This test used to take more than 10min because an instance of valgrind
was launched for each separate equivalence check. The list of
equivalences to checks are not given in a file, and only two valgrind
instances are run. The test takes less than 15sec.
* src/ltltest/equalsf.cc: New file.
* src/ltltest/Makefile.am (reduccmp, reductaustr): Build using
equalsf.cc.
* src/ltltest/reduccmp.test: Rewrite.
* src/ltltest/uwrm.test: Also rewrite, and use valgrind.
star_normal_form() used to be called under bounded
repetitions like [*0..4], but some of these rewritings
are only correct for [*0..]. For instance
(a*|1)[*] can be rewritten to 1[*]
but (a*|1)[*0..1] cannot be rewritten to 1[*0..1]
it would be correct to rewrite the latter as (a[+]|1)[*0..1],
canceling the empty word in a*.
Also (a*;b*)[*] can be rewritten to (a|b)[*]
but (a*;b*)[*0..1] cannot be rewritten to (a|b)[*0..1]
and it cannot either be rewritten to (a[+]|b[+])[*0..1].
This patch introduces a new function to implement
rewritings under bounded repetition.
* src/ltlvisit/snf.hh, src/ltlvisit/snf.cc (star_normal_form_unbounded):
New function.
* src/ltlvisit/simplify.cc: Use it.
* src/ltltest/reduccmp.test: Add tests.
* doc/tl/tl.tex: Document the rewritings implemented.
* src/ltlvisit/simplify.cc: Remove two incorrect rules, and
partially disable another one.
* doc/tl/tl.tex: Reflect the change.
* src/ltltest/reduccmp.test: Likewise.
* src/ltltest/equals.cc: Add safety checks to catch such errors in the
future.
* NEWS: Mention the bug.
And also speedup implication checks for Boolean expressions.
* src/ltlvisit/simplify.cc: Improve implication-based rules
rules for multops by checking one operand against all the
other at once (instead of one by one). Do not break
Boolean expressions while performing implication checks.
* src/ltlvisit/simplify.hh: Typo.
* src/ltltest/reduccmp.test: More tests.
When something like XFa & FXa is reduced, the subformulae XFa and FXa
are both rewritten separately to XFa, and then the vector of arguments
of the And operators, [XFa,XFa], is passed through a specialized loop
that searches of the form X(...) that can potentially be simplified with
some other terms. This loop converts the vector [XFa,XFa] into the set
{XFa,XFa}={XFa} and forgot to deal with the case where the insertion
would actually not add an existing subformula.
* src/ltlvisit/simplify.cc: Fix the code for Or, and And.
* src/ltltest/reduc0.test: New file, to test it.
* src/ltltest/Makefile.am (TESTS): Add it.
* src/ltltest/reduccmp.test: Add an extra test that does not
trigger the bug (because reduccmp.test uses more than basic
optimizations, and the implication-based simplifications are
already able to detect that XFa and FXa are equivalent).
* src/ltlparse/ltlparse.yy: Make all the above operators
right-associative. Also let `:' have precedence over `;'.
* src/ltltest/reduccmp.test: Adjust for the `:' precedence.
* doc/tl/tl.tex, NEWS: Document this.
* src/ltlvisit/simplify.cc (reduce_sere_ltl): New function,
to factor the code of the []-> and <>-> rewrittings.
* src/ltltest/reduccmp.test: Add more tests.
* doc/tl/tl.tex: Document these rewritings.
* src/ltlvisit/simplify.cc: Here.
* doc/tl/tl.tex: Document then.
* src/ltlast/bunop.hh (as_KleenStar): New helper function.
* src/ltltest/reduccmp.test: Add more tests.
* src/ltltest/reduc.cc: Also display the resulting formula
without reduce_size_stricly.
* src/ltlvisit/snf.cc, src/ltlvisit/snf.hh: New files.
* src/ltlvisit/Makefile.am: Distribute them.
* src/ltlvisit/simplify.cc, src/ltlvisit/simplify.hh: Call snf(f) for
all f[*].
* src/ltltest/reduccmp.test: Test it.
* doc/tl/tl.tex, doc/tl/tl.bib: Document it.
* src/ltlast/multop.hh, src/ltlast/multop.cc (all_but): New method
used to simplify the removal of one element of a multop.
* src/ltlvisit/simplify.cc: Implement the new rewriting rules.
* doc/tl/tl.tex: Document them.
* src/ltltest/reduccmp.test: Test them.
* src/ltlvisit/simplify.hh (ltl_simplifier_options): New option
reduce_size_stricly.
* src/ltlvisit/simplify.cc (simplify_visitor): Implement these
rules.
* src/ltltest/reduc.cc: Check with reduce_size_strictly unset or
set, but only use the latter result to check sizes.
* src/ltltest/reduccmp.test: Test them.
* doc/tl/tl.tex: Document them.
The actual rules are a bit more complex:
a & X(G(a&b...)&c...) = Ga & X(G(b...)&c...)
a | X(Fa | c) = F(a) | c
with the second rule being applied only if all XF can
be removed. See the documentation for an example.
* src/ltlvisit/simplify.cc: Implement these new rules.
* doc/tl/tl.tex: Document them.
* src/ltltest/reduccmp.test: Add test cases.
* src/ltlvisit/simplify.cc: Implement them here, and augment them
to support M, and W operators.
* src/ltltest/reduccmp.test: Add some tests.
* doc/tl/tl.tex (Simplifications Based on Implications): Document
these rules.
* doc/tl/tl.bib (babiak.12.tacas): New entry.
* src/ltlvisit/simplify.cc (simplify_visitor): Do it.
* src/ltltest/reduccmp.test: Add a test.
* doc/tl/tl.tex: Document it.
* src/ltlast/multop.cc: Fix the computation of is.accepting_eword
for Fusion. The Fusion operator never accepts [*0].
This fixes a bug where {(a&!a)[=2]} was translated either into an
universal automaton (with simplification turned off) or in an
empty automaton (with simplification turned on).
* src/tgbaalgos/ltl2tgba_fm.cc (ratexp_to_dfa::translate): Trim
the automaton.
(ratexp_to_dfa::succ, ratexp_to_dfa::get_label): Deal with trimed
states.
(ltl_trad_visitor::visit(unop::Closure)): Likewise.
* src/tgbatest/ltl2tgba.test, src/ltltest/reduccmp.test: New test
cases.
* src/ltlvisit/simplify.cc (syntactic_implication_aux): Refine
rules to deal with pure eventualities and purely universal
properties.
* src/ltltest/reduccmp.test: Add tests.
* src/tgbaalgos/ltl2tgba_fm.cc (ratexp_trad_visitor::visit): Fix
the translation of the Goto operator.
(ratexp_trad_visitor::next_to_concat): More comments.
* src/ltltest/reduccmp.test: Add a test case.
* src/ltlvisit/simplify.hh
(ltl_simplifier::negative_normal_form): Allow logical
unabbreviations during the NNF pass.
* src/ltlvisit/simplify.cc
(ltl_simplifier::negative_normal_form)
(negative_normal_form_visitor): Adjust.
(ltl_simplifier::simplify): Request unabbreviations.
* src/ltlvisit/reduce.cc (reduce): Remove most
of the code, leaving only a call ltl_simplifier
and some wrapper code to convert options.
* src/ltltest/reduccmp.test: Add more test cases.
* src/ltlvisit/simplify.cc (ltl_simplifier): Add more rewritings
for formulae that are both universal and eventual.
* src/ltltest/reduccmp.test: Add six more cases.
* src/ltlvisit/simplify.cc (ltl_simplifier): Since we are processing
the formula bottom-up, don't assume all trivial simplification have
been done.
* src/ltltest/reduccmp.test: More tests.
So far I have only checked these rewritings with reduccmp.test.
There are probably a few kinks to iron out.
* src/ltlvisit/simplify.cc: Reimplement most of the basic
rewriting rules, leaving some FIXME comments for dubious ones.
* src/ltlast/multop.cc, src/ltlast/multop.hh: Ignore NULL
pointers in the vector.
* src/ltlvisit/reduce.cc (reduce): Do not call basic_reduce().
* src/ltltest/reduccmp.test: Adjust tests.
* src/ltlvisit/reduce.cc (reduce_visitor): Always reduce "a M b"
to "a & b" if "a" is a pure eventual formula, remove the
constraint on "b".
* src/ltltest/reduccmp.test: Add two tests.
clone for M.
* src/ltlvisit/contain.cc (reduce_tau03_visitor): Remove
the stronger rules for R and M. They were wrong.
* src/ltltest/reduccmp.test: Test a simpple counterexample.
* src/ltlvisit/reduce.cc (reduce_visitor): Add the following
implication rewriting rules:
a M (b M c) = a M c if a implies b.
a M (b R c) = a M c if a implies b.
a R (b R c) = a R c if a implies b.
a R (b M c) = b M c if b implies a.
a M (b M c) = b M c if b implies a.
The latter rule was fixed from an incorrectly copied&pasted
rule for a M (b R c) = b R c if b implies a (this is wrong).
* src/ltltest/reduccmp.test: Add more tests.
* src/ltlvisit/basicreduce.cc: Perform the following reductions:
(a R b) | Gb = a R b
(a M b) | Gb = a R b
(a U b) & Fb = a U b
(a W b) & Fb = a U b
* src/ltltest/reduccmp.test: Test them.
* src/ltlvisit/basicreduce.cc: Perform the following reductions:
(a U b) & (c W b) = (a & c) U b
(a W b) & (c W b) = (a & c) W b
(a R b) | (c M b) = (a | c) R b
(a M b) | (c M b) = (a | c) M b
* src/ltltest/reduccmp.test: Test them.
* src/ltlvisit/basicreduce.cc: Perform the following reductions.
a R (b & F(a)) = a M b
a M (b & F(a)) = a M b
a R Fa = Fa
a M Fa = Fa
a R b & Fa = a M b
a R b & a M c = a M (b & c)
a M b & a M c = a M (b & c)
* src/ltltest/reduccmp.test: More tests.
* src/ltlvisit/basicreduce.cc: Perform the following reductions.
a U (b | Ga) = a W b
a W (b | Ga) = a W b
a U b | Ga = a W b
a U b | a W c = a W (b | c)
a W b | a W c = a W (b | c)
a U Ga = Ga
a W Ga = Ga
* src/ltltest/reduccmp.test: More tests.
* 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.
