On 9.3.2009, at 6:58, Duncan Collis wrote: > So - it looks like I can work around the problem by either omitting a > small part of the survey or by outputting at not-quite 1:1000, but it > seems to me that something is wrong here!
It is a bug, I discovered it yesterday too. The new snapshot has changed (improved) the MetaPost and TeX macros, so there should be a problem. m.