diff --git a/ChangeLog b/ChangeLog index de0dc5b7a..cdb33c9b7 100644 --- a/ChangeLog +++ b/ChangeLog @@ -1,3 +1,9 @@ +2009-11-05 Alexandre Duret-Lutz + + * src/tgbatest/ltlcounter.test (run): Only construct small + formulae (i.e. n<=2) under valgrind. The test case is too + slow otherwise. + 2009-11-04 Alexandre Duret-Lutz Fix spurious failure of style.test. diff --git a/src/tgbatest/ltlcounter.test b/src/tgbatest/ltlcounter.test index b1fb2df0f..cd2583ec0 100755 --- a/src/tgbatest/ltlcounter.test +++ b/src/tgbatest/ltlcounter.test @@ -31,13 +31,15 @@ lcl=$lcdir/LTLcounterLinear.pl lcc=$lcdir/LTLcounterCarry.pl lccl=$lcdir/LTLcounterCarryLinear.pl +run='run 0' + check_formula() { # First, check the satisfiability of the formula with Spot - run 0 ../ltl2tgba -e -x -f "$1" >/dev/null + $run ../ltl2tgba -e -x -f "$1" >/dev/null # Also check the satisfiability of the degeneralized formula - run 0 ../ltl2tgba -e -D -x -f "$1" >/dev/null - run 0 ../ltl2tgba -e -DS -x -f "$1" >/dev/null + $run ../ltl2tgba -e -D -x -f "$1" >/dev/null + $run ../ltl2tgba -e -DS -x -f "$1" >/dev/null } # Kristin Y. Rozier reported that the formulae with n=10 were badly @@ -59,4 +61,10 @@ do check_formula "$f" f=`"$lccl" $i` check_formula "$f" + + # Only run the first two formulae under valgrind, + # it is too slow otherwise. + if test $i = 2; then + run=: + fi done