Re: [libstdc++] New implementation of concept checking

[Phil Edwards <pedwards at disaster dot jaj dot com>]

[lists trimmed]

On Fri, Apr 06, 2001 at 01:47:53AM +0100, Jason Merrill wrote:
> > 2)  Since large chunks of libstdc++ are built -fno-implicit-templates,
> > the instantiations resulting from the typedefs become unresolved symbols
> > during final executable linking.
> 
> So they are referenced, then.  That would seem to be the bug...

Yeah, I'd forgotten that while many of the checks simply do some typedefs in
order to introduce the code, one check generates a template function call.
It's that function call which caused the symbol to be emitted, not the
other typedefs.  This is something of a change from the old implementation.

(The function itself declares a single variable in order to introduce the
checking code, and then returns.)


Phil

-- 
pedwards at disaster dot jaj dot com  |  pme at sources dot redhat dot com
devphil at several other less interesting addresses in various dot domains
The gods do not protect fools.  Fools are protected by more capable fools.