mtlb_uint16

Description

Matlab and Scilab uint16 behave differently :

• Beyong the main interval bounds, Scilab's uint16() wraps, while Matlab's mtlb_uint16() saturates.
• For non integer values, Scilab's uint16() truncates the fractional part, while Matlab's mtlb_uint16() rounds to the nearest integer.

x	uint16(x)	mtlb_uint16(x)
%nan	0	0
-%inf	0	0
-2	65534	0
-1	65535	0
0	0	0
2	2	2
10.2	10	10
10.5	10	11
10.51	10	11
...	...	...
65535	65535	65535
65536	0	65535
65537	1	65535
%inf	65535	65535

mtlb_uint16 has not to be used for hand coded functions. The function mtlb_uint16(A) is used by mfile2sci to replace uint16(A) when it is not possible to know what is the input while porting Matlab code to Scilab. This function then determines the correct semantic at run time.