chore: Enable workspace lint for all workspace members (#20577)

## Which issue does this PR close?

N/A

## Rationale for this change

Several workspace members did not opt into workspace-level lints. This
was almost certainly an oversight.

## What changes are included in this PR?

* Enable workspace-level lints for all workspace members
* Fix up code to fix various linter errors

This PR pulls out some of the changes made in #20250, although I didn't
actually use that PR.

## Are these changes tested?

Yes.

## Are there any user-facing changes?

No.

Author: neilconway

datafusion-cli/Cargo.toml

@@ -69,6 +69,9 @@ rustyline = "17.0"
 tokio = { workspace = true, features = ["macros", "parking_lot", "rt", "rt-multi-thread", "signal", "sync"] }
 url = { workspace = true }
 
+[lints]
+workspace = true
+
 [dev-dependencies]
 ctor = { workspace = true }
 insta = { workspace = true }

datafusion-cli/src/functions.rs

@@ -233,7 +233,7 @@ impl TableProvider for ParquetMetadataTable {
         self
     }
 
-    fn schema(&self) -> arrow::datatypes::SchemaRef {
+    fn schema(&self) -> SchemaRef {
         self.schema.clone()
     }
 
@@ -482,7 +482,7 @@ impl TableProvider for MetadataCacheTable {
         self
     }
 
-    fn schema(&self) -> arrow::datatypes::SchemaRef {
+    fn schema(&self) -> SchemaRef {
         self.schema.clone()
     }
 
@@ -600,7 +600,7 @@ impl TableProvider for StatisticsCacheTable {
         self
     }
 
-    fn schema(&self) -> arrow::datatypes::SchemaRef {
+    fn schema(&self) -> SchemaRef {
         self.schema.clone()
     }
 
@@ -735,7 +735,7 @@ impl TableProvider for ListFilesCacheTable {
         self
     }
 
-    fn schema(&self) -> arrow::datatypes::SchemaRef {
+    fn schema(&self) -> SchemaRef {
         self.schema.clone()
     }
 

datafusion-cli/src/highlighter.rs

@@ -38,7 +38,8 @@ pub struct SyntaxHighlighter {
 
 impl SyntaxHighlighter {
     pub fn new(dialect: &config::Dialect) -> Self {
-        let dialect = dialect_from_str(dialect).unwrap_or(Box::new(GenericDialect {}));
+        let dialect =
+            dialect_from_str(dialect).unwrap_or_else(|| Box::new(GenericDialect {}));
         Self { dialect }
     }
 }

datafusion-cli/src/object_storage.rs

@@ -64,6 +64,21 @@ pub async fn get_s3_object_store_builder(
     url: &Url,
     aws_options: &AwsOptions,
     resolve_region: bool,
+) -> Result<AmazonS3Builder> {
+    // Box the inner future to reduce the future size of this async function,
+    // which is deeply nested in the CLI's async call chain.
+    Box::pin(get_s3_object_store_builder_inner(
+        url,
+        aws_options,
+        resolve_region,
+    ))
+    .await
+}
+
+async fn get_s3_object_store_builder_inner(
+    url: &Url,
+    aws_options: &AwsOptions,
+    resolve_region: bool,
 ) -> Result<AmazonS3Builder> {
     let AwsOptions {
         access_key_id,
@@ -209,7 +224,7 @@ impl CredentialsFromConfig {
 
 #[derive(Debug)]
 struct S3CredentialProvider {
-    credentials: aws_credential_types::provider::SharedCredentialsProvider,
+    credentials: SharedCredentialsProvider,
 }
 
 #[async_trait]

datafusion-cli/src/object_storage/instrumented.rs

@@ -110,7 +110,7 @@ pub enum InstrumentedObjectStoreMode {
 }
 
 impl fmt::Display for InstrumentedObjectStoreMode {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "{self:?}")
     }
 }
@@ -383,7 +383,7 @@ impl InstrumentedObjectStore {
 }
 
 impl fmt::Display for InstrumentedObjectStore {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         let mode: InstrumentedObjectStoreMode =
             self.instrument_mode.load(Ordering::Relaxed).into();
         write!(
@@ -490,7 +490,7 @@ pub enum Operation {
 }
 
 impl fmt::Display for Operation {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         write!(f, "{self:?}")
     }
 }
@@ -508,7 +508,7 @@ pub struct RequestDetails {
 }
 
 impl fmt::Display for RequestDetails {
-    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
+    fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
         let mut output_parts = vec![format!(
             "{} operation={:?}",
             self.timestamp.to_rfc3339(),

datafusion-cli/src/print_format.rs

@@ -259,7 +259,7 @@ mod tests {
     fn print_csv_no_header() {
         let output = PrintBatchesTest::new()
             .with_format(PrintFormat::Csv)
-            .with_batches(split_batch(three_column_batch()))
+            .with_batches(split_batch(&three_column_batch()))
             .with_header(WithHeader::No)
             .run();
         assert_snapshot!(output, @r"
@@ -273,7 +273,7 @@ mod tests {
     fn print_csv_with_header() {
         let output = PrintBatchesTest::new()
             .with_format(PrintFormat::Csv)
-            .with_batches(split_batch(three_column_batch()))
+            .with_batches(split_batch(&three_column_batch()))
             .with_header(WithHeader::Yes)
             .run();
         assert_snapshot!(output, @r"
@@ -288,7 +288,7 @@ mod tests {
     fn print_tsv_no_header() {
         let output = PrintBatchesTest::new()
             .with_format(PrintFormat::Tsv)
-            .with_batches(split_batch(three_column_batch()))
+            .with_batches(split_batch(&three_column_batch()))
             .with_header(WithHeader::No)
             .run();
         assert_snapshot!(output, @r"
@@ -302,7 +302,7 @@ mod tests {
     fn print_tsv_with_header() {
         let output = PrintBatchesTest::new()
             .with_format(PrintFormat::Tsv)
-            .with_batches(split_batch(three_column_batch()))
+            .with_batches(split_batch(&three_column_batch()))
             .with_header(WithHeader::Yes)
             .run();
         assert_snapshot!(output, @r"
@@ -317,7 +317,7 @@ mod tests {
     fn print_table() {
         let output = PrintBatchesTest::new()
             .with_format(PrintFormat::Table)
-            .with_batches(split_batch(three_column_batch()))
+            .with_batches(split_batch(&three_column_batch()))
             .with_header(WithHeader::Ignored)
             .run();
         assert_snapshot!(output, @r"
@@ -334,7 +334,7 @@ mod tests {
     fn print_json() {
         let output = PrintBatchesTest::new()
             .with_format(PrintFormat::Json)
-            .with_batches(split_batch(three_column_batch()))
+            .with_batches(split_batch(&three_column_batch()))
             .with_header(WithHeader::Ignored)
             .run();
         assert_snapshot!(output, @r#"[{"a":1,"b":4,"c":7},{"a":2,"b":5,"c":8},{"a":3,"b":6,"c":9}]"#);
@@ -344,7 +344,7 @@ mod tests {
     fn print_ndjson() {
         let output = PrintBatchesTest::new()
             .with_format(PrintFormat::NdJson)
-            .with_batches(split_batch(three_column_batch()))
+            .with_batches(split_batch(&three_column_batch()))
             .with_header(WithHeader::Ignored)
             .run();
         assert_snapshot!(output, @r#"
@@ -358,7 +358,7 @@ mod tests {
     fn print_automatic_no_header() {
         let output = PrintBatchesTest::new()
             .with_format(PrintFormat::Automatic)
-            .with_batches(split_batch(three_column_batch()))
+            .with_batches(split_batch(&three_column_batch()))
             .with_header(WithHeader::No)
             .run();
         assert_snapshot!(output, @r"
@@ -371,7 +371,7 @@ mod tests {
     fn print_automatic_with_header() {
         let output = PrintBatchesTest::new()
             .with_format(PrintFormat::Automatic)
-            .with_batches(split_batch(three_column_batch()))
+            .with_batches(split_batch(&three_column_batch()))
             .with_header(WithHeader::Yes)
             .run();
         assert_snapshot!(output, @r"
@@ -633,7 +633,7 @@ mod tests {
     }
 
     /// Slice the record batch into 2 batches
-    fn split_batch(batch: RecordBatch) -> Vec<RecordBatch> {
+    fn split_batch(batch: &RecordBatch) -> Vec<RecordBatch> {
         assert!(batch.num_rows() > 1);
         let split = batch.num_rows() / 2;
         vec![

datafusion-cli/src/print_options.rs

@@ -115,7 +115,7 @@ impl PrintOptions {
         row_count: usize,
         format_options: &FormatOptions,
     ) -> Result<()> {
-        let stdout = std::io::stdout();
+        let stdout = io::stdout();
         let mut writer = stdout.lock();
 
         self.format.print_batches(
@@ -137,7 +137,7 @@ impl PrintOptions {
             query_start_time,
         );
 
-        self.write_output(&mut writer, formatted_exec_details)
+        self.write_output(&mut writer, &formatted_exec_details)
     }
 
     /// Print the stream to stdout using the specified format
@@ -153,7 +153,7 @@ impl PrintOptions {
             ));
         };
 
-        let stdout = std::io::stdout();
+        let stdout = io::stdout();
         let mut writer = stdout.lock();
 
         let mut row_count = 0_usize;
@@ -179,13 +179,13 @@ impl PrintOptions {
             query_start_time,
         );
 
-        self.write_output(&mut writer, formatted_exec_details)
+        self.write_output(&mut writer, &formatted_exec_details)
     }
 
     fn write_output<W: io::Write>(
         &self,
         writer: &mut W,
-        formatted_exec_details: String,
+        formatted_exec_details: &str,
     ) -> Result<()> {
         if !self.quiet {
             writeln!(writer, "{formatted_exec_details}")?;
@@ -237,11 +237,11 @@ mod tests {
 
         let mut print_output: Vec<u8> = Vec::new();
         let exec_out = String::from("Formatted Exec Output");
-        print_options.write_output(&mut print_output, exec_out.clone())?;
+        print_options.write_output(&mut print_output, &exec_out)?;
         assert!(print_output.is_empty());
 
         print_options.quiet = false;
-        print_options.write_output(&mut print_output, exec_out.clone())?;
+        print_options.write_output(&mut print_output, &exec_out)?;
         let out_str: String = print_output
             .clone()
             .try_into()
@@ -253,7 +253,7 @@ mod tests {
         print_options
             .instrumented_registry
             .set_instrument_mode(InstrumentedObjectStoreMode::Trace);
-        print_options.write_output(&mut print_output, exec_out.clone())?;
+        print_options.write_output(&mut print_output, &exec_out)?;
         let out_str: String = print_output
             .clone()
             .try_into()

datafusion-examples/examples/ffi/ffi_example_table_provider/Cargo.toml

@@ -28,6 +28,9 @@ datafusion = { workspace = true }
 datafusion-ffi = { workspace = true }
 ffi_module_interface = { path = "../ffi_module_interface" }
 
+[lints]
+workspace = true
+
 [lib]
 name = "ffi_example_table_provider"
 crate-type = ["cdylib", 'rlib']

datafusion-examples/examples/ffi/ffi_module_interface/Cargo.toml

@@ -21,6 +21,9 @@ version = "0.1.0"
 edition = "2024"
 publish = false
 
+[lints]
+workspace = true
+
 [dependencies]
 abi_stable = "0.11.3"
 datafusion-ffi = { workspace = true }

datafusion-examples/examples/ffi/ffi_module_loader/Cargo.toml

@@ -21,6 +21,9 @@ version = "0.1.0"
 edition = "2024"
 publish = false
 
+[lints]
+workspace = true
+
 [dependencies]
 abi_stable = "0.11.3"
 datafusion = { workspace = true }