|
Boost : |
From: Paul A. Bristow (pbristow_at_[hidden])
Date: 2000-07-18 11:40:30
After all that flurry of disputation
Should I use this sort of confusing macro to allow specification, or not?
// Do we need to be able to switch exceptions off/on completely?
#if _HAS_EXCEPTIONS // Used by Dinkumware/Microsoft Standard Library.
#define _THROW0() throw ()
#else /* no exceptions */
#define _THROW0()
#endif /* _HAS_EXCEPTIONS */
and/or
#if _HAS_EXCEPTIONS
#if _USES_EXCEPTION_SPECIFICATIONS_WELL // What should I call this?
#define _THROW0() throw ()
Paul
> -----Original Message-----
> From: David Abrahams [mailto:abrahams_at_[hidden]]
> Sent: Sunday, July 16, 2000 2:18 PM
> To: boost_at_[hidden]
> Subject: Re: [boost] Re: Standard Math Constants?
>
>
>
> ----- Original Message -----
> From: "Paul A. Bristow" <pbristow_at_[hidden]>
>
>
> > I like the boost process - it catches mistypes
> >
> > > Shouldn't the value type be 'T' and not hard-wired
> 'double'? Yes, of
> > course.
> >
> > and comes up with excellent suggestions!
> >
> > When I get back from holiday, I will try to post a first draft,
> > using this proposal.
> >
> > Unless there are other suggestions?
> > Or are there contrary views on the possible costs of this?
> > Do current compilers handle this efficienctly enough?
>
> Speaking of efficiency, you'd probably do well to leave off the
> exception-specifications. That will result in a pessimization on many
> compilers. Since the functions are inline, most compilers
> will be able to
> deduce that they can't throw anyway.
>
> -Dave
>
>
> --------------------------------------------------------------
> ----------
> Get great brand name shoes with just the click of a mouse. Check out
> the huge selection at Zappos.com, the Web's Most Popular Store!
> http://click.egroups.com/1/6994/4/_/9351/_/963753253/
> --------------------------------------------------------------
> ----------
>
>
>
>
Boost list run by bdawes at acm.org, gregod at cs.rpi.edu, cpdaniel at pacbell.net, john at johnmaddock.co.uk