GMime charset fixes

I've just committed some fixes to GMime's svn repository fixing bugs:

http://bugzilla.gnome.org/show_bug.cgi?id=423760

(explains the problem in depth)

and

http://bugzilla.gnome.org/show_bug.cgi?id=342196

(the later being especially of interest to a Pan user who reported it)


I'd like to hear comments/suggestions about the newly added APIs and the
changes happening to the header encoding/decoding functions (I'm more
interested in what you guys think of the new decoding behavior - you
probably don't care for about the encoding behavior as long as it works
and users are happy ;-)

To sum up, the new decoding behavior is that
g_mime_utils_header_decode_text/phrase() (and the old
g_mime_utils_8bit_header_decode_*() functions which simply wrap the 2
aforementioned functions) will now GUARANTEE strings returned from them
are in UTF-8 format (currently there is no such guarantee and in fact,
it's pretty bad because it may be in mixed charsets, depending on
whether or not any rfc2047 encoded-word tokens were successfully
decoded/converted into UTF-8 or not)

so... yea.

Of particular interest will probably be the forcing of text into UTF-8
if/when it doesn't want to be converted into UTF-8 seamlessly (using
either the charset declared in the rfc2047 encoded-word token or using
any of the user-specified preferred charsets).


Feel free to pester me with questions if any code is unclear/clearly
broken/dubious or whatever.


Hope to hear from you guys,

Jeff