Re: [xml] xmllint: Why does it convert UTF-8 to numeric entity refs?

["Peter Jacobi" <pj walter-graphtek com>]

Hi David,

I assume your test input is:

>     <?xml version="1.0"?>
>     <!DOCTYPE doc [
>     <!ELEMENT doc (test)+>
>     <!ELEMENT test (#PCDATA)>
>     <!ENTITY ccedil "&#231;">
>     <!ATTLIST test lang CDATA #IMPLIED>
>     ]>
>     <doc>
>         <test lang="français">UTF-8 character: ç</test>
>         <test lang="fran&#xE7;ais">numeric ref: &#xE7;</test>
>         <test lang="fran&ccedil;ais">entity ref: &ccedil;</test>
>     </doc>

As you noted,
xmllint --noent test-utf8.xml
gives the output:

>     <test lang="fran&#xE7;ais">UTF-8 character: &#xE7;</test>
>     <test lang="fran&#xE7;ais">numeric ref: &#xE7;</test>
>     <test lang="fran&#xE7;ais">entity ref: &#xE7;</test>

You can solve half of your problem by giving a seemingly redundant option

xmllint --noent --encode test-utf8.xml
gives the output:
     <test lang="fran&#xE7;ais">UTF-8 character: ç</test>
     <test lang="fran&#xE7;ais">numeric ref: ç</test>
     <test lang="fran&#xE7;ais">entity ref: ç</test>

So, if most of your accented character are in PCDATA, this will do it.

ATTN Daniel: Of course it's equivalent from an XML point of view, but
doesn't you find it somewhat disturbing that the occurrence of numerical
entities depend on source charset and a seemingly redundant option?

Regards,
Peter Jacobi