[unicode]{hyperref} results in garbled page numbers in Acrobat Reader 9.4.2 02/11/2011 Linux

Martin Scharrer

unread,

Mar 16, 2011, 12:09:33 PM3/16/11

to

When `hyperref` is loaded with the `unicode` option the page number
titles (i.e. the string representation of the page number which can
also be e.g. a roman number etc.) contains random unicode characters
like "1o最尙ထ鴋瘃" with the Linux version of Acrobat Reader 9.4.2
02/11/2011. When the PDF is reloaded (e.g. using CTRL+R) the
characters change to another random sequence. They disappear when the
`unicdoe` option is removed.

This seems like a bug of the Reader not of `hyperref`. I guess some
buffer overflow occurs and the displayed symbols are just some random
memory content.

A minimal example is:

\documentclass{article}
\usepackage[unicode]{hyperref}
\begin{document}
A\newpage A\newpage A\newpage A\newpage A\newpage
A\newpage A\newpage A\newpage A\newpage A\newpage
\end{document}

The result looks like on this screen-shot: http://i.imgur.com/WHE7e.png

Could some of you test this with the Windows version and post their
results. I'm planning to open a bug report, but like to find out
first if this is a Linux specific issue or not.

I'm using hyperref 2011/03/09 v6.82d with a manual installed and
freshly updated TeXLive 2010 (under Ubuntu Linux 10.10).

Thanks,
Martin Scharrer

toto

unread,

Mar 17, 2011, 2:46:01 AM3/17/11

to

Hello,

If it can help, it works perfectly with an updated mactex distribution
(under Mac OS X 10.5.8) and Acroread 9.4.2.

Sincerely yours,

GL

unread,

Mar 17, 2011, 3:56:34 AM3/17/11

to

Le 16/03/2011 17:09, Martin Scharrer a écrit :
> When `hyperref` is loaded with the `unicode` option the page number
> titles (i.e. the string representation of the page number which can
> also be e.g. a roman number etc.) contains random unicode characters
> like "1o最尙ထ鴋瘃" with the Linux version of Acrobat Reader 9.4.2
> 02/11/2011. When the PDF is reloaded (e.g. using CTRL+R) the
> characters change to another random sequence. They disappear when the
> `unicdoe` option is removed.
>

> A minimal example is:
>
> \documentclass{article}
> \usepackage[unicode]{hyperref}
> \begin{document}
> A\newpage A\newpage A\newpage A\newpage A\newpage
> A\newpage A\newpage A\newpage A\newpage A\newpage
> \end{document}

And I have this one for bookmarks : \ must be "escaped" : \\
and it seems this is quite new (I never had such bookmarks text
before...)

\documentclass{article}
\usepackage [T1]{fontenc}
\PassOptionsToPackage {unicode}{hyperref}
\usepackage{bookmark}

\begin{document}
\section{Test \string\test} % single \ => Test ॥猀琀
\section{Test \string\\test}% double \\ OK
\end{document}

Regards.

Heiko Oberdiek

unread,

Mar 17, 2011, 7:27:43 AM3/17/11

to

GL <goua...@gmail.com> wrote:

> And I have this one for bookmarks : \ must be "escaped" : \\
> and it seems this is quite new (I never had such bookmarks text
> before...)

It isn't new. Specifying the backslash directly allows low level
stuff. But hat requires knowledge about PDF strings, the encoding, ...
Therefore it is not recommended.

> \documentclass{article}
> \usepackage [T1]{fontenc}
> \PassOptionsToPackage {unicode}{hyperref}
> \usepackage{bookmark}
>
> \begin{document}

> \section{Test \string\test} % single \ => Test ???

> \section{Test \string\\test}% double \\ OK

The recommended way is the LaTeX way:

\section{Test \textbackslash test}

--
Heiko Oberdiek