This is the mail archive of the
mailing list for the libstdc++ project.
Re: libstdc++ related boostrap failure
On Wed, Feb 19, 2003 at 05:50:51PM -0600, Benjamin Kosnik wrote:
> > A) export _Atomic_add_mutex from the library, like Loren's patch did, or
> > B) for the generic case, move the definition of the atomic routines to
> > misc-inst.cc, along with the helper variables. This makes everything
> > nice and tidy and private, but kills any chance of inlining.
> >I tend to lean to 'A'.
> I too think 'A' is the best option.
> You should probably add _Atomic_add_mutex_once, and
> __gthread_atomic_add_mutex_once as well, no?
Yes, I figured FreeBSD must have a mutex initializer, else we'd be seeing
more errors. :-)
> This doesn't matter for the people who don't use the generic routines
> (most people except for arm linux), and the generated abi baselines can
> be made so that mis-configured systems (urp, i486/linux configured as
> i386/linux) can be caught by system integrators.
This took an embarassingly long time to solve.
I would therefore like to posit that computing's central challenge, viz. "How
not to make a mess of it," has /not/ been met.
- Edsger Dijkstra, 1930-2002