Closed jessealama closed 7 years ago
This is very mysterious! I'm afraid I couldn't reproduce it myself. Does it still occur for you on GHC 7.6?
If it's still a problem for you, the only thing I can think of is that it could be a problem with the FFI interface to MiniSat. You could try going into Haskell/Sat.hs and changing "unsafe" to "safe" in the declarations of s_clause, s_clause1, s_clause2, s_clause3, s_solve and s_simplify. But I'm doubtful that that will help.
It turns out we were using an old version of MiniSat which did an out-of-bounds array read! Paradox is now updated to the latest version of MiniSat which should hopefully fix the problem.
Paradox segfaults on Mac OS X 10.7, even with the latest ghc. The difficulty arises when dealing with "non-trivial" problems whose smallest countermodel has size, say, 5 or more. I don't know how to characterize the TPTP problems on which paradox crashes. All I can say is that when theories start to get a little complex, paradox becomes more likely to crash on Mac OS X. Even more strange, with a single theory paradox sometimes crashes and sometimes terminates.
I attach the TPTP problem on which paradox exhibits this strange behavior.