Re: [Gtkpod-devel] Faulty UTF-8 to UTF-16 conversion

[Jorg S. <Jor...@gm...>]

Javier Kohen wrote:
> There is a bug in DB artwork writer, where it assumes that for a given
> string S, with S8 being its UTF-8 representation and S16 its UTF-16LE
> representation, 2 * len(S8) = len(S16). Unfortunately this very nice
> property only holds for those characters that can be represented with
> one byte in UTF-8, that is, the ASCII range.

As far as I can see the strings processed in this function where 
filenames of thumbnails and therefore in ASCII. Your point is noted, 
though, and I have committed your patch.

> I find it strange that the padding size falls in the range 1 to 4.
> Shouldn't it be 0 to 3?

Yes -- I've fixed that too. The parsers are sufficiently robust to 
accept either 0 or 4 it seems -- as long as the padding makes the length 
  a multiple of 4.

> The second patch replaces redundant calls to utf16_strlen with the
> length value returned by g_utf8_to_utf16.

Thanks -- the patch looks fine to me and I have submitted it.

> I hope I can eventually get rid of the bug that's causing the copy of
> the database on the iPod to become slightly corrupted after unmount :(

I don't think this is related to libgpod, however :-(

JCS.