From 646ba41cf81c60ae921408abc6fedcccbc0e95e5 Mon Sep 17 00:00:00 2001
From: Abdo <abdo@abdnh.net>
Date: Fri, 5 Jan 2024 07:22:52 +0300
Subject: [PATCH] Ignore some extra chars in no-combining search (#2929)

* Ignore some extra chars in no-combining search

* Construct new string

* Update rslib/src/text.rs (dae)
---
 rslib/src/text.rs | 50 +++++++++++++++++++++++++++++++++++++++++------
 1 file changed, 44 insertions(+), 6 deletions(-)

diff --git a/rslib/src/text.rs b/rslib/src/text.rs
index 3d681a1ac..c8fae8dd9 100644
--- a/rslib/src/text.rs
+++ b/rslib/src/text.rs
@@ -380,22 +380,60 @@ pub(crate) fn ensure_string_in_nfc(s: &mut String) {
     }
 }
 
+static EXTRA_NO_COMBINING_REPLACEMENTS: phf::Map<char, &str> = phf::phf_map! {
+'€'  =>  "E",
+'Æ'  =>  "AE",
+'Ð'  =>  "D",
+'Ø'  =>  "O",
+'Þ'  =>  "TH",
+'ß'  =>  "s",
+'æ'  =>  "ae",
+'ð'  =>  "d",
+'ø'  =>  "o",
+'þ'  =>  "th",
+'Đ'  =>  "D",
+'đ'  =>  "d",
+'Ħ'  =>  "H",
+'ħ'  =>  "h",
+'ı'  =>  "i",
+'ĸ'  =>  "k",
+'Ł'  =>  "L",
+'ł'  =>  "l",
+'Ŋ'  =>  "N",
+'ŋ'  =>  "n",
+'Œ'  =>  "OE",
+'œ'  =>  "oe",
+'Ŧ'  =>  "T",
+'ŧ'  =>  "t",
+'Ə'  =>  "E",
+'ǝ'  =>  "e",
+'ɑ'  =>  "a",
+};
+
 /// Convert provided string to NFKD form and strip combining characters.
 pub(crate) fn without_combining(s: &str) -> Cow<str> {
     // if the string is already normalized
     if matches!(is_nfkd_quick(s.chars()), IsNormalized::Yes) {
         // and no combining characters found, return unchanged
-        if !s.chars().any(is_combining_mark) {
+        if !s
+            .chars()
+            .any(|c| is_combining_mark(c) || EXTRA_NO_COMBINING_REPLACEMENTS.contains_key(&c))
+        {
             return s.into();
         }
     }
 
     // we need to create a new string without the combining marks
-    s.chars()
-        .nfkd()
-        .filter(|c| !is_combining_mark(*c))
-        .collect::<String>()
-        .into()
+    let mut out = String::with_capacity(s.len());
+    for chr in s.chars().nfkd().filter(|c| !is_combining_mark(*c)) {
+        if let Some(repl) = EXTRA_NO_COMBINING_REPLACEMENTS.get(&chr) {
+            out.push_str(repl);
+        } else {
+            out.push(chr);
+        }
+    }
+
+    out.into()
 }
 
 /// Check if string contains an unescaped wildcard.