3
0
Fork 0
mirror of https://github.com/Z3Prover/z3 synced 2026-05-16 23:25:36 +00:00
z3/src/smt/smt_context.h
Ilana Shapiro 19e95f40af
Add global backbones to parallel architecture (#9343)
* setting up new backbone experiment

* fix phase scores bug

* debug crash from negated atoms

* backbone thread/global backbones in progress, does NOT compile yet

* debug, still need to add backbones worker as a new thread

* setting up complicated condition variable thing for backbones worker thread

* debug

* debug lock contention

* it's a little messy, but change how i'm checking backbones by initiating with batch check

* don't split on global backbones, share global backbones once detected. still need to prune search tree with backbones

* close global backbone branches in search tree

* fix backbone ranking (take average of bb age over cubes and incorporate hits/num cubes the bb appears in

* add stats to backbone experiment

* gate the backbones experiment by local vs global

* update stats and fix bug about unsat core size=1 means global backbone

* phase negation ablation

* unforce phase ablation

* reset ablations

* add todo notes

* fix backbone aging

* first draft of Janota Alg 7

* process exactly 10 bb candidates in each batch

* fixing the Janota algorithm

* add backbone stats for Janota algorithm

* fix bug about global backbones not being checked unless local is also true

* hopefully fix bug about closing global backbones in search tree

* fix another bug in janota alg

* report random seed for debug

* print random seed for debug

* refactor janota alg code, still can't repro the crash

* fix some bugs in the janota algorithm

* try to fix weird memory leak thing with ramon/linux

* revert fix, it didn't work

* add second backbones thread

* increase chunk size when undef

* fix how the 2 backbone threads work on batches (they each race to finish the same batch). this was very complicated to code due to thread synchronization and while it runs there may be bugs

* update how we report stats for backbones

* first draft of doing the bb threads in neg and pos mode, needs revising

* fix some bugs in the positive version of the bb check, still need to review

* debug some more things in the positive bb worker

* keep bb candidates sorted, increase batch and chunk size

* try to resolve a couple of bugs

* fix very bad bug about backbones workers not doing anything

* ablate positive backbone thread

* fix how we record backbones in positive mode (shouldn't impact previous run)

* clarify code about adding found backbones

* add back the positive bb thread

* try to fix the random segfault bug + ablate the postiive bb thread again

* clean up logs

* share clauses with bb threads

* fixes

Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>

* resolve deadlock

* add comment about SAT bb case

* todo comments

* complete TODOs in code, still need to debug bb threads

* debug bb threads, add bb_positive thread back in

* ablate bb_positive thread

* style

* configure num bb threads as param

* enable sat and unsat mode

Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>

* updates

Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>

* remove while true

Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>

* updates

* try to fix rewriter_exception bug

* possibly reduce code under lock when only 1 bb thread

* add some copilot-suggested optimizations

* add copilot suggestions to fix condition variable synchronization with bb threads

* revert changes that are too messy with the code

* ablate collect clauses

* ablate condition variable logic changes

* ablate reset batch

* revert ablation

* remove m_batch_in_progress that makes the bb threads wait until both have exited the batch after one signals cancel (can be long if one is stuck in ctx check)

* sharing theory lemmas

* finish setup for search tree thread modes, and fix local bb setup to pull from the global pool

* variable renames

* update bb hyperparams after copilot (hopefully??) ran tuning experiments

* fix possible AST manager bug

* ablate collect clauses

* remove bb collect shared clauses

* fix local bb experiment bug and reinstate collect clauses for global bb

* local bb cands are thread-local ablation

* remove thread-local local bb ablation

* fix bug in nonthread-local bb experiment

* fix more nonthreadlocal bb bugs

* try to fix local bb bug

* AST manager mismatch bugfix

* attempt to fix another canonicalization bug

* try another bugfix

* try another bugfix

* try yet another bugfix

* thread local bb ablation

* ablate force phase

* ablate set activity

* undo ablations since apparently it's not forcing phase or boosting activities

* remove old experiments

* try guarding m_birthdate size

* try to fix several bugs including with m_birthdate initialization and how we're storing original phases

* one more bugfix

* remove local bb experiment after negative signals on experiments, and change bb ranking to VSIDS scores as opposed to phase

* select bb polarity based on phase, not VSIDS

* first attempt with codex. Codex notes:
What changed:

  - Each tree node now tracks:
      - active worker count
      - lease epoch
      - cancel epoch
  - get_cube() now hands each worker an explicit lease: (node, epoch, cancel_epoch).
  - try_split() and backtrack() now operate on that lease, and the batch manager releases the worker’s lease under the tree lock before mutating the
    node.
  - If another worker closes the leased node or subtree, the batch manager cancels only the workers whose current leased nodes are now closed.
  - Workers detect canceled leases after check(), reset their local cancel flag, abandon the stale lease, and continue instead of turning that into a
    global exception.
  - The “reopen immediately into the open queue” policy is preserved. I did not add a barrier waiting for all workers on a node to finish.
  - Active-worker accounting is now separate from the open/active/closed scheduling status, so reopening a node no longer erases the fact that other
    workers are still on it.

  I also updated search_tree bookkeeping so:

  - closure bumps node cancel/lease epochs
  - active-node counting uses actual active-worker presence, not just status == active

* fix smts bugfix git merge issues with backtrack

* fix(parallel-smt): gate split/backtrack by lease epoch

What it changes:

  - util/search_tree.h
      - bumps node epoch on split
      - threads epoch through should_split(...) and try_split(...)
      - always records effort, but only split/reopen if the lease epoch still matches
  - smt/smt_parallel.cpp
      - requires is_lease_valid(..., lease.epoch) before backtrack(...)
      - passes lease.epoch into m_search_tree.try_split(...)

* clean up code and add some comments

* fix bug about backtracking condition being too strict: The epoch guard should not block backtrack(...) the same way it blocks try_split(...). A stale worker that proves UNSAT for n should still be able to
  close n, and that closure should then cancel the other workers on n and its subtree.

  I changed smt/smt_parallel.cpp accordingly:

  - try_split(...) still uses epoch to reject stale structural splits
  - backtrack(...) no longer requires is_lease_valid(..., epoch); it only requires that the lease is not already canceled

  So the intended asymmetry is now restored:

  - stale split: reject
  - stale unsat/backtrack: allow closure, then cancel affected workers

* ablate to no backtracking on stale leases

* fix merge

* revert codex change about exception handling

* fix linux bugs

* ablate backtrack gating

* attempt to fix linux crashes

* ablate backtracking on global bb

* the rare bb bug appears to be from creating the synthetic lease for a bb node and then backtracking on the synthetic lease. this is an attempt to fix it

* clean up code

* try to fix bug about active worker counts/lease accounting. current policy should hold: - stale leases: release/decrement
  - canceled leases: do not release/decrement (just ignore since we have an invariant that canceled leases mean closed nodes that are never revisited

* delay premature root activation

* fix major semantic bug about threads continually choosing the root if their lease is reset

* fix cancellation to unknown status

* fix very bad bug about all threads needing to start at the root

* ablate active ranking: now nodes are only reopened if they are truly inactive (active worker count is 0)

* fix some bugs about leases

* ablate adding static effort only

* fix some bugs about leases

* don't explode effort for portfolio nodes

* fix: still accumulate per-node effort, but don't over-accumulate on portfolio solves

* restore dynamically scaled effort

* clean up merge from cherry pick

* tighten which nodes we detect for proven global bb closure (only detect nonclosed nodes)

* fix cancel to unknown exception on bb code

* lease cancellation doens't touch rlimit now, it just sets max conflicts to 0. also fix a VERY BAD BUG about effort never being updated until all leases are done on a node, which meant we never left the root

* cross-thread modification of max conflicts is unsafe, so create an atomic lease canceled variable that's ch
ecked in ctx where max conflicts is also checked

* move atomic lease check in the context to the more global get_cancel_flag function

* Fix new SIGSEV. issue: The root cause: get_cancel_flag() is called from within propagation loops (mid-BCP, mid-equality-propagation, mid-atom-propagation). When it returns true there, the solver exits early and leaves the context in an intermediate state —
  propagation queues partially processed, theory state potentially inconsistent with boolean state.

  For the global cancel (m.limit().cancel()), this is harmless: the worker exits entirely and the context is destroyed. Intermediate state doesn't matter.

  For a lease cancel, the context is reused — the worker gets a new cube and calls ctx->check() again on the same context object. Re-entering check() on a context interrupted mid-propagation causes it to access that corrupted intermediate
  state → SIGSEGV.

  The m_max_conflicts check is the only checkpoint that's safe for re-entry: it only fires post-conflict-resolution, pre-decision, when propagation queues are empty and theory state is consistent.

  Fix: Remove m_lease_canceled from get_cancel_flag(). Keep it only at safe, between-phase checkpoints where the context is in a known-consistent state. The result is two safe checkpoints for m_lease_canceled: after each conflict (post-resolution, queues empty) and before each theory final check (not yet entered the theory). Neither interrupts the solver mid-mutation. The SIGSEGV should be
   gone, and NIA performance should improve because long theory final checks (where NIA burns most time) are now preemptable before they start.

* fix new inconsistent theory bug: The problem is returning FC_GIVEUP from inside final_check() after some theories have already run final_check_eh() and pushed propagations into the queue. Those pending propagations reference context state that gets invalidated on the next check() call → SIGSEGV. The fix: check m_lease_canceled before entering final_check() in bounded_search(), never from inside it. That way the context is always in a clean pre-final-check state when we bail out. This is safe: decide() returned false (all variables assigned, no pending propagations), theories haven't been touched yet, context is in a fully consistent state. For NIA, this is still a meaningful win — we avoid entering expensive arithmetic final checks entirely when the lease is already canceled.

* ablate lease cancel check in ctx final theory check due to crash (??)

* gate bb-specific code behind param

* try some possible bugfixes for the sigsev

* ablate some bugfixes

* remove second lease cancel check in smt_context, not sure it's safe. only check where we do the max conflicts check

* restore exception handling logic to master branch

* restore reslimit cancels since the bug appears to be latent

* add bookkeeping for race condition of multiple lease cancels on a single node (messes with reslimit)

* restore unrelated code to master

---------

Signed-off-by: Nikolaj Bjorner <nbjorner@microsoft.com>
Co-authored-by: Ilana Shapiro <ilanashapiro@Ilanas-MacBook-Pro.local>
Co-authored-by: Ilana Shapiro <ilanashapiro@Ilanas-MBP.lan1>
Co-authored-by: Nikolaj Bjorner <nbjorner@microsoft.com>
Co-authored-by: Ilana Shapiro <ilanashapiro@Mac.localdomain>
Co-authored-by: Ilana Shapiro <ilanashapiro@Ilanas-MBP.localdomain>
Co-authored-by: Ilana Shapiro <ilanashapiro@Mac.lan1>
2026-04-20 18:22:07 +02:00

1921 lines
65 KiB
C++

/*++
Copyright (c) 2006 Microsoft Corporation
Module Name:
smt_context.h
Abstract:
Logical context
Author:
Leonardo de Moura (leonardo) 2008-02-18.
Revision History:
--*/
#pragma once
#include <atomic>
#include "ast/quantifier_stat.h"
#include "ast/simplifiers/dependent_expr_state.h"
#include "smt/smt_clause.h"
#include "smt/smt_setup.h"
#include "smt/smt_enode.h"
#include "smt/smt_cg_table.h"
#include "smt/smt_b_justification.h"
#include "smt/smt_eq_justification.h"
#include "smt/smt_justification.h"
#include "smt/smt_bool_var_data.h"
#include "smt/smt_clause_proof.h"
#include "smt/smt_theory.h"
#include "smt/smt_quantifier.h"
#include "smt/smt_statistics.h"
#include "smt/smt_conflict_resolution.h"
#include "smt/smt_relevancy.h"
#include "smt/smt_case_split_queue.h"
#include "smt/smt_almost_cg_table.h"
#include "smt/smt_failure.h"
#include "smt/smt_types.h"
#include "smt/dyn_ack.h"
#include "ast/ast_smt_pp.h"
#include "smt/watch_list.h"
#include "util/trail.h"
#include "util/ref.h"
#include "util/timer.h"
#include "util/statistics.h"
#include "smt/fingerprints.h"
#include "smt/proto_model/proto_model.h"
#include "smt/theory_user_propagator.h"
#include "model/model.h"
#include "solver/progress_callback.h"
#include "solver/assertions/asserted_formulas.h"
#include <tuple>
// there is a significant space overhead with allocating 1000+ contexts in
// the case that each context only references a few expressions.
// Using a map instead of a vector for the literals can compress space
// consumption.
#define USE_BOOL_VAR_VECTOR 1
namespace smt {
class model_generator;
class context;
struct oom_exception : public z3_error {
oom_exception() : z3_error(ERR_MEMOUT) {}
};
struct enode_pp {
context const& ctx;
enode* n;
enode_pp(enode* n, context const& ctx): ctx(ctx), n(n) {}
};
struct replay_unit {
expr_ref m_unit;
bool m_sign;
bool m_relevant;
};
class context {
friend class model_generator;
friend class lookahead;
friend class parallel;
public:
statistics m_stats;
std::ostream& display_last_failure(std::ostream& out) const;
std::string last_failure_as_string() const;
void set_reason_unknown(char const* msg) { m_unknown = msg; }
void set_progress_callback(progress_callback *callback);
protected:
ast_manager & m;
smt_params & m_fparams;
params_ref m_params;
::statistics m_aux_stats;
setup m_setup;
unsigned m_relevancy_lvl;
timer m_timer;
region m_region;
asserted_formulas m_asserted_formulas;
th_rewriter m_rewriter;
scoped_ptr<quantifier_manager> m_qmanager;
scoped_ptr<model_generator> m_model_generator;
scoped_ptr<relevancy_propagator> m_relevancy_propagator;
theory_user_propagator* m_user_propagator;
random_gen m_random;
bool m_flushing; // (debug support) true when flushing
mutable unsigned m_lemma_id;
progress_callback * m_progress_callback;
unsigned m_next_progress_sample;
clause_proof m_clause_proof;
fingerprint_set m_fingerprints;
expr_ref_vector m_b_internalized_stack; // stack of the boolean expressions already internalized.
// Remark: boolean expressions can also be internalized as
// enodes. Examples: boolean expression nested in an
// uninterpreted function.
expr_ref_vector m_e_internalized_stack; // stack of the expressions already internalized as enodes.
quantifier_ref_vector m_l_internalized_stack;
ptr_vector<justification> m_justifications;
unsigned m_final_check_idx = 0; // circular counter used for implementing fairness
bool m_is_auxiliary = false; // used to prevent unwanted information from being logged.
class parallel* m_par = nullptr;
unsigned m_par_index = 0;
bool m_internalizing_assertions = false;
lbool m_internal_completed = l_undef;
scoped_ptr<dependent_expr_simplifier> m_simplifier;
scoped_ptr<base_dependent_expr_state> m_fmls;
svector<double> m_lit_scores[2];
svector<unsigned> m_birthdate;
// -----------------------------------
//
// Equality & Uninterpreted functions
//
// -----------------------------------
enode * m_true_enode;
enode * m_false_enode;
app2enode_t m_app2enode; // app -> enode
ptr_vector<enode> m_enodes;
plugin_manager<theory> m_theories; // mapping from theory_id -> theory
ptr_vector<theory> m_theory_set; // set of theories for fast traversal
vector<enode_vector> m_decl2enodes; // decl -> enode (for decls with arity > 0)
enode_vector m_empty_vector;
cg_table m_cg_table;
struct new_eq {
enode * m_lhs;
enode * m_rhs;
eq_justification m_justification;
new_eq(enode * lhs, enode * rhs, eq_justification const & js):
m_lhs(lhs), m_rhs(rhs), m_justification(js) {}
};
svector<new_eq> m_eq_propagation_queue;
struct new_th_eq {
theory_id m_th_id;
theory_var m_lhs;
theory_var m_rhs;
new_th_eq(theory_id id, theory_var l, theory_var r):m_th_id(id), m_lhs(l), m_rhs(r) {}
};
svector<new_th_eq> m_th_eq_propagation_queue;
svector<new_th_eq> m_th_diseq_propagation_queue;
#ifdef Z3DEBUG
svector<new_th_eq> m_propagated_th_eqs;
svector<new_th_eq> m_propagated_th_diseqs;
svector<enode_pair> m_diseq_vector;
#endif
enode * m_is_diseq_tmp { nullptr }; // auxiliary enode used to find congruent equality atoms.
tmp_enode m_tmp_enode;
ptr_vector<almost_cg_table> m_almost_cg_tables; // temporary field for is_ext_diseq
// -----------------------------------
//
// Boolean engine
//
// -----------------------------------
#if USE_BOOL_VAR_VECTOR
svector<bool_var> m_expr2bool_var; // expr id -> bool_var
#else
u_map<bool_var> m_expr2bool_var;
#endif
ptr_vector<expr> m_bool_var2expr; // bool_var -> expr
signed_char_vector m_assignment; //!< mapping literal id -> assignment lbool
vector<watch_list> m_watches; //!< per literal
unsigned_vector m_lit_occs; //!< occurrence count of literals
svector<bool_var_data> m_bdata; //!< mapping bool_var -> data
svector<double> m_activity;
clause_vector m_aux_clauses;
clause_vector m_lemmas;
vector<clause_vector> m_clauses_to_reinit;
vector<replay_unit> m_units_to_reassert;
literal_vector m_assigned_literals;
typedef std::pair<clause*, literal_vector> tmp_clause;
vector<tmp_clause> m_tmp_clauses;
unsigned m_qhead { 0 };
unsigned m_simp_qhead { 0 };
int m_simp_counter { 0 }; //!< can become negative
scoped_ptr<case_split_queue> m_case_split_queue;
double m_bvar_inc { 1.0 };
bool m_phase_cache_on { true };
unsigned m_phase_counter { 0 }; //!< auxiliary variable used to decide when to turn on/off phase caching
bool m_phase_default { false }; //!< default phase when using phase caching
// A conflict is usually a single justification. That is, a justification
// for false. If m_not_l is not null_literal, then m_conflict is a
// justification for l, and the conflict is union of m_not_l and m_conflict;
// m_empty_clause is set to ensure that an empty clause generated in deep scope
// levels survives to the base level.
b_justification m_conflict;
literal m_not_l;
scoped_ptr<conflict_resolution> m_conflict_resolution;
proof_ref m_unsat_proof;
literal_vector m_atom_propagation_queue;
obj_map<expr, unsigned> m_cached_generation;
obj_hashtable<expr> m_cache_generation_visited;
dyn_ack_manager m_dyn_ack_manager;
// -----------------------------------
//
// Model generation
//
// -----------------------------------
proto_model_ref m_proto_model;
model_ref m_model;
const char * m_unknown;
void mk_proto_model();
void reset_model() { m_model = nullptr; m_proto_model = nullptr; }
// -----------------------------------
//
// Unsat core extraction
//
// -----------------------------------
typedef u_map<expr *> literal2assumption;
literal_vector m_assumptions;
literal2assumption m_literal2assumption; // maps an expression associated with a literal to the original assumption
expr_ref_vector m_unsat_core;
unsigned m_last_position_log { 0 };
svector<size_t> m_last_positions;
// -----------------------------------
//
// Theory case split
//
// -----------------------------------
uint_set m_all_th_case_split_literals;
vector<literal_vector> m_th_case_split_sets;
u_map< vector<literal_vector> > m_literal2casesplitsets; // returns the case split literal sets that a literal participates in
// ----------------------------------
//
// Value initialization
//
// ----------------------------------
vector<std::pair<expr_ref, expr_ref>> m_values;
void initialize_value(expr* var, expr* value);
void initialize_values();
// -----------------------------------
//
// Accessors
//
// -----------------------------------
public:
ast_manager & get_manager() const {
return m;
}
th_rewriter & get_rewriter() {
return m_rewriter;
}
smt_params & get_fparams() {
return m_fparams;
}
params_ref const & get_params() {
return m_params;
}
void updt_params(params_ref const& p);
bool get_cancel_flag();
void set_internal_completed() {
if (m_internal_completed == l_undef)
m_internal_completed = l_true;
}
region & get_region() {
return m_region;
}
bool relevancy() const {
return relevancy_lvl() > 0;
}
unsigned relevancy_lvl() const;
enode * get_enode(expr const * n) const {
SASSERT(e_internalized(n));
return m_app2enode[n->get_id()];
}
void get_specrels(func_decl_set& rels) const;
/**
\brief Similar to get_enode, but returns 0 if n is to e_internalized.
*/
enode * find_enode(expr const * n) const {
return m_app2enode.get(n->get_id(), 0);
}
void reset_bool_vars() {
m_expr2bool_var.reset();
}
bool_var get_bool_var(expr const * n) const {
return m_expr2bool_var[n->get_id()];
}
bool_var get_bool_var(enode const * n) const {
return get_bool_var(n->get_expr());
}
bool_var get_bool_var_of_id(unsigned id) const {
return m_expr2bool_var[id];
}
bool_var get_bool_var_of_id_option(unsigned id) const {
return m_expr2bool_var.get(id, null_bool_var);
}
#if USE_BOOL_VAR_VECTOR
void set_bool_var(unsigned id, bool_var v) {
m_expr2bool_var.setx(id, v, null_bool_var);
}
#else
void set_bool_var(unsigned id, bool_var v) {
if (v == null_bool_var) {
m_expr2bool_var.erase(id);
}
else {
m_expr2bool_var.insert(id, v);
}
}
#endif
clause_vector const& get_lemmas() const { return m_lemmas; }
literal get_literal(expr * n) const;
bool has_enode(bool_var v) const {
return m_bdata[v].is_enode();
}
enode * bool_var2enode(bool_var v) const {
SASSERT(m_bdata[v].is_enode());
return m_app2enode[m_bool_var2expr[v]->get_id()];
}
bool_var enode2bool_var(enode const * n) const {
SASSERT(n->is_bool());
SASSERT(n != m_false_enode);
return get_bool_var_of_id(n->get_owner_id());
}
literal enode2literal(enode const * n) const {
SASSERT(n->is_bool());
return n == m_false_enode ? false_literal : literal(enode2bool_var(n));
}
unsigned get_num_bool_vars() const {
return m_b_internalized_stack.size();
}
bool_var_data & get_bdata(bool_var v) {
return m_bdata[v];
}
bool_var_data const & get_bdata(bool_var v) const {
return m_bdata[v];
}
lbool get_lit_assignment(unsigned lit_idx) const {
return static_cast<lbool>(m_assignment[lit_idx]);
}
lbool get_assignment(literal l) const {
return get_lit_assignment(l.index());
}
lbool get_assignment(bool_var v) const {
return get_assignment(literal(v));
}
literal_vector const & assigned_literals() const {
return m_assigned_literals;
}
watch_list const& get_watch(literal l) const {
return m_watches[l.index()];
}
lbool get_assignment(expr * n) const;
// Similar to get_assignment, but returns l_undef if n is not internalized.
lbool find_assignment(expr * n) const;
lbool get_assignment(enode * n) const;
void get_assignments(expr_ref_vector& assignments);
b_justification get_justification(bool_var v) const {
return get_bdata(v).justification();
}
void set_justification(bool_var v, bool_var_data& d, b_justification const& j);
bool has_th_justification(bool_var v, theory_id th_id) const {
b_justification js = get_justification(v);
return js.get_kind() == b_justification::JUSTIFICATION && js.get_justification()->get_from_theory() == th_id;
}
void set_random_seed(unsigned s) { m_random.set_seed(s); }
int get_random_value() { return m_random(); }
bool is_searching() const { return m_searching; }
svector<double> const & get_activity_vector() const { return m_activity; }
double get_activity(bool_var v) const { return m_activity[v]; }
void set_activity(bool_var v, double act) { m_activity[v] = act; }
void activity_changed(bool_var v, bool increased) {
if (increased) {
m_case_split_queue->activity_increased_eh(v);
}
else {
m_case_split_queue->activity_decreased_eh(v);
}
}
bool is_assumption(bool_var v) const {
return get_bdata(v).m_assumption;
}
bool is_assumption(literal l) const {
return is_assumption(l.var());
}
bool is_marked(bool_var v) const {
return get_bdata(v).m_mark;
}
void set_mark(bool_var v) {
SASSERT(!is_marked(v));
get_bdata(v).m_mark = true;
}
void unset_mark(bool_var v) {
SASSERT(is_marked(v));
get_bdata(v).m_mark = false;
}
/**
\brief Return the scope level when v was assigned.
*/
unsigned get_assign_level(bool_var v) const {
return get_bdata(v).m_scope_lvl;
}
unsigned get_assign_level(literal l) const {
return get_assign_level(l.var());
}
/**
\brief Return the scope level when v was internalized.
*/
unsigned get_intern_level(bool_var v) const {
return get_bdata(v).get_intern_level();
}
theory * get_theory(theory_id th_id) const {
return m_theories.get_plugin(th_id);
}
ptr_vector<theory> const& theories() const { return m_theories.plugins(); }
ptr_vector<theory>::const_iterator begin_theories() const {
return m_theories.begin();
}
ptr_vector<theory>::const_iterator end_theories() const {
return m_theories.end();
}
unsigned get_scope_level() const {
return m_scope_lvl;
}
unsigned get_base_level() const {
return m_base_lvl;
}
bool at_base_level() const {
return m_scope_lvl == m_base_lvl;
}
unsigned get_search_level() const {
return m_search_lvl;
}
bool at_search_level() const {
return m_scope_lvl == m_search_lvl;
}
bool tracking_assumptions() const {
return !m_assumptions.empty() && m_search_lvl > m_base_lvl;
}
expr * bool_var2expr(bool_var v) const {
return m_bool_var2expr[v];
}
void literal2expr(literal l, expr_ref & result) const {
if (l == true_literal)
result = m.mk_true();
else if (l == false_literal)
result = m.mk_false();
else if (l.sign())
result = m.mk_not(bool_var2expr(l.var()));
else
result = bool_var2expr(l.var());
}
expr_ref literal2expr(literal l) const {
expr_ref result(m);
literal2expr(l, result);
return result;
}
bool is_true(enode const * n) const {
return n == m_true_enode;
}
bool is_false(enode const * n) const {
return n == m_false_enode;
}
unsigned get_num_enodes_of(func_decl const * decl) const {
unsigned id = decl->get_small_id();
return id < m_decl2enodes.size() ? m_decl2enodes[id].size() : 0;
}
enode_vector const& enodes_of(func_decl const * d) const {
unsigned id = d->get_small_id();
return id < m_decl2enodes.size() ? m_decl2enodes[id] : m_empty_vector;
}
enode_vector::const_iterator begin_enodes_of(func_decl const * decl) const {
unsigned id = decl->get_small_id();
return id < m_decl2enodes.size() ? m_decl2enodes[id].begin() : nullptr;
}
enode_vector::const_iterator end_enodes_of(func_decl const * decl) const {
unsigned id = decl->get_small_id();
return id < m_decl2enodes.size() ? m_decl2enodes[id].end() : nullptr;
}
ptr_vector<enode> const& enodes() const { return m_enodes; }
ptr_vector<enode>::const_iterator begin_enodes() const {
return m_enodes.begin();
}
ptr_vector<enode>::const_iterator end_enodes() const {
return m_enodes.end();
}
unsigned get_generation(quantifier * q) const {
return m_qmanager->get_generation(q);
}
/**
\brief Return true if the logical context internalized universal quantifiers.
*/
bool internalized_quantifiers() const {
return !m_qmanager->empty();
}
/**
\brief Return true if the logical context internalized or will internalize universal quantifiers.
*/
bool has_quantifiers() const {
return m_asserted_formulas.has_quantifiers();
}
fingerprint * add_fingerprint(void * data, unsigned data_hash, unsigned num_args, enode * const * args, expr* def = nullptr) {
return m_fingerprints.insert(data, data_hash, num_args, args, def);
}
theory_id get_var_theory(bool_var v) const {
return get_bdata(v).get_theory();
}
/**
* flag to toggle quantifier tracing.
*/
bool m_coming_from_quant { false };
friend class set_var_theory_trail;
void set_var_theory(bool_var v, theory_id tid);
bool has_sls_model();
// -----------------------------------
//
// Backtracking support
//
// -----------------------------------
protected:
trail_stack m_trail_stack;
#ifdef Z3DEBUG
bool m_trail_enabled { true };
#endif
public:
template<typename TrailObject>
void push_trail(const TrailObject & obj) {
SASSERT(m_trail_enabled);
m_trail_stack.push(obj);
}
void push_trail_ptr(trail * ptr) {
m_trail_stack.push_ptr(ptr);
}
trail_stack& get_trail_stack() {
return m_trail_stack;
}
protected:
unsigned m_scope_lvl { 0 };
unsigned m_base_lvl { 0 };
unsigned m_search_lvl { 0 }; // It is greater than m_base_lvl when assumptions are used. Otherwise, it is equals to m_base_lvl
struct scope {
unsigned m_assigned_literals_lim;
unsigned m_aux_clauses_lim;
unsigned m_justifications_lim;
unsigned m_units_to_reassert_lim;
};
struct base_scope {
unsigned m_lemmas_lim;
unsigned m_simp_qhead_lim;
unsigned m_inconsistent;
};
svector<scope> m_scopes;
svector<base_scope> m_base_scopes;
void push_scope();
unsigned pop_scope_core(unsigned num_scopes);
void pop_scope(unsigned num_scopes);
void unassign_vars(unsigned old_lim);
void remove_watch_literal(clause * cls, unsigned idx);
void remove_cls_occs(clause * cls);
void del_clause(bool log, clause * cls);
void del_clauses(clause_vector & v, unsigned old_size);
void del_justifications(ptr_vector<justification> & justifications, unsigned old_lim);
bool is_unit_clause(clause const * c) const;
bool is_empty_clause(clause const * c) const;
void cache_generation(unsigned new_scope_lvl);
void cache_generation(clause const * cls, unsigned new_scope_lvl);
void cache_generation(unsigned num_lits, literal const * lits, unsigned new_scope_lvl);
void cache_generation(expr * n, unsigned new_scope_lvl);
void reset_cache_generation();
void reinit_clauses(unsigned num_scopes, unsigned num_bool_vars);
void reassert_units(unsigned units_to_reassert_lim);
public:
// \brief exposed for PB solver to participate in GC
void remove_watch(bool_var v);
// -----------------------------------
//
// Internalization
//
// -----------------------------------
public:
bool b_internalized(expr const * n) const {
return get_bool_var_of_id_option(n->get_id()) != null_bool_var;
}
bool lit_internalized(expr const * n) const {
return m.is_false(n) || (m.is_not(n) ? b_internalized(to_app(n)->get_arg(0)) : b_internalized(n));
}
bool e_internalized(expr const * n) const {
return m_app2enode.get(n->get_id(), 0) != 0;
}
unsigned get_num_b_internalized() const {
return m_b_internalized_stack.size();
}
expr * get_b_internalized(unsigned idx) const {
return m_b_internalized_stack.get(idx);
}
unsigned get_num_e_internalized() const {
return m_e_internalized_stack.size();
}
expr * get_e_internalized(unsigned idx) const {
return m_e_internalized_stack.get(idx);
}
/**
\brief Return the position (in the assignment stack) of the decision literal at the given scope level.
*/
unsigned get_decision_literal_pos(unsigned scope_lvl) const {
SASSERT(scope_lvl > m_base_lvl);
return m_scopes[scope_lvl - 1].m_assigned_literals_lim;
}
protected:
unsigned m_generation { 0 }; //!< temporary variable used during internalization
public:
bool binary_clause_opt_enabled() const {
return !m.proofs_enabled() && m_fparams.m_binary_clause_opt;
}
protected:
bool_var_data & get_bdata(expr const * n) {
return get_bdata(get_bool_var(n));
}
bool_var_data const & get_bdata(expr const * n) const {
return get_bdata(get_bool_var(n));
}
typedef std::pair<expr *, bool> expr_bool_pair;
void ts_visit_child(expr * n, bool gate_ctx, svector<expr_bool_pair> & todo, bool & visited);
bool ts_visit_children(expr * n, bool gate_ctx, svector<expr_bool_pair> & todo);
svector<expr_bool_pair> ts_todo;
char_vector tcolors;
char_vector fcolors;
bool should_internalize_rec(expr* e) const;
void top_sort_expr(expr* const* exprs, unsigned num_exprs, svector<expr_bool_pair> & sorted_exprs);
void internalize_rec(expr * n, bool gate_ctx);
void internalize_deep(expr * n);
void internalize_deep(expr* const* n, unsigned num_exprs);
void assert_default(expr * n, proof * pr);
void assert_distinct(app * n, proof * pr);
void internalize_formula(expr * n, bool gate_ctx);
void internalize_eq(app * n, bool gate_ctx);
void internalize_distinct(app * n, bool gate_ctx);
bool internalize_theory_atom(app * n, bool gate_ctx);
void internalize_quantifier(quantifier * q, bool gate_ctx);
obj_map<enode, quantifier*> m_lambdas;
bool has_lambda();
void internalize_lambda(quantifier * q);
void internalize_formula_core(app * n, bool gate_ctx);
void set_merge_tf(enode * n, bool_var v, bool is_new_var);
friend class set_enode_flag_trail;
public:
void set_enode_flag(bool_var v, bool is_new_var);
protected:
void internalize_term(app * n);
void internalize_ite_term(app * n);
bool internalize_theory_term(app * n);
void internalize_uninterpreted(app * n);
friend class mk_bool_var_trail;
class mk_bool_var_trail : public trail {
context& ctx;
public:
mk_bool_var_trail(context& ctx) :ctx(ctx) {}
void undo() override { ctx.undo_mk_bool_var(); }
};
mk_bool_var_trail m_mk_bool_var_trail;
void undo_mk_bool_var();
friend class mk_enode_trail;
class mk_enode_trail : public trail {
context& ctx;
public:
mk_enode_trail(context& ctx) :ctx(ctx) {}
void undo() override { ctx.undo_mk_enode(); }
};
mk_enode_trail m_mk_enode_trail;
void undo_mk_enode();
friend class mk_lambda_trail;
class mk_lambda_trail : public trail {
context& ctx;
public:
mk_lambda_trail(context& ctx) :ctx(ctx) {}
void undo() override { ctx.undo_mk_lambda(); }
};
mk_lambda_trail m_mk_lambda_trail;
void undo_mk_lambda();
void apply_sort_cnstr(app * term, enode * e);
bool simplify_aux_clause_literals(unsigned & num_lits, literal * lits, literal_buffer & simp_lits);
bool simplify_aux_lemma_literals(unsigned & num_lits, literal * lits);
void mark_for_reinit(clause * cls, unsigned scope_lvl, bool reinternalize_atoms);
unsigned get_max_iscope_lvl(unsigned num_lits, literal const * lits) const;
bool use_binary_clause_opt(literal l1, literal l2, bool lemma) const;
int select_learned_watch_lit(clause const * cls) const;
int select_watch_lit(clause const * cls, int starting_at) const;
void add_watch_literal(clause * cls, unsigned idx);
proof * mk_clause_def_axiom(unsigned num_lits, literal * lits, expr * root_gate);
public:
void mk_gate_clause(unsigned num_lits, literal * lits);
void mk_gate_clause(literal l1, literal l2);
void mk_gate_clause(literal l1, literal l2, literal l3);
void mk_gate_clause(literal l1, literal l2, literal l3, literal l4);
protected:
void mk_root_clause(unsigned num_lits, literal * lits, proof * pr);
void mk_root_clause(literal l1, literal l2, proof * pr);
void mk_root_clause(literal l1, literal l2, literal l3, proof * pr);
void add_and_rel_watches(app * n);
void add_or_rel_watches(app * n);
void add_implies_rel_watches(app* n);
void add_ite_rel_watches(app * n);
void mk_not_cnstr(app * n);
void mk_and_cnstr(app * n);
void mk_or_cnstr(app * n);
void mk_implies_cnstr(app* n);
void mk_iff_cnstr(app * n, bool sign);
void mk_ite_cnstr(app * n);
bool track_occs() const { return m_fparams.m_phase_selection == PS_OCCURRENCE; }
void dec_ref(literal l);
void inc_ref(literal l);
void remove_lit_occs(clause const& cls, unsigned num_bool_vars);
void add_lit_occs(clause const& cls);
ast_pp_util m_lemma_visitor;
void dump_lemma(unsigned n, literal const* lits);
void dump_axiom(unsigned n, literal const* lits);
public:
void ensure_internalized(expr* e);
void internalize(expr * n, bool gate_ctx);
void internalize(expr* const* exprs, unsigned num_exprs, bool gate_ctx);
void internalize(expr * n, bool gate_ctx, unsigned generation);
clause * mk_clause(unsigned num_lits, literal * lits, justification * j, clause_kind k = CLS_AUX, clause_del_eh * del_eh = nullptr);
void mk_clause(literal l1, literal l2, justification * j);
void mk_clause(literal l1, literal l2, literal l3, justification * j);
void mk_th_clause(theory_id tid, unsigned num_lits, literal * lits, unsigned num_params, parameter * params, clause_kind k);
void mk_th_axiom(theory_id tid, unsigned num_lits, literal * lits, unsigned num_params = 0, parameter * params = nullptr) {
mk_th_clause(tid, num_lits, lits, num_params, params, CLS_TH_AXIOM);
}
void mk_th_axiom(theory_id tid, literal l1, unsigned num_params = 0, parameter * params = nullptr);
void mk_th_axiom(theory_id tid, literal l1, literal l2, unsigned num_params = 0, parameter * params = nullptr);
void mk_th_axiom(theory_id tid, literal l1, literal l2, literal l3, unsigned num_params = 0, parameter * params = nullptr);
void mk_th_axiom(theory_id tid, literal_vector const& ls, unsigned num_params = 0, parameter * params = nullptr) {
mk_th_axiom(tid, ls.size(), ls.data(), num_params, params);
}
void mk_th_lemma(theory_id tid, literal l1, literal l2, unsigned num_params = 0, parameter * params = nullptr) {
literal ls[2] = { l1, l2 };
mk_th_lemma(tid, 2, ls, num_params, params);
}
void mk_th_lemma(theory_id tid, literal l1, literal l2, literal l3, unsigned num_params = 0, parameter * params = nullptr) {
literal ls[3] = { l1, l2, l3 };
mk_th_lemma(tid, 3, ls, num_params, params);
}
void mk_th_lemma(theory_id tid, unsigned num_lits, literal * lits, unsigned num_params = 0, parameter * params = nullptr) {
mk_th_clause(tid, num_lits, lits, num_params, params, CLS_TH_LEMMA);
}
void mk_th_lemma(theory_id tid, literal_vector const& ls, unsigned num_params = 0, parameter * params = nullptr) {
mk_th_lemma(tid, ls.size(), ls.data(), num_params, params);
}
/*
* Provide a hint to the core solver that the specified literals form a "theory case split".
* The core solver will enforce the condition that exactly one of these literals can be
* assigned 'true' at any time.
* We assume that the theory solver has already asserted the disjunction of these literals
* or some other axiom that means at least one of them must be assigned 'true'.
*/
void mk_th_case_split(unsigned num_lits, literal * lits);
/*
* Provide a hint to the branching heuristic about the priority of a "theory-aware literal".
* Literals marked in this way will always be branched on before unmarked literals,
* starting with the literal having the highest priority.
*/
void add_theory_aware_branching_info(bool_var v, double priority, lbool phase);
public:
// helper function for trail
void undo_th_case_split(literal l);
bool propagate_th_case_split(unsigned qhead);
bool_var mk_bool_var(expr * n);
enode * mk_enode(app * n, bool suppress_args, bool merge_tf, bool cgc_enabled);
void attach_th_var(enode * n, theory * th, theory_var v);
template<typename Justification>
justification * mk_justification(Justification const & j) {
justification * js = new (get_region()) Justification(j);
SASSERT(js->in_region());
if (js->has_del_eh())
m_justifications.push_back(js);
return js;
}
// -----------------------------------
//
// Engine
//
// -----------------------------------
protected:
lbool m_last_search_result { l_undef };
failure m_last_search_failure { UNKNOWN };
ptr_vector<theory> m_incomplete_theories; //!< theories that failed to produce a model
bool m_searching { false };
unsigned m_num_conflicts;
unsigned m_num_conflicts_since_restart;
unsigned m_num_conflicts_since_lemma_gc;
unsigned m_num_restarts;
unsigned m_num_simplifications;
unsigned m_restart_threshold;
unsigned m_restart_outer_threshold;
unsigned m_luby_idx;
double m_agility;
unsigned m_lemma_gc_threshold;
void assign_core(literal l, b_justification j, bool decision = false);
void trace_assign(literal l, b_justification j, bool decision) const;
public:
void assign(literal l, const b_justification & j, bool decision = false) {
SASSERT(l != false_literal);
SASSERT(l != null_literal);
switch (get_assignment(l)) {
case l_false:
set_conflict(j, ~l);
break;
case l_undef:
assign_core(l, j, decision);
break;
case l_true:
return;
}
}
void assign(literal l, justification * j, bool decision = false) {
assign(l, j ? b_justification(j) : b_justification::mk_axiom(), decision);
}
friend class set_true_first_trail;
void set_true_first_flag(bool_var v);
bool try_true_first(bool_var v) const { return get_bdata(v).try_true_first(); }
bool assume_eq(enode * lhs, enode * rhs);
bool is_shared(enode * n) const;
bool is_beta_redex(enode* p, enode* n) const;
void assign_eq(enode * lhs, enode * rhs, eq_justification const & js) {
push_eq(lhs, rhs, js);
}
/**
\brief Force the given phase next time we case split v.
This method has no effect if phase caching is disabled.
*/
void force_phase(bool_var v, bool phase) {
bool_var_data & d = get_bdata(v);
d.m_phase_available = true;
d.m_phase = phase;
}
void force_phase(literal l) {
force_phase(l.var(), !l.sign());
}
bool contains_instance(quantifier * q, unsigned num_bindings, enode * const * bindings);
bool add_instance(quantifier * q, app * pat, unsigned num_bindings, enode * const * bindings, expr* def, unsigned max_generation,
unsigned min_top_generation, unsigned max_top_generation, vector<std::tuple<enode *, enode*>> & used_enodes /*gives the equalities used for the pattern match, see mam.cpp for more info*/);
void set_global_generation(unsigned generation) { m_generation = generation; }
#ifdef Z3DEBUG
bool slow_contains_instance(quantifier const * q, unsigned num_bindings, enode * const * bindings) const {
return m_fingerprints.slow_contains(q, q->get_id(), num_bindings, bindings);
}
#endif
void add_eq(enode * n1, enode * n2, eq_justification js);
protected:
void push_new_th_eq(theory_id th, theory_var lhs, theory_var rhs);
void push_new_th_diseq(theory_id th, theory_var lhs, theory_var rhs);
friend class add_eq_trail;
void remove_parents_from_cg_table(enode * r1);
void reinsert_parents_into_cg_table(enode * r1, enode * r2, enode * n1, enode * n2, eq_justification js);
void invert_trans(enode * n);
theory_var get_closest_var(enode * n, theory_id th_id);
void merge_theory_vars(enode * r2, enode * r1, eq_justification js);
void propagate_bool_enode_assignment(enode * r1, enode * r2, enode * n1, enode * n2);
void propagate_bool_enode_assignment_core(enode * source, enode * target);
void undo_add_eq(enode * r1, enode * n1, unsigned r2_num_parents);
void restore_theory_vars(enode * r2, enode * r1);
void push_eq(enode * lhs, enode * rhs, eq_justification const & js) {
if (lhs->get_root() != rhs->get_root()) {
SASSERT(lhs->get_expr()->get_sort() == rhs->get_expr()->get_sort());
m_eq_propagation_queue.push_back(new_eq(lhs, rhs, js));
}
}
void push_new_congruence(enode * n1, enode * n2, bool used_commutativity) {
SASSERT(n1->m_cg == n2);
// if (is_relevant(n1)) mark_as_relevant(n2);
push_eq(n1, n2, eq_justification::mk_cg(used_commutativity));
}
bool add_diseq(enode * n1, enode * n2);
void assign_quantifier(quantifier * q);
void set_conflict(const b_justification & js, literal not_l);
void set_conflict(const b_justification & js) {
set_conflict(js, null_literal);
}
public:
void set_conflict(justification * js) {
SASSERT(js);
set_conflict(b_justification(js));
}
bool inconsistent() const {
return m_conflict != null_b_justification ||
m_asserted_formulas.inconsistent();
}
bool has_case_splits();
unsigned get_num_conflicts() const {
return m_num_conflicts;
}
static bool is_eq(enode const * n1, enode const * n2) { return n1->get_root() == n2->get_root(); }
bool is_diseq(enode * n1, enode * n2) const;
bool is_diseq_slow(enode * n1, enode * n2) const;
bool is_ext_diseq(enode * n1, enode * n2, unsigned depth);
enode * get_enode_eq_to(func_decl * f, unsigned num_args, enode * const * args);
bool guess(bool_var var, lbool phase);
protected:
bool m_has_case_split = true;
bool decide();
void update_phase_cache_counter();
#define ACTIVITY_LIMIT 1e100
#define INV_ACTIVITY_LIMIT 1e-100
void rescale_bool_var_activity();
public:
void inc_bvar_activity(bool_var v, double inc = 1.0) {
double & act = m_activity[v];
act += m_bvar_inc * inc;
if (act > ACTIVITY_LIMIT)
rescale_bool_var_activity();
m_case_split_queue->activity_increased_eh(v);
TRACE(case_split, tout << "v" << v << " " << m_bvar_inc << " -> " << act << "\n";);
}
protected:
void decay_bvar_activity() {
m_bvar_inc *= m_fparams.m_inv_decay;
}
bool simplify_clause(clause& cls);
unsigned simplify_clauses(clause_vector & clauses, unsigned starting_at);
void simplify_clauses();
/**
\brief Return true if the give clause is justifying some literal.
*/
bool is_justifying(clause * cls) const {
for (unsigned i = 0; i < 2; ++i) {
b_justification js;
js = get_justification((*cls)[i].var());
if (js.get_kind() == b_justification::CLAUSE && js.get_clause() == cls)
return true;
}
return false;
}
bool can_delete(clause * cls) const {
if (cls->in_reinit_stack())
return false;
return !is_justifying(cls);
}
void del_inactive_lemmas();
void del_inactive_lemmas1();
void del_inactive_lemmas2();
bool more_than_k_unassigned_literals(clause * cls, unsigned k);
void asserted_inconsistent();
bool validate_assumptions(expr_ref_vector const& asms);
void init_assumptions(expr_ref_vector const& asms);
void init_clause(expr_ref_vector const& clause);
lbool decide_clause();
void reset_tmp_clauses();
void reset_assumptions();
void add_theory_assumptions(expr_ref_vector & theory_assumptions);
lbool mk_unsat_core(lbool result);
bool should_research(lbool result);
void validate_unsat_core();
void init_search();
void end_search();
lbool search();
void inc_limits();
bool restart(lbool& status, unsigned curr_lvl);
void tick(unsigned & counter) const;
lbool bounded_search();
final_check_status final_check();
void check_proof(proof * pr);
void forget_phase_of_vars_in_current_level();
virtual bool resolve_conflict();
void add_scores(unsigned n, literal const *lits);
// -----------------------------------
//
// Propagation
//
// -----------------------------------
protected:
bool bcp();
bool propagate_eqs();
bool propagate_atoms();
void push_new_th_diseqs(enode * r, theory_var v, theory * th);
void propagate_bool_var_enode(bool_var v);
bool is_relevant_core(expr * n) const { return m_relevancy_propagator->is_relevant(n); }
bool_vector m_relevant_conflict_literals;
void record_relevancy(unsigned n, literal const* lits);
void restore_relevancy(unsigned n, literal const* lits);
public:
// event handler for relevancy_propagator class
void relevant_eh(expr * n);
bool is_relevant(expr * n) const {
return !relevancy() || is_relevant_core(n);
}
bool is_relevant(enode * n) const {
return is_relevant(n->get_expr());
}
bool is_relevant(bool_var v) const {
return is_relevant(bool_var2expr(v));
}
bool is_relevant(literal l) const {
SASSERT(l != true_literal && l != false_literal);
return is_relevant(l.var());
}
bool is_relevant_core(literal l) const {
return is_relevant_core(bool_var2expr(l.var()));
}
void mark_as_relevant(expr * n) { m_relevancy_propagator->mark_as_relevant(n); m_relevancy_propagator->propagate(); }
void mark_as_relevant(enode * n) { mark_as_relevant(n->get_expr()); }
void mark_as_relevant(bool_var v) { mark_as_relevant(bool_var2expr(v)); }
void mark_as_relevant(literal l) { mark_as_relevant(l.var()); }
template<typename Eh>
relevancy_eh * mk_relevancy_eh(Eh const & eh) {
return m_relevancy_propagator->mk_relevancy_eh(eh);
}
void add_relevancy_eh(expr * source, relevancy_eh * eh) { m_relevancy_propagator->add_handler(source, eh); }
void add_relevancy_dependency(expr * source, expr * target) { m_relevancy_propagator->add_dependency(source, target); }
void add_rel_watch(literal l, relevancy_eh * eh) { m_relevancy_propagator->add_watch(bool_var2expr(l.var()), !l.sign(), eh); }
void add_rel_watch(literal l, expr * n) { m_relevancy_propagator->add_watch(bool_var2expr(l.var()), !l.sign(), n); }
protected:
lbool get_assignment_core(expr * n) const;
void propagate_relevancy(unsigned qhead);
bool propagate_theories();
void propagate_th_eqs();
void propagate_th_diseqs();
bool can_theories_propagate() const;
bool propagate();
void add_rec_funs_to_model();
public:
bool can_propagate() const;
// -----------------------------------
//
// Value extraction and solving
//
// -----------------------------------
public:
bool get_value(enode * n, expr_ref & value);
void solve_for(vector<solution>& sol);
// -----------------------------------
//
// Pretty Printing
//
// -----------------------------------
protected:
ast_mark m_pp_visited;
ast_mark & get_pp_visited() const { return const_cast<ast_mark&>(m_pp_visited); }
public:
void display_enode_defs(std::ostream & out) const;
void display_bool_var_defs(std::ostream & out) const;
void display_asserted_formulas(std::ostream & out) const;
enode_pp pp(enode* n) { return enode_pp(n, *this); }
std::ostream& display_literal(std::ostream & out, literal l) const;
std::ostream& display_detailed_literal(std::ostream & out, literal l) const { return smt::display(out, l, m, m_bool_var2expr.data()); }
void display_literal_info(std::ostream & out, literal l) const;
std::ostream& display_literals(std::ostream & out, unsigned num_lits, literal const * lits) const;
std::ostream& display_literals(std::ostream & out, literal_vector const& lits) const {
return display_literals(out, lits.size(), lits.data());
}
std::ostream& display_literal_smt2(std::ostream& out, literal lit) const;
std::ostream& display_literals_smt2(std::ostream& out, literal l1, literal l2) const { literal ls[2] = { l1, l2 }; return display_literals_smt2(out, 2, ls); }
std::ostream& display_literals_smt2(std::ostream& out, unsigned num_lits, literal const* lits) const;
std::ostream& display_literals_smt2(std::ostream& out, literal_vector const& ls) const { return display_literals_smt2(out, ls.size(), ls.data()); }
std::ostream& display_literal_verbose(std::ostream & out, literal lit) const;
std::ostream& display_literals_verbose(std::ostream & out, unsigned num_lits, literal const * lits) const;
std::ostream& display_literals_verbose(std::ostream & out, literal_vector const& lits) const {
return display_literals_verbose(out, lits.size(), lits.data());
}
void display_watch_list(std::ostream & out, literal l) const;
void display_watch_lists(std::ostream & out) const;
std::ostream& display_clause_detail(std::ostream & out, clause const * cls) const;
std::ostream& display_clause(std::ostream & out, clause const * cls) const;
std::ostream& display_clause_smt2(std::ostream & out, clause const& cls) const;
std::ostream& display_clauses(std::ostream & out, ptr_vector<clause> const & v) const;
std::ostream& display_binary_clauses(std::ostream & out) const;
void display_assignment(std::ostream & out) const;
void display_eqc(std::ostream & out) const;
void display_app_enode_map(std::ostream & out) const;
void display_expr_bool_var_map(std::ostream & out) const;
void display_relevant_exprs(std::ostream & out) const;
void display_theories(std::ostream & out) const;
void display_eq_detail(std::ostream & out, enode * n) const;
void display_parent_eqs(std::ostream & out, enode * n) const;
void display_hot_bool_vars(std::ostream & out) const;
void display_lemma_as_smt_problem(std::ostream & out, unsigned num_antecedents, literal const * antecedents, literal consequent = false_literal, symbol const& logic = symbol::null) const;
unsigned display_lemma_as_smt_problem(unsigned num_antecedents, literal const * antecedents, literal consequent = false_literal, symbol const& logic = symbol::null) const;
void display_lemma_as_smt_problem(std::ostream & out, unsigned num_antecedents, literal const * antecedents,
unsigned num_antecedent_eqs, enode_pair const * antecedent_eqs,
literal consequent = false_literal, symbol const& logic = symbol::null, enode* x = nullptr, enode* y = nullptr) const;
unsigned display_lemma_as_smt_problem(unsigned num_antecedents, literal const * antecedents,
unsigned num_antecedent_eqs, enode_pair const * antecedent_eqs,
literal consequent = false_literal, symbol const& logic = symbol::null) const;
std::string mk_lemma_name() const;
void display_assignment_as_smtlib2(std::ostream& out, symbol const& logic = symbol::null) const;
void display_normalized_enodes(std::ostream & out) const;
void display_enodes_lbls(std::ostream & out) const;
void display_decl2enodes(std::ostream & out) const;
void display_subexprs_info(std::ostream & out, expr * n) const;
void display_var_occs_histogram(std::ostream & out) const;
void display_num_min_occs(std::ostream & out) const;
void display_profile_res_sub(std::ostream & out) const;
void display_profile(std::ostream & out) const;
std::ostream& display(std::ostream& out, b_justification j) const;
std::ostream& display_compact_j(std::ostream& out, b_justification j) const;
// -----------------------------------
//
// Debugging support
//
// -----------------------------------
protected:
#ifdef Z3DEBUG
bool is_watching_clause(literal l, clause const * cls) const;
bool check_clause(clause const * cls) const;
bool check_clauses(clause_vector const & v) const;
bool check_watch_list(literal l) const;
bool check_watch_list(unsigned l_idx) const;
bool check_bin_watch_lists() const;
bool check_enode(enode * n) const;
bool check_enodes() const;
bool check_invariant() const;
bool check_eqc_bool_assignment() const;
bool check_missing_clause_propagation(clause_vector const & v) const;
bool check_missing_bin_clause_propagation() const;
bool check_missing_eq_propagation() const;
bool check_missing_congruence() const;
bool check_missing_bool_enode_propagation() const;
bool check_missing_propagation() const;
bool check_relevancy(expr_ref_vector const & v) const;
bool check_relevancy() const;
bool check_bool_var_vector_sizes() const;
bool check_th_diseq_propagation() const;
bool check_missing_diseq_conflict() const;
#endif
// -----------------------------------
//
// Introspection
//
// -----------------------------------
unsigned get_lemma_avg_activity() const;
void display_literal_num_occs(std::ostream & out) const;
void display_num_assigned_literals_per_lvl(std::ostream & out) const;
// -----------------------------------
//
// Auxiliary
//
// -----------------------------------
void init();
void flush();
config_mode get_config_mode(bool use_static_features) const;
virtual void setup_context(bool use_static_features);
void setup_components();
void pop_to_base_lvl();
void pop_to_search_lvl();
#ifdef Z3DEBUG
bool already_internalized_theory(theory * th) const;
bool already_internalized_theory_core(theory * th, expr_ref_vector const & s) const;
#endif
bool check_preamble(bool reset_cancel);
struct search_completion {
context& ctx;
search_completion(context& ctx) : ctx(ctx) { ctx.m_search_finalized = false; }
~search_completion() { if (!ctx.m_search_finalized) ctx.m_last_search_failure = CANCELED; }
};
bool m_search_finalized = true;
lbool check_finalize(lbool r);
// -----------------------------------
//
// API
//
// -----------------------------------
void assert_expr_core(expr * e, proof * pr);
// copy plugins into a fresh context.
void copy_plugins(context& src, context& dst);
/*
\brief Utilities for consequence finding.
*/
typedef hashtable<unsigned, u_hash, u_eq> index_set;
//typedef uint_set index_set;
u_map<index_set> m_antecedents;
obj_map<expr, expr*> m_var2orig;
u_map<expr*> m_assumption2orig;
obj_map<expr, expr*> m_var2val;
void extract_fixed_consequences(literal lit, index_set const& assumptions, expr_ref_vector& conseq);
void extract_fixed_consequences(unsigned& idx, index_set const& assumptions, expr_ref_vector& conseq);
void display_consequence_progress(std::ostream& out, unsigned it, unsigned nv, unsigned fixed, unsigned unfixed, unsigned eq);
unsigned delete_unfixed(expr_ref_vector& unfixed);
unsigned extract_fixed_eqs(expr_ref_vector& conseq);
expr_ref antecedent2fml(index_set const& ante);
literal mk_diseq(expr* v, expr* val);
void validate_consequences(expr_ref_vector const& assumptions, expr_ref_vector const& vars,
expr_ref_vector const& conseq, expr_ref_vector const& unfixed);
bool validate_justification(bool_var v, bool_var_data const& d, b_justification const& j);
void justify(literal lit, index_set& s);
void extract_cores(expr_ref_vector const& asms, vector<expr_ref_vector>& cores, unsigned& min_core_size);
void preferred_sat(literal_vector& literals);
void display_partial_assignment(std::ostream& out, expr_ref_vector const& asms, unsigned min_core_size);
void log_stats();
void copy_user_propagator(context& src, bool copy_registered);
public:
context(ast_manager & m, smt_params & fp, params_ref const & p = params_ref());
virtual ~context();
/**
\brief Return a new context containing the same theories and simplifier plugins, but with an empty
set of assertions.
If l == 0, then the logic of this context is used in the new context.
If p == 0, then this->m_params is used
*/
context * mk_fresh(symbol const * l = nullptr, smt_params * smtp = nullptr, params_ref const & p = params_ref());
static void copy(context& src, context& dst, bool override_base = false);
/**
\brief Translate context to use new manager m.
*/
app * mk_eq_atom(expr * lhs, expr * rhs);
bool set_logic(symbol const& logic) { return m_setup.set_logic(logic); }
void register_plugin(theory * th);
void add_asserted(expr* e);
void assert_expr(expr * e);
void assert_expr(expr * e, proof * pr);
void internalize_assertions();
void push();
void pop(unsigned num_scopes);
lbool check(unsigned num_assumptions = 0, expr * const * assumptions = nullptr, bool reset_cancel = true);
lbool check(expr_ref_vector const& cube, vector<expr_ref_vector> const& clauses);
lbool get_consequences(expr_ref_vector const& assumptions, expr_ref_vector const& vars, expr_ref_vector& conseq, expr_ref_vector& unfixed);
lbool find_mutexes(expr_ref_vector const& vars, vector<expr_ref_vector>& mutexes);
lbool preferred_sat(expr_ref_vector const& asms, vector<expr_ref_vector>& cores);
lbool setup_and_check(bool reset_cancel = true);
void reduce_assertions();
bool resource_limits_exceeded();
failure get_last_search_failure() const;
proof * get_proof();
conflict_resolution& get_cr() { return *m_conflict_resolution.get(); }
void get_relevant_labels(expr* cnstr, buffer<symbol> & result);
void get_relevant_labeled_literals(bool at_lbls, expr_ref_vector & result);
void get_relevant_literals(expr_ref_vector & result);
void get_guessed_literals(expr_ref_vector & result);
void internalize_assertion(expr * n, proof * pr, unsigned generation);
void internalize_proxies(expr_ref_vector const& asms, vector<std::pair<expr*,expr_ref>>& asm2proxy);
void internalize_instance(expr * body, proof * pr, unsigned generation) {
internalize_assertion(body, pr, generation);
if (relevancy())
m_case_split_queue->internalize_instance_eh(body, generation);
}
unsigned get_unsat_core_size() const {
return m_unsat_core.size();
}
expr * get_unsat_core_expr(unsigned idx) const {
return m_unsat_core.get(idx);
}
expr_ref_vector const& unsat_core() const { return m_unsat_core; }
void get_levels(ptr_vector<expr> const& vars, unsigned_vector& depth);
expr_ref_vector get_trail(unsigned max_level);
void get_model(model_ref & m);
void set_model(model* m) { m_model = m; }
bool update_model(bool refinalize);
bool validate_model();
unsigned get_num_asserted_formulas() const { return m_asserted_formulas.get_num_formulas(); }
unsigned get_asserted_formulas_last_level() const { return m_asserted_formulas.get_formulas_last_level(); }
expr * get_asserted_formula(unsigned idx) const { return m_asserted_formulas.get_formula(idx); }
proof * get_asserted_formula_proof(unsigned idx) const { return m_asserted_formulas.get_formula_proof(idx); }
void get_asserted_formulas(ptr_vector<expr>& r) const { m_asserted_formulas.get_assertions(r); }
//proof * const * get_asserted_formula_proofs() const { return m_asserted_formulas.get_formula_proofs(); }
void get_assertions(ptr_vector<expr> & result) { m_asserted_formulas.get_assertions(result); }
void get_units(expr_ref_vector& result);
bool clause_proof_active() const { return m_clause_proof.is_enabled(); }
clause_proof& get_clause_proof() { return m_clause_proof; }
void register_on_clause(void* ctx, user_propagator::on_clause_eh_t& on_clause) {
m_clause_proof.register_on_clause(ctx, on_clause);
}
/*
* user-propagator
*/
void user_propagate_init(
void* ctx,
user_propagator::push_eh_t& push_eh,
user_propagator::pop_eh_t& pop_eh,
user_propagator::fresh_eh_t& fresh_eh);
void user_propagate_register_final(user_propagator::final_eh_t& final_eh) {
if (!m_user_propagator)
throw default_exception("user propagator must be initialized");
m_user_propagator->register_final(final_eh);
}
void user_propagate_register_fixed(user_propagator::fixed_eh_t& fixed_eh) {
if (!m_user_propagator)
throw default_exception("user propagator must be initialized");
m_user_propagator->register_fixed(fixed_eh);
}
void user_propagate_register_eq(user_propagator::eq_eh_t& eq_eh) {
if (!m_user_propagator)
throw default_exception("user propagator must be initialized");
m_user_propagator->register_eq(eq_eh);
}
void user_propagate_register_diseq(user_propagator::eq_eh_t& diseq_eh) {
if (!m_user_propagator)
throw default_exception("user propagator must be initialized");
m_user_propagator->register_diseq(diseq_eh);
}
void user_propagate_register_expr(expr* e) {
if (!m_user_propagator)
throw default_exception("user propagator must be initialized");
m_user_propagator->add_expr(e, true);
}
void user_propagate_register_created(user_propagator::created_eh_t& r) {
if (!m_user_propagator)
throw default_exception("user propagator must be initialized");
m_user_propagator->register_created(r);
}
void user_propagate_register_decide(user_propagator::decide_eh_t& r) {
if (!m_user_propagator)
throw default_exception("user propagator must be initialized");
m_user_propagator->register_decide(r);
}
void user_propagate_register_on_binding(user_propagator::binding_eh_t& t) {
m_user_propagator->register_on_binding(t);
}
void register_on_binding(std::function<bool(quantifier* q, expr* inst)>& f) {
m_qmanager->register_on_binding(f);
}
void user_propagate_initialize_value(expr* var, expr* value);
bool watches_fixed(enode* n) const;
bool has_split_candidate(bool_var& var, bool& is_pos);
bool decide_user_interference(bool_var& var, bool& is_pos);
void assign_fixed(enode* n, expr* val, unsigned sz, literal const* explain);
void assign_fixed(enode* n, expr* val, literal_vector const& explain) {
assign_fixed(n, val, explain.size(), explain.data());
}
void assign_fixed(enode* n, expr* val, literal explain) {
assign_fixed(n, val, 1, &explain);
}
bool is_fixed(enode* n, expr_ref& val, literal_vector& explain);
void display(std::ostream & out) const;
void display_unsat_core(std::ostream & out) const;
void collect_statistics(::statistics & st) const;
void display_statistics(std::ostream & out) const;
void display_istatistics(std::ostream & out) const;
// -----------------------------------
//
// Macros
//
// -----------------------------------
public:
unsigned get_num_macros() const { return m_asserted_formulas.get_num_macros(); }
unsigned get_first_macro_last_level() const { return m_asserted_formulas.get_first_macro_last_level(); }
func_decl * get_macro_func_decl(unsigned i) const { return m_asserted_formulas.get_macro_func_decl(i); }
func_decl * get_macro_interpretation(unsigned i, expr_ref & interp) const { return m_asserted_formulas.get_macro_interpretation(i, interp); }
quantifier * get_macro_quantifier(func_decl * f) const { return m_asserted_formulas.get_macro_quantifier(f); }
void insert_macro(func_decl * f, quantifier * m, proof * pr, expr_dependency * dep) { m_asserted_formulas.insert_macro(f, m, pr, dep); }
};
struct pp_lit {
context & ctx;
literal lit;
pp_lit(context & ctx, literal lit) : ctx(ctx), lit(lit) {}
};
inline std::ostream & operator<<(std::ostream & out, pp_lit const & pp) {
return pp.ctx.display_detailed_literal(out, pp.lit);
}
struct pp_lits {
context & ctx;
literal const *lits;
unsigned len;
pp_lits(context & ctx, unsigned len, literal const *lits) : ctx(ctx), lits(lits), len(len) {}
pp_lits(context & ctx, literal_vector const& ls) : ctx(ctx), lits(ls.data()), len(ls.size()) {}
};
inline std::ostream & operator<<(std::ostream & out, pp_lits const & pp) {
out << "{";
bool first = true;
for (unsigned i = 0; i < pp.len; ++i) {
if (first) { first = false; } else { out << " or\n"; }
pp.ctx.display_detailed_literal(out, pp.lits[i]);
}
return out << "}";
}
struct enode_eq_pp {
context const& ctx;
enode_pair const& p;
enode_eq_pp(enode_pair const& p, context const& ctx): ctx(ctx), p(p) {}
};
std::ostream& operator<<(std::ostream& out, enode_eq_pp const& p);
std::ostream& operator<<(std::ostream& out, enode_pp const& p);
};