refactor(library/tactic): reorganize tactic API, add assumption_tactic

Signed-off-by: Leonardo de Moura <leonardo@microsoft.com>

src/library/tactic/goal.h

@@ -18,9 +18,11 @@ class goal {
     list<std::pair<name, expr>> m_hypotheses;
     expr                        m_conclusion;
 public:
+    goal() {}
     goal(list<std::pair<name, expr>> const & h, expr const & c);
     list<std::pair<name, expr>> const & get_hypotheses() const { return m_hypotheses; }
     expr const & get_conclusion() const { return m_conclusion; }
+    operator bool() const { return m_conclusion; }
 };
 
 /**

src/library/tactic/proof_state.cpp

@@ -10,7 +10,6 @@ Author: Leonardo de Moura
 
 namespace lean {
 void swap(proof_state & s1, proof_state & s2) {
-    swap(s1.m_env, s2.m_env);
     swap(s1.m_goals, s2.m_goals);
     swap(s1.m_menv, s2.m_menv);
     swap(s1.m_proof_builder, s2.m_proof_builder);
@@ -20,22 +19,22 @@ void swap(proof_state & s1, proof_state & s2) {
 static name g_main("main");
 
 proof_state to_proof_state(environment const & env, context const & ctx, expr const & t) {
-    auto p           = to_goal(env, ctx, t);
+    auto gfn         = to_goal(env, ctx, t);
-    goal g           = p.first;
+    goal g           = gfn.first;
-    goal_proof_fn fn = p.second;
+    goal_proof_fn fn = gfn.second;
-    return proof_state(env,
+    proof_builder p  = mk_proof_builder(
-                       goals(mk_pair(g_main, g)),
+        [=](proof_map const & m, environment const &, assignment const &) -> expr {
-                       metavar_env(),
+            expr p = find(m, g_main);
-                       mk_proof_builder([=](proof_map const & m, environment const &, assignment const &) -> expr {
+            if (!p)
-                               expr p = find(m, g_main);
+                throw tactic_exception(sstream() << "failed to find proof for '" << g_main << "' goal");
-                               if (!p)
+            return fn(p);
-                                   throw tactic_exception(sstream() << "failed to find proof for '" << g_main << "' goal");
+        });
-                               return fn(p);
+    justification_builder j = mk_justification_builder(
-                           }),
+        [](name const & n, justification const & j, environment const &, assignment const &) -> justification {
-                       mk_justification_builder([](name const & n, justification const & j, environment const &, assignment const &) -> justification {
+            if (n != g_main)
-                               if (n != g_main)
+                throw tactic_exception(sstream() << "unknown goal name '" << n << "'");
-                                   throw tactic_exception(sstream() << "unknown goal name '" << n << "'");
+            return j;
-                               return j;
+        });
-                           }));
+    return proof_state(goals(mk_pair(g_main, g)), metavar_env(), p, j);
 }
 }

src/library/tactic/proof_state.h

@@ -13,17 +13,17 @@ Author: Leonardo de Moura
 namespace lean {
 typedef list<std::pair<name, goal>> goals;
 class proof_state {
-    environment                 m_env;
     goals                       m_goals;
     metavar_env                 m_menv;
     proof_builder               m_proof_builder;
     justification_builder       m_justification_builder;
 public:
     proof_state() {}
-    proof_state(environment const & env, list<std::pair<name, goal>> const & gs, metavar_env const & menv, proof_builder const & p, justification_builder const & j):
+    proof_state(list<std::pair<name, goal>> const & gs, metavar_env const & menv, proof_builder const & p, justification_builder const & j):
-        m_env(env), m_goals(gs), m_menv(menv), m_proof_builder(p), m_justification_builder(j) {}
+        m_goals(gs), m_menv(menv), m_proof_builder(p), m_justification_builder(j) {}
+    proof_state(proof_state const & s, goals const & gs, proof_builder const & p):
+        m_goals(gs), m_menv(s.m_menv), m_proof_builder(p), m_justification_builder(s.m_justification_builder) {}
     friend void swap(proof_state & s1, proof_state & s2);
-    environment const & get_environment() const { return m_env; }
     list<std::pair<name, goal>> const & get_goals() const { return m_goals; }
     metavar_env const & get_menv() const { return m_menv; }
     proof_builder const & get_proof_builder() const { return m_proof_builder; }
@@ -32,4 +32,18 @@ public:
 void swap(proof_state & s1, proof_state & s2);
 
 proof_state to_proof_state(environment const & env, context const & ctx, expr const & t);
+
+template<typename F>
+goals map_goals(proof_state const & s, F && f) {
+    return map_filter(s.get_goals(), [=](std::pair<name, goal> const & in, std::pair<name, goal> & out) -> bool {
+            goal new_goal = f(in.first, in.second);
+            if (new_goal) {
+                out.first  = in.first;
+                out.second = new_goal;
+                return true;
+            } else {
+                return false;
+            }
+        });
+}
 }

src/library/tactic/tactic.cpp

@@ -4,17 +4,37 @@ Released under Apache 2.0 license as described in the file LICENSE.
 
 Author: Leonardo de Moura
 */
+#include <utility>
+#include "util/interrupt.h"
+#include "library/tactic/tactic_exception.h"
 #include "library/tactic/tactic.h"
 
 namespace lean {
 tactic_result::~tactic_result() {
 }
 void tactic_result::interrupt() {
-    m_result = true;
+    m_interrupted = true;
 }
 void tactic_result::request_interrupt() {
-    std::lock_guard<std::mutex> lock(m_mutex);
+    {
-    interrupt();
+        std::lock_guard<std::mutex> lock(m_mutex);
+        interrupt();
+    }
+    ::lean::request_interrupt();
+}
+
+proof_state_ref tactic_result::next(environment const & env, state const & s) {
+    try {
+        return next_core(env, s);
+    } catch (interrupted &) {
+        if (m_interrupted) {
+            // promote to tactic_exception if tactic is marked as interrupted
+            throw tactic_exception("interrupted");
+        } else {
+            // if tactic was not interrupted, then continue propagating interrupted exception
+            throw;
+        }
+    }
 }
 
 tactic_cell::~tactic_cell() {
@@ -28,25 +48,6 @@ tactic & tactic::operator=(tactic && s) {
     LEAN_MOVE_REF(tactic, s);
 }
 
-class id_tactic : public tactic_cell {
-public:
-    class result : public tactic_result {
-        proof_state_ref m_r;
-    public:
-        result(proof_state const & s):m_r(new proof_state(s)) {}
-
-        virtual proof_state_ref next() {
-            return proof_state_ref(std::move(m_r));
-        }
-    };
-
-    virtual tactic_result_ref operator()(proof_state const & s) const {
-        return tactic_result_ref(new result(s));
-    }
-};
-
-tactic idtac() { return tactic(new id_tactic()); }
-
 class then_tactic : public tactic_cell {
     tactic m_t1;
     tactic m_t2;
@@ -67,9 +68,9 @@ public:
     public:
         result(tactic_result_ref && r1, tactic const & t2):m_r1(std::move(r1)), m_t2(t2) {}
 
-        virtual proof_state_ref next() {
+        virtual proof_state_ref next_core(environment const & env, state const & s) {
             if (m_r2) {
-                proof_state_ref s2 = m_r2->next();
+                proof_state_ref s2 = m_r2->next(env, s);
                 if (s2)
                     return s2;
                 m_r2.release();
@@ -77,20 +78,64 @@ public:
             }
             lean_assert(!m_r2);
             lean_assert(!m_s1);
-            proof_state_ref new_s1(m_r1->next());
+            proof_state_ref new_s1(m_r1->next(env, s));
             if (!new_s1)
                 return proof_state_ref(); // done, m_r1 has no more solutions
             m_s1.swap(new_s1);
             tactic_result_ref r2 = m_t2(proof_state(*m_s1));
             exclusive_update([&]() { m_r2.swap(r2); }); // must protect because interrupt() may be invoked from different thread
-            return m_r2->next();
+            return m_r2->next(env, s);
         }
     };
 
-    virtual tactic_result_ref operator()(proof_state const & s) const {
+    virtual tactic_result_ref operator()(proof_state const & s) {
         return tactic_result_ref(new result(m_t1(s), m_t2));
     }
 };
 
 tactic then(tactic const & t1, tactic const & t2) { return tactic(new then_tactic(t1, t2)); }
+
+tactic id_tactic() { return mk_tactic([](environment const &, state const &, proof_state const & s) -> proof_state { return s; }); }
+
+tactic fail_tactic() { return mk_tactic([](environment const &, state const &, proof_state const &) -> proof_state { throw tactic_exception("failed"); }); }
+
+tactic now_tactic() {
+    return mk_tactic([](environment const &, state const &, proof_state const & s) -> proof_state {
+            if (!empty(s.get_goals()))
+                throw tactic_exception("nowtac failed");
+            return s;
+        });
+}
+
+tactic assumption_tactic() {
+    return mk_tactic([](environment const &, state const &, proof_state const & s) -> proof_state {
+            list<std::pair<name, expr>> proofs;
+            goals new_goals = map_goals(s, [&](name const & ng, goal const & g) -> goal {
+                    expr const & c  = g.get_conclusion();
+                    expr pr;
+                    for (auto const & p : g.get_hypotheses()) {
+                        check_interrupted();
+                        if (p.second == c) {
+                            pr = mk_constant(p.first);
+                            break;
+                        }
+                    }
+                    if (pr) {
+                        proofs.emplace_front(ng, pr);
+                        return goal();
+                    } else {
+                        return g;
+                    }
+                });
+            proof_builder p     = s.get_proof_builder();
+            proof_builder new_p = mk_proof_builder([=](proof_map const & m, environment const & env, assignment const & a) -> expr {
+                    proof_map new_m(m);
+                    for (auto const & np : proofs) {
+                        new_m.insert(np.first, np.second);
+                    }
+                    return p(new_m, env, a);
+                });
+            return proof_state(s, new_goals, new_p);
+        });
+}
 }

src/library/tactic/tactic.h

@@ -8,6 +8,7 @@ Author: Leonardo de Moura
 #include <algorithm>
 #include <memory>
 #include <mutex>
+#include "library/state.h"
 #include "library/tactic/proof_state.h"
 
 namespace lean {
@@ -16,8 +17,8 @@ class tactic_result;
 typedef std::unique_ptr<tactic_result> tactic_result_ref;
 
 class tactic_result {
-    std::mutex    m_mutex;
+    std::mutex        m_mutex;
-    bool          m_result;
+    std::atomic<bool> m_interrupted;
 protected:
     template<typename F>
     void exclusive_update(F && f) {
@@ -26,12 +27,12 @@ protected:
     }
     virtual void interrupt();
     void propagate_interrupt(tactic_result_ref & r) { if (r) r->interrupt(); }
+    virtual proof_state_ref next_core(environment const & env, state const & s) = 0;
 public:
-    tactic_result():m_result(false) {}
+    tactic_result():m_interrupted(false) {}
-    bool interrupted() const { return m_result; }
     void request_interrupt();
+    proof_state_ref next(environment const & env, state const & s);
     virtual ~tactic_result();
-    virtual proof_state_ref next() = 0;
 };
 
 class tactic_cell {
@@ -39,7 +40,7 @@ class tactic_cell {
     MK_LEAN_RC();
 public:
     virtual ~tactic_cell();
-    virtual tactic_result_ref operator()(proof_state const & s) const = 0;
+    virtual tactic_result_ref operator()(proof_state const & s) = 0;
 };
 
 class tactic {
@@ -54,9 +55,47 @@ public:
     tactic & operator=(tactic const & s);
     tactic & operator=(tactic && s);
 
-    tactic_result_ref operator()(proof_state const & s) const { return m_ptr->operator()(s); }
+    tactic_result_ref operator()(proof_state const & s) { return m_ptr->operator()(s); }
 };
 
-tactic idtac();
+template<typename F>
+class simple_tactic_cell : public tactic_cell {
+    F m_f;
+public:
+    simple_tactic_cell(F && f):m_f(f) {}
+
+    class result : public tactic_result {
+        simple_tactic_cell * m_cell;
+        proof_state          m_in;
+    public:
+        result(simple_tactic_cell * c, proof_state const & in):m_cell(c), m_in(in) {
+            lean_assert(m_cell);
+            m_cell->inc_ref();
+        }
+        virtual ~result() { if (m_cell) m_cell->dec_ref(); }
+        virtual proof_state_ref next_core(environment const & env, state const & io) {
+            if (m_cell) {
+                proof_state_ref r(new proof_state(m_cell->m_f(env, io, m_in)));
+                m_cell = nullptr;
+                return std::move(r);
+            } else {
+                return proof_state_ref();
+            }
+        }
+    };
+
+    virtual tactic_result_ref operator()(proof_state const & s) {
+        return tactic_result_ref(new result(this, s));
+    }
+};
+
+
+template<typename F>
+tactic mk_tactic(F && f) { return tactic(new simple_tactic_cell<F>(std::forward<F>(f))); }
+
+tactic id_tactic();
+tactic fail_tactic();
+tactic now_tactic();
+tactic assumption_tactic();
 tactic then(tactic const & t1, tactic const & t2);
 }

src/tests/library/tactic/tactic.cpp

@@ -5,6 +5,8 @@ Released under Apache 2.0 license as described in the file LICENSE.
 Author: Leonardo de Moura
 */
 #include "util/test.h"
+#include "kernel/builtin.h"
+#include "library/all/all.h"
 #include "library/tactic/goal.h"
 #include "library/tactic/proof_builder.h"
 #include "library/tactic/justification_builder.h"
@@ -12,6 +14,23 @@ Author: Leonardo de Moura
 #include "library/tactic/tactic.h"
 using namespace lean;
 
+static void tst1() {
+    environment env;
+    import_all(env);
+    env.add_var("p", Bool);
+    env.add_var("q", Bool);
+    expr p = Const("p");
+    expr q = Const("q");
+    context ctx;
+    ctx = extend(ctx, "H1", p);
+    ctx = extend(ctx, "H2", q);
+    // proof_state s1 = to_proof_state(env, ctx, mk_var(0));
+    // tactic t = then(assumption_tactic(), now_tactic());
+    // tactic_result_ref r = t(s1);
+    // proof_state_ref s2 = r->next();
+}
+
 int main() {
+    tst1();
     return has_violations() ? 1 : 0;
 }