Re: [RFC, WIP] tree-ssa-strlen optimization pass

[Paolo Carlini <paolo dot carlini at oracle dot com>]

Hi,
> I'm fairly sure it is much less common than the narrow versions,
> we don't even handle wprintf and wscanf attributes.
I know. Actually, the reason I decided to say something, is that I 
noticed a few times in the past this "inequality" between char and 
wchar_t. Frankly, I instinctively think it's unfortunate, if asked, I 
would rather recommend systematically working on *pairs* of 
optimizations. I don't know if somebody ever discussed this idea in the 
open for GCC, as a matter of "policy" so to speak, but I also don't know 
what the other compilers are doing, I don't have much more to say at the 
moment. I hope somebody will give the general idea a thought.

Thanks,
Paolo.