Andres Noetzli [Wed, 9 Jun 2021 19:53:31 +0000 (12:53 -0700)]
Reorder ITE rewrites (#6723)
Fixes #6717. Commit
11c1fba added some
new rewrites for ITE. Due to the new rewrites taking precedence over
existing rewrites, it could happen that some of the previous rewrites
did not apply anymore even though they would have further simplified the
ITE. In the example from the issue, (ite c c true) was rewritten to
(or (not T) T) instead of (ite T true true) and then true. The
commit fixes the issue by moving rewrites resulting in
conjunctions/disjunctions to the end.
Gereon Kremer [Wed, 9 Jun 2021 19:25:04 +0000 (21:25 +0200)]
Make squasing more robust (#6713)
This PR makes squashing olds commits in the docs-cleanup CI job more robust: it makes sure that the squash commit has a proper commit date and that we gracefully handle if there is nothing to squash.
Andrew Reynolds [Wed, 9 Jun 2021 18:52:13 +0000 (13:52 -0500)]
Integrate learned literal manager into preprocessing pass context (#6718)
Towards "learned rewrite" preprocessing pass / improvements to Int <-> BV.
Andrew Reynolds [Wed, 9 Jun 2021 17:50:31 +0000 (12:50 -0500)]
Fixes related to printing tuples, define-fun commands in smt2 (#6679)
This PR fixes some initial issues with printing datatypes in smt2.
It does not yet address further issues on printing define-fun as a result of dump=raw-benchmark, which requires deeper refactoring of the implementation of the API / SmtEngine.
Gereon Kremer [Wed, 9 Jun 2021 17:15:30 +0000 (19:15 +0200)]
Push complex check inside GetInstantiationsCommand (#6715)
This PR pushes a rather complex check from the CommandExecutor inside the GetInstantiationsCommand.
The aim is to only use the instFormatMode option in the library (command.cpp) but not the main driver (command_executor.cpp).
Andres Noetzli [Wed, 9 Jun 2021 16:40:10 +0000 (09:40 -0700)]
Update options for SMT-COMP (#6704)
This commit removes obsolete options for BV and strings logics, and
updates QF_NIA to spend more time on our best configuration.
Co-authored-by: Gereon Kremer nafur42@gmail.com
Co-authored-by: Mathias Preiner mathias.preiner@gmail.com
Ouyancheng [Wed, 9 Jun 2021 15:53:16 +0000 (08:53 -0700)]
[Optimization] support for push/pop (#6706)
Use CDList for optimization objectives so that optimization solver supports push and pop (just use SmtEngine's push/pop).
SmtEngine::resetAssertions will also clears the optimization objectives, so no need to have the reset objectives function.
Haniel Barbosa [Wed, 9 Jun 2021 14:34:22 +0000 (11:34 -0300)]
Removing spurious HO checks (#6707)
Gereon Kremer [Wed, 9 Jun 2021 14:03:04 +0000 (16:03 +0200)]
Require statistics for regression (#6714)
This PR makes a new regression explicitly require statistics.
Aina Niemetz [Wed, 9 Jun 2021 07:20:14 +0000 (00:20 -0700)]
docs: Migrate separation logic theory reference. (#6702)
This migrates page https://cvc4.github.io/separation-logic.
Andres Noetzli [Wed, 9 Jun 2021 07:09:26 +0000 (00:09 -0700)]
docs: Fix `Kind` description (#6712)
This commit changes the Kind description not to include C/C++
preprocessor statements and updates the kind of bit-vector addtion.
It also marks some of the information as internal to exclude it from the
public documentation.
Andrew Reynolds [Tue, 8 Jun 2021 23:47:40 +0000 (18:47 -0500)]
Fix for 2 dimensional dependent bounded quantifiers (#6710)
This updates 2-dim dependent bounded quantifiers to not map constants to terms when computing ranges, when the type of the variable is closed enumerable. This is require to fix an incorrect model (possible solution unsoundness) issue in the reduction of str.indexof_re.
Fixes the 1st, 4th and 5th benchmarks from #6653. Fixes #6635.
Andrew Reynolds [Tue, 8 Jun 2021 21:50:41 +0000 (16:50 -0500)]
Add learned literal manager utility (#6709)
This is a simple class to maintain a list of literals that we have learned that may be useful during preprocessing.
This is work towards a "learned rewrite" preprocessing pass / better support for Int/BV translations during preprocessing.
Gereon Kremer [Tue, 8 Jun 2021 15:21:45 +0000 (17:21 +0200)]
Fix statistics option handler (#6703)
This PR fixes a typo in the option handler for the statistics options, which lead to options not properly propagating.
Gereon Kremer [Tue, 8 Jun 2021 13:43:13 +0000 (15:43 +0200)]
Make env hold a pointer to the original options to properly initialize subsolvers (#6705)
This PR extends the Env class to hold a pointer to the original options that are owned by the api::Solver. These original options will be used to properly initialize subsolvers as using the current options leads to subtle issues as setDefaults is not (in general) idempotent.
Gereon Kremer [Tue, 8 Jun 2021 07:05:35 +0000 (09:05 +0200)]
Remove `binary_name` option (#6693)
The binary_name is solely used as a temporary storage to pass the data from the options parser back to the runCvc5 method where it is put in a static variable. This PR gets rid of the option and the public option getter in favor of directly storing the program name in the static variable using an additional argument to parseOptions().
Alex Ozdemir [Tue, 8 Jun 2021 06:42:59 +0000 (23:42 -0700)]
Change output of getRealValue to a fraction. (#6692)
Our documentation for `toPythonObj` says that real values are represented as Fractions. However, getRealValue yields a float approximation thereof.
We should probably stick to Fractions, since they allow us to precisely capture values in LRA. Also, that's more aligned with the C++ API, which returns a string representation of the (unapproximated) Rational.
Also, document some (potential) weirdness with calling mkReal on floating-point values.
Andrew Reynolds [Tue, 8 Jun 2021 06:29:10 +0000 (01:29 -0500)]
Make TheoryUF compatible with central equality engine (#6695)
Work towards central equality engine.
This does some minor reorganization to TheoryUF, related to UF+cardinality constraints that makes it compatible with central equality engine. In particular, preNotifyFact is removed in favor of adding conflicts after cardinality constraints are added to the equality engine. This ensures that the central equality engine can explain conflicts that involve cardinality constraints (which will no longer be the responsibility of UF when --ee-mode=central).
Andrew Reynolds [Tue, 8 Jun 2021 06:11:28 +0000 (01:11 -0500)]
Fix str.update reduction (#6696)
Fixes the 2nd benchmark from #6653.
Andrew Reynolds [Mon, 7 Jun 2021 17:44:20 +0000 (12:44 -0500)]
(proof-new) Fix missing connection in trust substitution proofs (#6685)
This PR fixes a missing connection in trust substitution proofs, which was the cause of open proofs when solved equalities from ppAssert were not justified by proofs.
Also distinguishes TRUST_SUBS_EQ from TRUST_SUBS_MAP for clarity.
Gereon Kremer [Mon, 7 Jun 2021 17:00:10 +0000 (19:00 +0200)]
Remove `Options::wasSetByUser()` (#6682)
This PR removes the next heavily specialized template function Options::wasSetByUser() in favor of direct access to the *WasSetByUser flags.
Andrew Reynolds [Mon, 7 Jun 2021 14:09:59 +0000 (09:09 -0500)]
(proof-new) Lazy proof chain debug names (#6680)
Gereon Kremer [Sun, 6 Jun 2021 11:09:40 +0000 (13:09 +0200)]
Support public option modules (#6691)
This PR adds the possibility to make option modules public. As shown on the example of the main driver options, this allows to get rid of the wrappers from options_public.h. We plan to make only very few option modules public (i.e. main and parser).
Andres Noetzli [Sat, 5 Jun 2021 03:22:57 +0000 (20:22 -0700)]
Remove unwanted side effects in `SPLIT_EQ_STRIP_L` (#6689)
Fixes #6681. When checking whether SPLIT_EQ_STRIP_L applies, we were
using sripSymbolicLength, which changes its inputs depending on how
many elements of the concatenation it could strip. However, one of the
arguments, pfxv0, was reused across multiple checks if the strip did
not result in a rewrite. Later invocations were wrong as a result. This
commit changes the call to stripSymbolicLength() to use a new copy of
the vector instead.
Andrew Reynolds [Fri, 4 Jun 2021 21:05:43 +0000 (16:05 -0500)]
Miscellaneous changes from central ee branch (#6687)
Minor reorganization to make calls to theory engine from combination engine / shared solver more organized.
yoni206 [Fri, 4 Jun 2021 17:01:45 +0000 (10:01 -0700)]
pow2: header file for pow2 solver (#6676)
This PR adds a header file for the pow2 solver. It also includes an empty test file, to trigger compilation of the header file. The next PR will include implementations and tests.
Mathias Preiner [Fri, 4 Jun 2021 13:30:35 +0000 (06:30 -0700)]
bv: Enable bitblast solver by default. (#6660)
This commit enables the new bitblast solver by default. This commit also fixes model generation for Boolean variables when --bitblast=eager is enabled.
Fixes #3958, #5396, #5736, #5743, #5947.
Gereon Kremer [Fri, 4 Jun 2021 13:17:57 +0000 (15:17 +0200)]
Add missing dereference (#6684)
Andres Noetzli [Fri, 4 Jun 2021 00:41:40 +0000 (17:41 -0700)]
Fix handling of start index in `str.indexof_re` (#6674)
Fixes #6636, fixes #6637. When the start index was non-zero, the result of
str.indexof_re was not properly restricted to be greater or equal to
the start index. This commit fixes the issue by making the eager
reduction lemma more precise. Additionally, the commit fixes an issue
with the lower bound of the length of the match in str.indexof_re.
Gereon Kremer [Fri, 4 Jun 2021 00:14:27 +0000 (02:14 +0200)]
Some cleanup in `mkoptions.py` (#6667)
This PR does some general cleanup in the mkoptions.py script: removal of obsolete code and some fixes to the comments.
Aina Niemetz [Fri, 4 Jun 2021 00:07:05 +0000 (17:07 -0700)]
docs: Migrate datatypes theory reference. (#6662)
This migrates https://cvc4.github.io/datatypes.
yoni206 [Thu, 3 Jun 2021 23:57:29 +0000 (16:57 -0700)]
Adding unit tests for the datatypes python API (#6658)
This commit adds unit tests that are translated from `datatype_api_black.cpp`. One API function is also added to the python API.
This is the last part of the python api unit tests for datatypes.
Co-authored-by: Andres Noetzli <andres.noetzli@gmail.com>
Andrew Reynolds [Thu, 3 Jun 2021 22:49:32 +0000 (17:49 -0500)]
Simplify automatic set-logic in smt2 parser (#6678)
This simplifies the smt2 parser so that automatically setting the logic is done directly instead of being buffered as a command.
This prevents spurious errors for features that require (A) checking the logic is set and (B) fully intialize the underlying SMT engine. `declare-pool` is an example of an smt2 command where the user will get an error (instead of a warning) when set-logic is not used due to setting the logic, after fully initing SMT engine and then executing the buffered set-logic command.
Note this should also make dump=raw-benchmark more accurate (no set-logic is included when dumping benchmarks with no set-logic command).
Andres Noetzli [Thu, 3 Jun 2021 22:26:52 +0000 (15:26 -0700)]
[GitHub Actions] Make caching of dependencies depend on image version (#6677)
Our Clang builds started to fail to link when the environment changed. This commit changes our CI to only use cached dependencies if the build environment has not changed.
yoni206 [Thu, 3 Jun 2021 14:39:55 +0000 (07:39 -0700)]
Renaming pow2 to p2 in regression tests (#6675)
We plan to add a unary pow2 operator to cvc5, that is obtained from the binary operator pow by fixing the first argument to 2.
An initial working branch is here: https://github.com/yoni206/cvc5/tree/pow2
This PR does the first step, which is to rename some uninterpreted symbols in regression tests from pow2 to p2, to avoid clashing with the new operator.
Andres Noetzli [Wed, 2 Jun 2021 21:37:48 +0000 (14:37 -0700)]
Remove references to `bv-div-zero-const` in docs (#6672)
The bv-div-zero-const option has been removed and is now always
enabled, so this commit updates the documentation of the kinds to
reflect that. It also makes the language to describe the special cases a
bit more uniform.
Andrew Reynolds [Wed, 2 Jun 2021 21:16:40 +0000 (16:16 -0500)]
Fixes for printing define-fun-rec (#6673)
Andres Noetzli [Wed, 2 Jun 2021 19:41:14 +0000 (12:41 -0700)]
Remove option to ignore negative memberships (#6665)
Fixes #6661. The option `--strings-inm` could be used to ignore negative
membership constraints. However, this option made the string solver
model-unsound or produced incorrect models if the user provided a
benchmark that actually contained negative membership constraints. The
solver did not check for negative membership constraints and did not
warn the user about them. Because the option is not really being used,
this commit removes it.
yoni206 [Wed, 2 Jun 2021 19:30:46 +0000 (12:30 -0700)]
Adding getters to the python API and testing them (#6652)
This PR adds missing API functions from the cpp Term API to the python API.
Corresponding tests are translated from term_black.cpp.
Aina Niemetz [Wed, 2 Jun 2021 17:16:26 +0000 (10:16 -0700)]
Remove redundant logic ALL_SUPPORTED. (#6664)
yoni206 [Wed, 2 Jun 2021 16:47:09 +0000 (09:47 -0700)]
Move `toPythonObj` tests to the new API unit test directory (#6656)
This is the last test file that we move from the old directory to the new one, and so the old directory is deleted.
Gereon Kremer [Wed, 2 Jun 2021 16:01:08 +0000 (18:01 +0200)]
Use proper variable name (#6670)
This PR fixes the driver which used an incorrect variable name in competition mode.
Andrew Reynolds [Wed, 2 Jun 2021 15:28:36 +0000 (10:28 -0500)]
Fix unsat core proofs (#6655)
Fixes cases of satisfiable unsat cores when proofs are enabled.
Unfortunately, this bug was also preventing us from doing the final scope check for all proof checking. As this was not being tested, this PR uncovers that proof checking is now failing on 6 regressions. I'm disabling proof checking here and will address these issues on later PRs.
Andres Noetzli [Wed, 2 Jun 2021 13:50:51 +0000 (06:50 -0700)]
Make `STRINGS_CTN_DECOMPOSE` an explicit conflict (#6663)
Fixes #6643. The STRINGS_CTN_DECOMPOSE inference is always a conflict
but we sometimes sent it as an inference. To make sure that the
inference manager actually recognizes the inference as a conflict, this
commit ensures that the conclusion is always false and modifies the
explanation accordingly.
Gereon Kremer [Wed, 2 Jun 2021 13:21:22 +0000 (15:21 +0200)]
Remove `Options::operator[]` (#6649)
This PR removes the next heavily specialized template function Options::operator[] in favor of direct access to the option data.
Gereon Kremer [Wed, 2 Jun 2021 12:11:05 +0000 (14:11 +0200)]
Move public wrapper functions out of options class (#6600)
This PR moves options wrapper functions out of the Options class. These wrapper functions are meant to be called by "external" code that should not access the options modules. This PR thereby significantly reduces the interface of the Options class.
Gereon Kremer [Wed, 2 Jun 2021 11:55:40 +0000 (13:55 +0200)]
Fix issues with double negation in circuit propagator (#6669)
This PR fixes a subtle issue with double negations when producing proofs in the circuit propagator.
Adds the test case as a new regression, as well as some similar instances.
Fixes cvc5/cvc5-projects#277.
Gereon Kremer [Wed, 2 Jun 2021 10:11:13 +0000 (12:11 +0200)]
Fix issues when poly is disabled (#6668)
Recent changes introduced issues when libpoly is disabled.
Gereon Kremer [Wed, 2 Jun 2021 06:55:24 +0000 (08:55 +0200)]
Make `Options::assign()` specializations free functions (#6648)
This PR removes the next two heavily specialized template functions. Both Options::assign() and Options::assignBool() are only used within options.cpp now and there is thus no reason to keep them in the public interface. Furthermore, we can just make them properly named functions instead of template functions.
Gereon Kremer [Wed, 2 Jun 2021 06:29:02 +0000 (08:29 +0200)]
Do manual squash cleanup for docs (#6646)
This PR abandons the attempt to do the cleanup in a single rebase command, and instead squashes the old commits manually. The current solution does not handle conflicts properly. The new approach (which seems to be more robust) proceeds as follows (to squash $first..$last):
- checkout $last
- soft reset to $first (checkout $first, but keep changes in working copy)
- commit to squashed commit
- cherry-pick $last..HEAD from main branch
Aina Niemetz [Wed, 2 Jun 2021 00:21:56 +0000 (17:21 -0700)]
docs: Migrate input languages page. (#6659)
This migrates page https://cvc4.github.io/input-languages.
Aina Niemetz [Wed, 2 Jun 2021 00:05:52 +0000 (17:05 -0700)]
docs: Restructure index page, fix style issue. (#6657)
yoni206 [Tue, 1 Jun 2021 21:24:43 +0000 (14:24 -0700)]
Some additions to the datatypes python API (#6640)
This commit makes the following additions, in order to sync the python API with the cpp API.
1. adding `getName` functions to datatypes related classes
2. allowing `mkDatatypeSorts` with 1 or 2 arguments (previously allowed only 2).
3. In case there is a second argument to `mkDatatypeSorts`, we make sure it is a set.
4. Corresponding changes to the tests.
Andrew Reynolds [Tue, 1 Jun 2021 18:58:49 +0000 (13:58 -0500)]
Use top-level substitutions in ITE simp (#6651)
With this PR, we use the central top-level substitutions instance in the ITE simplification preprocessing pass. Previously the ITE simplification pass maintained its own copy of the substitutions.
Since the top-level substitutions now are the authority on models, this led to model failures after this change:
ac8cf53#diff-30677c5a1752b1d0e83ef25fd2cfb8949576ea42cf7821fe0ac00ebbd0122f8aL276.
This PR corrects the issue.
This is required for SMT-COMP.
Andrew Reynolds [Tue, 1 Jun 2021 14:59:41 +0000 (09:59 -0500)]
Disable timeout regressions (#6650)
Disables two regressions that have been timing out causing nightlies to fail.
yoni206 [Tue, 1 Jun 2021 09:40:54 +0000 (02:40 -0700)]
FP value support in python API (#6644)
This PR adds new is* functions from the cpp API to the python API.
In particular, it adds getFloatingPointValue() function from the cpp API.
A test (translated from term_black.cpp) is added.
getFloatingPointValue() returns a tuple, and so this requires importing an instance of tuples into cython.
Gereon Kremer [Mon, 31 May 2021 14:05:39 +0000 (16:05 +0200)]
Remove Options::ref() (#6647)
This PR gets rid of the templated Options::ref() method (and all its specializations for every option).
Andres Noetzli [Mon, 31 May 2021 04:13:37 +0000 (21:13 -0700)]
Remove invalid options from run scripts (#6645)
This commit removes some of the options in the run scripts that are not
supported anymore: `--bv-div-zero-const` and `--rewrite-divk`. Both of
those options are effectively enabled by default in cvc5.
yoni206 [Mon, 31 May 2021 00:56:14 +0000 (17:56 -0700)]
Update `toPythonObj` to use new getters -- part 1 (#6623)
Following #6496 , this PR adds new getters to the python API, as well as tests for them. This makes toPythonObj simpler.
A future PR will add more getters to the python API.
Co-authored-by: Gereon Kremer nafur42@gmail.com
Andres Noetzli [Mon, 31 May 2021 00:35:08 +0000 (17:35 -0700)]
Compute model values for nested sequences in order (#6631)
Fixes #6337 (the other benchmarks in this issue are either solved
correctly or time out after the changes in #6615) and fixes #5665.
While computing the model for a nested equivalence class containing
seq.unit, we were looking up the representative of the argument in
(seq.unit (seq.unit j)) and the representative was simpliy (seq.unit j). However, we had assigned (seq.unit 0) to (seq.unit j) earlier.
A second equivalence class of type (Seq (Seq Int)) and length 1 was
later assigned (seq.unit (seq.unit 0)) and we didn't detect that
(seq.unit (seq.unit j)) and (seq.unit (seq.unit 0)) have the same
value. This was incorrect because we do not allow assigning the same
value to different equivalence classes. In this case, it led to one of
the assertions being false.
This commit fixes the issues in two ways: it ensures that types are
processed in ascending order of nesting (e.g., (Seq Int) terms are
processed before (Seq (Seq Int)) terms) and it changes the procedure
to look up the representative in the model instead of the theory state
to take into account the model values assigned to the elements of
sequences.
cc @yoni206
Gereon Kremer [Sat, 29 May 2021 07:09:34 +0000 (09:09 +0200)]
Remove `Options::set()` method (#6556)
This PR gets rid of the Options::set() method, replacing it by direct access to the options data.
This method was only used internally and did nothing except for resolving the options data from the option tag type via template specializations (via ref()), which is no longer necessary.
Ouyancheng [Fri, 28 May 2021 21:00:16 +0000 (14:00 -0700)]
(Optimization) remove popObjective, add resetObjectives, rename pushObjective => addObjective (#6634)
In order for OptimizationSolver to support pushing & popping,
we could remove popObjective because it might be difficult to handle cases like:
optSlv->pushObjective(...);
optSlv->push();
optSlv->popObjective();
optSlv->pop();
In this case we need to add back the popped objective...
If push/pop is supported, pop does not bring back objectives if you resetObjective
but it will revert the objs you add.
just like assertFormula and resetAssertions.
yoni206 [Fri, 28 May 2021 20:28:56 +0000 (13:28 -0700)]
Python API: bugfix + translating tests from cpp unit tests (#6559)
This PR fixes an issue in the python API for datatypes, and also introduces tests translated from https://github.com/cvc5/cvc5/blob/master/test/unit/api/datatype_api_black.cpp
The next PR will translate more tests and will also introduce missing functions in the python API for datatypes.
Gereon Kremer [Fri, 28 May 2021 19:46:54 +0000 (21:46 +0200)]
Add non-templated method to set option defaults (#6540)
This PR replaces the templated Options::setDefault() methods by new non-templated free functions options::{module}::setDefault{option}().
These methods should be used instead of the common if (!set by user) { set option value } pattern.
Andres Noetzli [Fri, 28 May 2021 14:10:15 +0000 (07:10 -0700)]
Disable `--jh-rlv-order` for slow regressions (#6633)
This commit adds --no-jh-rlv-order to two string regressions that take
over 2 minutes to run in debug after #6613, which increases the overall
regression runtime significantly.
Andres Noetzli [Fri, 28 May 2021 03:28:20 +0000 (20:28 -0700)]
`STRINGS_CTN_DECOMPOSE`: Avoid multiple conflicts (#6632)
Fixes #5508. `STRINGS_CTN_DECOMPOSE` could be triggered multiple times
by the same term, which resulted in an assertion failure. This commit
returns immediately after the first conflict to avoid the assertion
failure.
Andrew Reynolds [Thu, 27 May 2021 23:22:28 +0000 (18:22 -0500)]
Fix regular expression aggressive elim (#6627)
Fixes #6620, fixes #6622. Fixes cvc5/cvc5-projects#254.
The benchmarks from the 2 issues timeout, a regression is added for the projects issue.
Andres Noetzli [Thu, 27 May 2021 22:42:10 +0000 (15:42 -0700)]
Fix `str.replace_re` and `str.replace_re_all` (#6615)
Fixes #6057. The reductions of `str.replace_re` and `str.replace_re_all`
were not correctly enforcing that the operations replace the _first_
occurrence of some regular expression in a string. This commit fixes the
issue by introducing a new operator `str.indexof_re(s, r, n)`, which,
analoguously to `str.indexof`, returns the index of the first match of
the regular expression `r` in `s`. The commit adds basic rewrites for
evaluating the operator as well as its reduction. Additionally, it
converts the reductions of `str.replace_re` and `str.replace_re_all` to
use that new operator. This simplifies the reductions of the two
operators and ensures that the semantics are consistent between the two.
Ouyancheng [Thu, 27 May 2021 22:02:48 +0000 (15:02 -0700)]
Add Lexicographic + Pareto Optimizations (#6626)
Lexicographic optimizations: Optimize the objectives one-by-one, in the order they are pushed.
Pareto optimizations: Optimize the objectives to the extend that further optimization on any objective will worsen the other objective.
Units tests are of course added.
Lexicographic optimization is using iterative implementation, pretty similar to the naive box optimization.
Andrew Reynolds [Thu, 27 May 2021 21:28:58 +0000 (16:28 -0500)]
Update proof namespaces (#6614)
This removes namespace theory from proof utilities, and moves MethodId to its own file in src/proof/.
Andrew Reynolds [Thu, 27 May 2021 21:18:28 +0000 (16:18 -0500)]
Fix CEGQI for datatypes with Boolean subfields (#6630)
Fixes a solution soundness issue caused by allowing ineligible terms of kind BOOLEAN_TERM_VARIABLE to appear in instantiations.
This also corrects the expected solution on a benchmark that had an incorrect status.
Fixes #6603.
Andrew Reynolds [Thu, 27 May 2021 20:02:43 +0000 (15:02 -0500)]
Fix spurious assertion for trivial abduction (#6629)
Fixes 2nd benchmark from #6605.
Aina Niemetz [Thu, 27 May 2021 19:08:24 +0000 (12:08 -0700)]
FP: Rename FLOATINGPOINT_PLUS to FLOATINGPOINT_ADD. (#6628)
This is to make it consistent with the name of the SMT-LIB operator
(fp.add).
Andres Noetzli [Thu, 27 May 2021 17:11:57 +0000 (10:11 -0700)]
Return `REWRITE_AGAIN` after rewriting bvcomp (#6624)
This commit fixes an assertion failure in the rewriter on some of the
SMT-LIB QF_ABVFP benchmarks (the regression in this commit is the
minified version of
`non_incremental/QF_ABVFP/
20170428-Liew-KLEE/imperial_gsl_benchmarks_statistics_klee.x86_64/query.14.smt2`).
The problem was that after applying the `BvComp` rewrite, the bit-vector
rewriter was returning `REWRITE_DONE` instead of `REWRITE_AGAIN`. The
rewrite simplifies expressions of the form `bvcomp(t, c)` where `c` is a
constant of bit-width 1. If `c` is zero, then the rewrite returns
`bvnot(t)`. This node can potentially be rewritten further, e.g., if `t`
is `bvnot(x)`. This commit fixes the response and adds the corresponding
tests.
Ouyancheng [Thu, 27 May 2021 07:53:58 +0000 (00:53 -0700)]
Add support for Box optimization (#6599)
Add support for box optimization -- independently optimize each goal as if the other goals do not exist.
Single minimize() / maximize() now maintains the pushed / popped context.
Of course unit tests are here as well.
Gereon Kremer [Thu, 27 May 2021 07:38:12 +0000 (09:38 +0200)]
Avoid uploading docs if they did not change (#6621)
Fixes an oversight from #6601.
Andrew Reynolds [Thu, 27 May 2021 06:44:55 +0000 (01:44 -0500)]
Enable new justification heuristic by default (#6613)
This enables the new implementation of justification heuristic by default.
Fixes #5454, fixes #5785. Fixes wishues 114, 115, 149, 160.
Gereon Kremer [Wed, 26 May 2021 20:30:19 +0000 (22:30 +0200)]
Use references instead of getter functions (#6597)
This PR follows a suggestions of @ajreynol to use public references instead of getter functions to access the individual option modules.
Andres Noetzli [Wed, 26 May 2021 14:30:17 +0000 (07:30 -0700)]
More precise includes of `Node` constants (#6617)
We store constants, e.g., BitVector and Rational, in our node infrastructure. As a result, we were indirectly including some headers in almost all files, e.g., the GMP headers. This commit changes that by forward-declaring the classes for the constants. As a result, we have to include headers like util/rational.h explicitly when we use Rational but it saves about 3 minutes in compile time (CPU time).
The commit changes RoundingMode from an enum to an enum class such that it can be forward declared.
Gereon Kremer [Wed, 26 May 2021 06:30:45 +0000 (08:30 +0200)]
Add more examples to the documentation (#6569)
This PR adds (most of) the existing examples to the documentation, and does some other minor updates on the documentation. Some details:
- for consistency, all cpp examples are moved from examples/api to examples/api/cpp
- add capabilities for SMT-LIB examples, and two simple smt2 examples
- more docs/examples/*.rst files
- two new documentation categories: installation (how to obtain, compile and install cvc5) and binary (about the cvc5 binary)
Gereon Kremer [Wed, 26 May 2021 06:20:03 +0000 (08:20 +0200)]
Ensure proper types in unit tests (#6598)
This PR fixes a type mixup (the usual long vs long long mixup on macos systems) in the unit tests for the new api::Term getters.
Fixes #6594.
Gereon Kremer [Wed, 26 May 2021 06:10:51 +0000 (08:10 +0200)]
Reduce size of sphinx-gh output (#6601)
This PR reduces the disk size of the docs generated by make sphinx-gh.
Apart from reformatting the cmake source, we now not only remove the _sources folder, but also .doctrees (essentially the sphinx cache) and _static/fonts/ (the fonts that are actually used live in _static/css/fonts).
In combination, this now reduces the disk size from about 20MB (sphinx) to less than 6MB (sphinx-gh).
Furthermore this PR only uploads the generated documentation if it differs from whatever we currently have for master.
This is relevant to make the docs-ci repository smaller (which already has more than 5GB...)
Aina Niemetz [Tue, 25 May 2021 21:05:40 +0000 (14:05 -0700)]
api docs: Fix and tweak style for home and top links. (#6618)
Andres Noetzli [Tue, 25 May 2021 13:51:56 +0000 (06:51 -0700)]
Replace deprecated calls to `std::allocator` (#6606)
Fixes #6453. This commit replaces all calls to std::allocator with calls to
std::allocator_traits. Strictly speaking, allocate() and deallocate() are
not deprecated but the commit replaces them for uniformity.
Andres Noetzli [Tue, 25 May 2021 03:21:16 +0000 (20:21 -0700)]
[Unit tests] Fix path of Java bindings (#6616)
Currently, when configuring cvc5 with Java bindings, CMake complains
about `get_filename_component(CVC5_JNI_PATH ${CVC5_JAR_PATH} DIRECTORY)`
not using the correct number of arguments in the Java unit tests. The
issue is that `${CVC5_JAR_PATH}` is empty. The value of
`${CVC5_JAR_PATH}` was computed in the Java API bindings but then not
shared with the rest of the build system. Because `${CVC5_JAR_PATH}` is
not used anywhere else, this commit moves the computation of
`${CVC5_JAR_PATH}` to the unit tests. The commit also ensures that the
API subdirectories are processed before the test subdirectories.
Andrew Reynolds [Mon, 24 May 2021 23:59:28 +0000 (18:59 -0500)]
Fix non-fixed length case in re-elim (#6612)
Fixes followup issues from #6604.
Andrew Reynolds [Mon, 24 May 2021 19:32:48 +0000 (14:32 -0500)]
Implementation of the new justification heuristic (#6465)
This adds the new implementation of the justification heuristic. It does not enable this strategy yet.
A followup PR will activate this strategy within DecisionEngine.
Andrew Reynolds [Mon, 24 May 2021 18:51:09 +0000 (13:51 -0500)]
Move proof utilities to src/proof/ (#6611)
This moves all generic proof utilites from src/expr/ and src/theory/ to src/proof/.
It also changes the include for term conversion proof generator to conv_proof_generator in preparation to rename this utility on a followup PR (to avoid confusion with the use of "Term").
Andrew Reynolds [Mon, 24 May 2021 16:15:12 +0000 (11:15 -0500)]
Fix re-elim length requirement for symbolic RE memberships (#6609)
Fixes #6604.
Previously, re-elim was solution unsound for cases where the LHS and a component of the RHS were both empty. This ensures a length requirement is given for the LHS to ensure proper containment.
Andrew Reynolds [Mon, 24 May 2021 15:55:17 +0000 (10:55 -0500)]
Fix instance of no rewrite in extended rewriter (#6610)
Fixes #6545.
An assertion failure was being raised indicating that we were reporting a rewrite that was not changing the original term.
Andrew Reynolds [Mon, 24 May 2021 14:39:56 +0000 (09:39 -0500)]
Better formalization of regular expression unfolding skolems (#6602)
This replaces our previous formalization of RE unfolding skolems with a more explicit one that is amenable to external proof conversion. It adds a few associated utility methods to SkolemManager required for LFSC proof conversion for RE_UNFOLD_POS.
It also changes the order of equalities in the RE_UNFOLD_POS rule, which simplifies LFSC proof checking.
Andrew Reynolds [Fri, 21 May 2021 21:46:48 +0000 (16:46 -0500)]
Fix tests of unsat cores (#6593)
This updates all regressions that pass check-unsat-cores to enable check-unsat-cores. This includes any incremental benchmark, which was disabled in run_regression.py previously.
It adds --no-check-unsat-cores to a few corner benchmarks that were previously disabled based on --incremental.
It also reverts a change to when proofs are disabled: options like sygus-inference should not permit proofs (or unsat cores).
Andrew Reynolds [Fri, 21 May 2021 21:27:19 +0000 (16:27 -0500)]
Fix and refactor relevant domain (#6528)
In a handcrafted case, one can make the body of quantified formula another quantified formula when special attributes are used. The relevant domain utility was not robust to this case, leading to instantiations with free variables.
This fixes the issue and also updates its style to use a term context stack, which also avoids a tree traversal of the bodies of quantified formulas in this utility.
Fixes #6476. The benchmark from that issue now times out.
Andrew Reynolds [Fri, 21 May 2021 21:07:50 +0000 (16:07 -0500)]
Formalize shared selectors as skolem functions (#6591)
This is work towards properly printing shared selectors in external proofs.
Andrew Reynolds [Fri, 21 May 2021 20:48:39 +0000 (15:48 -0500)]
Minor simplification to boolean proof checker (#6590)
Andrew Reynolds [Fri, 21 May 2021 20:16:20 +0000 (15:16 -0500)]
(proof-new) Minor documentation sync (#6592)
Andrew Reynolds [Fri, 21 May 2021 18:04:49 +0000 (13:04 -0500)]
Add utility to get all types occurring in a term (#6588)
Required for external proof conversions.
Andrew Reynolds [Fri, 21 May 2021 17:15:28 +0000 (12:15 -0500)]
Update to sygus standard output for check-synth responses (#6521)
This PR does two things:
(1) It eliminates the ad-hoc implementation of printSynthSolutions and removes it from the API. Now, printing response to a check-synth command is done in a more standard way, using the API + symbol manager. This is analogous to recent refactoring to get-model.
(2) It updates cvc5's output in response to check-synth to be compliant with the upcoming sygus 2.1 standard. The standard has changed slightly: responses to check-synth are now closed in parentheses, mirroring the smt2 response to get-model.
It also removes the unused command GetSynthSolutionCommand.
Andres Noetzli [Fri, 21 May 2021 16:05:45 +0000 (09:05 -0700)]
Support braced-init-lists with `mkNode()` (#6580)
This commit adds support for braced-init-lists in calls to `mkNode()`,
e.g., `mkNode(REGEXP_EMPTY, {})`. Previously, such a call would result
in a node of kind `REGEXP_EMPTY` with a single null node as a child
because the compiler chose the `mkNode(Kind kind, TNode child1)` variant
and converted `{}` to a node using the default constructor. This commit
adds an overload of `mkNode()` that takes an `initializer_list<TNode>`
to allow this use case. It also adds a `mkNode()` overload with zero children
for convenience and removes the 4- and 5-children variants because they
saw little use. Finally, it makes the default constructor of `NodeTemplate`
explicit to avoid accidental conversions.
Aina Niemetz [Fri, 21 May 2021 10:19:10 +0000 (03:19 -0700)]
api docs: Tweak and fix style. (#6582)
This fixes the style for narrow layouts. It further fixes and improves
the style for the menu.