regression: native emoji rendered as corrupted characters in omnichannel PDF transcript

ee/packages/pdf-worker/src/templates/ChatTranscript/markup/emoji.spec.ts

@@ -0,0 +1,19 @@
+import { unicodeToShortname } from './emoji';
+
+describe('unicodeToShortname', () => {
+	it('converts emoji without variation selector to shortname', () => {
+		expect(unicodeToShortname('\u{1F44D}')).toBe(':+1:');
+	});
+
+	it('converts emoji with variation selector to shortname', () => {
+		expect(unicodeToShortname('\u{1F44D}\u{FE0F}')).toBe(':+1:');
+	});
+
+	it('converts skin tone emoji to shortname', () => {
+		expect(unicodeToShortname('\u{1F44D}\u{1F3FD}')).toBe(':+1_tone3:');
+	});
+
+	it('returns the input unchanged when no shortname exists', () => {
+		expect(unicodeToShortname('not an emoji')).toBe('not an emoji');
+	});
+});

ee/packages/pdf-worker/src/templates/ChatTranscript/markup/emoji.ts

@@ -13,17 +13,19 @@ const firstShortcode = (hexcode: string): string | undefined => {
 	return Array.isArray(entry) ? entry[0] : entry;
 };
 
+const stripVariationSelectors = (unicode: string): string => unicode.replace(/\p{Variation_Selector}/gu, '');
+
 const buildMap = (): Map<string, string> => {
 	const map = new Map<string, string>();
 	for (const emoji of data as EmojibaseEntry[]) {
 		const code = firstShortcode(emoji.hexcode);
 		if (code) {
-			map.set(emoji.emoji, `:${code}:`);
+			map.set(stripVariationSelectors(emoji.emoji), `:${code}:`);
 		}
 		for (const skin of emoji.skins ?? []) {
 			const skinCode = firstShortcode(skin.hexcode);
 			if (skinCode) {
-				map.set(skin.emoji, `:${skinCode}:`);
+				map.set(stripVariationSelectors(skin.emoji), `:${skinCode}:`);
 			}
 		}
 	}
@@ -34,5 +36,5 @@ export const unicodeToShortname = (unicode: string): string => {
 	if (!unicodeToShort) {
 		unicodeToShort = buildMap();
 	}
-	return unicodeToShort.get(unicode) ?? unicode;
+	return unicodeToShort.get(stripVariationSelectors(unicode)) ?? unicode;
 };