2a7b1162af
This PR tries to remove from functional induction principles hypotheses that have been matched, as we expect the corresponding pattern to be more useful. This avoids duplicate hypotheses due to the way `match` refines hypotheses. Fixes #6281.
47 lines
1.3 KiB
Text
47 lines
1.3 KiB
Text
set_option linter.unusedVariables false
|
|
|
|
def bar (n : Nat) : Bool :=
|
|
if h : n = 0 then
|
|
true
|
|
else
|
|
match n with -- NB: the elaborator adds `h` as an discriminant
|
|
| m+1 => bar m
|
|
termination_by n
|
|
|
|
-- set_option pp.match false
|
|
-- #print bar
|
|
-- #check bar.match_1
|
|
-- #print bar.induct
|
|
|
|
-- NB: The induction theorem used to ahve two `h` in scope, as its original type mentioning `x`,
|
|
-- and a refined `h` mentioning `m+1`.
|
|
-- At some point we had a HEq between them, but not any more, thanks to proof irrelevance
|
|
-- Since #7110, we drop the shadowed `h`.
|
|
|
|
/--
|
|
info: bar.induct (motive : Nat → Prop) (case1 : motive 0) (case2 : ∀ (m : Nat), ¬m + 1 = 0 → motive m → motive m.succ)
|
|
(n : Nat) : motive n
|
|
-/
|
|
#guard_msgs in
|
|
#check bar.induct
|
|
|
|
def baz (n : Nat) (i : Fin (n+1)) : Bool :=
|
|
if h : n = 0 then
|
|
true
|
|
else
|
|
match n, i + 1 with
|
|
| 1, _ => true
|
|
| m+2, j => baz (m+1) ⟨j.1-1, by omega⟩
|
|
termination_by n
|
|
|
|
-- #print baz._unary
|
|
|
|
/--
|
|
info: baz.induct (motive : (n : Nat) → Fin (n + 1) → Prop) (case1 : ∀ (i : Fin (0 + 1)), motive 0 i)
|
|
(case2 : ¬1 = 0 → ∀ (i : Fin (1 + 1)), motive 1 i)
|
|
(case3 :
|
|
∀ (m : Nat), ¬m + 2 = 0 → ∀ (i : Fin (m.succ.succ + 1)), motive (m + 1) ⟨↑(i + 1) - 1, ⋯⟩ → motive m.succ.succ i)
|
|
(n : Nat) (i : Fin (n + 1)) : motive n i
|
|
-/
|
|
#guard_msgs in
|
|
#check baz.induct
|