- Relax restore_x() to handle backup/current size mismatches: when
backup is shorter (new columns added), call
move_non_basic_columns_to_bounds() to find a feasible solution.
- Fix 100x performance regression in nonlinear optimization: save LP
optimum before check_nla and return it as bound regardless of NLA
result, so opt_solver::check_bound() can validate via full re-solve
with accumulated NLA lemmas.
- Refactor theory_lra::maximize() into three helpers: max_with_lp(),
max_with_nl(), and max_result().
- Add mk_gt(theory_var, impq const&) overload for building blockers
from saved LP optimum values.
- Add BNH multi-objective optimization test (7/7 sat in <1s vs 1/7
in 30s before fix).
- Add restore_x test for backup size mismatch handling.
Fixes#8890
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
The NRA solver's check() uses cone-of-influence (COI) to select a subset
of constraints for nlsat. When nlsat returns l_true, the model is validated
against all constraints, but non-COI constraints can legitimately be
violated since nlsat only solved over the COI subset.
- Non-COI violations gracefully return l_undef (fallback to other strategies)
- COI violations still trigger an assertion (indicating a real nlsat bug)
Fixes#8883
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
Add checkpoint() call in gcd_prs() main loop so polynomial GCD
computation respects rlimit/timeout. Add cancellation callback to
HNF calculation so it can be interrupted when the solver is cancelled.
Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>
- Remove them all from m_fresh_k2xt_terms and m_row2fresh_defs
- Mark rows containing those vars in m_changed_rows for recalculation
- Move remove_irrelevant_fresh_defs() before the recalculate loop so all affected rows get recalculated
Signed-off-by: Lev Nachmanson <levnach@hotmail.com>
* Fix implicit conversion warnings: use UINT_MAX instead of -1 for unsigned variables
Replace implicit conversion from negative literal to unsigned type
with explicit UINT_MAX constant to eliminate compiler warnings.
Fixed 10 instances across 6 files:
- src/ast/rewriter/bv_rewriter.cpp: 1 instance
- src/ast/sls/sls_bv_tracker.h: 2 instances
- src/math/lp/dioph_eq.cpp: 3 instances
- src/math/lp/lp_primal_core_solver.h: 2 instances
- src/muz/transforms/dl_mk_array_instantiation.cpp: 1 instance
- src/muz/transforms/dl_mk_synchronize.cpp: 1 instance
These changes preserve the exact same runtime behavior (UINT_MAX
equals the wrapped value of -1 for unsigned types) while making
the code more explicit and warning-free.
* Update bv_rewriter.cpp
---------
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
The undo_fixed_column struct is region-allocated via trail_stack, so its
destructor is never called. When m_fixed_val contains a big number (one
that doesn't fit in a small int), the heap-allocated memory for the mpq
numerator/denominator was never freed.
Fix by calling m_fixed_val.reset() in undo() to explicitly free the
heap memory before the region deallocates the struct.
Verified with macOS leaks tool:
- Before: 126 leaks (6048 bytes) on convert-jpg2gif-query-1584.smt2
- After: 0 leaks on normal completion, 10 leaks on timeout (unfinished trail)
* Initial plan
* Add try_get_value for std::map and use it in var_register.h and dioph_eq.cpp
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
* Add try_get_value overload for unordered_map with custom hash and use in lar_solver.cpp
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
* Remove redundant try_get_value template overload
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
* Remove std::map include and try_get_value overload from lp_utils.h
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
* Update z3 codebase to use std::string_view (except z3++.h)
- Updated params.cpp/h to use string_view internally for parameter descriptions
- Updated trace.h/cpp to accept string_view for trace tag functions
- Updated hash.h/cpp to use string_view for string_hash function
- Updated all callers of string_hash to use string_view
- Properly handled nullptr to empty string_view conversions
- All tests passing
Co-authored-by: NikolajBjorner <3085284+NikolajBjorner@users.noreply.github.com>
* Add missing string_view includes to headers
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>
Issue #7502 shows that running nlsat eagerly during final check can block quantifier instantiation.
To give space for quantifier instances we introduce two levels for final check such that nlsat is only applied in the second and final level.
