> -----Original Message----- > From: Roland Schroth [mailto:[EMAIL PROTECTED] > > I think your idea could work if i had only one table starting at > the top of > the page. >
The mistake was mine... Didn't study the drawing closely enough to see that it was about *table* headers. Sorry 'bout that. > But (sorry that i did not mention that) i have a sequence of tables of > different lengths that therefore also may start in the middle > of a page. > There are tables being only a few rows long an others spanning 2 or 3 > pages. So i need something that will work for that case too. > A very tough one... I see a remote possiblity in performing some XSL trickery (resting on the assumption that the rows in the table body are generated by XML elements). If you define a row height globally in your stylesheet, as well as the region-body height... You know in advance how many rows will fit on a page... ( $regbh div $rowh ). Then, for each set of XML elements (corresponding to a table) you know in advance how many rows will be generated (counting the nodes). You also know that the first element of the set/table will generate an additional row to accomodate the header-text & the image --first cell has a rowspan of 4 - so if a table starts at the top of the page, you will be able to put ( ( $regbh div $rowh ) - 1 ) of the XML elements (rows) on the same page. If the table starts in the middle of a page and each table is being generated by a template, maybe the maximum possible number of rows that can still fit on the 'page' can be passed in as a parameter... (Keep in mind that this is all pure calculation, as there is no concept of a page-layout at the XSLT stage.) Then again... having to keep count of the elements and inserting a 'fake' header at every expected page-break is hardly an ideal situation. (Not to mention that it will lead to trouble if somehow the page-break doesn't end up where expected) For the moment, however, it's the closest thing to a solution I can come up with... Anyone? Cheerz, Andreas --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]