sat-minimize: allow different acceptances as input and output

* src/twaalgos/dtgbasat.cc (sat_minimize): Use any arbitrary
acceptance condition passed via the "acc" option.
* src/twa/acc.hh (mark_t::max_set): New method.
* src/tests/satmin2.test: Add two test cases with DRA as input
and DSA as output.

src/tests/satmin2.test

@@ -24,6 +24,8 @@ set -e
 # Skip if $SATSOLVE is not installed.
 (${SATSOLVER-glucose} --help >/dev/null 2>&1) || exit 77
 
+autfilt=../../bin/autfilt
+ltlfilt=../../bin/ltlfilt
 
 # This is a counterexample for one of the optimization idea we had for
 # the SAT-based minimization.
@@ -63,3 +65,49 @@ diff output expected
 ../../bin/ltl2tgba -BD -x sat-minimize "GF(a <-> XXb)" -H >out
 grep 'properties:.*state-acc' out
 grep 'properties:.*deterministic' out
+
+
+# DRA produced by ltl2dstar for GFp0 -> GFp1
+cat >test.hoa <<EOF
+HOA: v1
+States: 4
+properties: implicit-labels trans-labels no-univ-branch deterministic complete
+acc-name: Rabin 2
+Acceptance: 4 (Fin(0)&Inf(1))|(Fin(2)&Inf(3))
+Start: 0
+AP: 2 "p0" "p1"
+--BODY--
+State: 0 {0}
+1
+0
+3
+2
+State: 1 {1}
+1
+0
+3
+2
+State: 2 {0 3}
+1
+0
+3
+2
+State: 3 {1 3}
+1
+0
+3
+2
+--END--
+EOF
+
+# Let's try to find a smaller transition-based Streett automaton We
+# easily really check the expected automaton, because different SAT
+# solver (even different versions of glucose) can return different
+# automata.
+$autfilt -H --sat-minimize='acc="Fin(0)|Inf(1)"' test.hoa --stats=%s >output
+test `cat output` = 1
+
+# How about a state-based DSA?
+$autfilt -H --sat-minimize='state-based,acc="Fin(0)|Inf(1)"' test.hoa \
+	 --stats=%s > output
+test `cat output` = 3