From 3959dce5d16e274477b3690b14447942391bcd6b Mon Sep 17 00:00:00 2001
From: Max Crone <code@maxcrone.org>
Date: Mon, 25 Mar 2024 13:25:28 +0100
Subject: [PATCH] Replace unmaintained unidecode with deunicode

---
 Cargo.lock           | 14 +++++++-------
 Cargo.toml           |  2 +-
 src/comparison.rs    |  2 +-
 src/lib.rs           |  2 +-
 src/transliterate.rs | 19 ++++++++++---------
 5 files changed, 20 insertions(+), 19 deletions(-)

diff --git a/Cargo.lock b/Cargo.lock
index 18b20b1e..72b65778 100644
--- a/Cargo.lock
+++ b/Cargo.lock
@@ -61,6 +61,12 @@ dependencies = [
  "cfg-if",
 ]
 
+[[package]]
+name = "deunicode"
+version = "1.4.3"
+source = "registry+https://github.com/rust-lang/crates.io-index"
+checksum = "b6e854126756c496b8c81dec88f9a706b15b875c5849d4097a3854476b9fdf94"
+
 [[package]]
 name = "human_name"
 version = "2.0.3"
@@ -68,6 +74,7 @@ dependencies = [
  "alloc_counter",
  "compact_str",
  "crossbeam-utils",
+ "deunicode",
  "libc",
  "phf",
  "phf_codegen",
@@ -77,7 +84,6 @@ dependencies = [
  "unicode-case-mapping",
  "unicode-normalization",
  "unicode-segmentation",
- "unidecode",
 ]
 
 [[package]]
@@ -282,9 +288,3 @@ name = "unicode-segmentation"
 version = "1.10.0"
 source = "registry+https://github.com/rust-lang/crates.io-index"
 checksum = "0fdbf052a0783de01e944a6ce7a8cb939e295b1e7be835a1112c3b9a7f047a5a"
-
-[[package]]
-name = "unidecode"
-version = "0.3.0"
-source = "registry+https://github.com/rust-lang/crates.io-index"
-checksum = "402bb19d8e03f1d1a7450e2bd613980869438e0666331be3e073089124aa1adc"
diff --git a/Cargo.toml b/Cargo.toml
index 315661c2..3bbdf7f2 100644
--- a/Cargo.toml
+++ b/Cargo.toml
@@ -16,13 +16,13 @@ crossbeam-utils = "0.8"
 unicode-segmentation = "1.9"
 unicode-normalization = "0.1"
 unicode-case-mapping = "0.4"
-unidecode = "0.3"
 libc = { version = "0.2", optional = true }
 phf = "0.11"
 serde = { version = "1.0", features = ["derive"], optional = true }
 serde_json = { version = "1.0", optional = true }
 smallvec = { features = ["union"], version = "1.9" }
 compact_str = { version = "0.7.1", features = ["serde"] }
+deunicode = "1.4.3"
 
 [dev-dependencies]
 alloc_counter = "0.0"
diff --git a/src/comparison.rs b/src/comparison.rs
index f0efe1d1..f1f13175 100644
--- a/src/comparison.rs
+++ b/src/comparison.rs
@@ -43,7 +43,7 @@ impl Name {
     /// names and/or suffixes are present in both names, they must match as well.
     ///
     /// Transliterates everything to ASCII before comparison using the naive
-    /// algorithm of [unidecode](https://github.com/chowdhurya/rust-unidecode/)
+    /// algorithm of [deunicode](https://github.com/kornelski/deunicode/blob/main/README.md)
     /// (which ignores context), and ignores case, accents and combining marks.
     ///
     /// In the case of given and middle names, allows one name to be a prefix of
diff --git a/src/lib.rs b/src/lib.rs
index 275a844e..229b25f7 100644
--- a/src/lib.rs
+++ b/src/lib.rs
@@ -6,10 +6,10 @@
 #![cfg_attr(feature = "bench", feature(test))]
 
 extern crate crossbeam_utils;
+extern crate deunicode;
 extern crate smallvec;
 extern crate unicode_normalization;
 extern crate unicode_segmentation;
-extern crate unidecode;
 
 #[cfg(test)]
 #[cfg(feature = "bench")]
diff --git a/src/transliterate.rs b/src/transliterate.rs
index 3d916af6..f515b058 100644
--- a/src/transliterate.rs
+++ b/src/transliterate.rs
@@ -1,14 +1,12 @@
+use deunicode::deunicode_char;
 use std::str::Chars;
-use unidecode::unidecode_char;
 
 #[inline]
-fn transliterate(c: char) -> Chars<'static> {
-    let s = unidecode_char(c);
-    if s.is_empty() {};
+fn transliterate(c: char) -> Option<Chars<'static>> {
     // We should maybe use unicode case folding here as an initial pass,
     // but without a concrete motivating case (yet) it doesn't seem worth
     // the cost.
-    unidecode_char(c).chars()
+    deunicode_char(c).map(|s| s.chars())
 }
 
 #[inline]
@@ -41,14 +39,15 @@ fn ascii_to_upper_if_alpha(c: char) -> Option<char> {
 pub fn to_ascii_initial(c: char) -> Option<char> {
     match c {
         'A'..='Z' => Some(c),
-        _ => transliterate(c).find_map(ascii_to_upper_if_alpha),
+        _ => transliterate(c)?.find_map(ascii_to_upper_if_alpha),
     }
 }
 
 pub fn to_ascii_casefolded(text: &str) -> Option<impl Iterator<Item = char> + '_> {
     let mut result = text
         .chars()
-        .flat_map(transliterate)
+        .filter_map(transliterate)
+        .flatten()
         .filter_map(ascii_to_lower_if_alpha)
         .peekable();
 
@@ -63,7 +62,8 @@ pub fn to_ascii_casefolded(text: &str) -> Option<impl Iterator<Item = char> + '_
 pub fn to_ascii_casefolded_reversed(text: &str) -> Option<impl Iterator<Item = char> + '_> {
     let mut result = text
         .chars()
-        .flat_map(transliterate)
+        .filter_map(transliterate)
+        .flatten()
         .rev()
         .filter_map(ascii_to_lower_if_alpha)
         .peekable();
@@ -79,7 +79,8 @@ pub fn to_ascii_casefolded_reversed(text: &str) -> Option<impl Iterator<Item = c
 pub fn to_ascii_titlecase(s: &str) -> Option<String> {
     let mut result = s
         .chars()
-        .flat_map(transliterate)
+        .filter_map(transliterate)
+        .flatten()
         .filter_map(ascii_to_lower_if_alpha);
 
     result.next().map(|initial| {