[Octave-bug-tracker] [bug #45945] Missing Matlab's special support for integer literals

[anonymous]

URL:
  <http://savannah.gnu.org/bugs/?45945>

                 Summary: Missing Matlab's special support for integer
literals
                 Project: GNU Octave
            Submitted by: None
            Submitted on: Sat 12 Sep 2015 07:03:02 PM UTC
                Category: Interpreter
                Severity: 3 - Normal
                Priority: 5 - Normal
              Item Group: WTF, Matlab?!?
                  Status: None
             Assigned to: None
         Originator Name: Chris Wellons
        Originator Email: address@hidden
             Open/Closed: Open
         Discussion Lock: Any
                 Release: 3.8.2
        Operating System: GNU/Linux

    _______________________________________________________

Details:

In Octave:


octave:1> uint64(18446744073709551614)
ans = 18446744073709551615


But in Matlab I get:


ans = 18446744073709551614


This behavior is not documented by MathWorks, but from experimentation it
looks like the Matlab interpreter treats conversion functions as special when
the argument is an integer-like literal. It doesn't use an intermediate
double-precision value as you would expect. For example, make it any
expression at all and Octave and Matlab will agree on the output, due to the
intermediate double result:


octave:4> uint64(18446744073709551614 * 1)
ans = 18446744073709551615


It's weird, but it's at least a way to specify an integer literal, otherwise
absent from the language.

Very similarly, Octave and Matlab disagree on str2num here:


str2num('uint64(18446744073709551614)')


Ran into all this when I was trying to put large integer values in my code.





    _______________________________________________________

Reply to this item at:

  <http://savannah.gnu.org/bugs/?45945>

_______________________________________________
  Message sent via/by Savannah
  http://savannah.gnu.org/