From b805c5f10080912cbfb1fd2fd2733c48a92ce4f9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Erik=20B=C3=BCnnig?= <github@epbuennig.me>
Date: Tue, 30 May 2023 18:11:30 +0200
Subject: [PATCH] Unicode code point conversion (#1068) (#1132)

---
 library/src/compute/construct.rs | 57 ++++++++++++++++++++++++++++++++
 tests/typ/compute/construct.typ  | 21 ++++++++++++
 2 files changed, 78 insertions(+)

diff --git a/library/src/compute/construct.rs b/library/src/compute/construct.rs
index 169c53284..d11eb3980 100644
--- a/library/src/compute/construct.rs
+++ b/library/src/compute/construct.rs
@@ -461,6 +461,10 @@ cast_from_value! {
 /// - Floats are formatted in base 10 and never in exponential notation.
 /// - From labels the name is extracted.
 ///
+/// If you wish to convert from and to Unicode code points, see
+/// [`str.to-unicode`]($func/str.to-unicode) and
+/// [`str.from-unicode`]($func/str.from-unicode).
+///
 /// ## Example { #example }
 /// ```example
 /// #str(10) \
@@ -474,6 +478,11 @@ cast_from_value! {
 /// Category: construct
 /// Returns: string
 #[func]
+#[scope(
+    scope.define("to-unicode", to_unicode);
+    scope.define("from-unicode", from_unicode);
+    scope
+)]
 pub fn str(
     /// The value that should be converted to a string.
     value: ToStr,
@@ -553,6 +562,54 @@ fn int_to_base(mut n: i64, base: i64) -> EcoString {
     std::str::from_utf8(&digits[i..]).unwrap_or_default().into()
 }
 
+/// Converts a character into its corresponding code point.
+///
+/// ## Example
+/// ```example
+/// #str.to-unicode("a") \
+/// #"a\u{0300}".codepoints().map(str.to-unicode)
+/// ```
+///
+/// Display: String To Unicode
+/// Category: construct
+/// Returns: int
+#[func]
+pub fn to_unicode(
+    /// The character that should be converted.
+    value: char,
+) -> Value {
+    Value::Int(From::<u32>::from(value.into()))
+}
+
+#[func]
+/// Converts a unicode code point into its corresponding string.
+///
+/// ```example
+/// #str.from-unicode(97)
+/// ```
+///
+/// Display: Sting From Unicode
+/// Category: construct
+/// Returns: string
+pub fn from_unicode(
+    /// The code point that should be converted.
+    value: CodePoint,
+) -> Value {
+    Value::Str(format_str!("{}", value.0))
+}
+
+/// The numeric representation of a single unicode code point.
+struct CodePoint(char);
+
+cast_from_value! {
+    CodePoint,
+    v: i64 => {
+        Self(v.try_into().ok().and_then(|v: u32| v.try_into().ok()).ok_or_else(
+            || eco_format!("{:#x} is not a valid codepoint", v),
+        )?)
+    },
+}
+
 /// Create a label from a string.
 ///
 /// Inserting a label into content attaches it to the closest previous element
diff --git a/tests/typ/compute/construct.typ b/tests/typ/compute/construct.typ
index aea15b53e..f094b6b2f 100644
--- a/tests/typ/compute/construct.typ
+++ b/tests/typ/compute/construct.typ
@@ -82,6 +82,27 @@
 // Error: 18-19 base is only supported for integers
 #str(1.23, base: 2)
 
+---
+// Test the unicode function.
+#test(str.from-unicode(97), "a")
+#test(str.to-unicode("a"), 97)
+
+---
+// Error: 19-22 expected integer, found content
+#str.from-unicode([a])
+
+---
+// Error: 17-21 expected exactly one character
+#str.to-unicode("ab")
+
+---
+// Error: 19-21 0xffffffffffffffff is not a valid codepoint
+#str.from-unicode(-1) // negative values are not valid
+
+---
+// Error: 19-27 0x110000 is not a valid codepoint
+#str.from-unicode(0x110000) // 0x10ffff is the highest valid code point
+
 ---
 #assert(range(2, 5) == (2, 3, 4))