id	summary	reporter	owner	description	type	status	priority	milestone	component	version	resolution	keywords	cc	product
82	neercs assumes that terms send utf8	Pascal Terjan	Pascal Terjan	"In ansi.c, the character goes through caca_utf8_to_utf32 and only if bytes is 0 it is considered to be a 1 byte character to keep.

This does not really make sense as caca_utf8_to_utf32 returning 0 is not equivalent to invalid utf8 and bytes probably never 0 in this case.

====
If a null byte was reached before the expected end of the UTF-8 sequence, this function returns zero and the number of read bytes is set to zero.
This function never fails, but its behaviour with illegal UTF-8 sequences is undefined.
====
"	defect	new	major		core	SVN				libcaca