Incorrect "may contain" text for anyElement

[hcayless]

Elements that have <anyElement> in their content model, e.g. egXML and xenoData have the message "Character data only" in their "May contain" section. This must be a bug in the tagdocs processing. We probably never noticed it when Pure ODD came out.

[sydb]

Oops.

1. This is not a good ticket for Stylesheets group, as @hcayless has already fixed it (in 7b254fb)
2. However, Council should re-visit the use of the output “ANY”, which is being generated (in common/common_tagdocs.xsl at ~2922) whenever there is a descendant <anyElement>. But the descendant <anyElement> may well have a @require, so it is not really the case that ANY element is allowed. Furthermore, the <anyElement> may have a sibling (either within <sequence> or <alternate> — note that all our cases are actually <alternate>) that allows elements from a namespace other than the one(s) required by the <anyElement>, which we would have to take account of in the output.