FSet/OrderedType now includes an eq_dec, and hence become an extension of DecidableType

 After lots of hesitations, OrderedType now requires this "eq_dec" field, which
 is redundant (can be deduced from "compare"), but allows the subtyping relation
 DecidableType <= OrderedType, and hence WS <= S : ordered sets are now truly
 extensions of weak sets. Of course this change introduces a last-minute
 incompatibility, but:

  - There is a clear gain in term of functionnality / simplicity.
  - FSets 8.2 already needs some adaptations when compared with 8.1, so it's
    the right time to push such incompatible changes.
  - Transition shouldn't be too hard: the old OrderedType still exists under
    the name MiniOrderedType, and functor MOT_to_OT allows to convert from
    one to the other.

 Beware, for a FSetInterface.WS (resp. S) to be coercible to a DecidableType
 (resp. OrderedType), an eq_dec on sets is now required in these interfaces
 and in the implementations. In pratice, it is really easy to build from
 equal and equal_1 and equal_2.

 Some name changes : in FSetFacts, old WFacts now correspond to WFacts_fun,
 while WFacts now expects only one argument (WFacts M := WFacts_fun M.E M).
 Idem with WDecide, WProperties and WEqProperties.

git-svn-id: svn+ssh://scm.gforge.inria.fr/svn/coq/trunk@11693 85f007b7-540e-0410-9357-904b9bb8a0f7

1 files changed, 1 insertions, 1 deletions

diff --git a/theories/FSets/FMapList.v b/theories/FSets/FMapList.v
index e2ea68794..a99c6a908 100644
--- a/theories/FSets/FMapList.v
+++ b/theories/FSets/FMapList.v
@@ -402,7 +402,7 @@ Proof.
  elim (Sort_Inf_NotIn H6 H7).
  destruct H as (e'', hyp); exists e''; auto.
  apply MapsTo_eq with k; auto; order.
- apply H1 with k; destruct (eq_dec x k); auto.
+ apply H1 with k; destruct (X.eq_dec x k); auto.
 
 
  destruct (X.compare x x'); try contradiction; clear y.