“…Take ω i ∈ Γ(T * X i ) and f i ∈ C ∞ (X i ) with T i = ω i + f i e. Since T i ∞ ≤ C(n, K, D, β), we have ω i ∞ ≤ C(n, K, D, β). By [25,Theorem 4.9] and [27,Theorem 6.11] (see also [26,Proposition 4.2]), there exist a subsequence i(j), ω ∈ L 2 (T * X) and f ∈ H 1,2 (X) such that ω is differentiable at almost all point in X, ∇ω ∈ L 2 (T * X ⊗ T * X), ω i , f i converges to ω, f strongly in L 2 and ∇ω i , df i converges to ∇ω, df weakly in L 2 , respectively. By the lower semi-continuity of L ∞ norm [25, Proposition 3.64], we have f ∞ ≤ C(n, K, D, µ), ω ∞ ≤ C(n, K, D, µ) and T ∞ ≤ C(n, K, D, µ).…”