Kevin Ryde wrote:
While going through filename bytes versus utf8 in my own program I noticed this bit of GError which I think is currently concatting utf8 and bytes.
I was just about to apply this when I realized that Glib::filename_display_name(), which you use in the patch, is new in glib 2.6.0. So we can't use it unconditionally. Glib::filename_to_unicode() would be an alternative that's always available, but it's not quite the same.