On 23/06/10 19:39, Georg Datterl wrote: > Hi Craig, > > Give the first add in the classification an empty marker and each > other add the correct marker. Then grab the first marker. If a new > classification started on the page, you get the header, but grab the > empty marker. In each other case you have no header, but a correct > marker.
Really close, but still no cookie. Alas. Thanks so much for your efforts to help me out - I try to do the same on other lists (like PostgreSQL's) but here it's me who's the helpless newbie out of their depth. Your suggestion as I've implemented it mostly - but not quite - works out. Unfortunately the display of marked content must be in static-content in region-before, so headings placed by markers appear higher up the column than in-line headings that happen to be first on the page. This could be cropped away, but would result in different column lengths and therefore ragged bottoms - so back to square 1, really. If XSL-FO allowed static content and flowed content to be combined in one region, such that the flowed content flowed into space the static content hadn't used, this would be possible. However, because of markers etc the static content can change size & shape based on the flowed content, so it's really impossible to do and not surprising that the spec doesn't permit it. I could overlap the region-before and region-body, but because FO doesn't allow the region-before's content to "push" the region-body's out of the way, that'd just print the marker-added heading with the body content overtyped on it. Not a solution. Again, it seems to come down to the need for blocks that can be conditionally displayed depending on whether they're just after (or before?) a break. I am increasingly getting the feeling that XSL-FO just isn't the right tool for this job, and I'll have to try to bodge it with area tree output unless some way to achieve conditional blocks presents its self. Well, that or try to implement this in TeX(ML), since that offers a much more complete (and complex) set of tools for conditional formatting. -- Craig Ringer --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
