From 09d088b2b5c3295271e1f673bc8135a01eecc62e Mon Sep 17 00:00:00 2001
From: Henry Heino <46334387+personalizedrefrigerator@users.noreply.github.com>
Date: Sun, 12 May 2024 02:01:12 -0700
Subject: [PATCH] Mobile,Desktop: Fix nonbreaking spaces and CRLF break search
 for adjacent words (#10417)

---
 packages/lib/services/search/SearchEngine.test.ts | 14 ++++++++++++++
 packages/lib/services/search/SearchEngine.ts      |  4 ++++
 2 files changed, 18 insertions(+)

diff --git a/packages/lib/services/search/SearchEngine.test.ts b/packages/lib/services/search/SearchEngine.test.ts
index 3bad73880..e5a587195 100644
--- a/packages/lib/services/search/SearchEngine.test.ts
+++ b/packages/lib/services/search/SearchEngine.test.ts
@@ -340,6 +340,20 @@ describe('services/SearchEngine', () => {
 		expect((await engine.search('testing')).length).toBe(1);
 	}));
 
+	it('should use nonbreaking spaces as separators', (async () => {
+		await Note.save({
+			title: 'Test',
+			body: 'This is\u00A0a\u00A0test\r\nof different\r\nspace separators.',
+		});
+
+		await engine.syncTables();
+
+		expect((await engine.search('test')).length).toBe(1);
+		expect((await engine.search('different')).length).toBe(1);
+		expect((await engine.search('space')).length).toBe(1);
+		expect((await engine.search('separators')).length).toBe(1);
+	}));
+
 	it('should supports various query types', (async () => {
 		let rows;
 
diff --git a/packages/lib/services/search/SearchEngine.ts b/packages/lib/services/search/SearchEngine.ts
index fd4c85ae5..09a1c08ad 100644
--- a/packages/lib/services/search/SearchEngine.ts
+++ b/packages/lib/services/search/SearchEngine.ts
@@ -629,6 +629,10 @@ export default class SearchEngine {
 		// https://github.com/laurent22/joplin/issues/9694
 		normalizedText = htmlentitiesDecode(normalizedText);
 
+		// The FTS tokenizer doesn't understand some types of space,
+		// including nonbreaking spaces and CRLF.
+		normalizedText = normalizedText.replace(/\s/g, ' ');
+
 		return removeDiacritics(normalizedText.toLowerCase());
 	}