On Tue, Oct 11, 2011 at 04:15:03PM +0100, Joseph Wright wrote:

> On 10/10/2011 16:07, Bruno Le Floch wrote:
> > - Most encoding functions (as opposed to decoding) could be made
> > expandable. Is it worth it?
> > - There may be need to support other encodings. Which ones?
> > - It is not clear how utf-8 input/output should be treated in pdftex.
> 
> What is needed for PDF bookmark support? This seems to be the main use
> case for these things.

hyperref already reencodes bookmark strings with setting
pdfencoding=auto. The bookmark string is construced in
Unicode encoding. Then the reencoding to PDFDocEncoding is tried.
If successful the result string is used, otherwise the Unicode string.
For the reencoding stuff package stringenc is used and don't need
to be expandable for hyperref.

Yours sincerely
  Heiko Oberdiek