08eb78a5b2
This PR sets up the new integrated test/bench suite. It then migrates all benchmarks and some related tests to the new suite. There's also some documentation and some linting. For now, a lot of the old tests are left alone so this PR doesn't become even larger than it already is. Eventually, all tests should be migrated to the new suite though so there isn't a confusing mix of two systems.
47 lines
1.2 KiB
Text
47 lines
1.2 KiB
Text
/-!
|
|
# Issue #12242: grind fails when hypotheses contain metavariables
|
|
|
|
When `refine` introduces metavariables that appear in both hypotheses and the target,
|
|
`grind` should be able to close the goal. Previously, `abstractMVars` only abstracted
|
|
metavariables in the target, creating a disconnect between the target (using an fvar)
|
|
and hypotheses (still using the mvar).
|
|
-/
|
|
|
|
-- Original example from the issue
|
|
theorem grind_mvar_fail_2 (a res : Int) :
|
|
∃ x : Int,
|
|
(a = x) →
|
|
(x = res) →
|
|
(x = res) := by
|
|
refine ⟨?_, ?_⟩
|
|
rotate_left
|
|
intro h h2
|
|
grind
|
|
repeat constructor
|
|
|
|
-- Simpler version: mvar in hypothesis directly proves goal
|
|
theorem grind_mvar_simple (a : Nat) :
|
|
∃ x : Nat, (x = a) → (x = a) := by
|
|
refine ⟨?_, ?_⟩
|
|
rotate_left
|
|
intro h
|
|
grind
|
|
exact a
|
|
|
|
-- Multiple hypotheses sharing the same mvar
|
|
theorem grind_mvar_chain (a b c : Nat) :
|
|
∃ x : Nat, (a = x) → (x = b) → (a = b) := by
|
|
refine ⟨?_, ?_⟩
|
|
rotate_left
|
|
intro h1 h2
|
|
grind
|
|
exact a
|
|
|
|
-- Mvar in a more complex expression
|
|
theorem grind_mvar_complex (a b : Int) :
|
|
∃ x : Int, (a = x) → (x = b) → (a = b) := by
|
|
refine ⟨?_, ?_⟩
|
|
rotate_left
|
|
intro h1 h2
|
|
grind
|
|
exact a
|