Hi Kish, Thanks! The bug is due to the fact that all items have size zero. Fixed in the trunk.
Cheers Christian -- Christian Schulte, www.gecode.org/~schulte Professor of Computer Science, KTH, [email protected] Expert Researcher, SICS, [email protected] -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Kish Shen Sent: Monday, August 10, 2015 2:11 PM To: [email protected] Subject: [gecode-users] multi-dimensional binpacking gives fatal signal 11 Hi, I am gettng a fatal signal 11 (on Linux, Gecode 4.4.0) with the multi-dimensional binpacking for some arguments values. The simplest case that I can generate the problem is: binpacking(home, 2, l, b, s, c) l = {0, 0, l1, l2} (0 is where the IntVar has a singleton domain of 0) b = {0, b1} s = {0,0,0,0} c = {5,5} The binpacking constraint is called from ECLiPSe, and some of the arguments are generated. I haven't tried to call binpacking independently of ECLiPSe, but it does look like the problem is happening in Gecode: valgrind (a memory checker) detected the first problem in BinPacking::Pack::post(...) which is called by binpacking with a Conditional jump or move depends on uninitialised value(s) (Sorry, I don't know how to cut and paste from my xterm window in a VNC session to my Windows mailer). The first message may not be a real error, but the next, also in the Pack::post function, does seem to be an error: Invalid red of size 4. I haven't really written a stand-alone Gecode program, so I thought I should report this issue before trying to do so -- please let me know if you want me to do this. Thanks and cheers, Kish _______________________________________________ Gecode users mailing list [email protected] https://www.gecode.org/mailman/listinfo/gecode-users _______________________________________________ Gecode users mailing list [email protected] https://www.gecode.org/mailman/listinfo/gecode-users
