This module contains helpers useful to prove unfolding and/or equational theorems.
Simplify if-then-expressions in the goal target.
If useNewSemantics is true, the flag backward.split is ignored.
Equations
Instances For
Try to close goal using rfl with smart unfolding turned off.
If both sides are applications of the same function in declNames, the proof is Eq.refl of the
left-hand side unfolded once, instead of Eq.refl lhs. Example (#15987):
def f : Nat → Nat → Nat
| 0, x => x
| n+1, x => f n ((x * 1103515245) % 5)
theorem f.eq_2 (x n : Nat) : f (n+1) x = f n (x * 1103515245 % 5)
To check Eq.refl (f (n+1) x), the kernel must show f (n+1) x =?= f n (x * 1103515245 % 5).
Since both sides are applications of f, it first tries to show that the arguments are
definitionally equal, starting with the last one: x =?= x * 1103515245 % 5. This forces it to
reduce x * 1103515245 with a free x, which unfolds Nat.mul once per unit of 1103515245, and
fails with deep recursion. That comparison can never succeed anyway, since a recursive call is on
a smaller argument (here n vs n+1). The proof Eq.refl (Nat.brecOn (n+1) f._f x) avoids it:
the left-hand side is no longer an application of f, so the kernel unfolds both sides instead.
Equations
- One or more equations did not get rendered due to their size.
Instances For
Delta reduce the equation left-hand-side
Equations
- One or more equations did not get rendered due to their size.
Instances For
Equations
- Lean.Elab.Eqns.tryContradiction mvarId = mvarId.contradictionCore { genDiseq := true }