[pdftex] Re: not recognized fonts in Acrobat Reader

[Adrian Lanz]

On  9 Aug 2004, lanz at fowi.ethz.ch wrote:

> I have the following problem on a SUN/Solaris 2.9 (last patches
> installed) system with gcc 3.3.2 after compiling and installing the
> newest teTeX-beta 2.96.7.20040721:
>
> Acrobat Reader can not extract all embedded fonts from the pdf file
> generated through pdflatex. "Not all fonts" means, that if I use
> several fonts, only one (the last) font can not be extracted by
> AR. AR pops up an eror windows saying something like "unabale to
> extract the embedded font XXX", and replaces all occurencies of
> characters in this font with dots. ARs error occurs in different
> versions of AR under different platforms. xpdf or gv (under Solaris)
> display the same pdf file without problem. Also the "latex -> dvips
> -Ppdf -> ps2pdf" route produces a pdf file which can be read by AR.

It seems I have found what makes Acrobat Reader complain:

For the first (or only) font in the PDF file I get an extremly high
value for the /Length1 entry in /FontFile, in my case "/Length1
268436202". For the following fonts embedded in the PDF file, these
/Length1 values are ok.

(If I replace the wrong /Length1 value by a "reasonable" value, like
"/Length1 747" copied from a run with an non-broken pdflatex, Acrobat
Reader is happy.)

The following work-around also seems to solve the problem: In the
pdflatex-generated PDF file (*.pdf) replace all occurencies of
"/Subtype /Type1" with "/Subtype /Type1C".

Please comment!!! Thanks, Adrian.




Test case is very simple: just pdflatex

\documentclass{article}
\begin{document}
text
\end{document}

with the recent TeTeX-Beta distribution (temporarely switch off all
local and private TEXMF trees!).