Review the remaining uses of ascii formats and assess their legitimacy

[koolfy]

As we had some problem in the past with UTF-8, I think i would be a good measure to look where we still constrain some strings and variables to plain ACII, and make sure it's not safer/wiser to switch them to UTF when possible.

Some switchs might not be useful right now (but a good meaure for the future, if the codebase changes), others might be a bad idea or dangerous. Time and careful reading will tell.

Whatever happens, it will happen in a branch to see proper review and testing before it's merged.
If nohing really needs to change, I'll close this when I feel confident with our current variable format handling.