chore: bump MSRV to 1.84 (#7926)

mbrobbel · web-flow · commit 9c0cb9a56f00 · 2025-07-15T10:24:15.000+02:00
# Which issue does this PR close? None. # Rationale for this change - This allows us to keep up with dependencies bumping their MSRV (e.g. #7924) - parquet variant crates now use the workspace MSRV - #7395 is the next release and because this is a major release we can bump MSRV now for all the 56.x.y releases We can bump to 1.85 in #7835 to unblock #7270. # What changes are included in this PR? - Bump MSRV to 1.84 which was released more than 6 months ago - Removed half pins from CI # Are these changes tested? CI. # Are there any user-facing changes? Yes.
diff --git a/.github/workflows/rust.yml b/.github/workflows/rust.yml
@@ -30,7 +30,6 @@ on:
   pull_request:
 
 jobs:
-
   # Check workspace wide compile and test with default features for
   # mac
   macos:
@@ -54,7 +53,6 @@ jobs:
           # PyArrow tests happen in integration.yml.
           cargo test --workspace
 
-
   # Check workspace wide compile and test with default features for
   # windows
   windows:
@@ -84,8 +82,7 @@ jobs:
           # do not produce debug symbols to keep memory usage down
           export RUSTFLAGS="-C debuginfo=0"
           export PATH=$PATH:/d/protoc/bin
-          cargo test --workspace 
-
+          cargo test --workspace
 
   # Run cargo fmt for all crates
   lint:
@@ -121,15 +118,6 @@ jobs:
         uses: ./.github/actions/setup-builder
       - name: Install cargo-msrv
         run: cargo install cargo-msrv
-      - name: Downgrade arrow-pyarrow-integration-testing dependencies
-        working-directory: arrow-pyarrow-integration-testing
-        # Necessary because half 2.5 requires rust 1.81 or newer
-        run: |
-          cargo update -p half --precise 2.4.0
-      - name: Downgrade workspace dependencies
-        # Necessary because half 2.5 requires rust 1.81 or newer
-        run: |
-          cargo update -p half --precise 2.4.0
       - name: Check all packages
         run: |
           # run `cargo msrv verify --manifest-path "path/to/Cargo.toml"` to see problematic dependencies
diff --git a/Cargo.toml b/Cargo.toml
@@ -81,7 +81,7 @@ include = [
     "NOTICE.txt",
 ]
 edition = "2021"
-rust-version = "1.81"
+rust-version = "1.84"
 
 [workspace.dependencies]
 arrow = { version = "55.2.0", path = "./arrow", default-features = false }
@@ -102,7 +102,7 @@ arrow-string = { version = "55.2.0", path = "./arrow-string" }
 parquet = { version = "55.2.0", path = "./parquet", default-features = false }
 
 # These crates have not yet been released and thus do not use the workspace version
-parquet-variant = { version = "0.1.0", path = "./parquet-variant"}
+parquet-variant = { version = "0.1.0", path = "./parquet-variant" }
 parquet-variant-json = { version = "0.1.0", path = "./parquet-variant-json" }
 parquet-variant-compute = { version = "0.1.0", path = "./parquet-variant-json" }
 
diff --git a/README.md b/README.md
@@ -79,14 +79,10 @@ Planned Release Schedule
 
 ### Rust Version Compatibility Policy
 
-arrow-rs, parquet and object_store are built and tested with stable Rust, and will keep a rolling MSRV (minimum supported Rust version) that can only be updated in major releases on a need by basis (e.g. project dependencies bump their MSRV or a particular Rust feature is useful for us etc.). The new MSRV if selected will be at least 6 months old. The minor releases are guaranteed to have the same MSRV.
+arrow-rs and parquet are built and tested with stable Rust, and will keep a rolling MSRV (minimum supported Rust version) that can only be updated in major releases on a need by basis (e.g. project dependencies bump their MSRV or a particular Rust feature is useful for us etc.). The new MSRV if selected will be at least 6 months old. The minor releases are guaranteed to have the same MSRV.
 
 Note: If a Rust hotfix is released for the current MSRV, the MSRV will be updated to the specific minor version that includes all applicable hotfixes preceding other policies.
 
-E.g.
-
-in Apr 2025 we will release version 55.0.0 which might have a version bump. But the Rust version selected in this case will be at most version 1.81.
-
 ### Guidelines for `panic` vs `Result`
 
 In general, use panics for bad states that are unreachable, unrecoverable or harmful.
diff --git a/arrow-array/benches/union_array.rs b/arrow-array/benches/union_array.rs
@@ -15,11 +15,7 @@
 // specific language governing permissions and limitations
 // under the License.
 
-use std::{
-    hint,
-    iter::{repeat, repeat_with},
-    sync::Arc,
-};
+use std::{hint, iter::repeat_with, sync::Arc};
 
 use arrow_array::{Array, ArrayRef, Int32Array, UnionArray};
 use arrow_buffer::{NullBuffer, ScalarBuffer};
@@ -67,9 +63,8 @@ fn criterion_benchmark(c: &mut Criterion) {
                         fields,
                         type_ids.cycle().take(4096).collect(),
                         None,
-                        repeat(array_with_nulls())
-                            .take(with_nulls as usize)
-                            .chain(repeat(array_without_nulls()).take(without_nulls as usize))
+                        std::iter::repeat_n(array_with_nulls(), with_nulls as usize)
+                            .chain(std::iter::repeat_n(array_without_nulls(), without_nulls as usize))
                             .collect(),
                     )
                     .unwrap();
diff --git a/arrow-array/src/arithmetic.rs b/arrow-array/src/arithmetic.rs
@@ -420,13 +420,13 @@ native_type_float_op!(
     1.,
     unsafe {
         // Need to allow in clippy because
-        // current MSRV (Minimum Supported Rust Version) is `1.81.0` but this item is stable since `1.87.0`
+        // current MSRV (Minimum Supported Rust Version) is `1.84.0` but this item is stable since `1.87.0`
         #[allow(unnecessary_transmutes)]
         std::mem::transmute(-1_i32)
     },
     unsafe {
         // Need to allow in clippy because
-        // current MSRV (Minimum Supported Rust Version) is `1.81.0` but this item is stable since `1.87.0`
+        // current MSRV (Minimum Supported Rust Version) is `1.84.0` but this item is stable since `1.87.0`
         #[allow(unnecessary_transmutes)]
         std::mem::transmute(i32::MAX)
     }
@@ -437,13 +437,13 @@ native_type_float_op!(
     1.,
     unsafe {
         // Need to allow in clippy because
-        // current MSRV (Minimum Supported Rust Version) is `1.81.0` but this item is stable since `1.87.0`
+        // current MSRV (Minimum Supported Rust Version) is `1.84.0` but this item is stable since `1.87.0`
         #[allow(unnecessary_transmutes)]
         std::mem::transmute(-1_i64)
     },
     unsafe {
         // Need to allow in clippy because
-        // current MSRV (Minimum Supported Rust Version) is `1.81.0` but this item is stable since `1.87.0`
+        // current MSRV (Minimum Supported Rust Version) is `1.84.0` but this item is stable since `1.87.0`
         #[allow(unnecessary_transmutes)]
         std::mem::transmute(i64::MAX)
     }
diff --git a/arrow-array/src/array/list_array.rs b/arrow-array/src/array/list_array.rs
@@ -454,7 +454,7 @@ impl<OffsetSize: OffsetSizeTrait> From<FixedSizeListArray> for GenericListArray<
             _ => unreachable!(),
         };
 
-        let offsets = OffsetBuffer::from_lengths(std::iter::repeat(size).take(value.len()));
+        let offsets = OffsetBuffer::from_lengths(std::iter::repeat_n(size, value.len()));
 
         Self {
             data_type: Self::DATA_TYPE_CONSTRUCTOR(field.clone()),
diff --git a/arrow-array/src/array/list_view_array.rs b/arrow-array/src/array/list_view_array.rs
@@ -475,7 +475,7 @@ impl<OffsetSize: OffsetSizeTrait> From<FixedSizeListArray> for GenericListViewAr
             _ => unreachable!(),
         };
         let mut acc = 0_usize;
-        let iter = std::iter::repeat(size).take(value.len());
+        let iter = std::iter::repeat_n(size, value.len());
         let mut sizes = Vec::with_capacity(iter.size_hint().0);
         let mut offsets = Vec::with_capacity(iter.size_hint().0);
 
diff --git a/arrow-avro/src/reader/record.rs b/arrow-avro/src/reader/record.rs
@@ -338,7 +338,7 @@ impl Decoder {
                 moff.push_length(0);
             }
             Self::Fixed(sz, accum) => {
-                accum.extend(std::iter::repeat(0u8).take(*sz as usize));
+                accum.extend(std::iter::repeat_n(0u8, *sz as usize));
             }
             Self::Decimal128(_, _, _, builder) => builder.append_value(0),
             Self::Decimal256(_, _, _, builder) => builder.append_value(i256::ZERO),
diff --git a/arrow-buffer/src/buffer/immutable.rs b/arrow-buffer/src/buffer/immutable.rs
@@ -997,13 +997,13 @@ mod tests {
     #[should_panic(expected = "capacity overflow")]
     fn test_from_iter_overflow() {
         let iter_len = usize::MAX / std::mem::size_of::<u64>() + 1;
-        let _ = Buffer::from_iter(std::iter::repeat(0_u64).take(iter_len));
+        let _ = Buffer::from_iter(std::iter::repeat_n(0_u64, iter_len));
     }
 
     #[test]
     fn bit_slice_length_preserved() {
         // Create a boring buffer
-        let buf = Buffer::from_iter(std::iter::repeat(true).take(64));
+        let buf = Buffer::from_iter(std::iter::repeat_n(true, 64));
 
         let assert_preserved = |offset: usize, len: usize| {
             let new_buf = buf.bit_slice(offset, len);
@@ -1035,7 +1035,7 @@ mod tests {
 
     #[test]
     fn test_strong_count() {
-        let buffer = Buffer::from_iter(std::iter::repeat(0_u8).take(100));
+        let buffer = Buffer::from_iter(std::iter::repeat_n(0_u8, 100));
         assert_eq!(buffer.strong_count(), 1);
 
         let buffer2 = buffer.clone();
diff --git a/arrow-buffer/src/builder/mod.rs b/arrow-buffer/src/builder/mod.rs
@@ -26,7 +26,7 @@ pub use null::*;
 pub use offset::*;
 
 use crate::{ArrowNativeType, Buffer, MutableBuffer};
-use std::{iter, marker::PhantomData};
+use std::marker::PhantomData;
 
 /// Builder for creating a [Buffer] object.
 ///
@@ -214,7 +214,7 @@ impl<T: ArrowNativeType> BufferBuilder<T> {
     #[inline]
     pub fn append_n(&mut self, n: usize, v: T) {
         self.reserve(n);
-        self.extend(iter::repeat(v).take(n))
+        self.extend(std::iter::repeat_n(v, n))
     }
 
     /// Appends `n`, zero-initialized values
diff --git a/arrow-cast/src/cast/list.rs b/arrow-cast/src/cast/list.rs
@@ -24,7 +24,7 @@ pub(crate) fn cast_values_to_list<O: OffsetSizeTrait>(
     cast_options: &CastOptions,
 ) -> Result<ArrayRef, ArrowError> {
     let values = cast_with_options(array, to.data_type(), cast_options)?;
-    let offsets = OffsetBuffer::from_lengths(std::iter::repeat(1).take(values.len()));
+    let offsets = OffsetBuffer::from_lengths(std::iter::repeat_n(1, values.len()));
     let list = GenericListArray::<O>::new(to.clone(), offsets, values, None);
     Ok(Arc::new(list))
 }
diff --git a/arrow-cast/src/cast/mod.rs b/arrow-cast/src/cast/mod.rs
@@ -2167,7 +2167,7 @@ fn cast_numeric_to_binary<FROM: ArrowPrimitiveType, O: OffsetSizeTrait>(
 ) -> Result<ArrayRef, ArrowError> {
     let array = array.as_primitive::<FROM>();
     let size = std::mem::size_of::<FROM::Native>();
-    let offsets = OffsetBuffer::from_lengths(std::iter::repeat(size).take(array.len()));
+    let offsets = OffsetBuffer::from_lengths(std::iter::repeat_n(size, array.len()));
     Ok(Arc::new(GenericBinaryArray::<O>::new(
         offsets,
         array.values().inner().clone(),
diff --git a/arrow-data/src/data.rs b/arrow-data/src/data.rs
@@ -638,7 +638,7 @@ impl ArrayData {
                 ),
                 DataType::Union(f, mode) => {
                     let (id, _) = f.iter().next().unwrap();
-                    let ids = Buffer::from_iter(std::iter::repeat(id).take(len));
+                    let ids = Buffer::from_iter(std::iter::repeat_n(id, len));
                     let buffers = match mode {
                         UnionMode::Sparse => vec![ids],
                         UnionMode::Dense => {
diff --git a/arrow-flight/src/encode.rs b/arrow-flight/src/encode.rs
@@ -1695,9 +1695,9 @@ mod tests {
 
     #[tokio::test]
     async fn flight_data_size_even() {
-        let s1 = StringArray::from_iter_values(std::iter::repeat(".10 bytes.").take(1024));
+        let s1 = StringArray::from_iter_values(std::iter::repeat_n(".10 bytes.", 1024));
         let i1 = Int16Array::from_iter_values(0..1024);
-        let s2 = StringArray::from_iter_values(std::iter::repeat("6bytes").take(1024));
+        let s2 = StringArray::from_iter_values(std::iter::repeat_n("6bytes", 1024));
         let i2 = Int64Array::from_iter_values(0..1024);
 
         let batch = RecordBatch::try_from_iter(vec![
diff --git a/arrow-ord/src/sort.rs b/arrow-ord/src/sort.rs
@@ -1791,7 +1791,7 @@ mod tests {
                     None => {
                         builder
                             .values()
-                            .extend(std::iter::repeat(None).take(fixed_length as usize));
+                            .extend(std::iter::repeat_n(None, fixed_length as usize));
                         builder.append(false);
                     }
                 }
diff --git a/arrow-pyarrow-integration-testing/Cargo.toml b/arrow-pyarrow-integration-testing/Cargo.toml
@@ -25,7 +25,7 @@ authors = ["Apache Arrow <dev@arrow.apache.org>"]
 license = "Apache-2.0"
 keywords = ["arrow"]
 edition = "2021"
-rust-version = "1.81"
+rust-version = "1.84"
 publish = false
 
 [lib]
diff --git a/arrow-pyarrow-testing/Cargo.toml b/arrow-pyarrow-testing/Cargo.toml
@@ -38,9 +38,9 @@ homepage = "https://github.com/apache/arrow-rs"
 repository = "https://github.com/apache/arrow-rs"
 authors = ["Apache Arrow <dev@arrow.apache.org>"]
 license = "Apache-2.0"
-keywords = [ "arrow" ]
+keywords = ["arrow"]
 edition = "2021"
-rust-version = "1.81"
+rust-version = "1.84"
 publish = false
 
 
diff --git a/arrow-select/src/coalesce.rs b/arrow-select/src/coalesce.rs
@@ -730,7 +730,7 @@ mod tests {
         // The strings are designed to exactly fit into buffers that are powers of 2 long
         let batch = stringview_batch_repeated(100, [Some("This string is a power of two=32")]);
         let output_batches = Test::new()
-            .with_batches(std::iter::repeat(batch).take(20))
+            .with_batches(std::iter::repeat_n(batch, 20))
             .with_batch_size(900)
             .with_expected_output_sizes(vec![900, 900, 200])
             .run();
diff --git a/arrow-select/src/concat.rs b/arrow-select/src/concat.rs
@@ -1335,7 +1335,7 @@ mod tests {
         assert_eq!(data.buffers()[0].len(), 120);
         assert_eq!(data.buffers()[0].capacity(), 128); // Nearest multiple of 64
 
-        let a = StringArray::from_iter_values(std::iter::repeat("foo").take(100));
+        let a = StringArray::from_iter_values(std::iter::repeat_n("foo", 100));
         let b = StringArray::from(vec!["bingo", "bongo", "lorem", ""]);
 
         let a = concat(&[&a, &b]).unwrap();
@@ -1358,8 +1358,8 @@ mod tests {
         assert_eq!(data.buffers()[1].len(), 135);
         assert_eq!(data.buffers()[1].capacity(), 192); // Nearest multiple of 64
 
-        let a = LargeBinaryArray::from_iter_values(std::iter::repeat(b"foo").take(100));
-        let b = LargeBinaryArray::from_iter_values(std::iter::repeat(b"cupcakes").take(10));
+        let a = LargeBinaryArray::from_iter_values(std::iter::repeat_n(b"foo", 100));
+        let b = LargeBinaryArray::from_iter_values(std::iter::repeat_n(b"cupcakes", 10));
 
         let a = concat(&[&a, &b]).unwrap();
         let data = a.to_data();
diff --git a/arrow-select/src/filter.rs b/arrow-select/src/filter.rs
@@ -1449,12 +1449,11 @@ mod tests {
     #[test]
     fn test_slices() {
         // takes up 2 u64s
-        let bools = std::iter::repeat(true)
-            .take(10)
-            .chain(std::iter::repeat(false).take(30))
-            .chain(std::iter::repeat(true).take(20))
-            .chain(std::iter::repeat(false).take(17))
-            .chain(std::iter::repeat(true).take(4));
+        let bools = std::iter::repeat_n(true, 10)
+            .chain(std::iter::repeat_n(false, 30))
+            .chain(std::iter::repeat_n(true, 20))
+            .chain(std::iter::repeat_n(false, 17))
+            .chain(std::iter::repeat_n(true, 4));
 
         let bool_array: BooleanArray = bools.map(Some).collect();
 
diff --git a/arrow/benches/array_data_validate.rs b/arrow/benches/array_data_validate.rs
@@ -53,7 +53,7 @@ fn validate_benchmark(c: &mut Criterion) {
         b.iter(|| validate_utf8_array(&str_arr))
     });
 
-    let byte_array = BinaryArray::from_iter_values(std::iter::repeat(b"test").take(20000));
+    let byte_array = BinaryArray::from_iter_values(std::iter::repeat_n(b"test", 20000));
     c.bench_function("byte_array_to_string_array 20000", |b| {
         b.iter(|| StringArray::from(BinaryArray::from(byte_array.to_data())))
     });
diff --git a/arrow/benches/partition_kernels.rs b/arrow/benches/partition_kernels.rs
@@ -28,7 +28,7 @@ use arrow::{
 };
 use arrow_ord::partition::partition;
 use rand::distr::{Distribution, StandardUniform};
-use std::{hint, iter};
+use std::hint;
 
 fn create_array<T: ArrowPrimitiveType>(size: usize, with_nulls: bool) -> ArrayRef
 where
@@ -45,11 +45,10 @@ fn bench_partition(sorted_columns: &[ArrayRef]) {
 
 fn create_sorted_low_cardinality_data(length: usize) -> Vec<ArrayRef> {
     let arr = Int64Array::from_iter_values(
-        iter::repeat(1)
-            .take(length / 4)
-            .chain(iter::repeat(2).take(length / 4))
-            .chain(iter::repeat(3).take(length / 4))
-            .chain(iter::repeat(4).take(length / 4)),
+        std::iter::repeat_n(1, length / 4)
+            .chain(std::iter::repeat_n(2, length / 4))
+            .chain(std::iter::repeat_n(3, length / 4))
+            .chain(std::iter::repeat_n(4, length / 4)),
     );
     lexsort(
         &[SortColumn {
diff --git a/arrow/benches/string_run_iterator.rs b/arrow/benches/string_run_iterator.rs
@@ -29,7 +29,7 @@ fn build_strings_runs(
     let run_len = logical_array_len / physical_array_len;
     let mut values: Vec<String> = (0..physical_array_len)
         .map(|_| (0..string_len).map(|_| rng.random::<char>()).collect())
-        .flat_map(|s| std::iter::repeat(s).take(run_len))
+        .flat_map(|s| std::iter::repeat_n(s, run_len))
         .collect();
     while values.len() < logical_array_len {
         let last_val = values[values.len() - 1].clone();
diff --git a/arrow/src/util/bench_util.rs b/arrow/src/util/bench_util.rs
@@ -396,7 +396,7 @@ pub fn create_primitive_run_array<R: RunEndIndexType, V: ArrowPrimitiveType>(
                 take_len += 1;
                 run_len_extra -= 1;
             }
-            std::iter::repeat(V::Native::from_usize(s).unwrap()).take(take_len)
+            std::iter::repeat_n(V::Native::from_usize(s).unwrap(), take_len)
         })
         .collect();
     while values.len() < logical_array_len {
@@ -434,7 +434,7 @@ pub fn create_string_array_for_runs(
                 take_len += 1;
                 run_len_extra -= 1;
             }
-            std::iter::repeat(s).take(take_len)
+            std::iter::repeat_n(s, take_len)
         })
         .collect();
     while values.len() < logical_array_len {
diff --git a/parquet-variant-compute/Cargo.toml b/parquet-variant-compute/Cargo.toml
@@ -27,8 +27,7 @@ repository = { workspace = true }
 authors = { workspace = true }
 keywords = ["arrow", "parquet", "variant"]
 edition = { workspace = true }
-# parquet-variant needs newer version than workspace
-rust-version = "1.83"
+rust-version = { workspace = true }
 
 
 [dependencies]
diff --git a/parquet-variant-json/Cargo.toml b/parquet-variant-json/Cargo.toml
@@ -28,8 +28,7 @@ authors = { workspace = true }
 keywords = ["arrow", "parquet", "variant"]
 readme = "README.md"
 edition = { workspace = true }
-# parquet-variant needs newer version than workspace
-rust-version = "1.83"
+rust-version = { workspace = true }
 
 
 [dependencies]
diff --git a/parquet-variant/Cargo.toml b/parquet-variant/Cargo.toml
diff --git a/parquet/src/arrow/arrow_reader/statistics.rs b/parquet/src/arrow/arrow_reader/statistics.rs
diff --git a/parquet/src/arrow/arrow_writer/byte_array.rs b/parquet/src/arrow/arrow_writer/byte_array.rs
diff --git a/parquet/src/arrow/arrow_writer/levels.rs b/parquet/src/arrow/arrow_writer/levels.rs
diff --git a/parquet/src/arrow/arrow_writer/mod.rs b/parquet/src/arrow/arrow_writer/mod.rs
diff --git a/parquet/src/arrow/buffer/offset_buffer.rs b/parquet/src/arrow/buffer/offset_buffer.rs
diff --git a/parquet/src/column/writer/mod.rs b/parquet/src/column/writer/mod.rs
diff --git a/parquet/src/encodings/rle.rs b/parquet/src/encodings/rle.rs
diff --git a/parquet/src/file/metadata/writer.rs b/parquet/src/file/metadata/writer.rs
diff --git a/parquet/src/file/serialized_reader.rs b/parquet/src/file/serialized_reader.rs
diff --git a/parquet/src/util/bit_util.rs b/parquet/src/util/bit_util.rs
diff --git a/parquet/tests/arrow_reader/mod.rs b/parquet/tests/arrow_reader/mod.rs
diff --git a/parquet/tests/arrow_reader/statistics.rs b/parquet/tests/arrow_reader/statistics.rs

Original file line number	Diff line number	Diff line change
`@@ -338,7 +338,7 @@ impl Decoder {`
`338`	`338`	`moff.push_length(0);`
`339`	`339`	`}`
`340`	`340`	`Self::Fixed(sz, accum) => {`
`341`		`- accum.extend(std::iter::repeat(0u8).take(*sz as usize));`
	`341`	`+ accum.extend(std::iter::repeat_n(0u8, *sz as usize));`
`342`	`342`	`}`
`343`	`343`	`Self::Decimal128(_, _, _, builder) => builder.append_value(0),`
`344`	`344`	`Self::Decimal256(_, _, _, builder) => builder.append_value(i256::ZERO),`