Re: bitset ctor multiply defined

["Matt Austern" <austern@isolde.engr.sgi.com>]

On Jun 7,  1:50pm, Benjamin Kosnik wrote:
> Subject: Re: bitset ctor multiply defined
>
> > Basically, bits/std_bitset.h is trying to declare multiple ctors, all
> > taking a string, with varying default parameters.  The one on line 652
> > is always present, with another one around line 679 showing up inside
> > the #ifdef's.  They are both being defined, not just declared.
> >
> > Do we have any kind of feedback into SGI's implementation?
>
> I wish. For a while, Matt was on this list, and he may still lurk, but I
> don't know the exact status. If he hasn't responded by now, I'd say the
> answer is no. :(

Still here, and I'll take a look at the <bitset> problem.  I'm a bit
puzzled, since the compiler we run the most tests with (SGI MIPSpro
7.3, using the EDG 2.38 front end) does support member templates.
Perhaps we're inadvertently relying on an EDG extension; it wouldn't
be the first time.

I'll still mostly be in lurk mode for a while, though.  I just got back
from my honeymoon a couple of weeks ago, and I'm still trying to deal with
the backlog.  (I'm afraid I ended up just deleting most of this list's
traffic from May.)

			--Matt