Comments intermingled.Default reference orientation and lr-tb writing mode assumed.
> -----Original Message----- > From: J.Pietschmann [mailto:[EMAIL PROTECTED]] > Sent: May 5, 2002 1:34 PM > To: fop dev > Subject: Border properties > > Hello, > I tried to make something out of bug 684. Again, after > reading the spec in depth, I'm nearly biting pieces off > my keyboard. > In the tables.fo examples, the left edge of the table > content rectangle is the same as the edge of the reference > area, and left border (should I use "start edge border"?) > is tacked on so that it extends outside the reference area. What it really boils down to is, the start-indent and end-indent are content-rectangle to content-rectangle distances. If the start-indent is zero then borders and padding and space are outside the content rectangle of the reference area. Similarly for end-indent. > The upper and lower border, however, do not overlap previous > or following blocks, as expected. I would not expect this - see below. At least not with initial values. > Well 4.4.1 says > the start-edge of its allocation-rectangle ... offset from it > inward by a distance equal to the block-area's start-indent > plus its start-intrusion-adjustment (as defined below) > minus its border-start, padding-start, and space-start values... > given that the start-indent of the tables are zero (hopefully), > the behaviour regarding the border extending left beyond the > edge of the refernce area appears to be consistent with the spec, > albeit IMO a bit counter-intuitive, because not consistent with > what happens in BPD. The treatment is different; there is no BPD counterpart to start-indent and end-indent. The separation between content-rectangles is determined by padding+border+space (before & after). Borders and padding are <length-conditional> values so unless they are at the leading or trailing edge of a reference area they will definitely have the specified <length>. None of them are negative. The spaces _can_ be negative so this is the sole mechanism by which areas (including borders & padding) can overlap (and space-specifier resolution is involved of course). > Now, what is the problem bug 684 complains about? Does it mean > the border in BPD should be handled similarly to what happens in > IPD? Or does he mean something else? > And, of course: is my understanding on how borders/padding should > be handled resonable? I feel very confused. Join the club. :-) I constantly remind myself of what the definitions really mean in simple language. I also try to minimize use of allocation rectangles - I understand the concept but find it confusing. > BTW what happens if both start-indent and margin-start were > defined on the same block area? margin-* properties are absolute only (top, bottom, left, right). In any case, according to 5.3.2 and 5.3.1 the absolute properties take precedence. margin-left if explicitly specified has priority over start-indent. I took a quick look at table.fo (the FO) and I think this will probably help out. I have to admit if there is one area of the spec that I am not particularly familiar with it is tables - in this case I don't think there is any weirdness involved stemming from table border properties. Regards, Arved --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, email: [EMAIL PROTECTED]