Andres Noetzli [Sat, 30 Nov 2019 01:01:19 +0000 (17:01 -0800)]
Competition build: Skip parsing error regression (#3511)
Andrew Reynolds [Sat, 30 Nov 2019 00:24:30 +0000 (18:24 -0600)]
Fix fast SyGuS enumeration for interpreted constants (#3501)
Andrew Reynolds [Fri, 29 Nov 2019 15:31:54 +0000 (09:31 -0600)]
Check free variables in assertions when using SyGuS (#3504)
Andrew Reynolds [Wed, 27 Nov 2019 21:59:36 +0000 (15:59 -0600)]
Fix sygus inference for choice functions introduced at preprocess (#3500)
Haniel Barbosa [Wed, 27 Nov 2019 20:52:36 +0000 (17:52 -0300)]
Enable sygusRecFun by default and fixes SyGuS+RecFun+HO issues (#3502)
Andrew Reynolds [Wed, 27 Nov 2019 17:03:49 +0000 (11:03 -0600)]
Fix indexof range lemma (#3499)
Andrew Reynolds [Mon, 25 Nov 2019 23:59:51 +0000 (17:59 -0600)]
Better front-end type checking for SyGuS (#3496)
Andrew Reynolds [Fri, 22 Nov 2019 18:09:19 +0000 (12:09 -0600)]
Minor refactoring of compute model value for nl (#3489)
* Refactor compute model value for nl
* Format
Haniel Barbosa [Fri, 22 Nov 2019 14:50:12 +0000 (11:50 -0300)]
fixing stupid typo (#3488)
Haniel Barbosa [Thu, 21 Nov 2019 18:17:16 +0000 (15:17 -0300)]
hard limit for rec-fun eval (#3485)
Andrew Reynolds [Thu, 21 Nov 2019 15:54:12 +0000 (09:54 -0600)]
Evaluation unfolding for symbolic SyGuS constructors (#3483)
Haniel Barbosa [Wed, 20 Nov 2019 19:42:58 +0000 (16:42 -0300)]
Lazy evaluation via rec-funs of ITE expressions (#3482)
Andres Noetzli [Tue, 19 Nov 2019 01:59:22 +0000 (17:59 -0800)]
Fix reduction of `sqrt` (#3478)
Alex Ozdemir [Tue, 19 Nov 2019 01:07:52 +0000 (17:07 -0800)]
Add a few comments to ProofManager (#3477)
Alex Ozdemir [Tue, 19 Nov 2019 00:41:07 +0000 (16:41 -0800)]
Signature documentation update (#3476)
This comment was slightly out-of-date.
Andres Noetzli [Mon, 18 Nov 2019 20:28:54 +0000 (12:28 -0800)]
Use -Wimplicit-fallthrough (#3464)
This commit enables compiler warnings for implicit fallthroughs in
switch statements that are not explicitly marked as such. The commit
introduces a new macro `CVC4_FALLTHROUGH` that can be used to indicate
that a fallthrough is intentional. The commit fixes existing warnings
and a bug in the arithmetic rewriter for `abs` (the bug likely couldn't
be triggered easily because we rewrite `abs` to an `ite` while expanding
definitions).
To have the new macro also available in the parser, the commit changes
`src/base/check.h` to be visible to the parser (it includes
`cvc4_private_library.h` now instead of `cvc4_private.h`).
Andrew Reynolds [Mon, 18 Nov 2019 19:52:18 +0000 (13:52 -0600)]
Use standard sygus interface for abduction and rewrite rule synthesis (#3471)
Andrew Reynolds [Mon, 18 Nov 2019 19:13:14 +0000 (13:13 -0600)]
Improve interface for sygus datatype, fix utilities (#3473)
Andrew Reynolds [Mon, 18 Nov 2019 02:19:24 +0000 (20:19 -0600)]
Updates to the unit tests, api, and examples for datatypes (#3459)
* Updates to the unit tests, api, and examples for datatypes
* Format
Andres Noetzli [Sun, 17 Nov 2019 06:38:34 +0000 (22:38 -0800)]
Add support for ThreadSanitizer instrumentation (#3467)
This commit adds support for compiling CVC4 with ThreadSanitizer
instrumentation. This is useful for debugging issues when CVC4 is used
in a multi-threaded context (e.g. #3292).
Andrew Reynolds [Sat, 16 Nov 2019 04:52:58 +0000 (22:52 -0600)]
Use standard interface for sygus default grammar construction (#3466)
Andrew Reynolds [Fri, 15 Nov 2019 22:11:34 +0000 (16:11 -0600)]
Introduce SyGuS datatype API (#3465)
Andrew Reynolds [Fri, 15 Nov 2019 13:59:38 +0000 (07:59 -0600)]
Fix wrong kind in sygus version 1 parser (#3463)
Alex Ozdemir [Thu, 14 Nov 2019 18:39:05 +0000 (10:39 -0800)]
Use Shebang in cxxtestgen when appropriate (#3458)
CxxTest's generator has one of three names, depending on the CxxTest
version:
* `cxxtestgen.py`, a python program
* `cxxtestgen`, some kind of program with a shebang
* `cxxtestgen.pl`, a perl program
We were mistakenly assuming that the `cxxtestgen` form was always a
python program.
Now, if we find that form on the system, we assume that it is has a
shebang, and is executable.
Andres Noetzli [Wed, 13 Nov 2019 22:14:09 +0000 (14:14 -0800)]
Allow (set-logic ...) after (reset) (#3457)
Fixes #3353. #3062 introduced a flag that tracks whether we have seen a
`(set-logic ...)` command to improve the handling of `--force-logic`.
However, the flag was not set to `false` when `(reset)` was called. This
commit fixes the issue.
Andrew Reynolds [Wed, 13 Nov 2019 06:03:35 +0000 (00:03 -0600)]
Distinguish unknown status for model printing (#3454)
Andrew Reynolds [Wed, 13 Nov 2019 05:34:05 +0000 (23:34 -0600)]
Refactor non-linear extension for model-based refinement (#3452)
* Refactor non-linear extension for model-based refinement
* Format
* Minor
* Address
Andrew Reynolds [Mon, 11 Nov 2019 19:02:55 +0000 (13:02 -0600)]
Add missing utilities for Node-level Datatype API (#3451)
Andres Noetzli [Mon, 11 Nov 2019 18:07:29 +0000 (10:07 -0800)]
Fix mkConst<RoundingMode>() for Python bindings (#3447)
yoni206 [Mon, 11 Nov 2019 17:07:14 +0000 (09:07 -0800)]
Update README according to the new website (#3438)
Andrew Reynolds [Mon, 11 Nov 2019 16:09:00 +0000 (10:09 -0600)]
Eliminate remaining references to type/expr in datatype type rules. (#3450)
Andrew Reynolds [Sun, 10 Nov 2019 14:45:39 +0000 (08:45 -0600)]
Fix bugs related to sygus higher-order + recursive functions (#3448)
Andrew Reynolds [Sat, 9 Nov 2019 22:09:11 +0000 (16:09 -0600)]
Fixes in relations related to datatypes not passed by reference (#3449)
The current code is creating/destroying datatypes unnecessarily.
Mathias Preiner [Fri, 8 Nov 2019 21:41:21 +0000 (13:41 -0800)]
cmake: Disable C++ GNU extensions. (#3446)
Fixes #971.
Andrew Reynolds [Wed, 6 Nov 2019 23:12:29 +0000 (17:12 -0600)]
Move more string utility functions (#3398)
This is work towards splitting a "core solver" object from TheoryStrings.
This moves global functions from TheoryStrings to InferenceManager/SolverState, making them accessible in the future by modules that have references to these objects.
It also corrects an issue where we were maintaining two `d_conflict` fields.
Andrew Reynolds [Wed, 6 Nov 2019 17:11:28 +0000 (11:11 -0600)]
Migrate more datatype methods to the Node level (#3443)
This adds node-level interfaces for a few missing functions that will be necessary to have a Node-level API for datatypes.
Andres Noetzli [Wed, 6 Nov 2019 16:40:22 +0000 (08:40 -0800)]
[Regressions] Remove leading whitespace in output (#3444)
Andrew Reynolds [Wed, 6 Nov 2019 16:00:08 +0000 (10:00 -0600)]
Support for SyGuS PBE + recursive functions (#3433)
Andres Noetzli [Wed, 6 Nov 2019 00:26:04 +0000 (16:26 -0800)]
Remove casts to subclasses of Type in API (#3420)
Andrew Reynolds [Tue, 5 Nov 2019 23:37:37 +0000 (17:37 -0600)]
Separate model object in non-linear extension (#3426)
Andrew Reynolds [Tue, 5 Nov 2019 20:12:43 +0000 (14:12 -0600)]
Refactor type matcher utility (#3439)
Andres Noetzli [Tue, 5 Nov 2019 01:04:04 +0000 (17:04 -0800)]
[Regressions] Support for running w/ default args (#3436)
Andrew Reynolds [Mon, 4 Nov 2019 23:48:30 +0000 (17:48 -0600)]
Make check synth solution robust to auxiliary assertions (#3432)
Andrew Reynolds [Mon, 4 Nov 2019 22:22:51 +0000 (16:22 -0600)]
Fix ho extensionality in collect model info (#3435)
Andrew Reynolds [Mon, 4 Nov 2019 20:24:02 +0000 (14:24 -0600)]
Avoid non-well-founded sygus grammars (#3434)
Andrew Reynolds [Mon, 4 Nov 2019 18:23:34 +0000 (12:23 -0600)]
Make getSynthSolution return a Bool (#3306)
Andrew Reynolds [Mon, 4 Nov 2019 17:11:38 +0000 (11:11 -0600)]
Eliminate deprecated utility function from sygus (#3431)
Andrew Reynolds [Fri, 1 Nov 2019 22:06:33 +0000 (17:06 -0500)]
Fix non-termination in datatype type enumerator (#3369)
Andres Noetzli [Fri, 1 Nov 2019 18:58:57 +0000 (11:58 -0700)]
Fix and refactor TheoryStrings::checkFlatForms() (#3326)
This commit fixes some minor (performance) issues in
`TheoryStrings::checkFlatForms()`: The `inelig` vector was initialized
with copies of the `start` element instead of all the elements before
`start` and the `else` branch of `count == asize` was looping over all
elements from `1` instead of `start + 1`. Additionally, this commit
refactors the code to be a bit more readable.
Andrew Reynolds [Fri, 1 Nov 2019 12:09:35 +0000 (07:09 -0500)]
Eagerly beta reduce during sygus to builtin term conversion (#3418)
Andrew Reynolds [Fri, 1 Nov 2019 04:37:16 +0000 (23:37 -0500)]
Rename datatypes sygus solver (#3417)
Mathias Preiner [Thu, 31 Oct 2019 21:39:29 +0000 (14:39 -0700)]
Fix Unimplemented() macros missed in #3366. (#3424)
Mathias Preiner [Wed, 30 Oct 2019 22:27:10 +0000 (15:27 -0700)]
Unify CVC4_CHECK/CVC4_DCHECK/AlwaysAssert/Assert. (#3366)
Andrew Reynolds [Wed, 30 Oct 2019 00:42:50 +0000 (19:42 -0500)]
Split some generic utilities from the non-linear extension (#3419)
* Split arith util
* Cleaner
* cpp
* Format
* Minor
Andrew Reynolds [Mon, 28 Oct 2019 17:50:08 +0000 (12:50 -0500)]
Fix for non-linear models (#3410)
* Towards fix for non-linear models
* Format
* Fix
* More
* Improve
* Format
* More
Andres Noetzli [Mon, 28 Oct 2019 15:59:44 +0000 (08:59 -0700)]
Fix integer division rewrite (#3415)
Andrew Reynolds [Sun, 27 Oct 2019 22:47:56 +0000 (17:47 -0500)]
Fix collect model info for higher-order (#3409)
This ensures we add lemmas when collect model info fails for the higher order extension of UF. This fixes #3405 (that benchmark now answers unknown).
Andres Noetzli [Sun, 27 Oct 2019 18:06:40 +0000 (11:06 -0700)]
Fix global-declarations support (#3403)
Andrew Reynolds [Wed, 23 Oct 2019 23:15:28 +0000 (18:15 -0500)]
Fixes for SyGuS + regular expressions (#3313)
This commit fixes numerous issues involving the combination of SyGuS and regular expressions.
Combining SyGuS and regular expressions may involve constructing regular expressions that are neither variables nor builtin regular expression operators. The code was not robust for this case, either throwing spurious assertion failures or having incorrect behavior.
Andrew Reynolds [Wed, 23 Oct 2019 00:42:55 +0000 (19:42 -0500)]
Refactoring skolems for sets (#3381)
This refactors skolems introduced in the theory of sets. This is analogous to how skolems are treated for the theory of strings.
A key change that this commit enables is to identify "variable" sets based on those that weren't introduced by the SkolemCache (instead of via a check that their kind is `VARIABLE`, which is done currently and is error prone).
Aina Niemetz [Tue, 22 Oct 2019 21:03:09 +0000 (14:03 -0700)]
NodeValue: Eliminate redundant NBITS macros. (#3400)
Previously, the metakind header defined macros for the number of bits
reserved for fields in the NodeValue "header" (for the reference count,
the node kind, the number of children and the node id). These macros
were redundant, since the only one using them was the NodeValue itself,
which redefined them (while using them) as constants in the class.
Additionally, MAX_CHILDREN was defined (using these macros) not only
in the metakind header, but redefined in other places.
This commit defines the above values as constexpr members of the
NodeValue class and cleans up redundancy.
Andrew V. Jones [Sun, 20 Oct 2019 15:35:24 +0000 (16:35 +0100)]
Cleaning-up the declaration of wrapped functions/methods, which have no definitions (#3399)
Signed-off-by: Andrew V. Jones <andrewvaughanj@gmail.com>
makaimann [Fri, 18 Oct 2019 18:42:15 +0000 (11:42 -0700)]
Update overflow check to handle negative numbers (#3396)
Andrew Reynolds [Thu, 17 Oct 2019 21:40:18 +0000 (16:40 -0500)]
Move datatype utility functions to own file (#3397)
Andrew Reynolds [Wed, 16 Oct 2019 23:44:17 +0000 (18:44 -0500)]
Solver state for theory of strings (#3181)
This refactors the theory of strings to use a solver state object, which manages state information regarding assertions.
It also deletes some unused/undefined functions in theory_strings.h.
There are no major changes to the behavior of the code or its documentation in this PR.
This is work towards #1881.
Andres Noetzli [Tue, 15 Oct 2019 21:43:33 +0000 (14:43 -0700)]
Fix line numbers in templates (#3391)
This commit updates the line numbers in templates to address warnings
due to wrong line numbers.
Andres Noetzli [Tue, 15 Oct 2019 20:59:41 +0000 (13:59 -0700)]
Remove remaining references to Boost and Autotools (#3390)
This commit removes references to Boost and Autotools in the copyright
information and CMakeLists.txt.
Andres Noetzli [Tue, 15 Oct 2019 19:49:20 +0000 (12:49 -0700)]
Fix OOB access (#3383)
In theory_engine.cpp, we were calling `theoryOf()` with
`THEORY_SAT_SOLVER` as the theory id. However, `THEORY_SAT_SOLVER` is
defined as `THEORY_LAST` and thus out-of-bounds of the `d_theoryTable`
defined in theory_engine.h (which is of size `THEORY_LAST`. This commit
adds an assertion that detects the out-of-bound access and introduces a
method to turn a theory id into a string which correctly handles
`THEORY_SAT_SOLVER`.
Andres Noetzli [Tue, 15 Oct 2019 13:08:48 +0000 (06:08 -0700)]
Fix regression (#3393)
PR #3388 didn't disable the regression correctly (due to using `REQUIRE`
instead of `REQUIRES`). This commit fixes the issue.
Andres Noetzli [Mon, 14 Oct 2019 21:21:32 +0000 (14:21 -0700)]
Disable regression test for competition build (#3388)
This commit disables a regression test that was failing for the
competition build due to not emitting the expected error message.
Andrew Reynolds [Mon, 14 Oct 2019 20:53:55 +0000 (15:53 -0500)]
Remove benchmark (#3389)
Andrew Reynolds [Mon, 14 Oct 2019 20:01:05 +0000 (15:01 -0500)]
Support UF in default sygus grammars (#3319)
Andrew Reynolds [Mon, 14 Oct 2019 19:23:38 +0000 (14:23 -0500)]
Apply sygus repair constant techniques restricted to refinement lemmas (#3386)
Andrew Reynolds [Mon, 14 Oct 2019 18:24:27 +0000 (13:24 -0500)]
Ensure lemmas from sygus repair const are guarded (#3385)
Andrew Reynolds [Mon, 14 Oct 2019 17:03:41 +0000 (12:03 -0500)]
Minor refactor in strings rewriter (#3387)
Andrew Reynolds [Sun, 13 Oct 2019 23:22:28 +0000 (18:22 -0500)]
Eliminate negative constant coefficients in div/mod (#2929)
Fixes #1399.
Andrew Reynolds [Fri, 11 Oct 2019 19:00:56 +0000 (14:00 -0500)]
Check that logic is set when synth-fun command is encountered (#3384)
Andres Noetzli [Fri, 11 Oct 2019 08:09:09 +0000 (01:09 -0700)]
Add support for UBSan instrumentation (#3382)
This commit adds support for compiling CVC4 with UBSan instrumentation.
The commit also adds a dummy version of `AigBitblaster`. Previously,
when CVC4 was built without ABC, `AigBitblaster` was not fully defined
(the class was declared but the implementation was not being compiled).
This lead to missing RTTI information when compiling with UBSan
instrumentation.
Aina Niemetz [Fri, 11 Oct 2019 01:52:46 +0000 (18:52 -0700)]
Make order of theories explicit in the source code. (#3379)
Fixes #2517.
This makes the order of theories explicit in the source code rather than relying on the order defined via the build system. Previously, the build system ensured the order of the theories via the KINDS_FILES variable, which is a list of kinds files that is fed to code generation scripts (mkkind, mkmetakind, mkrewriter, mktheorytraits). The generated code critical to the order of theories w.r.t. soundess is the TheoryId enum, and the CVC4_FOR_EACH_THEORY macro. Ideally, we would want to get rid of the latter (ugly and error prone), which is not possible in the current configuration, and to be discussed in the future.
This PR moves the TheoryID enum and related functions to theory/theory_id.h, and the CVC4_FOR_EACH_THEORY macro to theory/theory_engine.cpp, the only place where it is used.
I ran it on whole SMT-LIB (non-incremental and incremental) and did not encounter any soundness issues. The only issue that did occur is not related to these changes, non-critical and known: #2993
Andrew Reynolds [Thu, 10 Oct 2019 19:47:03 +0000 (14:47 -0500)]
Warning instead of assertion for failing propagating instance (#3380)
Mathias Preiner [Thu, 10 Oct 2019 06:21:18 +0000 (23:21 -0700)]
test: Add TS_UTILS_EXPECT_ABORT macro for unit tests. (#3378)
TS_UTILS_EXPECT_ABORT can be used if an expression in a unit test is
expected to abort() instead of throwing an exception. This can happen if
CVC4_CHECK or CVC4_DCHECK fail.
Aina Niemetz [Wed, 9 Oct 2019 01:47:53 +0000 (18:47 -0700)]
NodeValue: Use 'using' instead of 'typedef'. (#3374)
Aina Niemetz [Wed, 9 Oct 2019 01:33:08 +0000 (18:33 -0700)]
NodeValue: Use fixed width return type for getRefCount(). (#3374)
Aina Niemetz [Tue, 8 Oct 2019 20:51:05 +0000 (13:51 -0700)]
Reorder NodeValue class according to our code style guidelines. (#3374)
Andres Noetzli [Wed, 9 Oct 2019 06:24:42 +0000 (23:24 -0700)]
Avoid printing success for `--force-logic` (#3363)
CVC4 was printing success when `--force-logic` was used because
internally, `--force-logic` generates a `SetBenchmarkLogicCommand`. This
caused issues with the SMT-COMP trace executor. This commit fixes the
behavior by muting the command if it was not issued by the user.
The issue was likely introduced with #3062.
Mathias Preiner [Wed, 9 Oct 2019 00:55:05 +0000 (17:55 -0700)]
cmake: Fix include of CVC4JavaTargets.cmake. (#3373)
Only include Java targets if Java bindings are enabled.
Aina Niemetz [Wed, 9 Oct 2019 00:24:49 +0000 (17:24 -0700)]
New C++ API: Term: Add missing checks for null. (#3364)
Co-Authored-By: Andres Noetzli <andres.noetzli@gmail.com>
Andrew Reynolds [Tue, 8 Oct 2019 23:06:16 +0000 (18:06 -0500)]
Limit cases of sygus inference based on type (#3370)
This makes `--sygus-inference` a no-op for inputs where there is a free function whose sort cannot be handled in a sygus grammar.
It also fixes an issue where skolem variables were not being treated as functions-to-synthesize.
Fixes #3250 and fixes #3356.
Andrew Reynolds [Tue, 8 Oct 2019 20:32:38 +0000 (15:32 -0500)]
Fix method for getting arithmetic function definition body (#3371)
Piotr Trojanek [Thu, 22 Aug 2019 21:22:51 +0000 (23:22 +0200)]
prefer prefix ++ operator for iterators
Detected with cppcheck static analyser, which said: (performance) Prefer
prefix ++/-- operators for non-primitive types. Reformat with clang-format
as needed.
Signed-off-by: Piotr Trojanek <piotr.trojanek@gmail.com>
Piotr Trojanek [Thu, 22 Aug 2019 21:12:20 +0000 (23:12 +0200)]
pass parameters by reference where it affects performance
Detected with cppcheck static analyser, which said: (performance) Function
parameter should be passed by reference. Reformat with clang-format as
needed.
Signed-off-by: Piotr Trojanek <piotr.trojanek@gmail.com>
Andres Noetzli [Tue, 8 Oct 2019 17:00:48 +0000 (10:00 -0700)]
[CVC Parser] Add support for regular expressions (#3346)
Andres Noetzli [Tue, 8 Oct 2019 16:05:50 +0000 (09:05 -0700)]
Disallow --proof and --incremental (#3332)
Ying Sheng [Tue, 8 Oct 2019 15:18:21 +0000 (08:18 -0700)]
Make ackermannization generally applicable rather than just BV (#3315)
The ackermannization process is currently already support general theories rather than specifically for BV. In this pull request, an option has been added to turn on ackermannization independently.
Aina Niemetz [Tue, 8 Oct 2019 06:27:44 +0000 (23:27 -0700)]
Build system: Add build type for incremental competition builds. (#3365)
Previously, competition builds for incremental tracks required to
manually pass in -DCVC4_SMTCOMP_APPLICATION_TRACK as compiler flag. This
introduces an additional build type for incremental competition builds
to simplify configuration for such builds.
Aina Niemetz [Tue, 8 Oct 2019 01:37:48 +0000 (18:37 -0700)]
New C++ API: Add Term::getId(). (#3360)
+ use explicit types in NodeValue
+ add unit test for Term::isParameterized()
Co-Authored-By: makaimann <makaim@stanford.edu>
Andres Noetzli [Tue, 8 Oct 2019 01:16:40 +0000 (18:16 -0700)]
[SMT2 Parser] Move code of `rewriterulesCommand` (#3334)
This commit moves the code in `rewriterulesCommand` in the SMT2 parser
to the `Smt2` class. Additionally, it creates a `boundVarList` rule to
reduce code duplication.
Andrew Reynolds [Sun, 6 Oct 2019 15:34:24 +0000 (10:34 -0500)]
Fix typo in regression (#3359)
Andrew Reynolds [Sun, 6 Oct 2019 07:33:25 +0000 (02:33 -0500)]
Fix str to int reduction (#3358)
This fixes a corner case of the str-to-int reduction for the case where the argument is the empty string.
This fixes #3357.
Andrew Reynolds [Fri, 4 Oct 2019 07:12:02 +0000 (02:12 -0500)]
Avoid duplicate lemmas in datatypes (#3310)
We previously were sending e.g. dt.size >= 0 lemmas when size terms are pre-registered, which can happen multiple times in a user context. This ensures we cache whether a lemma is sent in a user-context dependent way in the datatypes solver. This ensures we don't send the same lemma twice for dt.size >= 0 lemmas.