Hi,
+1
you guys are not up to date: current code (in the PR) has the floating point random functions as opt in (with a macro - default is off).
the discussion about 16 bit platforms was just about the floating part. the core random number generator should totally run on a 16 bit platform.
Is there more I can't see?
Christian
Hi Christian,
you guys are not up to date: current code (in the PR) has the floating point random functions as opt in (with a macro - default is off).
I'm aware of the current state of the PR, but your initial comment (which Kaspar commented on) sounded more generic. And I agree with Kaspar: we should be very, very careful with code that does not run on every supported architecture.
Cheers, Oleg
but it's probably not this case (with this code)...
In the initial PR, I was not sure about 16 bit validity of the code. So one should test this on 16 bit hardware. I have no such hardware. Nobody sad that this is _not_ running on 16 bit! I just had the fear!
the current code builds without any warnings on (-W -Wextra -pedantic): native msba2 msb-430
and is tested on: native msba2
Best Christian