Is it possible to strip only newlines from nodes containing mixed
content in XSLT 1.0?

I have a situation where incoming XML files occasionally have newlines
in the middle of text that need to be stripped, but the text may also
contain mixed content elements. For example:

<p>test <tag>more text</tag> text (TXT)
and more text after newline.</p>

This should be output as:

<p>test <tag>more text</tag> text (TXT) and more text after newline.</p>

However, if I run the XML through an identity transform and use
normalize-space on the text nodes, I get this:

<p>test<tag>more text</tag>text (TXT) and more text after newline.</p>

Note that the spaces around the nested <tag> element are gone.

The newline needs to be stripped out, but the spaces around the nested
element need to be retained.

============================
Douglas Rudder
XML Architect
Wolters Kluwer Health, Clinical Solutions
77 West Port Plaza, Suite 450
St. Louis, MO 63146
Phone: 314-216-2227
email: Doug.Rudder@xxxxxxxxxxxxxxxxx
============================