On 12/24/06, Joe Neeman <address@hidden> wrote:
>This is almost certainly because the height-estimation routines are
>overestimating the height of the systems prior to page breaking on the
>second pass. Can you try it (perhaps just the first couple of pages) with
>annotate-spacing = ##t to check? If the Y-extent-estimate interval is much
>larger than the Y-extent interval, this is the problem.
>Also, what version are you using?
Ah, that's the fact I always forget. Using 2.10.4 on OS X 10.4. Also
tried with similar results on 2.11.x.
Hard to tell if y-extent-estimate is overshooting. In many cases
properly respaced systems have the same margin of error as those that
aren't.
Judge for yourself. Two example pages with annotate-spacing added. The
first is not respaced as expected, the second is. Other pages not
included seem to vary wildly from these lengths.
http://www.foxchange.com/spamguy/extentestimate.pdf
Wow, that's some pretty appalling height-estimation! Could you send me the input file (uncondensed will do)?
I have found a bug and fixed it in both the
2.10 and 2.11 branches, so it will appear in the next 2.10 and 2.11 releases. I'm not sure if it completely solves the problem because my laptop doesn't have enough RAM to compile the full example :P. Just to be sure, could you check when the new version comes out?
(For the record, Will sent me the files off-list)