Dear all, V3.0.3 introduced some checking code to hunt down unexpected extension of memory usage which had been reported previously. Sure enough, this soon started to show by an increase of error reports. I think I have found the source(s) of the problem, at least the data sets sent to me which I used to reproduce it now run through. There's a 64 bit binary for testing available at http:/www.chevreux.org/tmp/mira_3.1.6_dev_linux-gnu_x86_64_static.tar.bz2 All users who had unexpected out of memory errors or the message "Would extend memory of AS_skim_edges? Shouldn't be ..." should try out this new version and please tell whether that particular problem is gone. Please note that this version is from the development tree and contains a number of other changes which are not documented yet, though I don't think they should affect current users of MIRA and things should run through pretty smoothly. Regards, Bastien