On 19 Jun 2002, Stephen J. Turnbull muttered drunkenly:
>>>>>> "Jan" == Jan Harkes <jaharkes_at_cs.cmu.edu> writes:
> 
>     Jan> On Sat, Jun 15, 2002 at 12:38:04AM +0100, Nix wrote:
>  
>     >> The fix is to take out -fno-exceptions, I think: g++ won't let
>     >> exception-specifications past under -fno-exceptions. (Yes, this
>     >> means -fno-exceptions in C++ on GNU/Linux and GNU/Hurd boxes is
>     >> pretty useless.)
> 
> [waaaay OT]: Nix, you're not doing a good job of selling C++, you
> know.  Worry not; I know the proverb about minds and parachutes.  ;-)

That's OK, I was talking nonsense: I'm just doing a good job of making
myself look a fool in front of dozens of people.

>     Jan> Problem, we actually need -fno-exceptions, because the
>     Jan> exception handling stuff that g++ adds conflicts with LWP,
>     Jan> our userspace threading code.

I'd understand the clashing, but even if you never throw or catch any
exceptions? (How odd. I think I can see how, but...)

-- 
`What happened?'
                 `Nick shipped buggy code!'
                                             `Oh, no dinner for him...'