From: Joerg Walter (jhr.walter_at_[hidden])
Date: 2002-06-28 00:00:56
----- Original Message -----
From: "Fernando Cacciola" <fcacciola_at_[hidden]>
Sent: Friday, June 28, 2002 12:44 AM
Subject: Re: [boost] uBLAS formal review
> > [snipped]
> > > The RingWithIdentity concept is a model of AdditiveAbelianGroup, but
> > > also a model of MultiplicativeAbelianGroup even though this isn't not
> > > reflected by the concept constraints().
> > Not exactly, some problems with the inverse.
> I see. But then, is MultiplicativeAbelianGroup ever used?
> > [snipped]
> > I notice that some headers and some jamfiles contain explicit
> > to
> > > supress warnings. IMHO, this is kind of cheating, actually. I would
> > > if warnings were removed with appropriate coding instead, whenever
> > possible
> > > at least.
> > The disabled compiler warnings usually are really dumb (like BCC telling
> > 'can't inline this function because it uses ...'). I'm rather sure,
> > not cheating here.
> > > Tests:
> > >
> > > I tried to compile test1 with bcc5.5, but I get this error while
> > > test12.cpp:
> > >
> > > matrix_et(559): E2034, Cannot convert 'const std::complex<float>' to
> > 'int'.
> > Sorry for breaking the Borland build. I've just rechecked that. Please
> > change the offending lines from something like
> > return const_iterator1 (*this, it1, it2, it2 != it2_end ? *it2 : 0);
> > to something like
> > return const_iterator1 (*this, it1, it2, it2 != it2_end ? *it2 :
> > value_type ());
> Great! I'll do that.
> BTW: I'm not entirely sure that value_type() initializes to '0' for
> types. I think this is specified by the std but I seem to recall that some
> compilers don't get it right. It might be safer to use value_type(0)
It's one of the few things we currently trust in ;-), that value_type()
works on every platform. If we'd get a bug report, we'd change.
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk