Re: [help-texinfo] Umlaut strangeness

[Gavin Smith]

On 5 June 2016 at 13:20, Gavin Smith <address@hidden> wrote:
>> Oh, that's great. Now it works with the \passthroughcharstrue patch
>> by Gavin. The umlauts now appear correctly both in the bookmarks
>> and on the page. Great, just like I wanted to have it.
>
> Do the bookmarks actually work when you click on them? They don't for me...

They did work in the Google Chrome PDF viewer, and sometimes they
worked in Okular (the KDE pdf viewer), and never in xpdf.

I tracked it down to the following in the PDF file:

/Names [(-1) 7 0 R (1) 18 0 R (2) 19 0 R (3) 20 0 R (Regular node) 22
0 R (Regular node 2) 26 0 R (Ubersicht) 30 0 R]

This lists the targets in the file: here I had three chapters,
"Regular node", "Regular node 2" and "Ubersicht".

When the chapter was called "Übersicht" instead (in Latin-1), this
/Names object in the PDF file would have the text "Übersicht" instead
(one byte's difference) but this led to all the hyperlinks in the file
ceasing to work. Evidently, Latin-1 isn't good to use in this context.

The definition of \pdfmkdest has to be fixed to output ASCII
destination names, something like this works:

===================================================================
--- texinfo.tex (revision 7201)
+++ texinfo.tex (working copy)
@@ -1311,8 +1311,9 @@
     % We have to set dummies so commands such as @code, and characters
     % such as \, aren't expanded when present in a section title.
     \indexnofonts
-    \turnoffactive
+    %\turnoffactive
     \makevalueexpandable
+    \passthroughcharsfalse
     \def\pdfdestname{#1}%
     \txiescapepdf\pdfdestname
     \safewhatsit{\pdfdest name{\pdfdestname} xyz}%


The PDF Reference Manual, Sixth Edition for version 1.7 says in
section 3.8.5, page 163, says that the encoding of the destination
names doesn't matter, so I don't know exactly what is going wrong
here.