chore: prune dependencies

.github/workflows/build.yml

@@ -48,7 +48,7 @@ jobs:
         run: cargo clippy --features azure,datafusion,s3,gcs,glue --tests
 
       - name: Spot-check build for native-tls features
-        run: cargo clippy --no-default-features --features azure,datafusion,s3-native-tls,gcs,glue-native-tls --tests
+        run: cargo clippy --no-default-features --features azure,datafusion,s3-native-tls,gcs,glue --tests
 
       - name: Check docs
         run: cargo doc --features azure,datafusion,s3,gcs,glue

Cargo.toml

@@ -28,6 +28,7 @@ arrow-ord = { version = "48.0.1" }
 arrow-row = { version = "48.0.1" }
 arrow-schema = { version = "48.0.1" }
 arrow-select = { version = "48.0.1" }
+object_store = { version = "0.7.1" }
 parquet = { version = "48.0.1" }
 
 # datafusion
@@ -46,7 +47,7 @@ serde_json = "1"
 # "stdlib"
 bytes = { version = "1" }
 chrono = { version = "0.4.31", default-features = false, features = ["clock"] }
-log = { version = "0.4" }
+tracing = { version = "0.1", features = ["log"] }
 regex = { version = "1" }
 thiserror = { version = "1" }
 url = { version = "2" }

crates/deltalake-aws/Cargo.toml

@@ -9,13 +9,15 @@ rusoto_core = { version = "0.47", default-features = false, optional = true }
 rusoto_credential = { version = "0.47" }
 rusoto_sts = { version = "0.47", default-features = false, optional = true }
 rusoto_dynamodb = { version = "0.47", default-features = false, optional = true }
-object_store = { version = "0.7.1", features = ["aws"]}
 lazy_static = "1"
 maplit = "1"
+
+# workspace dependencies
 async-trait = { workspace = true }
 bytes = { workspace = true }
 futures = { workspace = true }
-log = { workspace = true }
+tracing = { workspace = true }
+object_store = { workspace = true, features = ["aws"]}
 thiserror = { workspace = true }
 tokio = { workspace = true }
 regex = { workspace = true }

crates/deltalake-aws/src/lib.rs

@@ -5,14 +5,14 @@ pub mod logstore;
 pub mod storage;
 
 use lazy_static::lazy_static;
-use log::*;
 use regex::Regex;
 use std::{
     collections::HashMap,
     str::FromStr,
     sync::Arc,
     time::{Duration, SystemTime},
 };
+use tracing::debug;
 
 use deltalake_core::logstore::{logstores, LogStore, LogStoreFactory};
 use deltalake_core::storage::{factories, url_prefix_handler, ObjectStoreRef, StorageOptions};

crates/deltalake-aws/src/logstore.rs

@@ -9,7 +9,7 @@ use crate::{constants, CommitEntry, DynamoDbLockClient, UpdateLogEntryResult};
 
 use bytes::Bytes;
 use deltalake_core::{ObjectStoreError, Path};
-use log::*;
+use tracing::{debug, error, warn};
 use url::Url;
 
 use deltalake_core::logstore::*;

crates/deltalake-catalog-glue/Cargo.toml

@@ -5,17 +5,12 @@ edition = "2021"
 
 [dependencies]
 async-trait = { workspace = true }
-aws-config = "0.57.1"
-aws-sdk-glue = "0.35.0"
+aws-config = "1"
+aws-sdk-glue = "1"
 deltalake-core = { path = "../deltalake-core" }
 # This can depend on a lowest common denominator of core once that's released
 # deltalake_core = { version = "0.17.0" }
-log = "0.4"
 thiserror = { workspace = true }
 
 [dev-dependencies]
 tokio = { version = "1", features = ["macros", "rt-multi-thread"] }
-
-[features]
-default = []
-native-tls = []

crates/deltalake-catalog-glue/src/lib.rs

@@ -1,8 +1,7 @@
 //! Glue Data Catalog.
 //!
-use aws_config::SdkConfig;
+use aws_config::{BehaviorVersion, SdkConfig};
 use deltalake_core::data_catalog::{DataCatalog, DataCatalogError};
-use log::*;
 
 #[derive(thiserror::Error, Debug)]
 pub enum GlueError {
@@ -38,7 +37,7 @@ pub struct GlueDataCatalog {
 impl GlueDataCatalog {
     /// Creates a new GlueDataCatalog with environmental configuration
     pub async fn from_env() -> Result<Self, GlueError> {
-        let config = aws_config::load_from_env().await;
+        let config = aws_config::load_defaults(BehaviorVersion::latest()).await;
         let client = aws_sdk_glue::Client::new(&config);
         Ok(Self { client })
     }

crates/deltalake-core/Cargo.toml

@@ -68,13 +68,13 @@ tokio = { workspace = true, features = [
 
 # other deps (these should be organized and pulled into workspace.dependencies as necessary)
 cfg-if = "1"
+dashmap = "5"
 errno = "0.3"
 either = "1.8"
 fix-hidden-lifetime-bug = "0.2"
 hyper = { version = "0.14", optional = true }
-itertools = "0.11"
+itertools = "0.12"
 lazy_static = "1"
-log = "0"
 libc = ">=0.2.90, <1"
 num-bigint = "0.4"
 num-traits = "0.2.15"
@@ -83,7 +83,7 @@ once_cell = "1.16.0"
 parking_lot = "0.12"
 percent-encoding = "2"
 roaring = "0.10.1"
-tracing = { version = "0.1", optional = true }
+tracing = { workspace = true }
 rand = "0.8"
 z85 = "3.0.5"
 maplit = "1"
@@ -93,16 +93,8 @@ reqwest = { version = "0.11.18", default-features = false, features = [
     "rustls-tls",
     "json",
 ], optional = true }
-
-# Datafusion
-dashmap = "5"
-
 sqlparser = { version = "0.39", optional = true }
 
-# NOTE dependencies only for integration tests
-fs_extra = { version = "1.3.0", optional = true }
-tempdir = { version = "0", optional = true }
-
 [dev-dependencies]
 criterion = "0.5"
 ctor = "0"
@@ -114,7 +106,6 @@ pretty_assertions = "1.2.1"
 pretty_env_logger = "*"
 rand = "0.8"
 serial_test = "2"
-tempdir = "0"
 tempfile = "3"
 tokio = { version = "1", features = ["macros", "rt-multi-thread"] }
 utime = "0.3"
@@ -147,10 +138,10 @@ datafusion = [
 datafusion-ext = ["datafusion"]
 gcs = ["object_store/gcp"]
 # used only for integration testing
-integration_test = ["fs_extra", "tempdir"]
+integration_test = []
 json = ["parquet/json"]
 python = ["arrow/pyarrow"]
-unity-experimental = ["reqwest", "tracing", "hyper"]
+unity-experimental = ["reqwest", "hyper"]
 
 [[bench]]
 name = "read_checkpoint"

crates/deltalake-core/src/delta_datafusion/mod.rs

@@ -71,9 +71,9 @@ use datafusion_sql::planner::ParserOptions;
 use futures::TryStreamExt;
 
 use itertools::Itertools;
-use log::error;
 use object_store::ObjectMeta;
 use serde::{Deserialize, Serialize};
+use tracing::error;
 use url::Url;
 
 use crate::errors::{DeltaResult, DeltaTableError};

crates/deltalake-core/src/kernel/actions/types.rs

@@ -5,8 +5,8 @@ use std::str::FromStr;
 // use std::sync::Arc;
 
 // use roaring::RoaringTreemap;
-use log::warn;
 use serde::{Deserialize, Serialize};
+use tracing::warn;
 use url::Url;
 
 use super::super::schema::StructType;

crates/deltalake-core/src/kernel/expressions/mod.rs

@@ -1,9 +1,7 @@
 //! expressions.
 
-use std::{
-    collections::HashSet,
-    fmt::{Display, Formatter},
-};
+use std::collections::HashSet;
+use std::fmt::{Display, Formatter};
 
 use self::scalars::Scalar;
 

crates/deltalake-core/src/kernel/expressions/scalars.rs

@@ -1,9 +1,7 @@
 //! Scalar values for use in expressions.
 
-use std::{
-    cmp::Ordering,
-    fmt::{Display, Formatter},
-};
+use std::cmp::Ordering;
+use std::fmt::{Display, Formatter};
 
 use crate::kernel::schema::{DataType, PrimitiveType};
 

crates/deltalake-core/src/kernel/schema.rs

@@ -6,11 +6,11 @@ use std::hash::{Hash, Hasher};
 use std::sync::Arc;
 use std::{collections::HashMap, fmt::Display};
 
-use crate::kernel::DataCheck;
 use serde::{Deserialize, Serialize};
 use serde_json::Value;
 
 use super::error::Error;
+use crate::kernel::DataCheck;
 
 /// Type alias for a top level schema
 pub type Schema = StructType;

crates/deltalake-core/src/lib.rs

@@ -69,11 +69,6 @@
 #![allow(rustdoc::invalid_html_tags)]
 #![allow(clippy::nonminimal_bool)]
 
-#[cfg(all(feature = "glue", feature = "glue-native-tls"))]
-compile_error!(
-    "Features glue and glue-native-tls are mutually exclusive and cannot be enabled together"
-);
-
 pub mod data_catalog;
 pub mod errors;
 pub mod kernel;

crates/deltalake-core/src/logstore/mod.rs

@@ -22,8 +22,8 @@ use crate::{
     DeltaTableError,
 };
 use bytes::Bytes;
-use log::*;
 use object_store::{path::Path, Error as ObjectStoreError, ObjectStore};
+use tracing::{debug, warn};
 
 #[cfg(feature = "datafusion")]
 use datafusion::datasource::object_store::ObjectStoreUrl;

crates/deltalake-core/src/operations/convert_to_delta.rs

@@ -16,7 +16,6 @@ use futures::{
     future::{self, BoxFuture},
     TryStreamExt,
 };
-use log::{debug, info};
 use parquet::{
     arrow::async_reader::{ParquetObjectReader, ParquetRecordBatchStreamBuilder},
     errors::ParquetError,
@@ -29,6 +28,7 @@ use std::{
     str::{FromStr, Utf8Error},
     sync::Arc,
 };
+use tracing::{debug, info};
 
 /// Error converting a Parquet table to a Delta table
 #[derive(Debug, thiserror::Error)]

crates/deltalake-core/src/operations/create.rs

@@ -330,7 +330,7 @@ mod tests {
     use crate::operations::DeltaOps;
     use crate::table::config::DeltaConfigKey;
     use crate::writer::test_utils::get_delta_schema;
-    use tempdir::TempDir;
+    use tempfile::TempDir;
 
     #[tokio::test]
     async fn test_create() {
@@ -349,7 +349,7 @@ mod tests {
     #[tokio::test]
     async fn test_create_local_relative_path() {
         let table_schema = get_delta_schema();
-        let tmp_dir = TempDir::new_in(".", "tmp_").unwrap();
+        let tmp_dir = TempDir::new_in(".").unwrap();
         let relative_path = format!(
             "./{}",
             tmp_dir.path().file_name().unwrap().to_str().unwrap()
@@ -369,7 +369,7 @@ mod tests {
     #[tokio::test]
     async fn test_create_table_local_path() {
         let schema = get_delta_schema();
-        let tmp_dir = TempDir::new_in(".", "tmp_").unwrap();
+        let tmp_dir = TempDir::new_in(".").unwrap();
         let relative_path = format!(
             "./{}",
             tmp_dir.path().file_name().unwrap().to_str().unwrap()

crates/deltalake-core/src/operations/optimize.rs

@@ -30,13 +30,13 @@ use futures::future::BoxFuture;
 use futures::stream::BoxStream;
 use futures::{Future, StreamExt, TryStreamExt};
 use itertools::Itertools;
-use log::debug;
 use num_cpus;
 use parquet::arrow::async_reader::{ParquetObjectReader, ParquetRecordBatchStreamBuilder};
 use parquet::basic::{Compression, ZstdLevel};
 use parquet::errors::ParquetError;
 use parquet::file::properties::WriterProperties;
 use serde::{Deserialize, Serialize};
+use tracing::debug;
 
 use super::transaction::{commit, PROTOCOL};
 use super::writer::{PartitionWriter, PartitionWriterConfig};

crates/deltalake-core/src/operations/write.rs

@@ -920,7 +920,7 @@ mod tests {
 
     #[tokio::test]
     async fn test_special_characters_write_read() {
-        let tmp_dir = tempdir::TempDir::new("test").unwrap();
+        let tmp_dir = tempfile::tempdir().unwrap();
         let tmp_path = std::fs::canonicalize(tmp_dir.path()).unwrap();
 
         let schema = Arc::new(ArrowSchema::new(vec![

crates/deltalake-core/src/operations/writer.rs

@@ -10,6 +10,7 @@ use object_store::{path::Path, ObjectStore};
 use parquet::arrow::ArrowWriter;
 use parquet::basic::Compression;
 use parquet::file::properties::WriterProperties;
+use tracing::debug;
 
 use crate::crate_version;
 use crate::errors::{DeltaResult, DeltaTableError};
@@ -380,7 +381,7 @@ impl PartitionWriter {
             self.write_batch(&batch.slice(offset, length))?;
             // flush currently buffered data to disk once we meet or exceed the target file size.
             if self.buffer.len() >= self.config.target_file_size {
-                log::debug!("Writing file with size {:?} to disk.", self.buffer.len());
+                debug!("Writing file with size {:?} to disk.", self.buffer.len());
                 self.flush_arrow_writer().await?;
             }
         }

crates/deltalake-core/src/protocol/checkpoints.rs

@@ -10,12 +10,12 @@ use arrow_schema::{ArrowError, Schema as ArrowSchema};
 use chrono::{Datelike, Utc};
 use futures::{StreamExt, TryStreamExt};
 use lazy_static::lazy_static;
-use log::*;
 use object_store::ObjectStore;
 use parquet::arrow::ArrowWriter;
 use parquet::errors::ParquetError;
 use regex::Regex;
 use serde_json::Value;
+use tracing::{debug, error};
 
 use super::{time_utils, ProtocolError};
 use crate::kernel::arrow::delta_log_schema_for_table;

crates/deltalake-core/src/protocol/mod.rs

@@ -12,7 +12,6 @@ mod time_utils;
 use arrow_schema::ArrowError;
 use futures::StreamExt;
 use lazy_static::lazy_static;
-use log::debug;
 use object_store::{path::Path, Error as ObjectStoreError, ObjectStore};
 use regex::Regex;
 use serde::{Deserialize, Serialize};
@@ -22,6 +21,7 @@ use std::collections::HashMap;
 use std::hash::{Hash, Hasher};
 use std::mem::take;
 use std::str::FromStr;
+use tracing::{debug, error};
 
 use crate::errors::{DeltaResult, DeltaTableError};
 use crate::kernel::{Add, CommitInfo, Metadata, Protocol, Remove};
@@ -266,7 +266,7 @@ impl Add {
             Ok(Some(stats)) => Ok(Some(stats)),
             Ok(None) => self.get_json_stats(),
             Err(e) => {
-                log::error!(
+                error!(
                     "Error when reading parquet stats {:?} {e}. Attempting to read json stats",
                     self.stats_parsed
                 );