C++ PATCH: Throttle inlining

[Gerald Pfeifer]

On Wed, 28 Mar 2001, Mark Mitchell wrote:
> My guess is that these things should have been emitted as weak
> symbols, but that they're not being emitted that way.

Yes, that's my theory as well. (Probably functions that are designated
inline but not actually inlined.)

> Would you mail me two .ii files that generate .o files with the same
> symbols, plus the target triplet and cc1plus flags?

Will do.

On Wed, 28 Mar 2001, Benjamin Kosnik wrote:
> ...sorry, a little confused. Is this the status?
>
> -O3 doesn't work
> -O2 works
> -O0 works
>
> ie, what's the point where it doesn't work anymore?

-O0 works, and everything above -O1 has these duplicate symbols.

Gerald