Hmm, I guess that has something to do with the layoutDimension mechanism  that I suggested to Glen to have a look at a few days ago .
It may very well be that with the changes due to the Knuth approach the layoutDimension stuff was broken. I didn't pay a lot of attention and decided to look at it again later.  http://marc.theaimsgroup.com/?t=110630668800003&r=1&w=2  http://email@example.com/msg00210.html On 08.04.2005 18:52:22 Luca Furini wrote: > > I was trying to make lists work in the most general situation when I > stumbled across this strange bug: it seems that the label-end() function > (implemented in fo/expr/labelEndFunction.java) does not compute the right > value. > > It should compute the end-indent of a list-item-label, which is defined to > be (7.28.3) > > ipd of the reference area > - (provisional-distance-between-starts > + start-indent > + start-intrusion-adjustment > - provisional-label-separation) > > It seems to me that the method LabelEndFunction.eval() computes rightly > the inner expression between the brackets, but fails to get the right > reference ipd. > > The overall effect is that the end-indent stored inside a > CommonMarginBlock object is bigger than the available ipd, and the stack > limit set by the BlockLM for the LineLM is < 0. > > Any idea about how to fix this (should it really be a bug)? > > Regards > Luca > > Jeremias Maerki