Re: HP, aCC and the standards

From: Dave Abrahams (dave_at_[hidden])
Date: 2003-02-05 13:14:11

In news:20030205161921.15120_at_[hidden],
Duane Murphy <duanemurphy_at_[hidden]> typed:
> --- At Wed, 5 Feb 2003 09:48:17 +0100, Toon Knapen wrote:
>
> > On Tuesday 04 February 2003 17:18, Tom Matelich wrote:
> > > Who is running those tests, and what version of aCC is that?
> >
> > I am. I'm using aCC version 53800 currently installed on an HPUX11i. You
> > might notice on the HP website that the latest release was 53700 so it's
> > newer than new ;-) Nevertheless, conformance is still bad.
>
> This is out of curiosity following this thread, but conformance to what
> is still bad? Arent the boost tests testing "conformance" to boost
> libraries or the ability to compile and execute boost libraries? This is
> quite different from "conforming" to the standard. I exepect that boost
> tries to stay within the confines of the standard, but clearly many boost
> libraries push the envelope.

What is "pushing the envelope?"  Either it's legal C++ or it's not.

> While many compiler vendors strive to be able to compile boost, is this
> really a measure of Standard C++ conformance?

A pretty good one.  That's why they use it as a test.

> p.s.
>
> With the low marks that aCC is getting, I can imagine that its Standard
> C++ conformance is probably very low as well.

Precisely.

-- 
Dave Abrahams
Boost Consulting
http://www.boost-consulting.com