Re: [help-texinfo] Umlaut strangeness

[Gavin Smith]

On 2 June 2016 at 21:56, Karl Berry <address@hidden> wrote:
> It is possible.  The idea is to output the bookmark strings in PDF
> Unicode string syntax, and starting with an FEFF marker.  For example,
> this is "A" (U+0041): <FEFF0041>

I've done some work on this and have made some progress, managed to
output this sequence when generating a PDF bookmark (all byte
sequences are hard-coded in my current work) and this shows up
correctly in the PDF outline when viewed in a PDF reader. However, the
UTF-16 encoding doesn't work well with texi2dvi. The problem is grep
not working well with the log file. The log file contains text like:

! pdfTeX warning (dest):name{FE FF 00 DCbersicht 2} has been
referenced but does not exist, replaced by a fixed one

where "FE FF 00 DC" means four bytes in the log file, and this leads
grep to think it is a binary file, and looking for \openout lines in
the log file to find auxiliary files gives an error message like:

"Binary file test2.log matches"

instead of giving the lines in the file that match. I haven't looked
any further into solutions.

> I seem to recall reading that PDFDocEncoding is a superset of Latin 1,
> but I haven't done the diff.  If so, @documentencoding ISO-8859-1 could
> use the \passthroughcharstrue hack, I think ...-karl

I haven't looked at it in detail but  it seems the PDFDocEncoding has
characters at values 80-9F, which are missing from Latin-1 (the space
was left for a set of control characters).