Re: [docbook-apps] Incorrect characters in FO / PDF under windows

[Peer Brink <brink@riege.com>]

Bob Stayton wrote:
> The problem is not with FOP, because the extra
> characters are in the FO file before FOP sees them.

I agree.

> The separator between the "1." and the title should be
> one non-breaking space character, which when encoded
> in the default UTF-8 output encoding should be a two character
> sequence C2 A0 (hexadecimal).   When you view the
> FO file with a typical text editor that assumes ISO-Latin-1
> encoding, you will see A-caret and a non-breaking space, the characters
> associated with C2 and A0, respectively, in ISO-Latin-1.

In the correct (Linux-created) Version I in fact find what you are describing:
c2 a0

But in the incorrect (windows-created) version there is
c3 82 c2 a0

How is this non-breaking space character inserted into the fo-file? Does the stylesheet has to care about the encoding? In other words: could it be a stylesheet-problem? Or is the transformer (saxon) responsible for the correct encoding?

> Are you using any kind of customization?

Yes. I'm using a stylesheet-customization and a DTD-customization. But the problem occurs also without these customizations when using original stylesheets and DTD.

> Exactly what command
> are you using to process your files?

I have written my own java-programm using javax.xml.transform.Transformer. And I'm setting encoding UTF-8 on the org.dom4j.io.OutputFormat before writing the transformed xml-document to disk.

> Does the FO output file
> have an encoding="xxx" in the XML declaration at the top?
> In my output from Saxon 6.5.3 I see:
> 
> <?xml version="1.0" encoding="utf-8"?>

My output-fo has the same XML declaration.

Thanks in advance for any help,
Peer.