From 917080a5c27b3310daab135f9bfdbc531cb54186 Mon Sep 17 00:00:00 2001 From: Joe Wilm Date: Sat, 17 Sep 2016 17:02:29 -0700 Subject: Move utf8 parsing into separate crate --- Cargo.toml | 3 +- codegen/src/ext/utf8.rs | 2 +- codegen/src/main.rs | 4 +- src/lib.rs | 3 +- src/utf8/mod.rs | 91 ----------------------- src/utf8/table.rs | 184 ---------------------------------------------- src/utf8/table.rs.in | 60 --------------- src/utf8/types.rs | 77 ------------------- utf8parse/Cargo.toml | 6 ++ utf8parse/src/lib.rs | 93 +++++++++++++++++++++++ utf8parse/src/table.rs | 184 ++++++++++++++++++++++++++++++++++++++++++++++ utf8parse/src/table.rs.in | 60 +++++++++++++++ utf8parse/src/types.rs | 77 +++++++++++++++++++ 13 files changed, 427 insertions(+), 417 deletions(-) delete mode 100644 src/utf8/mod.rs delete mode 100644 src/utf8/table.rs delete mode 100644 src/utf8/table.rs.in delete mode 100644 src/utf8/types.rs create mode 100644 utf8parse/Cargo.toml create mode 100644 utf8parse/src/lib.rs create mode 100644 utf8parse/src/table.rs create mode 100644 utf8parse/src/table.rs.in create mode 100644 utf8parse/src/types.rs diff --git a/Cargo.toml b/Cargo.toml index 1e38f87..494ece6 100644 --- a/Cargo.toml +++ b/Cargo.toml @@ -3,4 +3,5 @@ name = "vte" version = "0.1.0" authors = ["Joe Wilm "] -[dependencies] +[dependencies.utf8parse] +path = "./utf8parse" diff --git a/codegen/src/ext/utf8.rs b/codegen/src/ext/utf8.rs index 5b73081..17ebc22 100644 --- a/codegen/src/ext/utf8.rs +++ b/codegen/src/ext/utf8.rs @@ -13,7 +13,7 @@ use syntex_syntax::parse::PResult; use syntex_syntax::ptr::P; use syntex_syntax::tokenstream::TokenTree; -#[path="../../../src/utf8/types.rs"] +#[path="../../../utf8parse/src/types.rs"] mod types; use self::types::{State, Action, pack}; diff --git a/codegen/src/main.rs b/codegen/src/main.rs index 5f8d153..0d21577 100644 --- a/codegen/src/main.rs +++ b/codegen/src/main.rs @@ -17,7 +17,7 @@ fn main() { // Expand UTF8 parser state table let mut registry = syntex::Registry::new(); ext::utf8::register(&mut registry); - let src = &Path::new("../src/utf8/table.rs.in"); - let dst = &Path::new("../src/utf8/table.rs"); + let src = &Path::new("../utf8parse/src/table.rs.in"); + let dst = &Path::new("../utf8parse/src/table.rs"); registry.expand("utf8_state_table", src, dst).expect("expand utf8_stable_table ok"); } diff --git a/src/lib.rs b/src/lib.rs index 71953fd..f449e13 100644 --- a/src/lib.rs +++ b/src/lib.rs @@ -23,9 +23,10 @@ //! [`Parser`]: struct.Parser.html //! [`Perform`]: trait.Perform.html //! [Paul Williams' ANSI parser state machine]: http://vt100.net/emu/dec_ansi_parser +extern crate utf8parse as utf8; + mod table; mod definitions; -mod utf8; use definitions::{Action, State, unpack}; diff --git a/src/utf8/mod.rs b/src/utf8/mod.rs deleted file mode 100644 index 3d099b1..0000000 --- a/src/utf8/mod.rs +++ /dev/null @@ -1,91 +0,0 @@ -//! A table-driven UTF-8 Parser -//! -//! This module implements a table-driven UTF-8 parser which should -//! theoretically contain the minimal number of branches (1). The only branch is -//! on the `Action` returned from unpacking a transition. -use std::char; - -mod types; -use self::types::{State, Action, unpack}; - -mod table; -use self::table::TRANSITIONS; - -/// Handles codepoint and invalid sequence events from the parser. -pub trait Receiver { - /// Code point parsed - /// - /// Called with the codepoint - fn codepoint(&mut self, char); - - /// Invalid sequence encountered - fn invalid_sequence(&mut self); -} - -/// A parser for Utf8 Characters -/// -/// Repeatedly call `advance` with bytes to emit Utf8 characters -pub struct Parser { - point: u32, - state: State, -} - -/// Continuation bytes are masked with this value. -const CONTINUATION_MASK: u8 = 0b0011_1111; - -impl Parser { - /// Create a new Parser - pub fn new() -> Parser { - Parser { - point: 0, - state: State::Ground, - } - } - - pub fn advance(&mut self, receiver: &mut R, byte: u8) - where R: Receiver - { - let cur = self.state as usize; - let change = TRANSITIONS[cur][byte as usize]; - let (state, action) = unsafe { unpack(change) }; - - self.perform_action(receiver, byte, action); - self.state = state; - } - - fn perform_action(&mut self, receiver: &mut R, byte: u8, action: Action) - where R: Receiver - { - match action { - Action::InvalidSequence => { - self.point = 0; - receiver.invalid_sequence(); - }, - Action::EmitByte => { - receiver.codepoint(byte as char); - }, - Action::SetByte1 => { - let point = self.point | ((byte & CONTINUATION_MASK) as u32); - let c = unsafe { char::from_u32_unchecked(point) }; - self.point = 0; - - receiver.codepoint(c); - }, - Action::SetByte2 => { - self.point |= ((byte & CONTINUATION_MASK) as u32) << 6; - }, - Action::SetByte2Top => { - self.point |= ((byte & 0b0001_1111) as u32) << 6; - }, - Action::SetByte3 => { - self.point |= ((byte & CONTINUATION_MASK) as u32) << 12; - }, - Action::SetByte3Top => { - self.point |= ((byte & 0b0000_1111) as u32) << 12; - }, - Action::SetByte4 => { - self.point |= ((byte & 0b0000_0111) as u32) << 18; - }, - } - } -} diff --git a/src/utf8/table.rs b/src/utf8/table.rs deleted file mode 100644 index 5a1292b..0000000 --- a/src/utf8/table.rs +++ /dev/null @@ -1,184 +0,0 @@ -//! UTF-8 Parse Transition Table - -/// Transition table for parsing UTF-8. This is built from the grammar described -/// at https://tools.ietf.org/html/rfc3629#section-4 which I have copied and -/// formatted below. -/// -/// # UTF-8 Grammar -/// -/// ```ignore -/// UTF8-octets = *( UTF8-char ) -/// UTF8-char = UTF8-1 / UTF8-2 / UTF8-3 / UTF8-4 -/// UTF8-1 = %x00-7F -/// UTF8-2 = %xC2-DF UTF8-tail -/// UTF8-3 = %xE0 %xA0-BF UTF8-tail / -/// %xE1-EC 2( UTF8-tail ) / -/// %xED %x80-9F UTF8-tail / -/// %xEE-EF 2( UTF8-tail ) -/// UTF8-4 = %xF0 %x90-BF 2( UTF8-tail ) / -/// %xF1-F3 3( UTF8-tail ) / -/// %xF4 %x80-8F 2( UTF8-tail ) -/// UTF8-tail = %x80-BF -/// ``` -/// -/// Not specifying an action in this table is equivalent to specifying -/// Action::InvalidSequence. Not specifying a state is equivalent to specifying -/// state::ground. -pub static TRANSITIONS: [[u8; 256]; 8] = - [[16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, - 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, - 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, - 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 100u8, 98u8, - 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 101u8, - 98u8, 98u8, 118u8, 113u8, 113u8, 113u8, 119u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8], - [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8], - [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8], - [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, - 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, - 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, - 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, - 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, - 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8], - [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8], - [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, - 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8], - [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8], - [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, - 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, - 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8]]; diff --git a/src/utf8/table.rs.in b/src/utf8/table.rs.in deleted file mode 100644 index 2acafe7..0000000 --- a/src/utf8/table.rs.in +++ /dev/null @@ -1,60 +0,0 @@ -//! UTF-8 Parse Transition Table - -/// Transition table for parsing UTF-8. This is built from the grammar described -/// at https://tools.ietf.org/html/rfc3629#section-4 which I have copied and -/// formatted below. -/// -/// # UTF-8 Grammar -/// -/// ```ignore -/// UTF8-octets = *( UTF8-char ) -/// UTF8-char = UTF8-1 / UTF8-2 / UTF8-3 / UTF8-4 -/// UTF8-1 = %x00-7F -/// UTF8-2 = %xC2-DF UTF8-tail -/// UTF8-3 = %xE0 %xA0-BF UTF8-tail / -/// %xE1-EC 2( UTF8-tail ) / -/// %xED %x80-9F UTF8-tail / -/// %xEE-EF 2( UTF8-tail ) -/// UTF8-4 = %xF0 %x90-BF 2( UTF8-tail ) / -/// %xF1-F3 3( UTF8-tail ) / -/// %xF4 %x80-8F 2( UTF8-tail ) -/// UTF8-tail = %x80-BF -/// ``` -/// -/// Not specifying an action in this table is equivalent to specifying -/// Action::InvalidSequence. Not specifying a state is equivalent to specifying -/// state::ground. -pub static TRANSITIONS: [[u8; 256]; 8] = utf8_state_table! { - State::Ground => { - 0x00...0x7f => (State::Ground, Action::EmitByte), - 0xc2...0xdf => (State::Tail1, Action::SetByte2Top), - 0xe0 => (State::U3_2_e0, Action::SetByte3Top), - 0xe1...0xec => (State::Tail2, Action::SetByte3Top), - 0xed => (State::U3_2_ed, Action::SetByte3Top), - 0xee...0xef => (State::Tail2, Action::SetByte3Top), - 0xf0 => (State::Utf8_4_3_f0, Action::SetByte4), - 0xf1...0xf3 => (State::Tail3, Action::SetByte4), - 0xf4 => (State::Utf8_4_3_f4, Action::SetByte4), - }, - State::U3_2_e0 => { - 0xa0...0xbf => (State::Tail1, Action::SetByte2), - }, - State::U3_2_ed => { - 0x80...0x9f => (State::Tail1, Action::SetByte2), - }, - State::Utf8_4_3_f0 => { - 0x90...0xbf => (State::Tail2, Action::SetByte3), - }, - State::Utf8_4_3_f4 => { - 0x80...0x8f => (State::Tail2, Action::SetByte3), - }, - State::Tail3 => { - 0x80...0xbf => (State::Tail2, Action::SetByte3), - }, - State::Tail2 => { - 0x80...0xbf => (State::Tail1, Action::SetByte2), - }, - State::Tail1 => { - 0x80...0xbf => (State::Ground, Action::SetByte1), - }, -}; diff --git a/src/utf8/types.rs b/src/utf8/types.rs deleted file mode 100644 index 4c604f4..0000000 --- a/src/utf8/types.rs +++ /dev/null @@ -1,77 +0,0 @@ -//! Types supporting the UTF-8 parser -#![allow(non_camel_case_types)] -use std::mem; - -/// States the parser can be in. -/// -/// There is a state for each initial input of the 3 and 4 byte sequences since -/// the following bytes are subject to different conditions than a tail byte. -#[allow(dead_code)] -#[derive(Debug, Copy, Clone)] -pub enum State { - /// Ground state; expect anything - Ground = 0, - /// 3 tail bytes - Tail3 = 1, - /// 2 tail bytes - Tail2 = 2, - /// 1 tail byte - Tail1 = 3, - /// UTF8-3 starting with E0 - U3_2_e0 = 4, - /// UTF8-3 starting with ED - U3_2_ed = 5, - /// UTF8-4 starting with F0 - Utf8_4_3_f0 = 6, - /// UTF8-4 starting with F4 - Utf8_4_3_f4 = 7, -} - -/// Action to take when receiving a byte -#[allow(dead_code)] -#[derive(Debug, Copy, Clone)] -pub enum Action { - /// Unexpected byte; sequence is invalid - InvalidSequence = 0, - /// Received valid 7-bit ASCII byte which can be directly emitted. - EmitByte = 1, - /// Set the bottom continuation byte - SetByte1 = 2, - /// Set the 2nd-from-last continuation byte - SetByte2 = 3, - /// Set the 2nd-from-last byte which is part of a two byte sequence - SetByte2Top = 4, - /// Set the 3rd-from-last continuation byte - SetByte3 = 5, - /// Set the 3rd-from-last byte which is part of a three byte sequence - SetByte3Top = 6, - /// Set the top byte of a four byte sequence. - SetByte4 = 7, -} - -/// Convert a state and action to a u8 -/// -/// State will be the bottom 4 bits and action the top 4 -#[inline] -#[allow(dead_code)] -pub fn pack(state: State, action: Action) -> u8 { - ((action as u8) << 4) | (state as u8) -} - -/// Convert a u8 to a state and action -/// -/// # Unsafety -/// -/// If this function is called with a byte that wasn't encoded with the `pack` -/// function in this module, there is no guarantee that a valid state and action -/// can be produced. -#[inline] -pub unsafe fn unpack(val: u8) -> (State, Action) { - ( - // State is stored in bottom 4 bits - mem::transmute(val & 0x0f), - - // Action is stored in top 4 bits - mem::transmute(val >> 4), - ) -} diff --git a/utf8parse/Cargo.toml b/utf8parse/Cargo.toml new file mode 100644 index 0000000..882b108 --- /dev/null +++ b/utf8parse/Cargo.toml @@ -0,0 +1,6 @@ +[package] +name = "utf8parse" +version = "0.1.0" +authors = ["Joe Wilm "] + +[dependencies] diff --git a/utf8parse/src/lib.rs b/utf8parse/src/lib.rs new file mode 100644 index 0000000..9585642 --- /dev/null +++ b/utf8parse/src/lib.rs @@ -0,0 +1,93 @@ +//! A table-driven UTF-8 Parser +//! +//! This module implements a table-driven UTF-8 parser which should +//! theoretically contain the minimal number of branches (1). The only branch is +//! on the `Action` returned from unpacking a transition. +use std::char; + +mod types; +use self::types::{State, Action, unpack}; + +mod table; +use self::table::TRANSITIONS; + +/// Handles codepoint and invalid sequence events from the parser. +pub trait Receiver { + /// Called whenever a codepoint is parsed successfully + fn codepoint(&mut self, char); + + /// Called when an invalid_sequence is detected + fn invalid_sequence(&mut self); +} + +/// A parser for Utf8 Characters +/// +/// Repeatedly call `advance` with bytes to emit Utf8 characters +pub struct Parser { + point: u32, + state: State, +} + +/// Continuation bytes are masked with this value. +const CONTINUATION_MASK: u8 = 0b0011_1111; + +impl Parser { + /// Create a new Parser + pub fn new() -> Parser { + Parser { + point: 0, + state: State::Ground, + } + } + + /// Advance the parser + /// + /// The provider receiver will be called whenever a codepoint is completed or an invalid + /// sequence is detected. + pub fn advance(&mut self, receiver: &mut R, byte: u8) + where R: Receiver + { + let cur = self.state as usize; + let change = TRANSITIONS[cur][byte as usize]; + let (state, action) = unsafe { unpack(change) }; + + self.perform_action(receiver, byte, action); + self.state = state; + } + + fn perform_action(&mut self, receiver: &mut R, byte: u8, action: Action) + where R: Receiver + { + match action { + Action::InvalidSequence => { + self.point = 0; + receiver.invalid_sequence(); + }, + Action::EmitByte => { + receiver.codepoint(byte as char); + }, + Action::SetByte1 => { + let point = self.point | ((byte & CONTINUATION_MASK) as u32); + let c = unsafe { char::from_u32_unchecked(point) }; + self.point = 0; + + receiver.codepoint(c); + }, + Action::SetByte2 => { + self.point |= ((byte & CONTINUATION_MASK) as u32) << 6; + }, + Action::SetByte2Top => { + self.point |= ((byte & 0b0001_1111) as u32) << 6; + }, + Action::SetByte3 => { + self.point |= ((byte & CONTINUATION_MASK) as u32) << 12; + }, + Action::SetByte3Top => { + self.point |= ((byte & 0b0000_1111) as u32) << 12; + }, + Action::SetByte4 => { + self.point |= ((byte & 0b0000_0111) as u32) << 18; + }, + } + } +} diff --git a/utf8parse/src/table.rs b/utf8parse/src/table.rs new file mode 100644 index 0000000..5a1292b --- /dev/null +++ b/utf8parse/src/table.rs @@ -0,0 +1,184 @@ +//! UTF-8 Parse Transition Table + +/// Transition table for parsing UTF-8. This is built from the grammar described +/// at https://tools.ietf.org/html/rfc3629#section-4 which I have copied and +/// formatted below. +/// +/// # UTF-8 Grammar +/// +/// ```ignore +/// UTF8-octets = *( UTF8-char ) +/// UTF8-char = UTF8-1 / UTF8-2 / UTF8-3 / UTF8-4 +/// UTF8-1 = %x00-7F +/// UTF8-2 = %xC2-DF UTF8-tail +/// UTF8-3 = %xE0 %xA0-BF UTF8-tail / +/// %xE1-EC 2( UTF8-tail ) / +/// %xED %x80-9F UTF8-tail / +/// %xEE-EF 2( UTF8-tail ) +/// UTF8-4 = %xF0 %x90-BF 2( UTF8-tail ) / +/// %xF1-F3 3( UTF8-tail ) / +/// %xF4 %x80-8F 2( UTF8-tail ) +/// UTF8-tail = %x80-BF +/// ``` +/// +/// Not specifying an action in this table is equivalent to specifying +/// Action::InvalidSequence. Not specifying a state is equivalent to specifying +/// state::ground. +pub static TRANSITIONS: [[u8; 256]; 8] = + [[16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, + 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 16u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, + 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, + 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 67u8, 100u8, 98u8, + 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 98u8, 101u8, + 98u8, 98u8, 118u8, 113u8, 113u8, 113u8, 119u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8], + [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8], + [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8], + [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, + 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, + 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, + 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, + 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, + 32u8, 32u8, 32u8, 32u8, 32u8, 32u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8], + [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8], + [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, + 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 51u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8], + [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8], + [0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, + 82u8, 82u8, 82u8, 82u8, 82u8, 82u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, + 0u8, 0u8, 0u8, 0u8, 0u8, 0u8, 0u8]]; diff --git a/utf8parse/src/table.rs.in b/utf8parse/src/table.rs.in new file mode 100644 index 0000000..2acafe7 --- /dev/null +++ b/utf8parse/src/table.rs.in @@ -0,0 +1,60 @@ +//! UTF-8 Parse Transition Table + +/// Transition table for parsing UTF-8. This is built from the grammar described +/// at https://tools.ietf.org/html/rfc3629#section-4 which I have copied and +/// formatted below. +/// +/// # UTF-8 Grammar +/// +/// ```ignore +/// UTF8-octets = *( UTF8-char ) +/// UTF8-char = UTF8-1 / UTF8-2 / UTF8-3 / UTF8-4 +/// UTF8-1 = %x00-7F +/// UTF8-2 = %xC2-DF UTF8-tail +/// UTF8-3 = %xE0 %xA0-BF UTF8-tail / +/// %xE1-EC 2( UTF8-tail ) / +/// %xED %x80-9F UTF8-tail / +/// %xEE-EF 2( UTF8-tail ) +/// UTF8-4 = %xF0 %x90-BF 2( UTF8-tail ) / +/// %xF1-F3 3( UTF8-tail ) / +/// %xF4 %x80-8F 2( UTF8-tail ) +/// UTF8-tail = %x80-BF +/// ``` +/// +/// Not specifying an action in this table is equivalent to specifying +/// Action::InvalidSequence. Not specifying a state is equivalent to specifying +/// state::ground. +pub static TRANSITIONS: [[u8; 256]; 8] = utf8_state_table! { + State::Ground => { + 0x00...0x7f => (State::Ground, Action::EmitByte), + 0xc2...0xdf => (State::Tail1, Action::SetByte2Top), + 0xe0 => (State::U3_2_e0, Action::SetByte3Top), + 0xe1...0xec => (State::Tail2, Action::SetByte3Top), + 0xed => (State::U3_2_ed, Action::SetByte3Top), + 0xee...0xef => (State::Tail2, Action::SetByte3Top), + 0xf0 => (State::Utf8_4_3_f0, Action::SetByte4), + 0xf1...0xf3 => (State::Tail3, Action::SetByte4), + 0xf4 => (State::Utf8_4_3_f4, Action::SetByte4), + }, + State::U3_2_e0 => { + 0xa0...0xbf => (State::Tail1, Action::SetByte2), + }, + State::U3_2_ed => { + 0x80...0x9f => (State::Tail1, Action::SetByte2), + }, + State::Utf8_4_3_f0 => { + 0x90...0xbf => (State::Tail2, Action::SetByte3), + }, + State::Utf8_4_3_f4 => { + 0x80...0x8f => (State::Tail2, Action::SetByte3), + }, + State::Tail3 => { + 0x80...0xbf => (State::Tail2, Action::SetByte3), + }, + State::Tail2 => { + 0x80...0xbf => (State::Tail1, Action::SetByte2), + }, + State::Tail1 => { + 0x80...0xbf => (State::Ground, Action::SetByte1), + }, +}; diff --git a/utf8parse/src/types.rs b/utf8parse/src/types.rs new file mode 100644 index 0000000..4c604f4 --- /dev/null +++ b/utf8parse/src/types.rs @@ -0,0 +1,77 @@ +//! Types supporting the UTF-8 parser +#![allow(non_camel_case_types)] +use std::mem; + +/// States the parser can be in. +/// +/// There is a state for each initial input of the 3 and 4 byte sequences since +/// the following bytes are subject to different conditions than a tail byte. +#[allow(dead_code)] +#[derive(Debug, Copy, Clone)] +pub enum State { + /// Ground state; expect anything + Ground = 0, + /// 3 tail bytes + Tail3 = 1, + /// 2 tail bytes + Tail2 = 2, + /// 1 tail byte + Tail1 = 3, + /// UTF8-3 starting with E0 + U3_2_e0 = 4, + /// UTF8-3 starting with ED + U3_2_ed = 5, + /// UTF8-4 starting with F0 + Utf8_4_3_f0 = 6, + /// UTF8-4 starting with F4 + Utf8_4_3_f4 = 7, +} + +/// Action to take when receiving a byte +#[allow(dead_code)] +#[derive(Debug, Copy, Clone)] +pub enum Action { + /// Unexpected byte; sequence is invalid + InvalidSequence = 0, + /// Received valid 7-bit ASCII byte which can be directly emitted. + EmitByte = 1, + /// Set the bottom continuation byte + SetByte1 = 2, + /// Set the 2nd-from-last continuation byte + SetByte2 = 3, + /// Set the 2nd-from-last byte which is part of a two byte sequence + SetByte2Top = 4, + /// Set the 3rd-from-last continuation byte + SetByte3 = 5, + /// Set the 3rd-from-last byte which is part of a three byte sequence + SetByte3Top = 6, + /// Set the top byte of a four byte sequence. + SetByte4 = 7, +} + +/// Convert a state and action to a u8 +/// +/// State will be the bottom 4 bits and action the top 4 +#[inline] +#[allow(dead_code)] +pub fn pack(state: State, action: Action) -> u8 { + ((action as u8) << 4) | (state as u8) +} + +/// Convert a u8 to a state and action +/// +/// # Unsafety +/// +/// If this function is called with a byte that wasn't encoded with the `pack` +/// function in this module, there is no guarantee that a valid state and action +/// can be produced. +#[inline] +pub unsafe fn unpack(val: u8) -> (State, Action) { + ( + // State is stored in bottom 4 bits + mem::transmute(val & 0x0f), + + // Action is stored in top 4 bits + mem::transmute(val >> 4), + ) +} -- cgit