aboutsummaryrefslogtreecommitdiffstats
path: root/src/util
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--src/util/classify_character.rs72
-rw-r--r--src/util/mod.rs1
2 files changed, 73 insertions, 0 deletions
diff --git a/src/util/classify_character.rs b/src/util/classify_character.rs
new file mode 100644
index 0000000..b938502
--- /dev/null
+++ b/src/util/classify_character.rs
@@ -0,0 +1,72 @@
+//! Utilities to classify characters as whitespace, punctuation, or rest.
+
+use crate::unicode::PUNCTUATION;
+
+/// Character kinds.
+#[derive(Debug, PartialEq, Eq)]
+pub enum Kind {
+ /// Whitespace.
+ ///
+ /// ## Example
+ ///
+ /// ```markdown
+ /// > | **a_b_ c**.
+ /// ^ ^ ^
+ /// ```
+ Whitespace,
+ /// Punctuation.
+ ///
+ /// ## Example
+ ///
+ /// ```markdown
+ /// > | **a_b_ c**.
+ /// ^^ ^ ^ ^
+ /// ```
+ Punctuation,
+ /// Everything else.
+ ///
+ /// ## Example
+ ///
+ /// ```markdown
+ /// > | **a_b_ c**.
+ /// ^ ^ ^
+ /// ```
+ Other,
+}
+
+/// Classify whether a character code represents whitespace, punctuation, or
+/// something else.
+///
+/// Used for attention (emphasis, strong), whose sequences can open or close
+/// based on the class of surrounding characters.
+///
+/// > 👉 **Note** that eof (`None`) is seen as whitespace.
+///
+/// ## References
+///
+/// * [`micromark-util-classify-character` in `micromark`](https://github.com/micromark/micromark/blob/main/packages/micromark-util-classify-character/dev/index.js)
+pub fn classify(char: char) -> Kind {
+ // Unicode whitespace.
+ if char.is_whitespace() {
+ Kind::Whitespace
+ }
+ // Unicode punctuation.
+ else if PUNCTUATION.contains(&char) {
+ Kind::Punctuation
+ }
+ // Everything else.
+ else {
+ Kind::Other
+ }
+}
+
+/// Like [`classify`], but supports eof as whitespace.
+pub fn classify_opt(char_opt: Option<char>) -> Kind {
+ if let Some(char) = char_opt {
+ classify(char)
+ }
+ // EOF.
+ else {
+ Kind::Whitespace
+ }
+}
diff --git a/src/util/mod.rs b/src/util/mod.rs
index f51845c..022c7d6 100644
--- a/src/util/mod.rs
+++ b/src/util/mod.rs
@@ -1,5 +1,6 @@
//! Utilities used when processing markdown.
+pub mod classify_character;
pub mod decode_character_reference;
pub mod edit_map;
pub mod encode;