From 7b82a6f507e97935449ee71c4d07bc587c159d49 Mon Sep 17 00:00:00 2001
From: Abraham Egnor <abraham.egnor@mongodb.com>
Date: Tue, 3 Jun 2025 11:34:57 -0400
Subject: [PATCH] RUST-1798 use simd to optimize utf8 validation

Co-authored-by: Liyixin95 <601947961@qq.com>
---
 Cargo.lock       | 7 +++++++
 Cargo.toml       | 1 +
 src/raw/error.rs | 6 ++----
 src/raw/mod.rs   | 2 +-
 4 files changed, 11 insertions(+), 5 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index cf5216a3..4881faa0 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -140,6 +140,7 @@ dependencies = [
  "serde_json",
  "serde_path_to_error",
  "serde_with",
+ "simdutf8",
  "time",
  "uuid",
 ]
@@ -959,6 +960,12 @@ version = "1.3.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "0fda2ff0d084019ba4d7c6f371c95d8fd75ce3524c3cb8fb653a3023f6323e64"
 
+[[package]]
+name = "simdutf8"
+version = "0.1.5"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "e3a9fe34e3e7a50316060351f37187a3f546bce95496156754b601a5fa71b76e"
+
 [[package]]
 name = "strsim"
 version = "0.11.1"
diff --git a/Cargo.toml b/Cargo.toml
index 44c2eb37..b2eaf819 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -65,6 +65,7 @@ serde_with = { version = "3.1.0", optional = true }
 time = { version = "0.3.9", features = ["formatting", "parsing", "macros", "large-dates"] }
 bitvec = "1.0.1"
 serde_path_to_error = { version = "0.1.16", optional = true }
+simdutf8 = "0.1.5"
 
 [target.'cfg(all(target_arch = "wasm32", target_os = "unknown"))'.dependencies]
 js-sys = "0.3"
diff --git a/src/raw/error.rs b/src/raw/error.rs
index e6ab8fbc..a9912d90 100644
--- a/src/raw/error.rs
+++ b/src/raw/error.rs
@@ -1,5 +1,3 @@
-use std::str::Utf8Error;
-
 use crate::spec::ElementType;
 
 /// An error that occurs when attempting to parse raw BSON bytes.
@@ -44,7 +42,7 @@ pub enum ErrorKind {
     MalformedValue { message: String },
 
     /// Improper UTF-8 bytes were found when proper UTF-8 was expected.
-    Utf8EncodingError(Utf8Error),
+    Utf8EncodingError,
 }
 
 impl std::fmt::Display for Error {
@@ -60,7 +58,7 @@ impl std::fmt::Display for Error {
             ErrorKind::MalformedValue { message } => {
                 write!(f, "{}malformed value: {:?}", prefix, message)
             }
-            ErrorKind::Utf8EncodingError(e) => write!(f, "{}utf-8 encoding error: {}", prefix, e),
+            ErrorKind::Utf8EncodingError => write!(f, "{}utf-8 encoding error", prefix),
         }
     }
 }
diff --git a/src/raw/mod.rs b/src/raw/mod.rs
index d6bcc2c9..4e1dd0d6 100644
--- a/src/raw/mod.rs
+++ b/src/raw/mod.rs
@@ -271,7 +271,7 @@ fn read_lenencode(buf: &[u8]) -> Result<&str> {
 }
 
 fn try_to_str(data: &[u8]) -> Result<&str> {
-    std::str::from_utf8(data).map_err(|e| Error::new(ErrorKind::Utf8EncodingError(e)))
+    simdutf8::basic::from_utf8(data).map_err(|_| Error::new(ErrorKind::Utf8EncodingError))
 }
 
 fn usize_try_from_i32(i: i32) -> Result<usize> {