read_u16 seems to be broken

06 Oct 2010

See details here.

07 Oct 2010

Hi

Is the workaround to just lop off the least signficant nibble then eg. read_u16() & 0xfff0

Regards
Daniel

PS It would be nice to get a comment from the mbed team on this; I can't see a reason why you'd want to mirror the most significant nibble in the least significant one either.

08 Oct 2010 . Edited: 08 Oct 2010

Hi,

 

The read_u16 semantics define it ti return a normalised value, between 0 an 2^16-1. 

 

The implementation of this can be optimised for a 12-bit value to repeat the top nibble at the bottom, so thats how it has been implemented!

Thanks,
Chris