backwards headers vs. C++0x

[Benjamin Kosnik]

> How about leaving them in the distribution, but making people include the
> appropriate -I line if they want them?  The annoyance factor will help
> motivate people to clean up their code, and for those who can't or won't,
> it's easy enough to change the makefiles to add a new flag.

Yeah, I thought this and David's other suggestion (warning to error) 
were interesting. I'm more in favor of warning to error than making 
include paths more complicated for the poor build folks.

It would be interesting to do a scan of free software C++ code and see 
if people are using -Wno-deprecated at the moment. And if so, who and 
why. Is synopsys using any of these backwards headers?

Hmm. I wonder if there is a way to limit google code searches to just 
build bits?

-benjamin