On 28/06/2014 02:16, Bill Hart wrote: > MPIR-2.7.0 is compatible with GMP 5.x.y only, but as soon as we produce > a version compatible with GMP 6.0.0 we'll have to change this. > > Bill. > > > On 28 June 2014 02:35, <sisyph...@optusnet.com.au > <mailto:sisyph...@optusnet.com.au>> wrote: > > > > -----Original Message----- From: Brian Gladman > Sent: Saturday, June 28, 2014 5:45 AM > To: mpir-devel@googlegroups.com <mailto:mpir-devel@googlegroups.com> > Subject: Re: [mpir-devel] mpf_fits_uXXX returns false for -0.9 > > > On 27/06/2014 17:40, Bill Hart wrote: > > The word "truncate" here is misleading. Clearly by truncate, > we mean the > greatest integer <= op. > > I don't know what needs to be changed, the documentation or > the code. > I'll let Brian give his opinion on this. > > > Personally I would take the semantic view that negative floating > point > values cannot fit in an unsigned integer value. So I would > change the > documentation rather than the code (which explicitly returns > false for > negative values). > > > Some counter-points to Brian's view: > > 1) This would put the mpir functions at odds with the corresponding > mpfr functions - which have already been fixed (in trunk) such that > mpfr_fits_uXXX now returns true for values greater than -1; > > 2) gmp-6.0.0 already fixes the error such that mpf_fits_uXXX now > returns true for values greater than -1; > > I *think* my assertions are factually correct - apologies if that's > not so. (I haven't had time to double check.) > > It actually makes good senses to me that -0.9 fits unsigned data > types. If it didn't, then I'd also expect that (eg) 4294967295.9 > would not fit into an unsigned 32 bit type. > > Cheers, > Rob
If that's what MPFR and GMP do (GMP does, I haven't checked MPFR), then I think we should also do it. It's an easy fix that I have put in my MPIR repository. I will leave it up to Bill to decide if it goes into MPIR 2.7.0 Brian -- You received this message because you are subscribed to the Google Groups "mpir-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to mpir-devel+unsubscr...@googlegroups.com. To post to this group, send email to mpir-devel@googlegroups.com. Visit this group at http://groups.google.com/group/mpir-devel. For more options, visit https://groups.google.com/d/optout.