refactor: Bump MSRV to 1.84 for preparing next release (#1185)

Xuanwo · web-flow · commit 7eb00d5e1571 · 2025-04-09T11:55:26.000+08:00
## Which issue does this PR close?  None ## What changes are included in this PR? Bump MSRV to 1.84 to get our community ready for the next release. Rust 1.84 is released on 9 January, 2025 which is 3 months ago. We need this PR in advance since we still need time to play well with rust 1.84 like using new deps or better APIs and so so. ## Are these changes tested?  --------- Signed-off-by: Xuanwo <github@xuanwo.io>
diff --git a/.github/workflows/ci.yml b/.github/workflows/ci.yml
@@ -30,7 +30,7 @@ concurrency:
   cancel-in-progress: true
 
 env:
-  rust_msrv: "1.77.1"
+  rust_msrv: "1.84.0"
 
 jobs:
   check:
diff --git a/Cargo.toml b/Cargo.toml
@@ -36,7 +36,7 @@ homepage = "https://rust.iceberg.apache.org/"
 repository = "https://github.com/apache/iceberg-rust"
 license = "Apache-2.0"
 # Check the MSRV policy in README.md before changing this
-rust-version = "1.77.1"
+rust-version = "1.84"
 
 [workspace.dependencies]
 anyhow = "1.0.72"
@@ -92,13 +92,14 @@ serde_json = "1.0.138"
 serde_repr = "0.1.16"
 serde_with = "3.4"
 tempfile = "3.18"
-thrift = "0.17.0"  
+thrift = "0.17.0"
 tokio = { version = "1.44", default-features = false }
 typed-builder = "0.20"
 url = "2.5.4"
-uuid = { version = "1.13.1", features = ["v7"] }
+uuid = { version = "1.14", features = ["v7"] }
 volo-thrift = "0.10"
 hive_metastore = "0.1"
 tera = "1"
 zstd = "0.13.2"
 expect-test = "1"
+http = "1.1"
diff --git a/bindings/python/Cargo.lock b/bindings/python/Cargo.lock
diff --git a/bindings/python/Cargo.toml b/bindings/python/Cargo.toml
@@ -20,7 +20,7 @@ name = "pyiceberg_core_rust"
 version = "0.4.0"
 edition = "2021"
 homepage = "https://rust.iceberg.apache.org"
-rust-version = "1.77.1"
+rust-version = "1.84"
 # This crate is used to build python bindings, we don't want to publish it
 publish = false
 
diff --git a/crates/catalog/rest/Cargo.toml b/crates/catalog/rest/Cargo.toml
@@ -32,7 +32,7 @@ keywords = ["iceberg", "rest", "catalog"]
 # async-trait = { workspace = true }
 async-trait = { workspace = true }
 chrono = { workspace = true }
-http = "1.1.0"
+http = { workspace = true }
 iceberg = { workspace = true }
 itertools = { workspace = true }
 log = { workspace = true }
diff --git a/crates/catalog/rest/src/types.rs b/crates/catalog/rest/src/types.rs
@@ -129,12 +129,14 @@ pub(super) struct ListNamespaceResponse {
     pub(super) next_page_token: Option<String>,
 }
 
+#[allow(dead_code)]
 #[derive(Debug, Serialize, Deserialize)]
 pub(super) struct UpdateNamespacePropsRequest {
     removals: Option<Vec<String>>,
     updates: Option<HashMap<String, String>>,
 }
 
+#[allow(dead_code)]
 #[derive(Debug, Serialize, Deserialize)]
 pub(super) struct UpdateNamespacePropsResponse {
     updated: Vec<String>,
diff --git a/crates/catalog/s3tables/Cargo.toml b/crates/catalog/s3tables/Cargo.toml
@@ -20,7 +20,7 @@ name = "iceberg-catalog-s3tables"
 version = { workspace = true }
 edition = { workspace = true }
 homepage = { workspace = true }
-rust-version = "1.81.0"
+rust-version = { workspace = true }
 
 categories = ["database"]
 description = "Apache Iceberg Rust S3Tables Catalog"
diff --git a/crates/catalog/s3tables/src/catalog.rs b/crates/catalog/s3tables/src/catalog.rs
@@ -82,7 +82,7 @@ impl S3TablesCatalog {
                 )
             })?;
 
-        let file_io = FileIO::from_path(&format!("s3://{}", bucket_name))?
+        let file_io = FileIO::from_path(format!("s3://{}", bucket_name))?
             .with_props(&config.properties)
             .build()?;
 
diff --git a/crates/iceberg/src/arrow/reader.rs b/crates/iceberg/src/arrow/reader.rs
@@ -540,7 +540,7 @@ impl ArrowReader {
                     f.metadata()
                         .get(PARQUET_FIELD_ID_META_KEY)
                         .and_then(|field_id| i32::from_str(field_id).ok())
-                        .map_or(false, |field_id| {
+                        .is_some_and(|field_id| {
                             projected_fields.insert((*f).clone(), field_id);
                             leaf_field_ids.contains(&field_id)
                         })
@@ -992,7 +992,7 @@ fn project_column(
 type PredicateResult =
     dyn FnMut(RecordBatch) -> std::result::Result<BooleanArray, ArrowError> + Send + 'static;
 
-impl<'a> BoundPredicateVisitor for PredicateConverter<'a> {
+impl BoundPredicateVisitor for PredicateConverter<'_> {
     type T = Box<PredicateResult>;
 
     fn always_true(&mut self) -> Result<Box<PredicateResult>> {
diff --git a/crates/iceberg/src/catalog/mod.rs b/crates/iceberg/src/catalog/mod.rs
@@ -361,6 +361,7 @@ pub enum TableRequirement {
 /// TableUpdate represents an update to a table in the catalog.
 #[derive(Debug, Serialize, Deserialize, PartialEq, Clone)]
 #[serde(tag = "action", rename_all = "kebab-case")]
+#[allow(clippy::large_enum_variant)]
 pub enum TableUpdate {
     /// Upgrade table's format version
     #[serde(rename_all = "kebab-case")]
@@ -750,6 +751,7 @@ pub struct ViewCreation {
 /// ViewUpdate represents an update to a view in the catalog.
 #[derive(Clone, Debug, PartialEq, Serialize, Deserialize)]
 #[serde(tag = "action", rename_all = "kebab-case")]
+#[allow(clippy::large_enum_variant)]
 pub enum ViewUpdate {
     /// Assign a new UUID to the view
     #[serde(rename_all = "kebab-case")]
@@ -1125,7 +1127,7 @@ mod tests {
 {
     "type": "assert-ref-snapshot-id",
     "ref": "snapshot-name",
-    "snapshot-id": null 
+    "snapshot-id": null
 }
         "#,
             TableRequirement::RefSnapshotIdMatch {
@@ -1245,7 +1247,7 @@ mod tests {
 {
     "action": "assign-uuid",
     "uuid": "2cc52516-5e73-41f2-b139-545d41a4e151"
-}        
+}
         "#,
             TableUpdate::AssignUuid {
                 uuid: uuid!("2cc52516-5e73-41f2-b139-545d41a4e151"),
@@ -1260,7 +1262,7 @@ mod tests {
 {
     "action": "upgrade-format-version",
     "format-version": 2
-}        
+}
         "#,
             TableUpdate::UpgradeFormatVersion {
                 format_version: FormatVersion::V2,
@@ -1569,7 +1571,7 @@ mod tests {
         1,
         2
     ]
-}  
+}
         "#;
 
         let update = TableUpdate::RemoveSnapshots {
@@ -1629,7 +1631,7 @@ mod tests {
     "min-snapshots-to-keep": 2,
     "max-snapshot-age-ms": 3,
     "max-ref-age-ms": 4
-}        
+}
         "#;
 
         let update = TableUpdate::SetSnapshotRef {
@@ -1656,7 +1658,7 @@ mod tests {
         "prop1": "v1",
         "prop2": "v2"
     }
-}        
+}
         "#;
 
         let update = TableUpdate::SetProperties {
@@ -1680,7 +1682,7 @@ mod tests {
         "prop1",
         "prop2"
     ]
-}        
+}
         "#;
 
         let update = TableUpdate::RemoveProperties {
@@ -1742,7 +1744,7 @@ mod tests {
 {
     "action": "assign-uuid",
     "uuid": "2cc52516-5e73-41f2-b139-545d41a4e151"
-}        
+}
         "#,
             ViewUpdate::AssignUuid {
                 uuid: uuid!("2cc52516-5e73-41f2-b139-545d41a4e151"),
@@ -1757,7 +1759,7 @@ mod tests {
 {
     "action": "upgrade-format-version",
     "format-version": 1
-}        
+}
         "#,
             ViewUpdate::UpgradeFormatVersion {
                 format_version: ViewFormatVersion::V1,
@@ -1825,7 +1827,7 @@ mod tests {
 {
     "action": "set-location",
     "location": "s3://db/view"
-}        
+}
         "#,
             ViewUpdate::SetLocation {
                 location: "s3://db/view".to_string(),
@@ -1843,7 +1845,7 @@ mod tests {
         "prop1": "v1",
         "prop2": "v2"
     }
-}        
+}
         "#,
             ViewUpdate::SetProperties {
                 updates: vec![
@@ -1866,7 +1868,7 @@ mod tests {
         "prop1",
         "prop2"
     ]
-}        
+}
         "#,
             ViewUpdate::RemoveProperties {
                 removals: vec!["prop1".to_string(), "prop2".to_string()],
@@ -1895,7 +1897,7 @@ mod tests {
               "dialect" : "spark"
             } ]
     }
-}        
+}
         "#,
             ViewUpdate::AddViewVersion {
                 view_version: ViewVersion::builder()
@@ -1925,7 +1927,7 @@ mod tests {
 {
     "action": "set-current-view-version",
     "view-version-id": 1
-}        
+}
         "#,
             ViewUpdate::SetCurrentViewVersion { view_version_id: 1 },
         );
@@ -1938,7 +1940,7 @@ mod tests {
 {
     "action": "remove-partition-specs",
     "spec-ids": [1, 2]
-}        
+}
         "#,
             TableUpdate::RemovePartitionSpecs {
                 spec_ids: vec![1, 2],
@@ -1972,7 +1974,7 @@ mod tests {
                                 }
                         ]
                 }
-        } 
+        }
         "#,
             TableUpdate::SetStatistics {
                 statistics: StatisticsFile {
@@ -2002,7 +2004,7 @@ mod tests {
         {
                 "action": "remove-statistics",
                 "snapshot-id": 1940541653261589030
-        } 
+        }
         "#,
             TableUpdate::RemoveStatistics {
                 snapshot_id: 1940541653261589030,
diff --git a/crates/iceberg/src/delete_file_index.rs b/crates/iceberg/src/delete_file_index.rs
@@ -101,7 +101,6 @@ impl PopulatedDeleteFileIndex {
     /// 2. If the partition is empty and the delete file is not a positional delete,
     ///    it is added to the `global_delees` vector
     /// 3. Otherwise, the delete file is added to one of two hash maps based on its content type.
-
     fn new(files: Vec<DeleteFileContext>) -> PopulatedDeleteFileIndex {
         let mut eq_deletes_by_partition: HashMap<Struct, Vec<Arc<DeleteFileContext>>> =
             HashMap::default();
diff --git a/crates/iceberg/src/delete_vector.rs b/crates/iceberg/src/delete_vector.rs
@@ -80,7 +80,7 @@ impl Iterator for DeleteVectorIterator<'_> {
     }
 }
 
-impl<'a> DeleteVectorIterator<'a> {
+impl DeleteVectorIterator<'_> {
     pub fn advance_to(&mut self, pos: u64) {
         let hi = (pos >> 32) as u32;
         let lo = pos as u32;
diff --git a/crates/iceberg/src/inspect/metadata_table.rs b/crates/iceberg/src/inspect/metadata_table.rs
@@ -46,6 +46,8 @@ impl<'a> MetadataTable<'a> {
 
 #[cfg(test)]
 pub mod tests {
+    //! Sharable tests for the metadata table.
+
     use expect_test::Expect;
     use futures::TryStreamExt;
     use itertools::Itertools;
diff --git a/crates/iceberg/src/puffin/test_utils.rs b/crates/iceberg/src/puffin/test_utils.rs
@@ -55,10 +55,10 @@ pub(crate) fn empty_footer_payload() -> FileMetadata {
 }
 
 pub(crate) fn empty_footer_payload_bytes() -> Vec<u8> {
-    return serde_json::to_string::<FileMetadata>(&empty_footer_payload())
+    serde_json::to_string::<FileMetadata>(&empty_footer_payload())
         .unwrap()
         .as_bytes()
-        .to_vec();
+        .to_vec()
 }
 
 pub(crate) fn empty_footer_payload_bytes_length_bytes() -> [u8; 4] {
diff --git a/crates/iceberg/src/scan/mod.rs b/crates/iceberg/src/scan/mod.rs
@@ -453,7 +453,7 @@ impl TableScan {
             }
         });
 
-        return Ok(file_scan_task_rx.boxed());
+        Ok(file_scan_task_rx.boxed())
     }
 
     /// Returns an [`ArrowRecordBatchStream`].
@@ -591,6 +591,9 @@ pub(crate) struct BoundPredicates {
 
 #[cfg(test)]
 pub mod tests {
+    //! shared tests for the table scan API
+    #![allow(missing_docs)]
+
     use std::collections::HashMap;
     use std::fs;
     use std::fs::File;
diff --git a/crates/iceberg/src/spec/datatypes.rs b/crates/iceberg/src/spec/datatypes.rs
@@ -726,7 +726,7 @@ pub(super) mod _serde {
         Primitive(PrimitiveType),
     }
 
-    impl<'a> From<SerdeType<'a>> for Type {
+    impl From<SerdeType<'_>> for Type {
         fn from(value: SerdeType) -> Self {
             match value {
                 SerdeType::List {
@@ -947,7 +947,7 @@ mod tests {
         let record = r#"
         {
             "type": "struct",
-            "fields": [ 
+            "fields": [
                 {
                     "id": 1,
                     "name": "id",
@@ -960,7 +960,7 @@ mod tests {
                     "name": "data",
                     "required": false,
                     "type": "int"
-                } 
+                }
             ]
         }
         "#;
diff --git a/crates/iceberg/src/spec/sort.rs b/crates/iceberg/src/spec/sort.rs
@@ -147,7 +147,7 @@ impl SortOrderBuilder {
     /// Creates a new unbound sort order.
     pub fn build_unbound(&self) -> Result<SortOrder> {
         let fields = self.fields.clone().unwrap_or_default();
-        return match (self.order_id, fields.as_slice()) {
+        match (self.order_id, fields.as_slice()) {
             (Some(SortOrder::UNSORTED_ORDER_ID) | None, []) => Ok(SortOrder::unsorted_order()),
             (_, []) => Err(Error::new(
                 ErrorKind::Unexpected,
@@ -164,7 +164,7 @@ impl SortOrderBuilder {
                 order_id: maybe_order_id.unwrap_or(1),
                 fields: fields.to_vec(),
             }),
-        };
+        }
     }
 
     /// Creates a new bound sort order.
diff --git a/crates/iceberg/src/spec/values.rs b/crates/iceberg/src/spec/values.rs
diff --git a/crates/iceberg/src/writer/file_writer/track_writer.rs b/crates/iceberg/src/writer/file_writer/track_writer.rs
diff --git a/crates/integrations/datafusion/Cargo.toml b/crates/integrations/datafusion/Cargo.toml
diff --git a/deny.toml b/deny.toml
diff --git a/rust-toolchain.toml b/rust-toolchain.toml

Original file line number	Diff line number	Diff line change
`@@ -129,12 +129,14 @@ pub(super) struct ListNamespaceResponse {`
`129`	`129`	`pub(super) next_page_token: Option<String>,`
`130`	`130`	`}`
`131`	`131`
	`132`	`+#[allow(dead_code)]`
`132`	`133`	`#[derive(Debug, Serialize, Deserialize)]`
`133`	`134`	`pub(super) struct UpdateNamespacePropsRequest {`
`134`	`135`	`removals: Option<Vec<String>>,`
`135`	`136`	`updates: Option<HashMap<String, String>>,`
`136`	`137`	`}`
`137`	`138`
	`139`	`+#[allow(dead_code)]`
`138`	`140`	`#[derive(Debug, Serialize, Deserialize)]`
`139`	`141`	`pub(super) struct UpdateNamespacePropsResponse {`
`140`	`142`	`updated: Vec<String>,`
Original file line number	Diff line number	Diff line change
`@@ -82,7 +82,7 @@ impl S3TablesCatalog {`
`82`	`82`	`)`
`83`	`83`	`})?;`
`84`	`84`
`85`		`- let file_io = FileIO::from_path(&format!("s3://{}", bucket_name))?`
	`85`	`+ let file_io = FileIO::from_path(format!("s3://{}", bucket_name))?`
`86`	`86`	`.with_props(&config.properties)`
`87`	`87`	`.build()?;`
`88`	`88`