v0.1.3 release

rusticstuff · May 14, 2021 · 0798da0 · 0798da0
1 parent 5f020de
commit 0798da0
Show file tree

Hide file tree

Showing 7 changed files with 50 additions and 32 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,6 +1,10 @@
 # Changelog
 ## [Unreleased]
 
+## [0.1.3] - 2021-05-14
+### New features
+* Low-level streaming validation API in `simdutf8::basic::imp` 
+
 ## [0.1.2] - 2021-05-09
 ### New features
 * Aarch64 support (e.g. Apple Silicon, Raspberry Pi 4, ...) with nightly Rust and crate feature `aarch64_neon`
@@ -46,7 +50,8 @@
 ## [0.0.1] - 2021-04-20
 - Initial release.
 
-[Unreleased]: https://github.com/rusticstuff/simdutf8/compare/v0.1.2...HEAD
+[Unreleased]: https://github.com/rusticstuff/simdutf8/compare/v0.1.3...HEAD
+[0.1.3]: https://github.com/rusticstuff/simdutf8/compare/v0.1.2...v0.1.3
 [0.1.2]: https://github.com/rusticstuff/simdutf8/compare/v0.1.1...v0.1.2
 [0.1.1]: https://github.com/rusticstuff/simdutf8/compare/v0.1.0...v0.1.1
 [0.1.0]: https://github.com/rusticstuff/simdutf8/compare/v0.0.3...v0.1.0

diff --git a/Cargo.toml b/Cargo.toml
@@ -1,6 +1,6 @@
 [package]
 name = "simdutf8"
-version = "0.1.2"
+version = "0.1.3"
 authors = ["Hans Kratz <[email protected]>"]
 edition = "2018"
 description = "SIMD-accelerated UTF-8 validation."
@@ -11,7 +11,16 @@ readme = "README.md"
 keywords = ["utf-8", "unicode", "string", "validation", "simd"]
 categories = ["encoding", "algorithms", "no-std"]
 license = "MIT OR Apache-2.0"
-exclude = ["/.github", "/.vscode", "/bench", "/fuzzing", "/img", "expected-methods-*.txt"]
+exclude = [
+    "/.gitignore",
+    "/.github",
+    "/.vscode",
+    "/bench",
+    "/fuzzing",
+    "/img",
+    "/inlining",
+    "TODO.md",
+]
 
 [features]
 default = ["std"]

diff --git a/README.md b/README.md
@@ -28,12 +28,12 @@ This library has been thoroughly tested with sample data as well as fuzzing and
 Add the dependency to your Cargo.toml file:
 ```toml
 [dependencies]
-simdutf8 = { version = "0.1.2" }
+simdutf8 = { version = "0.1.3" }
 ```
 or on ARM64 with Rust Nightly:
 ```toml
 [dependencies]
-simdutf8 = { version = "0.1.2", features = ["aarch64_neon"] }
+simdutf8 = { version = "0.1.3", features = ["aarch64_neon"] }
 ```
 
 Use `simdutf8::basic::from_utf8()` as a drop-in replacement for `std::str::from_utf8()`.

diff --git a/src/basic.rs b/src/basic.rs
@@ -28,7 +28,7 @@ impl std::error::Error for Utf8Error {}
 /// Analogue to [`std::str::from_utf8()`].
 ///
 /// Checks if the passed byte sequence is valid UTF-8 and returns an
-/// [`std::str``] reference to the passed byte slice wrapped in `Ok()` if it is.
+/// [`std::str`] reference to the passed byte slice wrapped in `Ok()` if it is.
 ///
 /// # Errors
 /// Will return the zero-sized Err([`Utf8Error`]) on if the input contains invalid UTF-8.
@@ -43,7 +43,7 @@ pub fn from_utf8(input: &[u8]) -> Result<&str, Utf8Error> {
 /// Analogue to [`std::str::from_utf8_mut()`].
 ///
 /// Checks if the passed mutable byte sequence is valid UTF-8 and returns a mutable
-/// [`std::str``] reference to the passed byte slice wrapped in `Ok()` if it is.
+/// [`std::str`] reference to the passed byte slice wrapped in `Ok()` if it is.
 ///
 /// # Errors
 /// Will return the zero-sized Err([`Utf8Error`]) on if the input contains invalid UTF-8.
@@ -58,16 +58,18 @@ pub fn from_utf8_mut(input: &mut [u8]) -> Result<&mut str, Utf8Error> {
 /// Allows direct access to the platform-specific unsafe validation implementations.
 #[cfg(feature = "public_imp")]
 pub mod imp {
+    use crate::basic;
+
     /// A low-level interfacne for streaming validation of UTF-8 data. It is meant to be integrated
     /// in high-performance data processing pipelines.
     ///
     /// Data can be streamed in arbitrarily-sized chunks using the [`Self::update()`] method. There is
     /// no way to find out if the input so far was valid UTF-8 during the validation. Only when
     /// the validation is completed with the [`Self::finalize()`] method the result of the validation is
-    /// returned. Use [`ChunkedUtf8Validator`] is possible for highest performance.
+    /// returned. Use [`ChunkedUtf8Validator`] if possible for highest performance.
     ///
     /// This implementation requires CPU SIMD features specified by the module it resides in.
-    /// It is undefined behavior to call it if the required CPU features are not available which
+    /// It is undefined behavior to use it if the required CPU features are not available which
     /// is why all trait methods are `unsafe`.
     ///
     /// General usage:
@@ -123,13 +125,13 @@ pub mod imp {
         /// Finishes the validation and returns `Ok(())` if the input was valid UTF-8.
         ///
         /// # Errors
-        /// A [`crate::basic::Utf8Error`] is returned if the input was not valid UTF-8. No
+        /// A [`basic::Utf8Error`] is returned if the input was not valid UTF-8. No
         /// further information about the location of the error is provided.
         ///
         /// # Safety
         /// This implementation requires CPU SIMD features specified by the module it resides in.
         /// It is undefined behavior to call it if the required CPU features are not available.
-        unsafe fn finalize(self) -> core::result::Result<(), crate::basic::Utf8Error>;
+        unsafe fn finalize(self) -> core::result::Result<(), basic::Utf8Error>;
     }
 
     /// Like [`Utf8Validator`] this low-level API is for streaming validation of UTF-8 data.
@@ -146,7 +148,7 @@ pub mod imp {
     /// data passed to it.
     ///
     /// This implementation requires CPU SIMD features specified by the module it resides in.
-    /// It is undefined behavior to call it if the required CPU features are not available which
+    /// It is undefined behavior to use it if the required CPU features are not available which
     /// is why all trait methods are `unsafe`.
     pub trait ChunkedUtf8Validator {
         /// Creates a new validator.
@@ -175,7 +177,7 @@ pub mod imp {
         /// Finishes the validation and returns `Ok(())` if the input was valid UTF-8.
         ///
         /// # Errors
-        /// A [`crate::basic::Utf8Error`] is returned if the input was not valid UTF-8. No
+        /// A [`basic::Utf8Error`] is returned if the input was not valid UTF-8. No
         /// further information about the location of the error is provided.
         ///
         /// # Safety
@@ -184,7 +186,7 @@ pub mod imp {
         unsafe fn finalize(
             self,
             remaining_input: core::option::Option<&[u8]>,
-        ) -> core::result::Result<(), crate::basic::Utf8Error>;
+        ) -> core::result::Result<(), basic::Utf8Error>;
     }
 
     /// Includes the x86/x86-64 SIMD implementations.
@@ -201,7 +203,7 @@ pub mod imp {
         }
         /// Includes the validation implementation for SSE 4.2-compatible CPUs.
         ///
-        /// Using the provided functionality on CPUs which do not support AVX 2 is undefined
+        /// Using the provided functionality on CPUs which do not support SSE 4.2 is undefined
         /// behavior and will very likely cause a crash.
         pub mod sse42 {
             pub use crate::implementation::x86::sse42::validate_utf8_basic as validate_utf8;

diff --git a/src/compat.rs b/src/compat.rs
@@ -7,7 +7,7 @@
 //!
 //! The functions in this module also fail early: errors are checked on-the-fly as the string is processed and once
 //! an invalid UTF-8 sequence is encountered, it returns without processing the rest of the data.
-//! This comes at a performance penality compared to the [`crate::basic`] module even if the input is valid UTF-8.
+//! This comes at a slight performance penality compared to the [`crate::basic`] module if the input is valid UTF-8.
 
 use core::fmt::Display;
 use core::fmt::Formatter;

diff --git a/src/implementation/algorithm.rs b/src/implementation/algorithm.rs
@@ -5,6 +5,8 @@
 
 macro_rules! algorithm_simd {
     ($feat:expr) => {
+        use crate::{basic, compat};
+
         impl Utf8CheckAlgorithm<SimdU8Value> {
             #[cfg_attr(not(target_arch="aarch64"), target_feature(enable = $feat))]
             #[inline]
@@ -203,7 +205,7 @@ macro_rules! algorithm_simd {
         /// Validation implementation for CPUs supporting the SIMD extension (see module).
         ///
         /// # Errors
-        /// Return the zero-sized [`crate::basic::Utf8Error`] on failure.
+        /// Returns the zero-sized [`basic::Utf8Error`] on failure.
         ///
         /// # Safety
         /// This function is inherently unsafe because it is compiled with SIMD extensions
@@ -213,7 +215,7 @@ macro_rules! algorithm_simd {
         #[inline]
         pub unsafe fn validate_utf8_basic(
             input: &[u8],
-        ) -> core::result::Result<(), crate::basic::Utf8Error> {
+        ) -> core::result::Result<(), basic::Utf8Error> {
             use crate::implementation::helpers::SIMD_CHUNK_SIZE;
             let len = input.len();
             let mut algorithm = Utf8CheckAlgorithm::<SimdU8Value>::default();
@@ -250,7 +252,7 @@ macro_rules! algorithm_simd {
             }
             algorithm.check_incomplete_pending();
             if algorithm.has_error() {
-                Err(crate::basic::Utf8Error {})
+                Err(basic::Utf8Error {})
             } else {
                 Ok(())
             }
@@ -259,7 +261,7 @@ macro_rules! algorithm_simd {
         /// Validation implementation for CPUs supporting the SIMD extension (see module).
         ///
         /// # Errors
-        /// Return [`crate::compat::Utf8Error`] with detailed error information on failure.
+        /// Returns [`compat::Utf8Error`] with detailed error information on failure.
         ///
         /// # Safety
         /// This function is inherently unsafe because it is compiled with SIMD extensions
@@ -269,7 +271,7 @@ macro_rules! algorithm_simd {
         #[inline]
         pub unsafe fn validate_utf8_compat(
             input: &[u8],
-        ) -> core::result::Result<(), crate::compat::Utf8Error> {
+        ) -> core::result::Result<(), compat::Utf8Error> {
             validate_utf8_compat_simd0(input)
                 .map_err(|idx| crate::implementation::helpers::get_compat_error(input, idx))
         }
@@ -347,7 +349,7 @@ macro_rules! algorithm_simd {
             }
         }
 
-        /// Low-level implementation of the [`crate::basic::imp::Utf8Validator]` trait.
+        /// Low-level implementation of the [`basic::imp::Utf8Validator`] trait.
         ///
         /// This is implementation requires CPU SIMD features specified by the module it resides in.
         /// It is undefined behavior to call it if the required CPU features are not
@@ -371,7 +373,7 @@ macro_rules! algorithm_simd {
         }
 
         #[cfg(feature = "public_imp")]
-        impl crate::basic::imp::Utf8Validator for Utf8ValidatorImp {
+        impl basic::imp::Utf8Validator for Utf8ValidatorImp {
             #[cfg_attr(not(target_arch="aarch64"), target_feature(enable = $feat))]
             #[inline]
             #[must_use]
@@ -424,7 +426,7 @@ macro_rules! algorithm_simd {
 
             #[cfg_attr(not(target_arch="aarch64"), target_feature(enable = $feat))]
             #[inline]
-            unsafe fn finalize(mut self) -> core::result::Result<(), crate::basic::Utf8Error> {
+            unsafe fn finalize(mut self) -> core::result::Result<(), basic::Utf8Error> {
                 if self.incomplete_len != 0 {
                     for i in &mut self.incomplete_data[self.incomplete_len..] {
                         *i = 0
@@ -433,14 +435,14 @@ macro_rules! algorithm_simd {
                 }
                 self.algorithm.check_incomplete_pending();
                 if self.algorithm.has_error() {
-                    Err(crate::basic::Utf8Error {})
+                    Err(basic::Utf8Error {})
                 } else {
                     Ok(())
                 }
             }
         }
 
-        /// Low-level implementation of the [`crate::basic::imp::ChunkedUtf8Validator]` trait.
+        /// Low-level implementation of the [`basic::imp::ChunkedUtf8Validator`] trait.
         ///
         /// This is implementation requires CPU SIMD features specified by the module it resides in.
         /// It is undefined behavior to call it if the required CPU features are not
@@ -451,7 +453,7 @@ macro_rules! algorithm_simd {
         }
 
         #[cfg(feature = "public_imp")]
-        impl crate::basic::imp::ChunkedUtf8Validator for ChunkedUtf8ValidatorImp {
+        impl basic::imp::ChunkedUtf8Validator for ChunkedUtf8ValidatorImp {
             #[cfg_attr(not(target_arch="aarch64"), target_feature(enable = $feat))]
             #[inline]
             #[must_use]
@@ -480,7 +482,7 @@ macro_rules! algorithm_simd {
             unsafe fn finalize(
                 mut self,
                 remaining_input: core::option::Option<&[u8]>,
-            ) -> core::result::Result<(), crate::basic::Utf8Error> {
+            ) -> core::result::Result<(), basic::Utf8Error> {
                 use crate::implementation::helpers::SIMD_CHUNK_SIZE;
 
                 if let Some(mut remaining_input) = remaining_input {
@@ -505,7 +507,7 @@ macro_rules! algorithm_simd {
                 }
                 self.algorithm.check_incomplete_pending();
                 if self.algorithm.has_error() {
-                    Err(crate::basic::Utf8Error {})
+                    Err(basic::Utf8Error {})
                 } else {
                     Ok(())
                 }

diff --git a/src/lib.rs b/src/lib.rs
@@ -25,12 +25,12 @@
 //! Add the dependency to your Cargo.toml file:
 //! ```toml
 //! [dependencies]
-//! simdutf8 = { version = "0.1.2" }
+//! simdutf8 = { version = "0.1.3" }
 //! ```
 //! or on ARM64 with Rust Nightly:
 //! ```toml
 //! [dependencies]
-//! simdutf8 = { version = "0.1.2", features = ["aarch64_neon"] }
+//! simdutf8 = { version = "0.1.3", features = ["aarch64_neon"] }
 //! ```
 //!
 //! Use [`basic::from_utf8()`] as a drop-in replacement for `std::str::from_utf8()`.
@@ -87,7 +87,7 @@
 //!
 //! ### Access to low-level functionality
 //! If you want to be able to call a SIMD implementation directly, use the `public_imp` feature flag. The validation
-//! implementations are then accessible via [`basic::imp`] and [`compat::imp`].Traits facilitating streaming validation are available
+//! implementations are then accessible via [`basic::imp`] and [`compat::imp`]. Traits facilitating streaming validation are available
 //! there as well.
 //!
 //! ## Optimisation flags