Kshitij Bansal [Wed, 28 Nov 2012 21:52:56 +0000 (21:52 +0000)]
fix a potential race (have failed to reproduce)
Clark Barrett [Wed, 28 Nov 2012 21:26:43 +0000 (21:26 +0000)]
Fix for getValue. Now it can handle lambda applications
Morgan Deters [Wed, 28 Nov 2012 21:18:05 +0000 (21:18 +0000)]
Attempted "quick-fix" for QF_UF performance regression since Boolean terms added.
Sharing is turned on only when Boolean terms are detected during preprocessing. QF_UF problems (and others)
that don't use any Boolean terms won't have BV turned on.
(this commit was certified error- and warning-free by the test-and-commit script.)
Kshitij Bansal [Wed, 28 Nov 2012 20:18:46 +0000 (20:18 +0000)]
treat all get commands like getValue (send only to lastWinner)
Kshitij Bansal [Wed, 28 Nov 2012 18:37:51 +0000 (18:37 +0000)]
minor
Morgan Deters [Wed, 28 Nov 2012 18:00:10 +0000 (18:00 +0000)]
update to release notes
Morgan Deters [Wed, 28 Nov 2012 17:37:57 +0000 (17:37 +0000)]
Bug fix:
* Fix creation of bound variables in CVC native language parser
* This corrects a problem with misleading model output
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Wed, 28 Nov 2012 17:23:37 +0000 (17:23 +0000)]
fix: correct misleading comment in dump output
Morgan Deters [Tue, 27 Nov 2012 22:40:05 +0000 (22:40 +0000)]
Functions and predicates over Boolean now work with --check-models and output correct
models for such functions (though they are somewhat ugly at present).
There's still a problem with model extraction, but it's not Boolean terms' fault.
Sometimes checkModel() can report that the model is just fine, but if a user extracts
values with getValue(), they find problems with the model (i.e., it doesn't satisfy some
assertions). This appears to be due to an asymmetry between how checkModel() works and
how Model::getValue() works. I'll open a bugzilla report to discuss this after thinking
some more on it.
(this commit was certified error- and warning-free by the test-and-commit script.)
Kshitij Bansal [Tue, 27 Nov 2012 21:31:46 +0000 (21:31 +0000)]
fix in CommandSequence invoke : maintain success/failure. Fixes bug 465.
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Tue, 27 Nov 2012 19:18:12 +0000 (19:18 +0000)]
more mac fixes
Morgan Deters [Tue, 27 Nov 2012 19:05:13 +0000 (19:05 +0000)]
fix for some Mac builds
Kshitij Bansal [Tue, 27 Nov 2012 19:02:50 +0000 (19:02 +0000)]
Simplify --help=decision with only currently supported options
Add notice/warning when using incremental-mode + decision (it was
already disabled)
Some other minor cleanup
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Tue, 27 Nov 2012 18:09:16 +0000 (18:09 +0000)]
give warning at configure-time about unsupported language bindings, and don't advertise them in help listing for --enable-language-bindings
Morgan Deters [Tue, 27 Nov 2012 17:52:01 +0000 (17:52 +0000)]
do not turn on BV for QF_SAT
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Tue, 27 Nov 2012 05:52:21 +0000 (05:52 +0000)]
First chunk of boolean-terms support.
Passes simple tests and doesn't break existing functionality.
Still need some work merged in for models.
This version enables BV except for pure arithmetic (since we might otherwise need Boolean term support, which uses BV). Tonight's nightly regression run should tell us if/how that hurts performance.
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Tue, 27 Nov 2012 02:13:38 +0000 (02:13 +0000)]
Tuples and records merge. Resolves bug 270.
Also some fixes to parametric datatypes I found, and fixes for a handful of bugs, including some observed with --check-models --incremental on together.
(this commit was certified error- and warning-free by the test-and-commit script.)
Tim King [Tue, 27 Nov 2012 01:39:58 +0000 (01:39 +0000)]
Adding an example to show how to use arithmetic.
Tim King [Mon, 26 Nov 2012 21:38:28 +0000 (21:38 +0000)]
Improved implementation of Integer::length() with CLN enabled. Additional tests to sanity check both versions CLN and GMP versions of length.
Morgan Deters [Mon, 26 Nov 2012 20:05:02 +0000 (20:05 +0000)]
include new regression directories in summary test output
Morgan Deters [Mon, 26 Nov 2012 19:39:03 +0000 (19:39 +0000)]
rolling back r4625 for now (closes bug 464), Andy we should talk about this a bit more..
Dejan Jovanović [Mon, 26 Nov 2012 17:40:31 +0000 (17:40 +0000)]
fixup for incremental solving
Tim King [Mon, 26 Nov 2012 17:30:44 +0000 (17:30 +0000)]
Removing DioSolver::acceptableOriginalNodes(). This assertion was too strong, and was broken by r4620. This commit resolves bug463. Adding a previously triggering test case.
Dejan Jovanović [Mon, 26 Nov 2012 17:07:46 +0000 (17:07 +0000)]
Adding support for a master equality engine. Each theory gets the master equality engine through the setMasterEqualityEngine method. This is a read-only equality engine, so nothing should be added to it directly. Instead each equality engine that is of interest should have the master equality engine attached to it. To set when to use the master equality engine see TheoryEngine::finishInit().
Tim King [Mon, 26 Nov 2012 17:02:34 +0000 (17:02 +0000)]
Improving arithmetic debugging output.
Tim King [Mon, 26 Nov 2012 17:01:48 +0000 (17:01 +0000)]
Disabling test/regress/regress0/push-pop/bug396.smt2. This takes 2m to run in debug mode. This is too long for a regress0 test.
Morgan Deters [Mon, 26 Nov 2012 15:00:26 +0000 (15:00 +0000)]
don't include internal variables in model output
Morgan Deters [Mon, 26 Nov 2012 14:39:56 +0000 (14:39 +0000)]
some fixes to language bindings and function visibility
Morgan Deters [Mon, 26 Nov 2012 13:42:27 +0000 (13:42 +0000)]
Makefile fix for new versions of Make (thanks Clark for noticing this); see e.g. osdir.com/ml/bug-make-gnu/2011-04/msg00011.html
Tim King [Sun, 25 Nov 2012 01:02:49 +0000 (01:02 +0000)]
Adding a regression test from bug 462.
Tim King [Sun, 25 Nov 2012 00:49:25 +0000 (00:49 +0000)]
This commit fixes two incompleteness bugs (461, 459) introduced in r4611 dues to a problem of not correctly handling disequalities. Performance implications are uncertain.
Tim King [Sat, 24 Nov 2012 23:32:04 +0000 (23:32 +0000)]
Adds ensureConstraint(...) to ConstraintDatabase. This is a slightly optimized way of doing getConstraint() if the user already has a ValueCollection&.
François Bobot [Fri, 23 Nov 2012 14:37:40 +0000 (14:37 +0000)]
Example of rewrite rules use that comes from an harness test
Tim King [Wed, 21 Nov 2012 18:46:23 +0000 (18:46 +0000)]
- Removes getDeltaValueWithNonlinear() entirely
- Changes the signature of getDeltaValue to throw DeltaRationalException and ModelException
-- DeltaRationalExceptions can occur with non-linear arithmetic computations that cannot be done. (0 + delta) * (1 + delta) is not a DeltaRational.
-- ModelExceptions occur if getDeltaValue() is going to return a value that disagrees with its value in the model (due to non-linear arithmetic)
-- ModelExceptions also occur if getDeltaValue(n) is called on a variable arithmetic has never seen before.
- getEqualityStatus() now wraps and catches both of these exceptions. If either occurs, the equality status is EQUALITY_UNKNOWN. This allows arithmetic to handle terms it has never seen before in getEqualityStatus. This resolves bug 453.
- Removes the dead code rowImplication() entirely.
Tim King [Wed, 21 Nov 2012 18:36:29 +0000 (18:36 +0000)]
Adds a number of new capabilities to DeltaRational. This adds DeltaRationalException which can be called when an operation on a DeltaRational does not have well formed semantics. This allows for things like multiplying two DeltaRationals, which may or may not result in a DeltaRational.
Tim King [Wed, 21 Nov 2012 18:31:55 +0000 (18:31 +0000)]
Added debugging output to --check-models. I've found this output quite useful while debugging.
Kshitij Bansal [Mon, 19 Nov 2012 23:48:34 +0000 (23:48 +0000)]
Run lastWinner thread for all commands. Earlier behavior was to run
lastWinner only for (get-model) command, using thread0 otherwise.
Tim King [Mon, 19 Nov 2012 19:41:07 +0000 (19:41 +0000)]
Adding hand minimized test for bug 450.
Tim King [Mon, 19 Nov 2012 17:16:16 +0000 (17:16 +0000)]
Changed the splitting-on-demand lemmas of arithmetic to no longer contain the equality being split on. Instead of issuing 'x != y implies (x < y or x > y)', the lemma is now '(x <= y or x >= y)'. This resolves bug 450.
Tim King [Mon, 19 Nov 2012 01:18:10 +0000 (01:18 +0000)]
Fix for bug452.
Morgan Deters [Sun, 18 Nov 2012 21:53:36 +0000 (21:53 +0000)]
Disable predicate subtyping:
* remove from public interface (ExprManager, Type)
* CVC parser reports an unimplemented feature error if used
I didn't want to tear it out completely (from NodeManager, TypeNode,
type-checking, pre-processing, etc.) because that's a lot of hassle
and we'll add it back in after the release anyway. It *does* mean
that CVC4::Predicate is in the public interface, but that it can't be
used for anything (by users).
(this commit was certified error- and warning-free by the test-and-commit script.)
Andrew Reynolds [Sun, 18 Nov 2012 17:34:00 +0000 (17:34 +0000)]
support for quantifiers macros, bug fix for bug 454 involving E-matching Array select terms (thanks for the bug report Francois)
Clark Barrett [Sat, 17 Nov 2012 20:54:30 +0000 (20:54 +0000)]
Fixed last currently known bug in array models
Morgan Deters [Sat, 17 Nov 2012 18:57:16 +0000 (18:57 +0000)]
* enable previously-failing (now succeeding) datatype example that uses records
* some bindings cleanup
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Sat, 17 Nov 2012 12:55:23 +0000 (12:55 +0000)]
fix for language bindings (fixes debian build fail last night)
Morgan Deters [Sat, 17 Nov 2012 04:05:17 +0000 (04:05 +0000)]
* Fix for bug 445 agreed to in meeting 11/13/2012: always dump in ALL_SUPPORTED logic
* Java bindings fixes: fixed access to ostreams, iterators
* Make SmtEngine::setUserAttribute() (and others) take a const string&
* Also a few compliance fixes
(this commit was certified error- and warning-free by the test-and-commit script.)
Clark Barrett [Fri, 16 Nov 2012 20:14:07 +0000 (20:14 +0000)]
Fix for bug451
Dejan Jovanović [Fri, 16 Nov 2012 19:46:43 +0000 (19:46 +0000)]
fixing and refactoring the equality iterator
Morgan Deters [Fri, 16 Nov 2012 19:05:36 +0000 (19:05 +0000)]
Fix dumping of array-select expressions in CVC native language.
Thanks to Wei for the bug report.
Morgan Deters [Fri, 16 Nov 2012 17:52:37 +0000 (17:52 +0000)]
fix a compiler warning in models
Kshitij Bansal [Thu, 15 Nov 2012 22:39:58 +0000 (22:39 +0000)]
some fixes for --threads=1
Morgan Deters [Thu, 15 Nov 2012 22:26:10 +0000 (22:26 +0000)]
fix for "make examples"
Clark Barrett [Thu, 15 Nov 2012 21:27:33 +0000 (21:27 +0000)]
More fixes to model generation, with previously failing testcases
Also refactored some header file includes to reduce compile time
Liana Hadarean [Thu, 15 Nov 2012 20:57:22 +0000 (20:57 +0000)]
forgot to uncomment setLogicInternal for THEORY_BV
Liana Hadarean [Thu, 15 Nov 2012 20:52:09 +0000 (20:52 +0000)]
changed logic options so that justification is turned on for QF_ABV and QF_UFBV as well
Morgan Deters [Thu, 15 Nov 2012 18:12:40 +0000 (18:12 +0000)]
d_incomplete is context-dependent; we shouldn't be saving its value and restoring after a flipDecision().
(this commit was certified error- and warning-free by the test-and-commit script.)
Clark Barrett [Thu, 15 Nov 2012 15:03:53 +0000 (15:03 +0000)]
fuzz15 should have been fuzz14
Tim King [Thu, 15 Nov 2012 02:15:22 +0000 (02:15 +0000)]
Fix for bug 447.
Clark Barrett [Thu, 15 Nov 2012 02:14:42 +0000 (02:14 +0000)]
Fixed another AUFBV model bug. BV equality subtheory needed to do something
similar to arrays - limit the set of terms reported to those relevant in the
current context. Also removed collectModelInfo from sharedTermsDatabase -
doesn't seem to be needed any more.
Tim King [Thu, 15 Nov 2012 02:02:47 +0000 (02:02 +0000)]
Fixing comments in print_lambda.cvc.
Tim King [Wed, 14 Nov 2012 23:49:07 +0000 (23:49 +0000)]
Fix for bug 407. mkAnonymousFunction() in the parser no longer uses ':'. CVC printer now properly prints LAMBDAs. Model builing now gives bound variables names that can be parsed bypresentation language.
Andrew Reynolds [Wed, 14 Nov 2012 22:58:53 +0000 (22:58 +0000)]
replaced all static member data from rewrite rule triggers, added infrastructure for recognizing quantifier macros
Tim King [Wed, 14 Nov 2012 22:48:06 +0000 (22:48 +0000)]
Fixed a typo in r4576
Tim King [Wed, 14 Nov 2012 22:45:46 +0000 (22:45 +0000)]
Beautifying smt_engine.cpp.
Tim King [Wed, 14 Nov 2012 22:43:57 +0000 (22:43 +0000)]
Fix to bug449. Adds shared constants to the set of DeltaRationals that must be in the final total order.
Clark Barrett [Wed, 14 Nov 2012 21:54:25 +0000 (21:54 +0000)]
bug fixes to models, array rewriter with previously failing testcases
Kshitij Bansal [Wed, 14 Nov 2012 20:59:00 +0000 (20:59 +0000)]
Quantifiers enabled with portfolio, closing bug 423.
Kshitij Bansal [Wed, 14 Nov 2012 00:29:50 +0000 (00:29 +0000)]
fix a race problem. due to interrupt mechanism minisat returned true instead of undef.
Clark Barrett [Tue, 13 Nov 2012 22:51:27 +0000 (22:51 +0000)]
More bugfixes for models
Tim King [Tue, 13 Nov 2012 21:48:33 +0000 (21:48 +0000)]
SmtEngine::checkModel() should only be called if the result is sat or unknown because of incompleteness. Other unknown results are not safe to build models for, timeout, interrupted, etc.
Andrew Reynolds [Tue, 13 Nov 2012 21:16:26 +0000 (21:16 +0000)]
refactoring of quantifiers rewriter based on code review from yesterday, refactoring code out of instantiators in preparation for quantifiers equality engine
Liana Hadarean [Tue, 13 Nov 2012 20:38:40 +0000 (20:38 +0000)]
fixed failed bv regressions by refactoring out some rewrite rules from smt_engine.cpp
Liana Hadarean [Tue, 13 Nov 2012 04:49:34 +0000 (04:49 +0000)]
added support for division by zero for bit-vector division operators
Clark Barrett [Tue, 13 Nov 2012 02:47:09 +0000 (02:47 +0000)]
Relaxing too-strict assertion
Clark Barrett [Tue, 13 Nov 2012 02:28:05 +0000 (02:28 +0000)]
Fixed an array rewriting bug found by fuzzer
Clark Barrett [Tue, 13 Nov 2012 01:24:53 +0000 (01:24 +0000)]
Testcases for fixed bugs
Clark Barrett [Tue, 13 Nov 2012 01:24:27 +0000 (01:24 +0000)]
Fixed bug in array constant normalization found by fuzzer.
Tim King [Mon, 12 Nov 2012 23:50:29 +0000 (23:50 +0000)]
Improved error reporting for improperly using non-linear division in linear arithmetic.
Tim King [Mon, 12 Nov 2012 22:54:35 +0000 (22:54 +0000)]
Delta is now generated in arithmetic to keep consistent the total order of DeltaRational values for lowerbounds, upperbounds, assignments and disequalities. Throws LogicException when a non-linear term is asserted and the the LogicInfo isLinear() disagrees.
Liana Hadarean [Mon, 12 Nov 2012 19:38:47 +0000 (19:38 +0000)]
changed BitVector::unsignedRem to match the behavior of the bit-blasted circuit for division by 0; temporary fix for bug440
Morgan Deters [Mon, 12 Nov 2012 18:36:45 +0000 (18:36 +0000)]
Fix for bug 444, dealing with the placing of set-logic in dumping modes.
CVC4 now produces equivalent output when dumping the SMT1 and SMT2 attachments to that bug report.
This also fixes a memory leak in the new-variable-notification mechanism.
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Mon, 12 Nov 2012 18:34:32 +0000 (18:34 +0000)]
* Fix language bindings: various issues
** remove a number of warnings in bindings generation
** give appropriate names for operator-overloading
** make sure Java language bindings are built with -fno-strict-aliasing, to ensure the optimizer doesn't produce bad code
* Also remove BitVector::equals(), which wasn't used and was inconsistently implemented (operator==() is still there).
(this commit was certified error- and warning-free by the test-and-commit script.)
Andrew Reynolds [Mon, 12 Nov 2012 16:46:51 +0000 (16:46 +0000)]
minor bug fixes for quantifiers, added sort inference module (not ready to be used yet), added new totality lemma option for uf strong solver
Tim King [Sun, 11 Nov 2012 00:28:05 +0000 (00:28 +0000)]
Fixes for the arithmetic normal form and rewriter to handle arbitrary constants for total functions.
Tim King [Sat, 10 Nov 2012 23:56:46 +0000 (23:56 +0000)]
Beautifying integer_cln_imp.h
Tim King [Sat, 10 Nov 2012 23:55:34 +0000 (23:55 +0000)]
Fix to quantifier rewritting not being idempotent. See bug 441.
Tim King [Sat, 10 Nov 2012 20:27:00 +0000 (20:27 +0000)]
Removing rewriter call in SmtEngine::addFormula().
Clark Barrett [Sat, 10 Nov 2012 20:15:24 +0000 (20:15 +0000)]
Fixed missing \ in uflra/Makefile.ma
Fixed another model bug and added previously failing fuzz testcase
Tim King [Sat, 10 Nov 2012 19:01:49 +0000 (19:01 +0000)]
Fix for bug 439. Delta computation now includes disequality information.
Morgan Deters [Sat, 10 Nov 2012 17:08:54 +0000 (17:08 +0000)]
Change run-regression script to *additionally* run a second test with --check-models *if* the benchmark is sat/invalid (or is incremental, with at least one sat/invalid result).
This increases significantly the time to do a "make regress", as more tests are running.
We need to run both *with* and *without* --check-models, since that option disables certain preprocessing.
To turn off these extra tests during a make regress, you can set CVC4_REGRESSION_ARGS=--no-check-models.
To turn off the extra test for a *particular* regression benchmark, you can put this in the benchmark:
; COMMAND-LINE: --no-check-models
That might be necessary in e.g. nonlinear arithmetic benchmarks.
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Sat, 10 Nov 2012 16:41:01 +0000 (16:41 +0000)]
Updates to Clark's commit r4540:
* ALL_SUPPORTED/QF_ALL_SUPPORTED don't include nonlinear
* Change "Notice" to "Warning" when produce-models turned off due to non-linear
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Sat, 10 Nov 2012 16:08:37 +0000 (16:08 +0000)]
fix typo in language bindings
Clark Barrett [Sat, 10 Nov 2012 04:34:34 +0000 (04:34 +0000)]
Finally tracked down problems in regressions and fuzz results (this also
explains why my build was giving different answers from Dejan's build).
Problem was that if you run:
cvc4 --check-models foo.smt
it would fail, but if you run
cvc4 --check-models --produce-models foo.smt
it would succeed. Even though produce-models is supposed to be automatically
set when you set check-models, it seems this was happening too late. Fixed by
removing any distinction between produce-models and check-models in the
heuristic setting code. Kind of ugly though.
Also, disable models if nonlinear arithmetic is on.
Morgan Deters [Fri, 9 Nov 2012 21:25:07 +0000 (21:25 +0000)]
TheoryEngineModelBuilder::buildModel() is only called once with fullModel=true, within a SAT context. This fixes some outstanding model bugs.
Committing also a Clark-provided assertion the Model code to ensure the call is only done once per context.
(this commit was certified error- and warning-free by the test-and-commit script.)
Kshitij Bansal [Fri, 9 Nov 2012 21:11:47 +0000 (21:11 +0000)]
export null nodes (fixes a bug in portfolio model stuff)
Tim King [Fri, 9 Nov 2012 21:05:43 +0000 (21:05 +0000)]
Test that breaks arithmetic model building due to disequality terms.
Tim King [Fri, 9 Nov 2012 19:58:00 +0000 (19:58 +0000)]
Arithmetic problem that fails --check-models due incompleteness with multiplication.
Morgan Deters [Fri, 9 Nov 2012 19:18:58 +0000 (19:18 +0000)]
Bug-fix for a crash involving improperly-thrown exceptions; also, add LogicException for errors where the user uses a feature not permitted in the current logic (e.g., a quantifier in a QF logic)
Morgan Deters [Fri, 9 Nov 2012 18:41:24 +0000 (18:41 +0000)]
In non-linear logics, rewrite DIVISION, INTS_DIVISION, and INTS_MODULUS into ITEs of the form:
IF (denom = 0) THEN divByZero(num) ELSE (DIVISION_TOTAL num denom) ENDIF
where divByZero is an uninterpreted function symbol (there's one for each of the partial operators).
In linear logics, don't do any of this.
Bitvector partial functions to come, if this is successful for Tim.
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Fri, 9 Nov 2012 17:39:05 +0000 (17:39 +0000)]
another DISTCLEANFILES entry, for proper "make distclean" behavior (fixes "distclean" target error last night)