Merge branch 'master' of gitlab.mpi-sws.org:FP/iris-coq

algebra/dra.v

@@ -131,10 +131,24 @@ Proof.
       intuition eauto 10 using dra_disjoint_minus, dra_op_minus.
 Qed.
 Definition validityRA : cmraT := discreteRA validity_ra.
-Definition validity_update (x y : validityRA) :
+Lemma validity_update (x y : validityRA) :
   (∀ z, ✓ x → ✓ z → validity_car x ⊥ z → ✓ y ∧ validity_car y ⊥ z) → x ~~> y.
 Proof.
   intros Hxy. apply discrete_update.
   intros z (?&?&?); split_ands'; try eapply Hxy; eauto.
 Qed.
+
+Lemma to_validity_valid (x : A) :
+  ✓ to_validity x → ✓ x.
+Proof. intros. done. Qed.
+Lemma to_validity_op (x y : A) :
+  ✓ x → ✓ y → (✓ (x ⋅ y) → x ⊥ y) →
+  to_validity (x ⋅ y) ≡ to_validity x ⋅ to_validity y.
+Proof.
+  intros Hvalx Hvaly Hdisj. split; [split | done].
+  - simpl. auto.
+  - simpl. intros (_ & _ & ?).
+    auto using dra_op_valid, to_validity_valid.
+Qed.
+
 End dra.

algebra/sts.v

@@ -333,22 +333,13 @@ Lemma sts_op_auth_frag_up s T :
 Proof. intros; apply sts_op_auth_frag; auto using elem_of_up, closed_up. Qed.
 
 Lemma sts_op_frag S1 S2 T1 T2 :
-  T1 ∪ T2 ⊆ ∅ → sts.closed S1 T1 → sts.closed S2 T2 →
+  T1 ∩ T2 ⊆ ∅ → sts.closed S1 T1 → sts.closed S2 T2 →
   sts_frag (S1 ∩ S2) (T1 ∪ T2) ≡ sts_frag S1 T1 ⋅ sts_frag S2 T2.
 Proof.
-  (* Somehow I feel like a very similar proof muts have happened above, when
-     proving the DRA axioms. After all, we are just reflecting the operation here. *)
-  intros HT HS1 HS2; split; [split|constructor; solve_elem_of]; simpl.
-  - intros; split_ands; try done; []. constructor; last solve_elem_of.
-    by eapply closed_ne.
-  - intros (_ & _ & H). inversion_clear H. constructor; first done.
-    + move=>s /elem_of_intersection
-              [/(closed_disjoint _ _ HS1) Hs1 /(closed_disjoint _ _ HS2) Hs2].
-      solve_elem_of +Hs1 Hs2.
-    + move=> s1 s2 /elem_of_intersection [Hs1 Hs2] Hstep.
-      apply elem_of_intersection.
-      split; eapply closed_step; eauto; [|]; eapply framestep_mono, Hstep;
-        done || solve_elem_of.
+  intros HT HS1 HS2. rewrite /sts_frag.
+  (* FIXME why does rewrite not work?? *)
+  etransitivity; last eapply to_validity_op; try done; [].
+  intros Hval. constructor; last solve_elem_of. eapply closed_ne, Hval.
 Qed.
 
 (** Frame preserving updates *)

program_logic/sts.v

@@ -53,18 +53,19 @@ Section sts.
 
   (* The same rule as implication does *not* hold, as could be shown using
      sts_frag_included. *)
-  Lemma sts_ownS_weaken E γ S1 S2 T :
-    S1 ⊆ S2 → sts.closed S2 T →
-    sts_ownS γ S1 T ⊑ pvs E E (sts_ownS γ S2 T).
-  Proof. intros. by apply own_update, sts_update_frag. Qed.
+  Lemma sts_ownS_weaken E γ S1 S2 T1 T2 :
+    T1 ≡ T2 → S1 ⊆ S2 → sts.closed S2 T2 →
+    sts_ownS γ S1 T1 ⊑ pvs E E (sts_ownS γ S2 T2).
+  Proof. intros -> ? ?. by apply own_update, sts_update_frag. Qed.
 
-  Lemma sts_own_weaken E γ s S T :
-    s ∈ S → sts.closed S T →
-    sts_own γ s T ⊑ pvs E E (sts_ownS γ S T).
-  Proof. intros. by apply own_update, sts_update_frag_up. Qed.
+  Lemma sts_own_weaken E γ s S T1 T2 :
+    T1 ≡ T2 → s ∈ S → sts.closed S T2 →
+    sts_own γ s T1 ⊑ pvs E E (sts_ownS γ S T2).
+  (* FIXME for some reason, using "->" instead of "<-" does not work? *)
+  Proof. intros <- ? ?. by apply own_update, sts_update_frag_up. Qed.
 
   Lemma sts_ownS_op γ S1 S2 T1 T2 :
-    T1 ∪ T2 ⊆ ∅ → sts.closed S1 T1 → sts.closed S2 T2 →
+    T1 ∩ T2 ⊆ ∅ → sts.closed S1 T1 → sts.closed S2 T2 →
     sts_ownS γ (S1 ∩ S2) (T1 ∪ T2) ≡ (sts_ownS γ S1 T1 ★ sts_ownS γ S2 T2)%I.
   Proof. intros. by rewrite /sts_ownS -own_op sts_op_frag. Qed.