EML

This document
http://dev.nceas.ucsb.edu/knb/metacat/mobrien.1.4/xml
was inserted to metacat with the script wrapped in a cdata section, as in:
<inline>
<script>
<![CDATA[
function matchwo(a,b) {
if (a < b && a < 0) then {
return 1;
}
else {
return 0;
}
}
]]>
</script>
</inline>

When metacat returns the doc, the CDATA wrapper is no longer there, the < etc get resolved and result is no longer valid xml. As expected, if I replaced the < and & with references AND wrapped the script in CDATA, metacat had no problem returning the whole doc, since now it has escaped characters to replace, too(see docid=mobrien.1.5).

So it seems that metacat is altering the incoming doc by removing the CDATA delimiters before it stores the content. Does this mean that this is not an EML problem, but a metacat behavioral problem instead?

Sorry, that comment lost some text:

When metacat returned doc mobrien.1.4, the '<' etc that were originally enclosed in CDATA are now exposed, and the doc is no longer valid xml.

If instead of using a CDATA section, I replaced the offending characters with entity refs, (<) then predictably, these were resolved and the doc is no longer valid. See mobrien.1.6

Doing both (using entity refs inside a CDATA section, mobrien.1.5) preserved the document for viewing, but the returned xml doc is still altered by metacat - not a good thing.

Original Bugzilla ID was 3465