@@ -231,11 +231,29 @@ protected theorem div_le_iff' {x y z : ℝ≥0∞} (h1 : y ≠ 0) (h2 : y ≠
231231protected theorem mul_inv {a b : ℝ≥0 ∞} (ha : a ≠ 0 ∨ b ≠ ∞) (hb : a ≠ ∞ ∨ b ≠ 0 ) :
232232 (a * b)⁻¹ = a⁻¹ * b⁻¹ := by
233233 cases b
234- case top => simp_all only [Ne, not_true_eq_false, or_false, top_ne_zero, not_false_eq_true,
235- or_true, mul_top, inv_top, mul_zero]
234+ case top =>
235+ #adaptation_note
236+ /--
237+ It may be possible to change this back to
238+ `grind [mul_top, mul_zero, inv_top, ENNReal.inv_eq_zero]`
239+ on nightly-2025-12-14 (or ideally without needing the `mul_zero`),
240+ but if not just remove this note.
241+ The proof by `grind` is a hack, doing arithmetic reasoning (i.e. `mul_zero`) by e-matching.
242+ -/
243+ simp_all only [Ne, not_true_eq_false, or_false, top_ne_zero, not_false_eq_true, or_true,
244+ mul_top, inv_top, mul_zero]
236245 cases a
237- case top => simp_all only [Ne, top_ne_zero, not_false_eq_true, coe_ne_top, or_self,
238- not_true_eq_false, coe_eq_zero, false_or, top_mul, inv_top, zero_mul]
246+ case top =>
247+ #adaptation_note
248+ /--
249+ It may be possible to change this back to
250+ `grind [top_mul, zero_mul, inv_top, ENNReal.inv_eq_zero]`
251+ on nightly-2025-12-14 (or ideally without needing the `zero_mul`),
252+ but if not just remove this note.
253+ The proof by `grind` is a hack, doing arithmetic reasoning (i.e. `zero_mul`) by e-matching.
254+ -/
255+ simp_all only [Ne, top_ne_zero, not_false_eq_true, coe_ne_top, or_self, not_true_eq_false,
256+ coe_eq_zero, false_or, top_mul, inv_top, zero_mul]
239257 grind [_=_ coe_mul, coe_zero, inv_zero, = mul_inv, coe_ne_top, ENNReal.inv_eq_zero,
240258 =_ coe_inv, zero_mul, = mul_eq_zero, mul_top, mul_zero, top_mul]
241259
0 commit comments