diff --git a/Cargo.toml b/Cargo.toml
index 26b0f6d8..f1036157 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -8,7 +8,7 @@ categories = [ "algorithms", "data-structures", "science" ]
 license = "MIT OR Apache-2.0"
 name = "num-bigint"
 repository = "https://github.com/rust-num/num-bigint"
-version = "0.4.1"
+version = "0.4.2"
 readme = "README.md"
 build = "build.rs"
 exclude = ["/bors.toml", "/ci/*", "/.github/*"]
diff --git a/RELEASES.md b/RELEASES.md
index d6469a35..b06f4fa8 100644
--- a/RELEASES.md
+++ b/RELEASES.md
@@ -1,3 +1,11 @@
+# Release 0.4.2 (2021-09-03)
+
+- [Use explicit `Integer::div_ceil` to avoid the new unstable method.][219]
+
+**Contributors**: @catenacyber, @cuviper
+
+[219]: https://github.com/rust-num/num-bigint/pull/219
+
 # Release 0.4.1 (2021-08-27)
 
 - [Fixed scalar divide-by-zero panics.][200]
diff --git a/src/biguint.rs b/src/biguint.rs
index 271a8837..623823c8 100644
--- a/src/biguint.rs
+++ b/src/biguint.rs
@@ -395,7 +395,7 @@ impl Roots for BigUint {
                 // Try to guess by scaling down such that it does fit in `f64`.
                 // With some (x * 2ⁿᵏ), its nth root ≈ (ⁿ√x * 2ᵏ)
                 let extra_bits = bits - (core::f64::MAX_EXP as u64 - 1);
-                let root_scale = extra_bits.div_ceil(&n64);
+                let root_scale = Integer::div_ceil(&extra_bits, &n64);
                 let scale = root_scale * n64;
                 if scale < bits && bits - scale > n64 {
                     (self >> scale).nth_root(n) << root_scale
diff --git a/src/biguint/convert.rs b/src/biguint/convert.rs
index edeed84c..5cf05cb6 100644
--- a/src/biguint/convert.rs
+++ b/src/biguint/convert.rs
@@ -65,9 +65,8 @@ fn from_inexact_bitwise_digits_le(v: &[u8], bits: u8) -> BigUint {
     debug_assert!(!v.is_empty() && bits <= 8 && big_digit::BITS % bits != 0);
     debug_assert!(v.iter().all(|&c| BigDigit::from(c) < (1 << bits)));
 
-    let big_digits = (v.len() as u64)
-        .saturating_mul(bits.into())
-        .div_ceil(&big_digit::BITS.into())
+    let total_bits = (v.len() as u64).saturating_mul(bits.into());
+    let big_digits = Integer::div_ceil(&total_bits, &big_digit::BITS.into())
         .to_usize()
         .unwrap_or(core::usize::MAX);
     let mut data = Vec::with_capacity(big_digits);
@@ -580,9 +579,7 @@ pub(super) fn to_bitwise_digits_le(u: &BigUint, bits: u8) -> Vec<u8> {
     let last_i = u.data.len() - 1;
     let mask: BigDigit = (1 << bits) - 1;
     let digits_per_big_digit = big_digit::BITS / bits;
-    let digits = u
-        .bits()
-        .div_ceil(&u64::from(bits))
+    let digits = Integer::div_ceil(&u.bits(), &u64::from(bits))
         .to_usize()
         .unwrap_or(core::usize::MAX);
     let mut res = Vec::with_capacity(digits);
@@ -608,9 +605,7 @@ fn to_inexact_bitwise_digits_le(u: &BigUint, bits: u8) -> Vec<u8> {
     debug_assert!(!u.is_zero() && bits <= 8 && big_digit::BITS % bits != 0);
 
     let mask: BigDigit = (1 << bits) - 1;
-    let digits = u
-        .bits()
-        .div_ceil(&u64::from(bits))
+    let digits = Integer::div_ceil(&u.bits(), &u64::from(bits))
         .to_usize()
         .unwrap_or(core::usize::MAX);
     let mut res = Vec::with_capacity(digits);