[xslt] HTML vs. XHTML: different output when including a file with \r\n [WAS: xmllint vs. xsltproc: different output when including a file with \r\n]

["Boris Schaeling" <boris highscore de>]

On Sat, 23 Jan 2010 20:51:46 +0100, Boris Schaeling <boris highscore de>  
wrote:

> When I use "xmllint --xinclude" to include a text file with \r\n end  
> line characters xmllint inserts &#13; characters while "xsltproc  
> --xinclude" doesn't. I'm currently trying to find out (on the DocBook  
> mailing list; see  
> http://lists.oasis-open.org/archives/docbook/201001/msg00052.html) why  
> the output is different and what to do to make xmllint not generate  
> &#13; characters. Maybe someone here can tell me if this is a bug or if  
> there is a trick to change xmllint's output (I couldn't find a command  
> line option so far)?

As it turns out the problem is different. After some discussions on the  
DocBook mailing list it is now clear that generating XHTML leads to &#13;  
characters being inserted while no &#13; characters are inserted when  
generating HTML. Thus it depends on the xsl:output setting of the  
stylesheet used with xsltproc if &#13; characters are inserted or not  
(please forget what I wrote about xmllint; xmllint can be ignored). Bob  
Stayton explained this in his message to the DocBook mailing list:  
http://lists.oasis-open.org/archives/docbook/201001/msg00065.html

The question is now why \r becomes &#13; when generating XHTML but not  
when generating HTML? Are there any specifications which are different for  
XHTML and HTML when it comes to xincluding simple text files with \r\n end  
line markers?

Boris