[RFC] A QoI problem with filebuf and POD charT != char, wchar_t

[Paolo Carlini]

Pétur Runólfsson wrote:

>This fragment will cause bad_cast to be thrown if getloc() doesn't
>contain the correct codecvt, so it seems that the current behaviour is
>required by the standard.
>  
>
... but now it occurs to me: Pétur, what about providing a generic codecvt
facet similar to the specialization for <char, char, mbstate_t>, that 
is, doing
almost nothing but with do_always_noconv() returning true?

Would that be legal? Perhaps this is the correct way to go!

Paolo.