Morgan Deters [Sat, 1 Dec 2012 00:31:38 +0000 (00:31 +0000)]
definition-expansion fixed for get-model, resolves bug 411
Tim King [Sat, 1 Dec 2012 00:11:20 +0000 (00:11 +0000)]
Polishing API examples.
Tim King [Sat, 1 Dec 2012 00:08:38 +0000 (00:08 +0000)]
Adding SmtEngine::setLogic(const char* logic) so that smt.setLogic("QF_LRA"); works.
Tim King [Fri, 30 Nov 2012 23:34:47 +0000 (23:34 +0000)]
Fixes for stricter compilers Andy brought to my attention.
Tim King [Fri, 30 Nov 2012 23:10:10 +0000 (23:10 +0000)]
Changing the documentation of ARR_TABLE_FUN to say (internal symbol).
Morgan Deters [Fri, 30 Nov 2012 23:07:19 +0000 (23:07 +0000)]
all API examples now have java versions too; bitvectors gets built; also updated old-style copyrights in the examples
Morgan Deters [Fri, 30 Nov 2012 23:04:20 +0000 (23:04 +0000)]
incorporating some comments from Clark
Clark Barrett [Fri, 30 Nov 2012 22:44:26 +0000 (22:44 +0000)]
Fix assertion in smt_engine's getValue
Minor changes to RELASE-NOTES
Tim King [Fri, 30 Nov 2012 22:33:28 +0000 (22:33 +0000)]
Updating the combination.cpp example.
Tim King [Fri, 30 Nov 2012 22:28:46 +0000 (22:28 +0000)]
Committing tests to potentially discover an obscure CLN library issue on 32 bit platforms. The issue is discussed here: ginac.de/CLN/cln_3.html#SEC15.
Morgan Deters [Fri, 30 Nov 2012 22:28:06 +0000 (22:28 +0000)]
renaming --smtlib to --smtlib-strict; removing --smtlib2 option
Morgan Deters [Fri, 30 Nov 2012 21:43:11 +0000 (21:43 +0000)]
internal variables (skolems) aren't printed as part of the model
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Fri, 30 Nov 2012 21:31:12 +0000 (21:31 +0000)]
change detection/handling of output language more reasonably; fixes a nagging bug Tim found in API examples
(this commit was certified error- and warning-free by the test-and-commit script.)
Andrew Reynolds [Fri, 30 Nov 2012 20:38:45 +0000 (20:38 +0000)]
quantifiers now uses master equality engine, preparation work to cleanup code
Andrew Reynolds [Fri, 30 Nov 2012 20:21:39 +0000 (20:21 +0000)]
parametric datatypes fix related to non-ascribed type constructors introduced by decision procedure
Liana Hadarean [Fri, 30 Nov 2012 19:44:52 +0000 (19:44 +0000)]
added a simple API example example showing how to use the bit-vector theory.
Tim King [Fri, 30 Nov 2012 18:30:37 +0000 (18:30 +0000)]
Changes to SExpr to accept autoconversion from bool and const char*. Adding an example for combination.
Tim King [Fri, 30 Nov 2012 18:16:59 +0000 (18:16 +0000)]
Adding smtname level options for tlimit, rlimit, etc. Fix to the internal documentation in base_options.
Morgan Deters [Fri, 30 Nov 2012 18:11:48 +0000 (18:11 +0000)]
Partial fix for bug 435; still needs some effort.
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Fri, 30 Nov 2012 15:29:36 +0000 (15:29 +0000)]
Add some regressions for bug 438.
(this commit was certified error- and warning-free by the test-and-commit script.)
Morgan Deters [Fri, 30 Nov 2012 15:14:36 +0000 (15:14 +0000)]
fix rewrite-rules syntax in regression
Morgan Deters [Fri, 30 Nov 2012 14:07:11 +0000 (14:07 +0000)]
minor fix to release script
François Bobot [Fri, 30 Nov 2012 11:35:02 +0000 (11:35 +0000)]
fix the syntax of assert-rewrite/-propagation/-reduction by putting the pattern first
just after the bindings
Do it before the release in order to not break user files later
Kshitij Bansal [Thu, 29 Nov 2012 23:55:37 +0000 (23:55 +0000)]
reliable benchmark corresponding to bug468
Andrew Reynolds [Thu, 29 Nov 2012 23:28:29 +0000 (23:28 +0000)]
require type ascriptions for parametric datatype constructors (making them canonical), this fixes the followup issue of bug 438
Morgan Deters [Thu, 29 Nov 2012 23:08:25 +0000 (23:08 +0000)]
Fix for hidden symbols in library on Mac. It's a strange issue to do with
explicit template instantiation rules, -fvisibility=hidden, and the way that
Apple distributes libstdc++.
(this commit was certified error- and warning-free by the test-and-commit script.)
Andrew Reynolds [Thu, 29 Nov 2012 22:02:20 +0000 (22:02 +0000)]
fixes bug 438, incorporate subtypes into type unification when typechecking parameterized datatypes
Morgan Deters [Thu, 29 Nov 2012 21:34:16 +0000 (21:34 +0000)]
fix for andy: boolean terms stuff really shouldn't look at datatypes at all in this release
Morgan Deters [Thu, 29 Nov 2012 19:37:32 +0000 (19:37 +0000)]
minor documentation fix
Morgan Deters [Thu, 29 Nov 2012 18:15:56 +0000 (18:15 +0000)]
svn:ignore property
Clark Barrett [Thu, 29 Nov 2012 14:28:28 +0000 (14:28 +0000)]
Fix for nasty corner case found by fuzzer...
Kshitij Bansal [Thu, 29 Nov 2012 06:59:21 +0000 (06:59 +0000)]
Hack to support global variables for CVC language extended to export mechanism.
- Adds GlobalVarAttr node attribute
(this commit was certified error- and warning-free by the test-and-commit script.)
Clark Barrett [Thu, 29 Nov 2012 00:59:06 +0000 (00:59 +0000)]
Fixing function models with Boolean terms. Also, LAMBDA's should not be const.
Tim King [Wed, 28 Nov 2012 22:59:58 +0000 (22:59 +0000)]
Adding the helloworld.cpp example.
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