quick question: has anyone made some kind of encoding scheme that produces valid UTF-8 while avoiding null bytes and "?
It has been bothering me for a long while that in Lichat transmission of binary data has the base64 1/3rd overhead when transmitted in-band, which is the case for stuff like avatars and icons.
Hrm, then again I guess in UTF-8 anything above 127 gets the double-byte encoding, so the mapping would have to be 1-127 except for 34?