At 08:27 PM 7/21/01 +0300, Ilkka Hartikainen wrote:
>Why does FOP seem to render body-region first, then before-region and 
>finally after-region? I'm asking because this would solve all issues with 
>background colors and/or images. Putting a page-size svg (for example) in 
>the before region would create a background, but this does not work since 
>FOP renders before-region on top of body-region.
>I'm not sure what the candidate recommendation syas about this (if 
>anything), but please share any insight you might have. If someone has any 
>hints on changing the rendering order from source, I'd appreciate any help!

Boy, I don't know if I want to go down this road. :-)

The spec definitely talks about layering and conflict of marks - background 
and intrinsic marks. This applies to positioned areas (relative and 
absolute), not normal (stacked) areas. These would be, for example, fo:block 
with a "relative-position" of 'relative', or an fo:block-container with an 
"absolute-position" of 'absolute' or 'fixed'.

I see nothing in the spec that leads me to believe that it was ever intended 
that regions would overlap at all. The spec uses this language:

"NOTE: The body region should be sized and positioned within the 
fo:simple-page-master so that there is room
for the areas returned by the flow that is assigned to the fo:region-body 
and for any desired side regions, that is, fo:region-before, 
fo:region-after, fo:region-start and fo:region-end's that are to be placed 
on the same page. These side regions are positioned within the 
content-rectangle of the page-reference-area. The margins on the 
fo:region-body are used to position the region-viewport-area for the 
fo:region-body and to leave space for the other regions that surround the 

This tells me that regions are not supposed to overlap.

It is explicitly stated that all areas generated by fo:region-body are 
absolute (area class 'xsl-absolute'). If one assumes that this holds for all 
regions, the language which obtains in Section 4.9.6 could be stretched to 
suggest that, with this content model for fo:simple-page-master,


that areas in fo:region-body lie beneath (are rendered behind or under) 
areas in the outer regions. FOP, as you note, seems to do this already 
(although I doubt whether this was intentional).

Agin, IMO, this is all a bit of a stretch. Yes, areas generated by regions 
may be absolute, but then the z-index really applies to fo:title, fo:block, 
fo:inline, fo:leader, and children of fo:table. And the spec is quite clear 
about suggesting that regions ought not to overlap anyway. In summary, at 
this point in time I would avoid doing stuff like this, and if it happens to 
turn out to your satisfaction, enjoy it while it lasts, because it could 

Arved Sandstrom

Fairly Senior Software Type
e-plicity (
Wireless * B2B * J2EE * XML --- Halifax, Nova Scotia

To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]

Reply via email to