aboutsummaryrefslogtreecommitdiff
diff options
context:
space:
mode:
authorGravatar Jason Gross <jagro@google.com>2016-07-19 17:02:53 -0700
committerGravatar Jason Gross <jagro@google.com>2016-07-19 17:06:15 -0700
commit89b3547b231851c739c58c20e2f19073a5cb4c5b (patch)
treeb8703b75c04b6b010e44cba48195ad2a1db486e1
parent24c557e38fac865e059832a17d6ae1591284bc42 (diff)
Work around bad design in Coq
This is https://coq.inria.fr/bugs/show_bug.cgi?id=4949, [intuition] should not use [auto with *] by default
-rw-r--r--folkwisdom.md2
-rw-r--r--src/ModularArithmetic/ModularBaseSystemProofs.v2
2 files changed, 3 insertions, 1 deletions
diff --git a/folkwisdom.md b/folkwisdom.md
index 3c8d92036..d6d65f12e 100644
--- a/folkwisdom.md
+++ b/folkwisdom.md
@@ -280,6 +280,8 @@ Considerations not (yet) covered in this document include the following:
- how to write tactics that are debuggable
- how to write tactics that are robust against small changes
- reification: ltac, typeclasses, canonical structures (maybe reference an existing document)
+ - how `intuition` should never be used because it calls `auto with *`, which
+ leads to fragile documents, and `intuition auto` should be used instead
- performance of proofs and proof checking
- `simpl`, `cbn`, `cbv`
- `Qed slowness, `change` vs `rewrite`
diff --git a/src/ModularArithmetic/ModularBaseSystemProofs.v b/src/ModularArithmetic/ModularBaseSystemProofs.v
index c3b1b76b4..c06dcdf98 100644
--- a/src/ModularArithmetic/ModularBaseSystemProofs.v
+++ b/src/ModularArithmetic/ModularBaseSystemProofs.v
@@ -127,7 +127,7 @@ Section PseudoMersenneProofs.
Lemma add_rep : forall u v x y, u ~= x -> v ~= y -> BaseSystem.add u v ~= (x+y)%F.
Proof.
- autounfold; intuition. {
+ autounfold; intuition auto. {
unfold add.
auto using add_same_length.
}