"Malformed UTF-8 character" warnings in gtkdoc-fixxref

[James Henstridge <james daa com au>]

Anders recently filed a bug report that would be worth looking at before 
the next release:
   http://bugzilla.gnome.org/show_bug.cgi?id=101221

If you have a newer Perl (5.8 definitely, maybe 5.6 too) and are using a 
UTF-8 locale, you will get many warnings from perl about "Malformed 
UTF-8 character" warnings.  This is because the HTML output currently 
uses latin1, which doesn't validate as UTF-8 (unless I am mistaken, this 
would occur for any input that didn't validate against the locale's 
encoding even without a UTF-8 locale).

This problem can be worked around by telling perl to treat the input 
data as byte data rather than character data by adding the "use bytes;" 
pragma.  With this change, gtkdoc-fixxref runs without warnings for me.

I don't know that much about perl (especially these esoteric parts), so 
I would appreciate a second opinion about this (eg. compatibility with 
different perl versions).  For reference, SpamAssassin seems to use the 
following:
   eval "use bytes";

Which I guess might be for backward compat reasons.  I haven't tested 
this with gtk-doc though.

James.

--
Email: james daa com au              | Linux.conf.au   http://linux.conf.au/
WWW:   http://www.daa.com.au/~james/ | Jan 22-25   Perth, Western Australia.