DOCBOOK-APPS: xmllint bug?

[Bob Stayton <bobs at caldera dot com>]

In answering a previous question about where to
put the email tag in a bookinfo element, I decided
to test my answer.  xmllint says it is invalid.
It looks valid to me, and nsgmls thinks it is valid too.

Here is the shortened sample that should validate with
the DocBook XML 4.2 DTD:

<?xml version="1.0"?>
<!DOCTYPE book PUBLIC "-//OASIS//DTD DocBook XML V4.2//EN"
               "../docbook42/docbookx.dtd">
<book>
<bookinfo>
<title> .. </title>
<author>
   <firstname> .. </firstname>
   <surname> .. </surname>
   <email> .. </email>
</author>
</bookinfo>

</book>

Here is what xmllint (version 20423) says:

$ xmllint --noout --valid invalid.xml
invalid.xml:11: validity error: Element author content doesn't follow the DTD
Expecting ((personname | honorific | firstname | surname | lineage | othername |
 affiliation | authorblurb | contrib+) , (personblurb | email | address)*), got
(firstname surname email )
</author>
        ^

Here is the content model of author in 4.2:

((personname|(%person.ident.mix;)+),(personblurb|email|address)*)

and here is %person.ident.mix:

<!ENTITY % person.ident.mix 
             "honorific|firstname|surname|lineage|othername|affiliation 
             |authorblurb|contrib %local.person.ident.mix;">

The example has firstname and surname (and no personname)
to satisfy the group before the comma, followed by email which
is part of the group after the comma.  It should validate,
shouldn't it?  Am I missing something?

Or is xmllint misinterpreting the content model?
It seems to have dropped a set of parens, the ones around
the parameter entity.

-- 

Bob Stayton                                 400 Encinal Street
Publications Architect                      Santa Cruz, CA  95060
Technical Publications                      voice: (831) 427-7796
The SCO Group                               fax:   (831) 429-1887
                                            email: bobs@sco.com