From 79c3275f91f1c0867a1bfba3085c0682aa5486ef Mon Sep 17 00:00:00 2001
From: Titus Wormer <tituswormer@gmail.com>
Date: Wed, 22 Jun 2022 17:24:05 +0200
Subject: Add support for normalizing identifiers

---
 src/util/encode.rs               |  2 ++
 src/util/mod.rs                  |  1 +
 src/util/normalize_identifier.rs | 37 +++++++++++++++++++++++++++++++++++++
 3 files changed, 40 insertions(+)
 create mode 100644 src/util/normalize_identifier.rs

(limited to 'src/util')

diff --git a/src/util/encode.rs b/src/util/encode.rs
index f79c8ea..5762c22 100644
--- a/src/util/encode.rs
+++ b/src/util/encode.rs
@@ -21,6 +21,8 @@
 ///
 /// *   [`micromark-util-encode` in `micromark`](https://github.com/micromark/micromark/tree/main/packages/micromark-util-encode)
 pub fn encode(value: &str) -> String {
+    // To do: replacing 4 times might just be slow.
+    // Perhaps we can walk the chars.
     value
         .replace('&', "&amp;")
         .replace('"', "&quot;")
diff --git a/src/util/mod.rs b/src/util/mod.rs
index c3db267..ee58518 100644
--- a/src/util/mod.rs
+++ b/src/util/mod.rs
@@ -2,5 +2,6 @@
 
 pub mod decode_character_reference;
 pub mod encode;
+pub mod normalize_identifier;
 pub mod sanitize_uri;
 pub mod span;
diff --git a/src/util/normalize_identifier.rs b/src/util/normalize_identifier.rs
new file mode 100644
index 0000000..870fd33
--- /dev/null
+++ b/src/util/normalize_identifier.rs
@@ -0,0 +1,37 @@
+//! To do.
+
+/// To do.
+pub fn normalize_identifier(value: &str) -> String {
+    let mut codes = vec![];
+    let mut at_start = true;
+    let mut at_whitespace = true;
+
+    // Collapse markdown whitespace and trim it.
+    for char in value.chars() {
+        match char {
+            '\t' | '\r' | '\n' | ' ' => {
+                at_whitespace = true;
+            }
+            _ => {
+                if at_whitespace && !at_start {
+                    codes.push(' ');
+                }
+
+                codes.push(char);
+                at_start = false;
+                at_whitespace = false;
+            }
+        }
+    }
+
+    // To do: test if this matches unicode.
+    // Some characters are considered “uppercase”, but if their lowercase
+    // counterpart is uppercased will result in a different uppercase
+    // character.
+    // Hence, to get that form, we perform both lower- and uppercase.
+    codes
+        .iter()
+        .collect::<String>()
+        .to_uppercase()
+        .to_lowercase()
+}
-- 
cgit