Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Strange behaviour of file in !PDF

0 views
Skip to first unread message

Nat Queen

unread,
Mar 3, 2009, 5:50:11 PM3/3/09
to
Someone sent me a PDF file, which can be viewed normally in Windoze,
but not in our !PDF (I have !PDF 3.00.1.17). The file contains only
plain English text - no unusual symbols.

When viewed in !PDF, each character is replaced by a different one, as
if a simple Caesar cipher has been applied. For example, the word
"available" appears as "DYDLODEOH" - with each ASCII value reduced by
29 (but characters with lower ASCII values such as numerals 1-9 are
shifted upward to top-bit characters.

Has anyone ever seen anything like this? Is there a cure for this?

The file is viewable normally when using !GhostScr/!GView - but this
is less convenient, being slower to process each page.

Any ideas?

Nat

--
Dr. N.M. Queen | Phone: +44 121 414 6590 Fax: +44 121 414 3389
School of Mathematics | PGP-encrypted e-mail preferred. Public key at
Univ. of Birmingham | http://www.queen.clara.net and on keyservers.
Birmingham B15 2TT, UK | Info: http://www.queen.clara.net/pgp/pgp.html

Martin Wuerthner

unread,
Mar 3, 2009, 6:17:13 PM3/3/09
to
In message <eac77336...@clara.co.uk>
Nat Queen <n.m....@bham.ac.uk.invalid> wrote:

> Someone sent me a PDF file, which can be viewed normally in Windoze,
> but not in our !PDF (I have !PDF 3.00.1.17). The file contains only
> plain English text - no unusual symbols.

> When viewed in !PDF, each character is replaced by a different one, as
> if a simple Caesar cipher has been applied.

Yes, that is more or less what is happening here.

> For example, the word "available" appears as "DYDLODEOH" - with each
> ASCII value reduced by 29 (but characters with lower ASCII values such
> as numerals 1-9 are shifted upward to top-bit characters.

This is caused by the file using an embedded subset TrueType font,
i.e., a font that contains only the glyphs that are used in the file,
represented by a compact range of character codes. The code points
have been rearranged accordingly. !PDF does not interpret embedded
TrueType fonts at all and renders the character codes using a
substitute font, but the mapping from code points to actual characters
is part of the embedded font, which is ignored.

> Has anyone ever seen anything like this? Is there a cure for this?

Yes, using a PDF viewer that can interpret embedded TrueType fonts,
such as Ghostscript/GView, as you found out.

> The file is viewable normally when using !GhostScr/!GView - but this
> is less convenient, being slower to process each page.

Martin
--
---------------------------------------------------------------------
Martin Wuerthner MW Software http://www.mw-software.com/
ArtWorks 2 -- Designing stunning graphics has never been easier
spam...@mw-software.com [replace "spamtrap" by "info" to reply]

0 new messages