Bob Stayton <> writes:
> I just checked the latest schemas that Norm uploaded to OASIS, and
> it appears that this is indeed a bug.
>
> In all other instances of xi:include in the DB XInclude schema, the
> xi:include was part of a pattern named db.*.content or
> db.*.components, and then that pattern is used with a repeatable
> suffix in the element definition. But in set, we have:
>
> db.set.components =
> ((db.set | db.book)+ | db.article+)
> | db.xi.include
>
> element set {
> db.set.attlist,
> db.set.info,
> db.toc.components?,
> db.set.components,
> db.setindex.components?
> }
>
> That only allows a single xi:include in place of other elements. We
> must allow more than one, and we must allow a mix of xi:include and
> elements.
Yep. We don't do that in set because we forbid mixing books and articles
in a set.
Given how relaxed we are about such things in other places, perhaps the
straightforward fix is:
db.set.components =
(db.set | db.book | db.article)
| db.xi.include
element set {
db.set.attlist,
db.set.info,
db.toc.components?,
db.set.components+,
db.setindex.components?
}
Alternatively, we could work a little harder:
db.set.components =
((db.set | db.book | db.xi.include)+ | (db.article | db.xi.include)+)
element set {
db.set.attlist,
db.set.info,
db.toc.components?,
db.set.components+,
db.setindex.components?
}
Thoughts?
Be seeing you,
norm
--
Norman Walsh <> | To think is not enough; you must
http://www.oasis-open.org/docbook/ | think of something.--Jules Renard
Chair, DocBook Technical Committee |
