[Bugzilla] [Bug 15505] Samuel GOUGEON has commented the bug

[bugzilla.admin at scilab.org]

http://bugzilla.scilab.org/show_bug.cgi?id=15505
--------- Bug Summary -----------
Trouble with int64() around int64(%inf)-

---------- Changed by ------------
sgougeon at free.fr

---------- What changed ----------

--- Comment #3 from Samuel GOUGEON <sgougeon at free.fr> ---
(In reply to Stéphane MOTTELET from comment #2)
> The help page of "double" 
> 
> https://help.scilab.org/docs/6.0.1/en_US/double.html
> 
> is very informative about this.

Thanks (i am the author https://codereview.scilab.org/#/c/17908/5..14/scilab/modules/elementary_functions/help/en_US/floating_point/double.xml :).


(In reply to Stéphane MOTTELET from comment #1)
> int64(2^63-1) cannot be equal to int64(%inf), because 2^63-1 is evaluated as a double, with a mantissa which has too many digits. 

yes, sure

>Here 
> 
> --> 2^63-1
>  ans  =
> 
>    9.223372036854776480D+18
> 
> is stored as  9223372036854776480 hence greater than 9223372036854775807 !

Not on Windows:
-- format(25)
--> 2^63-1
 ans  =
   9223372036854775800.

When seeing all these traps, it looked to me that some inequality operators like %s_1_i() might be improved wrt int64/uint64 specific features.
They presently convert the integer into double, while, for instance, the double could be converted into int64|uint64.
To be analyzed.

-- 
You are receiving this mail because:
You are watching all bug changes.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.scilab.org/pipermail/bugzilla/attachments/20180327/ea55f5c8/attachment.htm>