Re: New file uploaded to boost

From: Daryle Walker (darylew_at_[hidden])
Date: 2000-11-08 13:27:21

on 11/7/00 12:54 AM, David Abrahams at abrahams_at_[hidden] wrote:

[SNIP]
> You seem to have dropped this borland-specific fix:
> 
> #ifndef __BORLANDC__
> friend D requires_difference_operator(const T& x, const T& y) {
> return x - y;
> }
> #endif
> 
> Was that intentional?

Yes.  I asked about the #define on this list.  You suggested that Mark
Rodgers knew about the issue and he said:

> Yes, right on both counts.  I put it in because it caused an error with
> BCC5.5.  It should have a #define in config.hpp if we figure out exactly what
> the error is.  However I think 5.5.1 has fixed the problem anyway, so we
> should probably just remove it, and drop 5.5 as a supported version.

[SNIP]
> In operators_test.cpp:
> 
> You are avoiding the use of base class chaining for a reason, I suppose? It
> wouldn't hurt to be explicit about it.

There was no particular reason, I just didn't think of doing it.  (And one
of my inspirations for the composite operator templates was that there was a
limit to base-class chaining.)  I added it in anyway.

[SNIP]
> I asked before, but you haven't explained: why does Wrapped use operator void*
> while Wrapped2 uses operator bool?

I think I answered this in another post.  (You may have not gotten it when
you sent your post.)

-- 
Daryle Walker
Mac, Internet, and Video Game Junkie
darylew AT mac DOT com