Emanuel Mair <[EMAIL PROTECTED]> writes:
> If I run something like gedit from a terminal and open a file
> requester I get a message saying: "Gtk-Message: The filename
> "someswedishcharacter.foo" could not be converted to UTF-8 (try to
> set the environment variable G_BROKEN_FILENAMES): Invalid byte
> sequence in conversion in-data" [my translation]. This is repeated
> once for each "broken" filename in the file requester. But that env
> var *is* set!
What is your locale, though? G_BROKEN_FILENAMES means that filenames
are in locale encoding. So you need to be in a Latin-1 locale, not a
UTF-8 locale, to use Latin-1 filenames.
You could also try renaming files to UTF-8. Appended is a quick
script that lets you do something like:
recode-file.sh ISO-8859-1 UTF-8 myfile
> "When I hear the word Unicode, I cock my gun." ;)
One thing about Unicode is that it makes all of us experience the pain
that Asia has been experiencing for years - part of the point is to
get the Asian locales to finally work properly. ;-)
Ironically enough, we don't default the Asian locales to Unicode
because there are so many CJK-specific hacks that they rely on that
need to be fixed in the Unicode case...
Other than that, Unicode lets us mix more than one language on a
single system, and handle things like word wrap correctly for all
languages.
Havoc
#! /bin/bash
set -e
## be sure we don't have side effects from locale we're running in
export LANG=C
export LC_ALL=C
SOURCE_ENCODING=$1
DEST_ENCODING=$2
FILENAME=$3
NEW_FILENAME=`echo -n "$FILENAME" | iconv -f $SOURCE_ENCODING -t $DEST_ENCODING -o -`
if ! test "$FILENAME" -ef "$NEW_FILENAME"; then
mv "$FILENAME" "$NEW_FILENAME"
else
echo "$FILENAME is unchanged"
fi
