bit vars while DateDiff may return 32bit
>> data (don't have the docs with me).
>>> Print will use 64bit to print the vars and 32bit with direct DateDiff
>> (you may confirm by using CLng to affect the DateDiff's in the second
>> Print)
>>> Remember that when you have the most significant bit (MSB) set,
>> Gambas will interpret the value as twos-complement sign and yield
>> negative values.
>>> Regards,
>>> zxMarce.
>>>
>>>
>>> On Aug 6, 2017, 10:21, at 10:21, Hans Lehmann <hans at ...3219...>
>> wrote:
>>>> Hello,
>>>> Why are two different results:
>>>>
>>>> Public Sub Main()
>>>>
>>>>     Dim a, b As Long
>>>>
>>>>     a = DateDiff("01/01/1900", "01/01/1935", gb.Second)
>>>>     b = DateDiff("01/01/1935", "01/01/1970 ", gb.Second)
>>>>
>>>>     Print a + b
>>>>     Print DateDiff("01/01/1900", "01/01/1935", gb.Second) +
>>>> DateDiff("01/01/1935", "01/01/1970 ", gb.Second)
>>>>
>>>> End
>>>>
>>>> Results:
>>>>
>>>> 2208988800     ' ok
>>>> -2085978496    ' complement?
>>>>
>>>> Best regards
>>>>
>>>> Hans
>>>>
>> ------------------------------------------------------------------------------
>>>> Check out the vibrant tech community on one of the world's most
>>>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>>>> _______________________________________________
>>>> Gambas-user mailing list
>>>> Gambas-user at lists.sourceforge.net
>>>> https://lists.sourceforge.net/lists/listinfo/gambas-user
>> ------------------------------------------------------------------------------
>>> Check out the vibrant tech community on one of the world's most
>>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>>> _______________________________________________
>>> Gambas-user mailing list
>>> Gambas-user at lists.sourceforge.net
>>> https://lists.sourceforge.net/lists/listinfo/gambas-user
>>
>>
>> ------------------------------------------------------------------------------
>> Check out the vibrant tech community on one of the world's most
>> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
>> _______________________________________________
>> Gambas-user mailing list
>> Gambas-user at lists.sourceforge.net
>> https://lists.sourceforge.net/lists/listinfo/gambas-user
> ------------------------------------------------------------------------------
> Check out the vibrant tech community on one of the world's most
> engaging tech sites, Slashdot.org! http://sdm.link/slashdot
> _______________________________________________
> Gambas-user mailing list
> Gambas-user at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/gambas-user