* spot/twaalgos/sccfilter.cc: If an inherently-weak automaton has
no rejecting cycle, reduce its acceptance to t instead of Büchi.
* spot/twa/acc.hh (operator==, operator<): Fix comparisons of
true acceptances.
* NEWS: Mention these two changes.
* spot/twaalgos/sccfilter.hh: Update documentation.
* spot/twaalgos/determinize.cc (tgba_determinize): The call
to scc_filter assume that the input BA is never reduced to t
acceptance. Call scc_filter with an extra option to ensure that.
* spot/twaalgos/postproc.cc (do_scc_filter): Adjust to add the
extra option when we want to build Büchi or coBuchi.
(ensure_ba): Don't mark trivial SCCs as accepting.
* tests/core/complement.test, tests/core/dstar.test,
tests/core/ltlsynt.test, tests/core/readsave.test,
tests/core/wdba2.test, tests/python/_product_susp.ipynb,
tests/python/automata-io.ipynb, tests/python/dualize.py,
tests/python/highlighting.ipynb, tests/python/intrun.py,
tests/python/setacc.py, tests/python/simstate.py,
tests/python/stutter-inv.ipynb, tests/python/zlktree.py: Adjust test
cases.
This massive set of changes was triggered by issue #546.
In addition to the better handling of !weak, this also adds some
weak properties in a few places.
* spot/twaalgos/product.cc (product_aux): Throw some exception
if an automaton with t or f acceptance has the !weak property. This
is a cheap sanity check to help detect algorithms that incorrectly
assumed !weak input would necessarily become !weak output.
* spot/twaalgos/hoa.cc (print_hoa): Likewise, also do not assume
that terminal implies very-weak.
* spot/parseaut/parseaut.yy: Add several diagnostics for similar
cases. E.g., a one-state automaton cannot be declared as !very-weak.
* tests/core/parseaut.test: Check those new diagnostics.
* spot/twa/twa.cc (twa::intersecting_run): Temporary remove the weak
property by setting it to maybe, not to false.
* spot/twaalgos/minimize.cc, spot/twaalgos/parity.cc,
spot/twaalgos/sccfilter.cc, spot/twaalgos/simulation.cc: Account for
the fact that these algorithm may in fact improve the weakness.
* spot/twaalgos/strength.cc: Only look at colors used by the
acceptance condition when deciding weakness.
* spot/twaalgos/synthesis.cc: Declare the strategy as weak.
* bin/randaut.cc: Add weak to automata with t/f acceptance.
* spot/kripke/kripke.hh: Make kripke structures as weak.
* tests/core/acc_word.test, tests/core/alternating.test,
tests/core/complement.test, tests/core/complete.test,
tests/core/ltlsynt.test, tests/core/randomize.test,
tests/core/readsave.test, tests/core/remfin.test,
tests/core/sccsimpl.test, tests/core/strength.test,
tests/core/wdba2.test, tests/ltsmin/kripke.test,
tests/python/automata-io.ipynb, tests/python/automata.ipynb,
tests/python/dbranch.py, tests/python/highlighting.ipynb,
tests/python/kripke.py, tests/python/ltsmin-dve.ipynb,
tests/python/mealy.py, tests/python/simstate.py: Adjust all these test
cases.
* NEWS: Mention the fixes.
Reported by Reuben Rowe.
* spot/twaalgos/complement.cc (complement): Remove the hard-coded
simul=0 option on automata with >32 states. In 2.10 simul=0 now
implies det-simul=0, causing the regression, and most importantly it
is not needed anymore, because we have other threashold like simul-max
and simul-trans-pruning in place.
* tests/core/complement.test: Add Reuben's automaton as test case.
* NEWS: Mention the fix.
This fixes a complementation bug reported by Juraj Major and Tereza
Šťastná.
* spot/twaalgos/remfin.cc (is_scc_tba_type): Fix the condition for
handling Fin-alone pairs.
* tests/core/complement.test: Add Juraj & Tereza's test case.
* NEWS: Mention it.
* bin/autfilt.cc: Fix the --help string for --complement, and also
merge edges in the resulting automaton, as suggested by František
Blahoudek.
* tests/core/complement.test: Adjust output and add František's
example.
Fixes#282.
* spot/misc/bddlt.hh (bdd_less_than_stable): New function.
* spot/twa/twagraph.cc (merge_edges): Use it.
* tests/core/genltl.test: Adjust, and add an extra test
for the behavior of #282.
* tests/core/complement.test, tests/core/degenid.test,
tests/core/ltldo.test, tests/core/prodor.test,
tests/core/readsave.test, tests/core/sbacc.test,
tests/python/atva16-fig2a.ipynb, tests/python/automata.ipynb,
tests/python/decompose.ipynb, tests/python/dualize.py,
tests/python/highlighting.ipynb, tests/python/piperead.ipynb,
tests/python/product.ipynb, tests/python/simstate.py,
tests/python/tra2tba.py: Adjust all expected outputs.
* NEWS: Mention the bug.
Fixes#262 again. Reported by Maximilien Colange.
* spot/twaalgos/simulation.cc: Use state numbers to order classes, not
their signatures. The problem was that even if two simulation of the
same automaton assign the same signature, the BDD identifier used for
that signature might be different, and therefore the ordering obtained
by using BDDs as keys in a map can be different. A side-effect of
this change is that the order of states in automata produced by
simulation-based reduction may change; many tests had to be updated.
* tests/core/ltl2tgba.test: Add a new test case based on Maximilien's
report.
* tests/core/complement.test, tests/core/det.test,
tests/core/parseaut.test, tests/core/prodor.test, tests/core/scc.test,
tests/python/atva16-fig2a.ipynb, tests/python/automata.ipynb,
tests/python/decompose.ipynb, tests/python/decompose_scc.py,
tests/python/highlighting.ipynb, tests/python/piperead.ipynb,
tests/python/sccinfo.py, tests/python/simstate.py,
tests/python/testingaut.ipynb, tests/python/word.ipynb: Update
test case for new order of states.
