On Thu, Jan 26, 2006 at 09:56:13PM +0100, Thomas Viehmann wrote:
> Jesus Climent wrote:
> >>"-u" will make the CDDB data work, but it seems that abcde still needs
> >>to validate/convert the data (potentially modified by users with
> >>whatever locale) before storing it into id strings.
>
> > Well, that might be a problem. In a previous discussion, we figured out that
> > is rather impossible to get the encoding of a file with mixed latin1/utf-8.
> > So
> > it is basically a user problem to get it right, and then pass the utf-8
> > stuff
> > to the tag/comments.
>
> > Or if i am mistaken, please, I would need a utf8 guru to give me a hand
> > here.
>
> I couln't claim I'm an expert, but the meager thing I could come up with is
> pipe through iconv before (-f utf8) and after (-t utf8) editing
> or
> check the result of iconv (potentially even if there was no editing
> if something failed
> let the user choose to fix the problem himself or strip the illegal
> chars (the latter is "iconv -c -f utf8 -t utf8")
>
> This seems like a lot of iconv, but I think that it's best to get this
> right and the data passed through iconv is very little, so this doesn't
> seem too bad in terms of perfomance...
>
> I could write a patch implementing above if it helps.
That would help!
Thanks!
--
Jesus Climent info:www.pumuki.org
Unix SysAdm|Linux User #66350|Debian Developer|2.6.15|Helsinki Finland
GPG: 1024D/86946D69 BB64 2339 1CAA 7064 E429 7E18 66FC 1D7F 8694 6D69
If a man cannot choose, he ceases to be a man.
--Minister (A clockwork orange)
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
