From 4a730440ba4490223291a22caf5de46fe5daf47c Mon Sep 17 00:00:00 2001
From: Lachlan Deakin <ljdgit@gmail.com>
Date: Tue, 6 Feb 2024 09:30:03 +1100
Subject: [PATCH] Add a global configuration `config::Config` with
 `validate_checksums` option

---
 CHANGELOG.md                                  |  4 +
 Cargo.toml                                    |  2 +-
 .../bytes_to_bytes/crc32c/crc32c_codec.rs     | 15 ++--
 src/config.rs                                 | 78 +++++++++++++++++++
 src/lib.rs                                    |  1 +
 5 files changed, 92 insertions(+), 8 deletions(-)
 create mode 100644 src/config.rs

diff --git a/CHANGELOG.md b/CHANGELOG.md
index 36c391f6..f1d59085 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -7,6 +7,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+### Added
+ - Add a global configuration `config::Config` accessible via `config::{get_config,get_config_mut}`
+   - Currently it exposes a single configuration option: `validate_checksums` (default: `true`)
+
 ## [0.11.5] - 2024-02-05
 
 ### Fixed
diff --git a/Cargo.toml b/Cargo.toml
index e47ee3e5..539fee1c 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "zarrs"
-version = "0.11.5"
+version = "0.11.6"
 authors = ["Lachlan Deakin <ljdgit@gmail.com>"]
 edition = "2021"
 rust-version = "1.71"
diff --git a/src/array/codec/bytes_to_bytes/crc32c/crc32c_codec.rs b/src/array/codec/bytes_to_bytes/crc32c/crc32c_codec.rs
index 35f82399..fc5f1643 100644
--- a/src/array/codec/bytes_to_bytes/crc32c/crc32c_codec.rs
+++ b/src/array/codec/bytes_to_bytes/crc32c/crc32c_codec.rs
@@ -67,14 +67,15 @@ impl BytesToBytesCodecTraits for Crc32cCodec {
         _parallel: bool,
     ) -> Result<Vec<u8>, CodecError> {
         if encoded_value.len() >= CHECKSUM_SIZE {
-            let decoded_value = &encoded_value[..encoded_value.len() - CHECKSUM_SIZE];
-            let checksum = crc32c::crc32c(decoded_value).to_le_bytes();
-            if checksum == encoded_value[encoded_value.len() - CHECKSUM_SIZE..] {
-                encoded_value.resize_with(encoded_value.len() - CHECKSUM_SIZE, Default::default);
-                Ok(encoded_value)
-            } else {
-                Err(CodecError::InvalidChecksum)
+            if crate::config::global_config().validate_checksums() {
+                let decoded_value = &encoded_value[..encoded_value.len() - CHECKSUM_SIZE];
+                let checksum = crc32c::crc32c(decoded_value).to_le_bytes();
+                if checksum != encoded_value[encoded_value.len() - CHECKSUM_SIZE..] {
+                    return Err(CodecError::InvalidChecksum);
+                }
             }
+            encoded_value.resize_with(encoded_value.len() - CHECKSUM_SIZE, Default::default);
+            Ok(encoded_value)
         } else {
             Err(CodecError::Other(
                 "CRC32C checksum decoder expects a 32 bit input".to_string(),
diff --git a/src/config.rs b/src/config.rs
new file mode 100644
index 00000000..9e089460
--- /dev/null
+++ b/src/config.rs
@@ -0,0 +1,78 @@
+//! Zarrs global configuration options.
+
+use std::sync::{OnceLock, RwLock, RwLockReadGuard, RwLockWriteGuard};
+
+/// Global configuration options for the zarrs crate.
+///
+/// Retrieve the global [`Config`] with [`global_config`] and modify it with [`global_config_mut`].
+///
+/// ## Configuration Options
+///
+/// ### Validate Checksums
+///  > default: [`true`]
+///
+/// If enabled, checksum codecs (e.g. `crc32c`) will validate that encoded data matches stored checksums, otherwise validation is skipped.
+/// Note that regardless of this configuration option, checksum codecs may skip validation when partial decoding.
+#[derive(Debug)]
+pub struct Config {
+    validate_checksums: bool,
+}
+
+#[allow(clippy::derivable_impls)]
+impl Default for Config {
+    fn default() -> Self {
+        Config {
+            validate_checksums: true,
+        }
+    }
+}
+
+impl Config {
+    /// Get the [validate checksums](#validate-checksums) configuration.
+    #[must_use]
+    pub fn validate_checksums(&self) -> bool {
+        self.validate_checksums
+    }
+
+    /// Set the [validate checksums](#validate-checksums) configuration.
+    pub fn set_validate_checksums(&mut self, validate_checksums: bool) {
+        self.validate_checksums = validate_checksums;
+    }
+}
+
+static CONFIG: OnceLock<RwLock<Config>> = OnceLock::new();
+
+/// Returns a reference to the global zarrs configuration.
+///
+/// # Panics
+/// This function panics if the underlying lock has been poisoned and might panic if the global config is already held by the current thread.
+pub fn global_config() -> RwLockReadGuard<'static, Config> {
+    CONFIG
+        .get_or_init(|| RwLock::new(Config::default()))
+        .read()
+        .unwrap()
+}
+
+/// Returns a mutable reference to the global zarrs configuration.
+///
+/// # Panics
+/// This function panics if the underlying lock has been poisoned and might panic if the global config is already held by the current thread.
+pub fn global_config_mut() -> RwLockWriteGuard<'static, Config> {
+    CONFIG
+        .get_or_init(|| RwLock::new(Config::default()))
+        .write()
+        .unwrap()
+}
+
+#[cfg(test)]
+mod tests {
+    use super::*;
+
+    #[test]
+    fn config_validate_checksums() {
+        assert!(global_config().validate_checksums());
+        global_config_mut().set_validate_checksums(false);
+        assert!(!global_config().validate_checksums());
+        global_config_mut().set_validate_checksums(true);
+    }
+}
diff --git a/src/lib.rs b/src/lib.rs
index 952ad77b..5033e96c 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -87,6 +87,7 @@
 pub mod array;
 pub mod array_subset;
 pub mod byte_range;
+pub mod config;
 pub mod group;
 pub mod metadata;
 pub mod node;