Excuse me for firing off a comment from the hip, but I have just looked at the putative fo:retrieve-table-marker FO. Prima facie, it looks to behave like fo:retrieve-marker, except with respect to table headers and footers.
The difference that immediately occurs to me is that fo:retrieve-marker can logically occurs after the layout of region-body, and, because the dimensions of those regions which are the targets of static-content are size-constrained by the applied master-page. This simplifies the resolution of the marker.
With fo:retrieve-table-marker, the possibility seems to exist that the formatting of a table-marker may change the region-body layout, and the page boundaries, to the extent that the source fo:marker may change. We have then another one of those awkward catch-22s of page layout.
I may have missed something in my brief scan of the text, and I would appreciate any light on the subject. If my knee-jerk response is correct, fo:retrieve-table-marker can just be added to the list of layout "nasties". It is hardly a show-stopper, given the similar unavoidable problems that already exist.
Another thing that occurs to me as a result of these considerations is that the editors might comment (even non-normatively) on such issues, and possible strategies for resolving them. Your own non-normative feedback would be much appreciated.
Peter -- Peter B. West <http://www.powerup.com.au/~pbwest/resume.html>
