Bjoern Hoehrmann <[EMAIL PROTECTED]> writes: > Hmm, maybe I should ask why you proposed to keep the old behavior of > encode_utf8 in the first place? The change would make more sense to > me if both encode("UTF-8" => ...) and encode_utf8(...) were changed.
I just wanted to keep the the ability still do the cheap encode/decode where we just end up tweaking the SvUTF8 flag. But we also have utf8::encode() so I would not really mind if Encode::encode_utf8() became strict. The other rationale for this is that I think of "utf8" as the perl extended version, while "UTF-8" is the standard interpretation. I could imagine having: encode("utf8", ...) # perl extended UTF-8 encode_utf8(...) # perl extended UTF-8 encode("UTF-8", ...) # the real thing. Seems almost sane to me. Regards, Gisle