XOM API for Processing XML with Java

[Wolfgang Hoschek <wolfgang.hoschek AT mac.com>]

On Nov 28, 2006, at 3:38 PM, Randall R Schulz wrote:

> Hi,
>
> I'm putting a String that sometimes contains newlines into a Text
> element. When I serialize those elements with an indent and a  
> MaxLength
> in force, the newlines appear to have been replaced with plain spaces
> and the text wraps into a block / blob. Not invoking
> Serializer.setMaxLength() doesn't stop this alteration, though of
> course now the text just wraps at the terminal boundary when I  
> print it
> to a terminal emulator. So apparently using Serializer.setIndent(n)
> with n > 0 is what triggers it.
>
> This is not critical, since this output mode (under the control of
> a "pretty-print" option) is meant only for human consumption (the
> program is actually an importer for an eXist database), but I'd  
> like to
> know if I can get the indenting without having text elements  
> altered in
> this way. Can I?

Currently, XOM's pretty printing (n > 0) is aggressively not  
preserving the XML Infoset. From the javadoc for setIndent(n):

     *   When this variable is set to a value greater than 0,
     *   the serializer does not preserve white space. Spaces,
     *   tabs, carriage returns, and line feeds can all be
     *   interchanged at the serializer's discretion, and additional
     *   white space may be added before and after tags.
     *   Carriage returns, line feeds, and tabs will not be
     *   escaped with numeric character references.

However, the doc goes on to mention that xml:space="preserve" can be  
used to selectively work around it:

     *   Inside elements with an <code>xml:space="preserve"</code>
     *   attribute, white space is preserved and no indenting
     *   takes place, regardless of the setting of the indent
     *   property, unless, of course, an
     *   <code>xml:space="default"</code> attribute overrides the
     *   <code>xml:space="preserve"</code> attribute.

Wolfgang.