1

I just encountered a SMTLIB problem in string theory that Z3 might have answered inconsistently. When envoking Z3 to solve the problem: with argument smt.string_solver=z3str3 it returns unsat; without any arguments it returns sat.

I also used CVC4 to solve the problem. It returned a solution, which seems to be a valid model as I checked it by manually replacing the variable assignments into the problem.

Since I'm trying to do a research using Z3, I would like to know if this is a known behavior of Z3. Thanks to anyone that could help! :)

Edit: I was using Z3 4.7.1 on WSL Ubuntu 16.04.

hsuhw
  • 11
  • 2

1 Answers1

0

I'd say that getting unsat or sat, depending on Z3's configuration, sounds like a bug to me. Check the Z3 issue tracker for issues that describe a similar behaviour, and if nothing pops up, file an issue there. Ideally with a minimal example to reproduce the problem with, your current one is rather long.

Malte Schwerhoff
  • 12,684
  • 4
  • 41
  • 71
  • Thanks! After a few checks, I found something similar to my case in the issue list and had earned PR merging. Would give the master branch a try. – hsuhw Oct 01 '18 at 08:46
  • 1
    For future reference, please link to the issue you found. Either by updating your post, by adding a comment, or by adding an answer (you can answer your own questions). You could also consider closing the question (accepting an answer, even yours) to indicate that it is no longer open. – Malte Schwerhoff Oct 01 '18 at 10:33