You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
A second program, which we believe has the same problem but terminates with a fixpoint error (which maybe makes the debugging easier), is the following:
However, the fixpoint error of the second program only occured with activated refinements and the activated bitfield domain of our PR (see #1623 ). Without our bitfield domain but still with activated refinements, the second program terminates normally without any problems. Therefore, we need a slightly different config file to produce this error: svcomp_bitfield.json
Minimal example where it misbehaves with not finding the fixpoint (decoupled from SV-COMP). Interestingly #1674 does not fix the issue, whereas #1675 does.
Activity
sim642 commentedon Feb 6, 2025
michael-schwarz commentedon Feb 6, 2025
sim642 commentedon Feb 7, 2025
michael-schwarz commentedon Feb 10, 2025
michael-schwarz commentedon Feb 11, 2025
michael-schwarz commentedon Feb 12, 2025
Minimal example where it misbehaves with not finding the fixpoint (decoupled from SV-COMP). Interestingly #1674 does not fix the issue, whereas #1675 does.
michael-schwarz commentedon Feb 12, 2025
Without loop unrolling.
michael-schwarz commentedon Feb 12, 2025
With @jerhard I traced the fixpoint error and it is in fact due to a mistake in Congruences (#1587)
The minimized program if someone wants to take a look again:
[-]Int domain refinement causes timeout and fixpoint error[/-][+]Int domain refinement causes fixpoint error[/+]Add test case for unreached fixpoint due to bug in congruences showin…