* src/tgba/tgbasafracomplement.cc
(tgba_safra_complement::tgba_safra_complement)
(tgba_safra_complement::succ_iter): Correct the declaration and
use of multiple acceptance conditions.
(state_complement::to_string): Output the L set, not U. The previous
code caused different states to share the same names, causing issues
with the text-based output (state with identical names get merged).
* src/tgba/tgbasafracomplement.hh
(tgba_safra_complement::acceptance_cond_vec_): Adjust type to
store BDDs.
* src/tgbatest/complementation.cc: Implement a new "-b" option
to output automata in Spot's syntax.
* src/tgbatest/complementation.test: Add a test-case supplied
by Martin Dieguez Lodeiro.
* THANKS: Add Martin.
* src/tgba/tgbasafracomplement.cc (print_safra_tree): Fix output in
case of hash collision. Use the actual states to get a number, not
their hash value.
(print_safra_automaton): Output a mapping of values to states names.
(safra_tree_automaton::get_sba): New method, used by
print_safra_automaton.
* src/evtgbaalgos/tgba2evtgba.cc (process_link): Fix prototype
to match tgba_reachable_iterator::process_link.
* src/ltlvisit/tunabbrev.hh: Add using super::visit, so that the
other visit() method are in scope when we overload one.
* src/tgba/tgbareduc.hh, src/tgba/tgbareduc.cc (start, end,
process_link): Remove these empty methods. The default
implementations are empty too, and process_link had the
wrong prototype.
* src/tgbaalgos/reductgba_sim.hh, src/tgbaalgos/reductgba_sim.cc
(start, end, process_link): Likewise.
This avoids a small regression on the size of degeneralized
automata of our usual list of literature formulae.
* src/tgba/tgbatba.hh, src/tgba/tgbatba.cc
(tgba_tba_proxy::union_acceptance_conditions_of_original_state):
New method.
* src/tgba/tgbatba.cc (tgba_tba_proxy_succ_iterator): In accepting
states, ignore only the last expected acceptance condition if its
common to all outgoing transitions AND if it is not used by any
outgoing transitions of the destination.
states).
* src/tgbaalgos/tgbatba.cc: When degeneralizing to SBA, remove the
acceptance conditions that are common to all outgoing transitions
of this state. This helps to make the degeneralization
idempotent.
* src/tgbatest/degenid.test: New test case.
* src/tgbatest/Makefile.am: Add it.
Reported by Tomáš Babiak <xbabiak@fi.muni.cz>.
* src/tgba/tgbatba.cc (tgba_tba_proxy): Replace the std::map used
to record outgoing transitions by an Sgi::hash_map, and keep the
order of these transitions in a separate list.
* src/tgbatest/degendet.test: New file.
* src/tgbatest/Makefile.am (TESTS): Add it.
* THANKS: Add Tomáš and convert to utf8.
* src/tgba/tgbaexplicit.cc, src/tgba/tgbaexplicit.hh: Merge
state_explicit and tgba_explicit::state. In the past,
state_explicit was a small object encapsulating a pointer to the
persistent tgba_explicit::state; and we used to clone() and
destroy() a lot of state_explicit instance. Now state_explicit is
persistent, and clone() and destroy() have no effects.
* src/tgba/tgbareduce.cc: Adjust, since this inherits from
tgbaexplicit and uses the internals of state_explicit.
* src/tgbatest/reductgba.cc: Fix deletion order for automata.
* src/tgba/tgba.hh (last_support_conditions_input_,
last_support_variables_input_): Make these protected, so
they can be zeroed by tgba_explicit.
The gain is not very impressive. The runtime of the first example
in iface/dve2/README (also in dve2check.test) is 15% faster.
* src/tgba/tgbaproduct.hh (tgba_succ_iterator_product): Move ...
* src/tgba/tgbaproduct.cc (tgba_succ_iterator_product,
tgba_succ_iterator_product_common): ... in these two classes.
(tgba_succ_iterator_product_kripke): New class to speedup
successor computation on Kripke structures. We can assume that
all the transitions leaving the same state have the same label.
(tgba_product::tgba_product, tgba_product::succ_iter): Use
tgba_succ_iterator_product_kripke when appropriate.
(tgba_product_init::tgba_product_init): Adjust for the case
where tgba_product did reverse its operands.
* src/tgba/tgbaexplicit.cc (tgba_explicit::compute_support_conditions):
Fix logic. This function has always been returning bddtrue instead
of the actual computed value...
* src/tgba/succiter.hh, src/tgba/tgba.hh,
src/tgba/tgbabddconcrete.hh, src/tgba/tgbaproduct.hh,
src/tgba/tgbaunion.hh, src/tgbaalgos/bfssteps.hh,
src/tgbaalgos/gtec/ce.cc, src/tgbaalgos/gtec/explscc.hh,
src/tgbaalgos/gtec/gtec.cc, src/tgbaalgos/replayrun.cc,
src/tgbaalgos/scc.cc, src/tgbaalgos/scc.hh: Update comments
to say that we "destroy" a state instead of "deleting" it.
Right now, destroy() just executes "delete this". But in a later
version, we will rewrite tgba_explicit so that it does not
allocate new states (and the destroy() method for explicit state
will do nothing).
* src/tgba/state.hh (state::destroy): New method, to replace
state::~state() in the future.
(shared_state_deleter): New function.
* src/evtgba/product.cc, src/evtgbaalgos/reachiter.cc,
src/evtgbaalgos/save.cc, src/evtgbaalgos/tgba2evtgba.cc,
src/tgba/tgba.cc, src/tgba/tgbaproduct.cc, src/tgba/tgbareduc.cc,
src/tgba/tgbasafracomplement.cc, src/tgba/tgbasgba.cc,
src/tgba/tgbatba.cc, src/tgba/tgbaunion.cc, src/tgba/wdbacomp.cc,
src/tgbaalgos/cutscc.cc, src/tgbaalgos/emptiness.cc,
src/tgbaalgos/gtec/ce.cc, src/tgbaalgos/gtec/explscc.cc,
src/tgbaalgos/gtec/gtec.cc, src/tgbaalgos/gtec/nsheap.cc,
src/tgbaalgos/gv04.cc, src/tgbaalgos/magic.cc,
src/tgbaalgos/minimize.cc, src/tgbaalgos/ndfs_result.hxx,
src/tgbaalgos/neverclaim.cc, src/tgbaalgos/powerset.hh,
src/tgbaalgos/reachiter.cc, src/tgbaalgos/reducerun.cc,
src/tgbaalgos/reductgba_sim.cc,
src/tgbaalgos/reductgba_sim_del.cc, src/tgbaalgos/replayrun.cc,
src/tgbaalgos/safety.cc, src/tgbaalgos/save.cc,
src/tgbaalgos/scc.cc, src/tgbaalgos/se05.cc,
src/tgbaalgos/tau03.cc, src/tgbaalgos/tau03opt.cc: Adjust to call
"s->destroy()" instead of "delete s".
* src/saba/sabacomplementtgba.cc, src/tgba/tgbakvcomplement.cc:
Pass shared_state_deleter to the shared_ptr constructor, so that
it calls destroy() instead of delete.
* src/misc/hash.hh (identity_hash): New function.
* src/tgba/tgbaexplicit.hh (tgba_explicit_number): Use
identity_hash<int> instead of std::tr1::hash<int> that does not
exist with g++-3.3.
* src/tgbaalgos/minimize.hh, src/tgbaalgos/minimize.cc
(minimize): Split into ...
(minimize_wdba, minimize_monitor): ... these two functions.
* src/tgbatest/ltl2tgba.cc (main): Adjust the call to
minimize_monitor.
* wrap/python/cgi-bin/ltl2tgba.in: Adjust the calls to
minimize_monitor and minimize_obligation.
* wrap/python/spot.i: Declare minimize_monitor, minimize_wdba,
minimize_obligations.
* src/tgba/tgbaexplicit.hh (tgba_explicit_string)
(tgba_explicit_formula, tgba_explicit_number): Add fake
declarations so that SWIG can see they inherits from tgba.
from the initial automaton.
* src/tgba/tgbaexplicit.cc, src/tgba/tgbaexplicit.hh:
Added class tgba_explicit_number, a tgba_explicit where states are
labelled by integers.
* src/tgbaalgos/powerset.hh, src/tgbaalgos/powerset.cc:
When building the deterministic automaton, keep track of which states
were created from an accepting state in the initial automaton.
The states are added to the new optional parameter (if not 0),
acc_list.
Use tgba_explicit_number instead of tgba_explicit_string to build
the result.
* src/tgbaalgos/scc.cc (spot): Small fix.
Print everything on std::cout.
This fixes a bug in the powerset of such TGBA on the minimize branch.
* src/tgba/tgbabddconcrete.cc (compute_support_conditions): Also
account for the conditions from the acceptance relations.
* rc/tgba/tgbabddconcretefactory.hh, rc/tgba/tgbabddconcretefactory.cc
(acceptance_conditions_support): New variable to hold the value
of bdd_support(acceptance_conditions_support).
* src/tgba/tgbabddconcretefactory.cc (finish): Update
data_.acceptance_conditions_support.
With this change the output of
ltl2tgba -f -x -k -DS "GF(p_1) & ... & GF(p_n)
uses less than (n+1)^2 transitions when it used
exactly (n+1)*(2^n) transitions before.
* src/tgba/tgbatba.cc (tgba_tba_proxy_succ_iterator): Merge
transitions going to the same states if they are both accepting or
if neither are.
(state_ptr_bool_t, state_ptr_bool_less_than): Helper type to
store a transition in tgba_tba_proxy_succ_iterator.
* src/tgba/tgbatba.cc, src/tgba/tgbatba.hh
(tgba_tba_proxy::transition_annotation): Remove. We cannot
implement this method if transitions are merged.
tgba_tba_proxy used to spend a lot of time (re)computing the
acceptance condition common to all outgoing transition of a state.
* src/tgba/tgbatba.hh (accmap_): New cache.
(common_acceptance_conditions_of_original_state): New method.
* src/tgba/tgbatba.cc (tgba_tba_proxy_succ_iterator::~sync)
Call common_acceptance_conditions_of_original_state() instead of
computing the result.
(~tgba_tba_proxy): Cleanup the cache.
(common_acceptance_conditions_of_original_state): Implement it.
ltl2ba.
* src/neverparse/neverclaimparse.yy: Accept multiple labels
for the same state. Honor accepting states. Forward parse
error from the parser used for guards. Accept "false" as a
single instruction for a state.
* src/neverparse/neverclaimscan.ll: Recognize "false" specifically,
and remove the ";" hack.
* src/tgba/tgbaexplicit.cc
(tgba_explicit_string::~tgba_explicit_string): Adjust not to
destroy a state twice.
* src/tgba/tgbaexplicit.hh
(tgba_explicit_string::add_state_alias): New function.
* src/tgbatest/defs.in (SPIN, LTL2BA): New variables.
* src/tgbatest/neverclaimread.test: Check error messages for
syntax errors in guards. Make sure we can read the output
of `spin -f' and `ltl2ba -f' on a few test formulae.
We do that because the declaration of this type, which is local to
src/tgba/tgbasafracomplement.cc has a member in an anonymous
namespace, and some versions of g++-4.2 issue a very annoying
warning about this legitimate code. See Bug 29365 on GCC's
Bugzilla. Report by Silien Hong <silien.hong@lip6.fr>.
* src/tgba/tgbasafracomplement.hh (safra_tree_automaton): Do not
forward declare.
(tgba_safra_complement): Use void* instead of
safra_tree_automaton*.
* src/tgba/tgbasafracomplement.cc: static_cast void* to
safra_tree_automaton* anywhere needed.
* src/ltlast/binop.cc, src/ltlast/binop.cc: Add support for
these new operators.
* src/ltlparse/ltlparse.yy, src/ltlparse/ltlscan.ll: Parse them.
* src/ltltest/reduccmp.test: Add new tests for W and M.
* src/ltlvisit/basicreduce.cc, src/ltlvisit/contain.cc,
src/ltlvisit/lunabbrev.cc, src/ltlvisit/nenoform.cc,
src/ltlvisit/randomltl.cc, src/ltlvisit/randomltl.hh,
src/ltlvisit/reduce.cc, src/ltlvisite/simpfg.cc,
src/ltlvisit/simpfg.hh, src/ltlvisit/syntimpl.cc,
src/ltlvisit/tostring.cc, src/tgba/formula2bdd.cc,
src/tgbaalgos/eltl2tgba_lacim.cc, src/tgbaalgos/ltl2taa.cc,
src/tgbaalgos/ltl2tgba_fm.cc, src/tgbaalgos/ltl2tgba_lacim.cc:
Add support for W and M.
* src/tgbatest/ltl2neverclaim.test: Test never claim output
using LBTT, this is more thorough. Also we cannot use -N
any more in the spotlbtt.test.
* src/tgbatests/ltl2tgba.cc: Define M and W for ELTL.
* src/tgbatest/ltl2neverclaim.test: Test W and M, and use
-DS instead of -N, because lbtt-translate does not want
to translate these operators for tools that masquerade as Spin.
degeneralization.
* src/tgba/tgbatba.cc (tgba_sba_proxy::tgba_tba_proxy): Build the
list of acceptance condition in the reverse order. The order is
still arbitrary, but the bdd_satone() call seems to output the
acceptance conditions that are more used first, and this helps the
degeneralization process.
* src/tgba/tgbatba.cc (tgba_sba_proxy::tgba_sba_proxy): Set
cycle_start_ to start in the accepting layer of the degeneralized
automaton if the initial state has an accepting self-loop.
Otherwise, starts at the level of the first acceptance condition
as previously.
(tgba_sba_proxy::get_init_state): Use cycle_start_.
* src/tgba/tgbatba.hh (tgba_tba_proxy::a_): Make it protected so
that we can use it in tgba_sba_proxy::tgba_sba_proxy.
(tgba_sba_proxy::cycle_start_, tgba_sba_proxy::get_init_state):
Declare.
* src/tgbatest/ltl2tgba.test: More tests.
* src/tgba/tgbatba.cc (tgba_tba_proxy_succ_iterator::sync_): Move
the optimization step added by the previous patch outside the
before the bddtrue check, so that it also applies to accepting
states in SBA.
an acceptance condition on all outgoing transitions.
This was motivated by experiments from Rdiger Ehlers, showing
that "ltl2ba -f 'a U (b U c)'" outperformed "ltl2tgba -f -N -R3 'a
U (b U c)'". With this change and the previous one, it is no
longer the case.
* src/tgba/tgbatba.cc (tgba_tba_proxy_succ_iterator::aut_): Store
a pointer to the source automaton and...
(tgba_tba_proxy_succ_iterator::sync_): ... use it in an extra
optimization step to gather the acceptance conditions common
to all outgoing transitions of the destination state, and pretend
they are on the current (ingoing) transition.
(tgba_tba_proxy::succ_iter): Pass the
source automaton to the constructed iterator.
* src/tgbatest/spotlbtt.test: Test -f -N -R3 -r7.
* src/tgbatest/ltl2tgba.test: Add a test case for 'a U (b U c)'.
* src/tgbaalgos/ltl2tgba_fm.cc, src/tgbaalgos/ltl2tgba_fm.hh:
Remove the containment option.
* src/tgbafromfile.cc, src/tgbafromfile.hh: Remove the
containment_ member.
* src/tgbatest/ltl2tgba.cc (syntax): Remove -c option for
FM algorithm, use it exclusively for TAA.
* src/tgba/tgbacomplement.hh, src/tgba/tgbacomplement.cc: Rename
as...
* src/tgba/tgbakvcomplement.hh, src/tgba/tgbakvcomplement.cc:
... these. It makes more sense since we also have
tgba_safra_complement.
* src/tgba/Makefile.am, src/tgbatest/complement.cc, NEWS: Adjust.
