Andreas L Delmelle schrieb:
Triggering layout sooner is the only way we are ever going to get FOP to accept arbitrarily large tables, without consuming massive amounts of heap.
It is little OT, but multipass would be another way - at the cost of runtime and diskspace/-access (if the memory footprint should be shrinked). I know it is not a issue now, but I sometimes think  of how the "formating objects for indexing" (section 6.10 of xsl 1.1) could be implemented and it always ends in a multipass solution because in contrast to page-number-citation it is hardly possible to guess how much space to reserve for the resulting index-page-citation-list.
Indexing would be a great help for my project - I do it currently by creating as much page-number-citations as needed but without beeing able to use any advanced features, so I end up with long lists of pagenumbers that could/should be merged to page-ranges. It is not realy a big problem because it is a "private" project, but I want to keep the door open for a better solution that is contained in the standard.
Nevertheless resolving as much as possible at the FO tree building stage and trigger an earlier layout process would be an advantage for most use cases.
gerhard  Yes, "thinking" is all what I do at present ;-)))