On Thu, 28 May 2020 19:33:12 +0200
Laslo Hunhold <[email protected]> wrote:

> On Thu, 28 May 2020 16:49:22 +0200
> Mattias Andrée <[email protected]> wrote:
> 
> Dear Mattias,
> 
> > > > So would you recommend an explicit cast to uint32_t, i.e.
> > > > 
> > > >    (uint32_t)1 << 16
> > > > 
> > > > to overcome this?      
> > > 
> > > Yes.    
> > 
> > Don't forget about 0x10FFFF, where you need UINT32_C(0x10FFFF).
> > Casting doesn't work here unless you also add the L suffix.  
> 
> thanks for your input on this. I've fixed these problems and also
> simplified the mask-operation as you've suggested.
> 
> With best regards
> 
> Laslo
> 

Looks good. But I wouldn't call the mask operation change
and optimisation (even though it is), than I would have
suggested `*cp = s[0] ^ lut[off].lower`. :)

Reply via email to