Re: Matlab compatibility or not Matlab compatibility

[Nicholas Jankowski]

On Feb 17, 2016 2:27 AM, "Julien Bect" <address@hidden> wrote:
>
> Le 16/02/2016 19:06, Nicholas Jankowski a écrit :
>>
>> ran on 2015b:
>>
>> >> script_test
>>
>> x =
>>     A object: 3-by-4-by-5
>>
>> ans =
>>      3     4     5
>>
>> ans =
>>      2
>>
>> x =
>>   3x4x5 B array with properties:
>>
>>     s: [3 4 5]
>>
>> ans =
>>      3     4     5
>>
>> ans =
>>      2
>
>
> Thank you Nicholas.
>
> No surprise then, the conclusion is still that Matlab's ndims() function doesn't not respect the size returned by the overloaded size() function.
>
> It looks like a bug to me, since it is I think the only size-related function which behaves like that.
>
> But, since Octave currently does the same, and since nobody complained about it, I won't try to fix it in the patch I'm working on (see bug #43925 and #44498).
>
> @++
>
>
>
>

Worth adding a "FIXME: but only if MATLAB ever fixes it" note in the source with an explanation?