Correct rebase on STM code. Thanks to E. Tassi for help on dealing with

latent universes. Now the universes in the type of a definition/lemma
are eagerly added to the environment so that later proofs can be checked
independently of the original (delegated) proof body.

- Fixed firstorder, ring to work correctly with universe polymorphism.
- Changed constr_of_global to raise an anomaly if side effects would be lost by
turning a polymorphic constant into a constr.
- Fix a non-termination issue in solve_evar_evar.
-

1 files changed, 2 insertions, 1 deletions

diff --git a/plugins/firstorder/ground.ml b/plugins/firstorder/ground.ml
index e0f4fa95f3..3faf7f8021 100644
--- a/plugins/firstorder/ground.ml
+++ b/plugins/firstorder/ground.ml
@@ -119,5 +119,6 @@ let ground_tac solver startseq gl=
 		  end
 	    with Heap.EmptyHeap->solver
       end gl in
-    wrap (List.length (pf_hyps gl)) true (toptac []) (startseq gl) gl
+    let seq, gl' = startseq gl in
+    wrap (List.length (pf_hyps gl)) true (toptac []) seq gl'