I'm the author / maintainer of a bunch of filesystem-related modules, and ran 
into a problem tonight when I realized that none of them really care about 
unicode.  I'd like to make them smarter, but I'm not sure what I should do.

For things that "write" the filesystem, the most logical thing to do would be 
to encode perl characters into utf8 octets, since that's what's happening now 
(with a warning).  But not everyone is using utf8, and surely the non-users 
would like their data in their native encoding.  Any recommendation on the 
best thing to do here?

Reading things is a whole 'nother can of worms.  Right now my modules are 
doing nothing, which seems wrong to me (the user shouldn't have to worry 
about octets vs. characters, he should always have characters).  The approach 
i took in angerwhale was to assume ascii until i could read a config setting 
that indicated the encoding (assuming utf8 after that).   Unfortunately the 
modules in question deal with metadata... so there's nowhere to store 
metadata metadata :)

Anyway, any advice would be appreciated.  I'm tempted to say "use utf8 or die" 
in the docs, but I think I could do better.  For reference, the modules in 
question are File::Attributes, File::Attributes::Simple, 
File::Attributes::Extended, and Directory::Scratch.

Thanks!

Regards,
Jonathan Rockway

-- 
package JAPH;use Catalyst qw/-Debug/;($;=JAPH)->config(name => do {
$,.=reverse qw[Jonathan tsu rehton lre rekca Rockway][$_].[split //,
";$;"]->[$_].q; ;for 1..4;$,=~s;^.;;;$,});$;->setup;

Attachment: pgpXlEkcqD9Sp.pgp
Description: PGP signature

Reply via email to