AVRO-3839: [Rust] Replace lazy_static crate with std::sync::OnceLock (#…

…2461) * AVRO-3839: [Rust] Replace lazy_static crate with std::sync::OnceLock Signed-off-by: Martin Tzvetanov Grigorov <[email protected]> * AVRO-3839: [Rust] Bump minimal Rust version to 1.70.0 Signed-off-by: Martin Tzvetanov Grigorov <[email protected]> --------- Signed-off-by: Martin Tzvetanov Grigorov <[email protected]>
apache · Dec 11, 2023 · 92512b2 · 92512b2
1 parent bc9dc05
commit 92512b2
Show file tree

Hide file tree

Showing 13 changed files with 273 additions and 155 deletions.
diff --git a/.github/workflows/test-lang-rust-ci.yml b/.github/workflows/test-lang-rust-ci.yml
@@ -50,7 +50,7 @@ jobs:
           - 'stable'
           - 'beta'
           - 'nightly'
-          - '1.65.0'  # MSRV
+          - '1.70.0'  # MSRV
         target:
           - x86_64-unknown-linux-gnu
           - wasm32-unknown-unknown
@@ -273,4 +273,4 @@ jobs:
       - name: Build
         run: |
           set -x
-          ./build.sh test
+          ./build.sh test
diff --git a/.github/workflows/test-lang-rust-clippy.yml b/.github/workflows/test-lang-rust-clippy.yml
@@ -47,7 +47,7 @@ jobs:
       matrix:
         rust:
           - 'stable'
-          - '1.65.0'  # MSRV
+          - '1.70.0'  # MSRV
     steps:
       - uses: actions/checkout@v4
       - uses: dtolnay/rust-toolchain@nightly

diff --git a/lang/rust/Cargo.lock b/lang/rust/Cargo.lock
diff --git a/lang/rust/Cargo.toml b/lang/rust/Cargo.toml
@@ -34,14 +34,13 @@ authors = ["Apache Avro team <[email protected]>"]
 license = "Apache-2.0"
 repository = "https://github.com/apache/avro"
 edition = "2021"
-rust-version = "1.65.0"
+rust-version = "1.70.0"
 keywords = ["avro", "data", "serialization"]
 categories = ["encoding"]
 documentation = "https://docs.rs/apache-avro"
 
 # dependencies used by more than one members
 [workspace.dependencies]
-lazy_static = { default-features = false, version = "1.4.0" }
 log = { default-features = false, version = "0.4.20" }
 serde = { default-features = false, version = "1.0.193", features = ["derive"] }
 serde_json = { default-features = false, version = "1.0.108", features = ["std"] }

diff --git a/lang/rust/avro/Cargo.toml b/lang/rust/avro/Cargo.toml
@@ -59,7 +59,6 @@ bigdecimal = { default-features = false, version = "0.4.2", features = ["std"] }
 bzip2 = { default-features = false, version = "0.4.4", optional = true }
 crc32fast = { default-features = false, version = "1.3.2", optional = true }
 digest = { default-features = false, version = "0.10.7", features = ["core-api"] }
-lazy_static = { workspace = true }
 libflate = { default-features = false, version = "2.0.0", features = ["std"] }
 log = { workspace = true }
 num-bigint = { default-features = false, version = "0.4.4" }

diff --git a/lang/rust/avro/README.md b/lang/rust/avro/README.md
@@ -656,7 +656,7 @@ assert!(SchemaCompatibility::can_read(&writers_schema, &readers_schema).is_err()
 
 ## Minimal supported Rust version
 
-1.65.0
+1.70.0
 
 ## License
 This project is licensed under [Apache License 2.0](https://github.com/apache/avro/blob/main/LICENSE.txt).

diff --git a/lang/rust/avro/src/rabin.rs b/lang/rust/avro/src/rabin.rs
@@ -20,22 +20,23 @@ use digest::{
     consts::U8, core_api::OutputSizeUser, generic_array::GenericArray, FixedOutput,
     FixedOutputReset, HashMarker, Output, Reset, Update,
 };
-use lazy_static::lazy_static;
+use std::sync::OnceLock;
 
 const EMPTY: i64 = -4513414715797952619;
 
-lazy_static! {
-    static ref FPTABLE: [i64; 256] = {
+fn fp_table() -> &'static [i64; 256] {
+    static FPTABLE_ONCE: OnceLock<[i64; 256]> = OnceLock::new();
+    FPTABLE_ONCE.get_or_init(|| {
         let mut fp_table: [i64; 256] = [0; 256];
         for i in 0..256 {
             let mut fp = i;
             for _ in 0..8 {
                 fp = (fp as u64 >> 1) as i64 ^ (EMPTY & -(fp & 1));
             }
-            fp_table[i as usize] = fp
+            fp_table[i as usize] = fp;
         }
         fp_table
-    };
+    })
 }
 
 /// Implementation of the Rabin fingerprint algorithm using the Digest trait as described in [schema_fingerprints](https://avro.apache.org/docs/current/spec.html#schema_fingerprints).
@@ -94,7 +95,7 @@ impl Update for Rabin {
     fn update(&mut self, data: &[u8]) {
         for b in data {
             self.result = (self.result as u64 >> 8) as i64
-                ^ FPTABLE[((self.result ^ *b as i64) & 0xff) as usize];
+                ^ fp_table()[((self.result ^ *b as i64) & 0xff) as usize];
         }
     }
 }

diff --git a/lang/rust/avro/src/schema.rs b/lang/rust/avro/src/schema.rs
@@ -18,7 +18,6 @@
 //! Logic for parsing and interacting with schemas in Avro format.
 use crate::{error::Error, types, util::MapHelper, AvroResult};
 use digest::Digest;
-use lazy_static::lazy_static;
 use regex_lite::Regex;
 use serde::{
     ser::{SerializeMap, SerializeSeq},
@@ -34,19 +33,36 @@ use std::{
     hash::Hash,
     io::Read,
     str::FromStr,
+    sync::OnceLock,
 };
 use strum_macros::{EnumDiscriminants, EnumString};
 
-lazy_static! {
-    static ref ENUM_SYMBOL_NAME_R: Regex = Regex::new(r"^[A-Za-z_][A-Za-z0-9_]*$").unwrap();
+fn enum_symbol_name_r() -> &'static Regex {
+    static ENUM_SYMBOL_NAME_ONCE: OnceLock<Regex> = OnceLock::new();
+    ENUM_SYMBOL_NAME_ONCE.get_or_init(|| Regex::new(r"^[A-Za-z_][A-Za-z0-9_]*$").unwrap())
+}
 
-    // An optional namespace (with optional dots) followed by a name without any dots in it.
-    static ref SCHEMA_NAME_R: Regex =
-        Regex::new(r"^((?P<namespace>([A-Za-z_][A-Za-z0-9_]*(\.[A-Za-z_][A-Za-z0-9_]*)*)?)\.)?(?P<name>[A-Za-z_][A-Za-z0-9_]*)$").unwrap();
+// An optional namespace (with optional dots) followed by a name without any dots in it.
+fn schema_name_r() -> &'static Regex {
+    static SCHEMA_NAME_ONCE: OnceLock<Regex> = OnceLock::new();
+    SCHEMA_NAME_ONCE.get_or_init(|| {
+        Regex::new(
+            r"^((?P<namespace>([A-Za-z_][A-Za-z0-9_]*(\.[A-Za-z_][A-Za-z0-9_]*)*)?)\.)?(?P<name>[A-Za-z_][A-Za-z0-9_]*)$",
+        )
+        .unwrap()
+    })
+}
 
-    static ref FIELD_NAME_R: Regex = Regex::new(r"^[A-Za-z_][A-Za-z0-9_]*$").unwrap();
+fn field_name_r() -> &'static Regex {
+    static FIELD_NAME_ONCE: OnceLock<Regex> = OnceLock::new();
+    FIELD_NAME_ONCE.get_or_init(|| Regex::new(r"^[A-Za-z_][A-Za-z0-9_]*$").unwrap())
+}
 
-    static ref NAMESPACE_R: Regex = Regex::new(r"^([A-Za-z_][A-Za-z0-9_]*(\.[A-Za-z_][A-Za-z0-9_]*)*)?$").unwrap();
+fn namespace_r() -> &'static Regex {
+    static NAMESPACE_ONCE: OnceLock<Regex> = OnceLock::new();
+    NAMESPACE_ONCE.get_or_init(|| {
+        Regex::new(r"^([A-Za-z_][A-Za-z0-9_]*(\.[A-Za-z_][A-Za-z0-9_]*)*)?$").unwrap()
+    })
 }
 
 /// Represents an Avro schema fingerprint
@@ -252,9 +268,9 @@ impl Name {
     }
 
     fn get_name_and_namespace(name: &str) -> AvroResult<(String, Namespace)> {
-        let caps = SCHEMA_NAME_R
+        let caps = schema_name_r()
             .captures(name)
-            .ok_or_else(|| Error::InvalidSchemaName(name.to_string(), SCHEMA_NAME_R.as_str()))?;
+            .ok_or_else(|| Error::InvalidSchemaName(name.to_string(), schema_name_r().as_str()))?;
         Ok((
             caps["name"].to_string(),
             caps.name("namespace").map(|s| s.as_str().to_string()),
@@ -285,10 +301,10 @@ impl Name {
             .filter(|ns| !ns.is_empty());
 
         if let Some(ref ns) = namespace {
-            if !NAMESPACE_R.is_match(ns) {
+            if !namespace_r().is_match(ns) {
                 return Err(Error::InvalidNamespace(
                     ns.to_string(),
-                    NAMESPACE_R.as_str(),
+                    namespace_r().as_str(),
                 ));
             }
         }
@@ -657,7 +673,7 @@ impl RecordField {
     ) -> AvroResult<Self> {
         let name = field.name().ok_or(Error::GetNameFieldFromRecord)?;
 
-        if !FIELD_NAME_R.is_match(&name) {
+        if !field_name_r().is_match(&name) {
             return Err(Error::FieldName(name));
         }
 
@@ -1617,7 +1633,7 @@ impl Parser {
         let mut existing_symbols: HashSet<&String> = HashSet::with_capacity(symbols.len());
         for symbol in symbols.iter() {
             // Ensure enum symbol names match [A-Za-z_][A-Za-z0-9_]*
-            if !ENUM_SYMBOL_NAME_R.is_match(symbol) {
+            if !enum_symbol_name_r().is_match(symbol) {
                 return Err(Error::EnumSymbolName(symbol.to_string()));
             }
 
@@ -6193,15 +6209,15 @@ mod tests {
         let name = Name::new(full_name);
         assert!(name.is_err());
         let expected =
-            Error::InvalidSchemaName(full_name.to_string(), SCHEMA_NAME_R.as_str()).to_string();
+            Error::InvalidSchemaName(full_name.to_string(), schema_name_r().as_str()).to_string();
         let err = name.map_err(|e| e.to_string()).err().unwrap();
         assert_eq!(expected, err);
 
         let full_name = "ns..record1";
         let name = Name::new(full_name);
         assert!(name.is_err());
         let expected =
-            Error::InvalidSchemaName(full_name.to_string(), SCHEMA_NAME_R.as_str()).to_string();
+            Error::InvalidSchemaName(full_name.to_string(), schema_name_r().as_str()).to_string();
         let err = name.map_err(|e| e.to_string()).err().unwrap();
         assert_eq!(expected, err);
         Ok(())