cairo and character encoding issues

[Allin C. <cot...@wf...>]

I believe there are a couple of things not quite right with the 
encoding to UTF-8 in the context of the cairo-based terminals.

* In gp_cairo_convert() (gp_cairo.c) there's an *unconditional* 
conversion of an input string, from the charset returned by 
gp_cairo_get_encoding() to UTF-8.  I'd suggest that conversion 
should be conditional on failure of g_utf8_validate() on the 
string in question.

Rationale: the supposed source charset may not be correct (see 
below).  If the string validates as UTF-8 it should be passed as 
is.  You get bamboozling error messages if g_convert() is called 
on a string that's already in UTF-8, but is misrepresented as 
being in some other encoding.

* The function gp_cairo_get_encoding() doesn't work properly 
unless the right encoding has been set explicitly by the user. The 
fallback in that function is g_get_charset(), but this won't do 
anything useful, in the sense of identifying the character set 
actually in use on the given system, unless there has been a prior 
call

setlocale (LC_ALL, "");

-- 
Allin Cottrell
Department of Economics
Wake Forest University, NC