Compiling boost with gcc 3.1

[Paolo Carlini]

Benjamin Kosnik wrote:

> Hey Craig.
>
> These regressions in boost are unfortunate: do you know what's up?
> Getting this fixed ASAP is a top priority, so if somebody can put in
> the analysis time I would definitely be willing to work on fixes.

I completely agree with Benjamin.

On the other hand, I'd like to warn (and reassure) that the Boost tests are not
"Standard C++ conformance" tests in that are *full* of specific tweakings for
specific versions of the various compilers (recently I noticed again this while
checking that call_traits are now 100% ok, that is from 4 (g++ <= 3.0.X) to 0
expected failures).

For sure 3.1 is not recognized in many of them and many failures are spurious.

Ciao, Paolo.