From 16de10fe2395002644d685fdfcf76823346d1cc4 Mon Sep 17 00:00:00 2001
From: Titus Wormer <tituswormer@gmail.com>
Date: Mon, 5 Sep 2022 12:00:33 +0200
Subject: Add support for getting `char`s from bytes

---
 src/construct/attention.rs | 31 ++++++++++---------------------
 src/util/slice.rs          | 22 ++++++++++++++++++++++
 2 files changed, 32 insertions(+), 21 deletions(-)

diff --git a/src/construct/attention.rs b/src/construct/attention.rs
index 526f58c..947a79b 100644
--- a/src/construct/attention.rs
+++ b/src/construct/attention.rs
@@ -81,8 +81,8 @@ use crate::resolve::Name as ResolveName;
 use crate::state::{Name as StateName, State};
 use crate::tokenizer::Tokenizer;
 use crate::util::classify_character::{classify_opt, Kind as CharacterKind};
-use crate::util::slice::Slice;
-use alloc::{string::String, vec, vec::Vec};
+use crate::util::slice::{char_after_index, char_before_index, Slice};
+use alloc::{vec, vec::Vec};
 
 /// Attentention sequence that we can take markers from.
 #[derive(Debug)]
@@ -234,28 +234,17 @@ fn get_sequences(tokenizer: &mut Tokenizer) -> Vec<Sequence> {
                 let end = index + 1;
                 let exit = &tokenizer.events[end];
 
-                let before_end = enter.point.index;
-                let before_start = if before_end < 4 { 0 } else { before_end - 4 };
-                let after_start = exit.point.index;
-                let after_end = if after_start + 4 > tokenizer.parse_state.bytes.len() {
-                    tokenizer.parse_state.bytes.len()
-                } else {
-                    after_start + 4
-                };
-
                 let marker = Slice::from_point(tokenizer.parse_state.bytes, &enter.point)
                     .head()
                     .unwrap();
-                let before = classify_opt(
-                    String::from_utf8_lossy(&tokenizer.parse_state.bytes[before_start..before_end])
-                        .chars()
-                        .last(),
-                );
-                let after = classify_opt(
-                    String::from_utf8_lossy(&tokenizer.parse_state.bytes[after_start..after_end])
-                        .chars()
-                        .next(),
-                );
+                let before = classify_opt(char_before_index(
+                    tokenizer.parse_state.bytes,
+                    enter.point.index,
+                ));
+                let after = classify_opt(char_after_index(
+                    tokenizer.parse_state.bytes,
+                    exit.point.index,
+                ));
                 let open = after == CharacterKind::Other
                     || (after == CharacterKind::Punctuation && before != CharacterKind::Other);
                 let close = before == CharacterKind::Other
diff --git a/src/util/slice.rs b/src/util/slice.rs
index 0734d78..d02a526 100644
--- a/src/util/slice.rs
+++ b/src/util/slice.rs
@@ -5,6 +5,28 @@ use crate::util::constant::TAB_SIZE;
 use alloc::string::String;
 use core::str;
 
+/// Get a [`char`][] right before `index` in bytes (`&[u8]`).
+///
+/// In most cases, markdown operates on ASCII bytes.
+/// In a few cases, it is unicode aware, so we need to find an actual char.
+pub fn char_before_index(bytes: &[u8], index: usize) -> Option<char> {
+    let start = if index < 4 { 0 } else { index - 4 };
+    String::from_utf8_lossy(&bytes[start..index]).chars().last()
+}
+
+/// Get a [`char`][] right at `index` in bytes (`&[u8]`).
+///
+/// In most cases, markdown operates on ASCII bytes.
+/// In a few cases, it is unicode aware, so we need to find an actual char.
+pub fn char_after_index(bytes: &[u8], index: usize) -> Option<char> {
+    let end = if index + 4 > bytes.len() {
+        bytes.len()
+    } else {
+        index + 4
+    };
+    String::from_utf8_lossy(&bytes[index..end]).chars().next()
+}
+
 /// A range between two points.
 #[derive(Debug)]
 pub struct Position<'a> {
-- 
cgit