Discussion #9524 reported systematic TPTP result polarity errors
(notably `Unsatisfiable/Theorem` cases returned as `Satisfiable`) and
timeout-path instability in the frontend. The dominant issue was
semantic: free variables in FOF/CNF formulas were parsed as constants
instead of implicitly universally quantified variables.
- **Parser semantics: free variables now follow TPTP rules**
- In `/home/runner/work/z3/z3/src/cmd_context/tptp_frontend.cpp`, free
variables encountered while parsing an annotated formula are now
collected in formula scope and wrapped with a top-level `forall`.
- This applies to both term and atomic-formula paths, so variable
occurrences are treated consistently.
- Explicitly bound quantifier variables continue to take precedence over
implicit ones.
- **Scoped implementation cleanup**
- Added scoped state for implicit-variable collection to avoid leaking
parser state across formulas.
- Kept variable binding order stable so quantifier construction is
deterministic.
- **Timeout-path robustness**
- Updated frontend exception catches to `const&` in the TPTP stream
entrypoint to make timeout/error handling behavior consistent with
thrown exception forms.
- **Regression tests**
- Extended `/home/runner/work/z3/z3/src/test/tptp.cpp` with focused
cases for:
- FOF free-variable implicit universal quantification.
- CNF free-variable implicit universal quantification.
```tptp
cnf(c1,axiom, p(X)).
cnf(c2,axiom, ~ p(a)).
```
This now maps to `forall X. p(X)` plus `~p(a)`, yielding `% SZS status
Unsatisfiable` as expected.
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
This PR targets the main TFF frontend parsing failures: bare numeric
atoms were being treated as uninterpreted terms (`U`) in formula
position, decimal literals were not parsed, and `$uminus` was not
recognized as an arithmetic builtin. These issues caused widespread
parse/type failures in arithmetic-heavy TPTP inputs.
- **Numeric atom parsing in formulas (TFF)**
- Added a shared numeric-literal parser path for term/formula contexts.
- In atomic formulas, numeric LHS now parses as arithmetic numerals
before `=`/`!=` handling, avoiding `U` vs `Int/Real` mismatches.
- **Decimal literal support**
- Extended numeral parsing to accept `d.d` forms and construct `Real`
numerals.
- Keeps existing integer (`d`) and rational (`d/d`) behavior on the same
code path.
- **`$uminus` builtin support**
- Added explicit handling for `$uminus(<arg>)` in term parsing.
- Enforces arity 1 and arithmetic-argument checks; maps directly to
arithmetic unary minus.
- **Focused regression coverage**
- Added/updated TPTP unit cases for:
- bare integer inequality: `31 != 12`
- decimal arithmetic literal usage
- `$uminus` in arithmetic predicates
Example of now-supported inputs:
```tptp
tff(c1,conjecture, 31 != 12).
tff(c2,conjecture, ~ $less(-3.25,-8.69)).
tff(c3,conjecture, $less($uminus(2),0)).
```
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
* Initial plan
* Refactor find_probe() to use std::optional
- Updated tactic_manager.h: Changed return type to std::optional<probe_info*>
- Updated tactic_manager.cpp: Modified implementation to return std::nullopt or probe pointer
- Updated api_tactic.cpp: Changed 2 call sites to use optional checks and dereference
- Updated tactic_cmds.cpp: Changed 1 call site to use optional check and dereference
- Build verified successfully
- Probe functionality tested with Python bindings
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
* Initial plan
* Refactor mk_and and mk_or call sites to use overloaded methods
Changed 130 call sites across 64 files to use vector overloads directly instead of manually passing .size() and .data()/.c_ptr()
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
* Revert mk_or changes for ptr_buffer/ptr_vector (no overload exists in ast_util.h)
* Fix compilation errors from mk_and/mk_or refactoring
Fixed type mismatches by:
- Removing m parameter for expr_ref_vector (ast_util.h has mk_and/mk_or(expr_ref_vector) overloads)
- Reverting changes for ref_buffer types (no overload exists in ast_util.h, only in ast.h for m.mk_and)
- Verified build succeeds and Z3 works correctly
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
* Fix test files to use correct mk_and/mk_or overloads
Changed test/doc.cpp and test/udoc_relation.cpp to use mk_and(expr_ref_vector) and mk_or(expr_ref_vector) without m parameter
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
---------
Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
* somewhaat failed attempt at declaring subterm predicate
I can't really figure out how to link the smt parser to the rest of the
machinenery, so I will stop here and try from the other side. I'll start
implmenting the logic and see if it brings me back to the parser.
* initial logic implmentation
Very primitive, but I don't like have that much work uncommitted.
* parser implementation
* more theory
* Working base
* subterm reflexivity
* a few optimization
Skip adding obvious equalities or disequality
* removed some optimisations
* better handling of backtracking
* stupid segfault
Add m_subterm to the trail
* Update src/smt/theory_datatype.h
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update src/ast/rewriter/datatype_rewriter.cpp
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update src/smt/theory_datatype.cpp
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update src/smt/theory_datatype.cpp
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* Update src/smt/theory_datatype.cpp
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
* review
* forgot to update `iterate_subterm`'s signature
* fix iterator segfault
* Remove duplicate include statement
Removed duplicate include of 'theory_datatype.h'.
* Replace 'optional' with 'std::option' in datatype_decl_plugin.h
* Add is_subterm_predicate matcher to datatype_decl_plugin
* Change std::option to std::optional for m_subterm
* Update pdecl.h
* Change has_subterm to use has_value method
* Update pdecl.cpp
---------
Co-authored-by: Nikolaj Bjorner <nbjorner@microsoft.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Add commands:
(prefer <formula>)
- will instruct case split queue to assign formula to true.
- prefer commands added within a scope are forgotten after leaving the scope.
(reset-preferences)
- resets asserted preferences. Has to be invoked at base level.
This provides functionality related to MathSAT and based on an ask by Tomáš Kolárik who is integrating the functionality with OpenSMT2
* Introduce X-macro-based trace tag definition
- Created trace_tags.def to centralize TRACE tag definitions
- Each tag includes a symbolic name and description
- Set up enum class TraceTag for type-safe usage in TRACE macros
* Add script to generate Markdown documentation from trace_tags.def
- Python script parses trace_tags.def and outputs trace_tags.md
* Refactor TRACE_NEW to prepend TraceTag and pass enum to is_trace_enabled
* trace: improve trace tag handling system with hierarchical tagging
- Introduce hierarchical tag-class structure: enabling a tag class activates all child tags
- Unify TRACE, STRACE, SCTRACE, and CTRACE under enum TraceTag
- Implement initial version of trace_tag.def using X(tag, tag_class, description)
(class names and descriptions to be refined in a future update)
* trace: replace all string-based TRACE tags with enum TraceTag
- Migrated all TRACE, STRACE, SCTRACE, and CTRACE macros to use enum TraceTag values instead of raw string literals
* trace : add cstring header
* trace : Add Markdown documentation generation from trace_tags.def via mk_api_doc.py
* trace : rename macro parameter 'class' to 'tag_class' and remove Unicode comment in trace_tags.h.
* trace : Add TODO comment for future implementation of tag_class activation
* trace : Disable code related to tag_class until implementation is ready (#7663).
* add prd
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
* missing text
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
* fix
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
* fix#7647
* fix#7647 - with respect to scope level
* initial stab at randomizer
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
* Create prd.yml
* missing text
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
* fix
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
* Update prd.yml
* allows setting simplifier factory independently of whether solver has been allocated.
Instances, such as #7484 can be solved faster by either having authors rewrite benchmarks or by using incremental pre-processing. You can add incremental pre-processing to the SMT solver by using the command
```
(set-simplifier (then simplify propagate-values solve-eqs elim-unconstrained simplify))
```
This command can be invoked any time prior to push or adding assertions.
The effect of the command is that it adds an incremental pre-processing pass to check-sat invocations that is potentially more powerful than the default pre-processing. The default pre-processing functionality is not touched mainly to avoid instability against functionality that is already built around its behavior.
* remove experiment from pr
---------
Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
