• Normalize single-character emojis by adding VS16 (U+FE0F) for consistent terminal alignment
• Implement emoji presentation detection using Unicode properties and grapheme segmentation
• Add comprehensive test coverage for emoji normalization with various emoji types

Implement emoji normalization with VS16 support

• Added normalizeEmoji() function to append VS16 (U+FE0F) to text-style emojis for consistent
 terminal width
• Implemented Unicode property detection using Intl.Segmenter and regex patterns for emoji
 classification
• Integrated emoji normalization into the enum list mapping logic
• Handles edge cases: existing variation selectors, presentation-default emojis, and trailing
 whitespace

@commitlint/cz-commitlint/src/services/getRuleQuestionConfig.ts

Add comprehensive emoji normalization test coverage

• Added new test case for emoji normalization covering single-character emojis (🛠, 🗑) and
 presentation-default emojis (⚙️)
• Extended existing emojiInHeader test with additional emoji types to verify VS16 insertion
 behavior
• Added inline comments documenting emoji presentation properties and expected normalization
 behavior
• Verified that emojis with and without VS16 render with correct column alignment

@commitlint/cz-commitlint/src/services/getRuleQuestionConfig.test.ts

Description

getRuleQuestionConfig.ts instantiates Intl.Segmenter, but @commitlint/cz-commitlint inherits
tsconfig.shared.json which sets compilerOptions.lib to only ["es2022"], so Intl.Segmenter is
not available in the configured TS standard library types. This will break TypeScript
compilation/typechecking for this package.

Code

@commitlint/cz-commitlint/src/services/getRuleQuestionConfig.ts[R17-19]

+const segmenter = new Intl.Segmenter("en", { granularity: "grapheme" });
+const isPresentation = /^\p{Emoji_Presentation}/u;
+const isEmojiBase = /^\p{Emoji}/u;

Evidence

The PR introduces a new usage of Intl.Segmenter, while the package’s tsconfig extends a shared
config that restricts the TS lib to es2022 only; no package-level override adds additional libs.

@commitlint/cz-commitlint/src/services/getRuleQuestionConfig.ts[17-19]
@commitlint/cz-commitlint/tsconfig.json[1-11]
tsconfig.shared.json[1-5]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
`@commitlint/cz-commitlint` now uses `Intl.Segmenter`, but the TypeScript project config only includes `lib: ["es2022"]`. This configuration does not provide typings for `Intl.Segmenter`, causing TypeScript compilation/typechecking failures.

## Issue Context
The package extends the monorepo shared tsconfig and does not override `compilerOptions.lib`.

## Fix Focus Areas
- tsconfig.shared.json[1-24]
- @commitlint/cz-commitlint/tsconfig.json[1-11]
- @commitlint/cz-commitlint/src/services/getRuleQuestionConfig.ts[17-19]

## Suggested fix
Prefer a package-local override to minimize monorepo-wide impact:
- In `@commitlint/cz-commitlint/tsconfig.json`, add a `compilerOptions.lib` that extends the existing lib with an Intl-capable lib that includes `Intl.Segmenter` (e.g. add the appropriate newer `es20xx`/`esnext` lib entries).

Alternative:
- Remove `Intl.Segmenter` usage and use a simpler code-point based check (`Array.from(...)`) since the feature is intended for *single-codepoint* emojis anyway.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

Description

normalizeEmoji appends U+FE0F to the end of any single grapheme cluster that matches \p{Emoji}
but not \p{Emoji_Presentation}; for multi-codepoint clusters (e.g., skin-tone modifiers or ZWJ
sequences), VS16 is appended after the whole cluster rather than after the base emoji code point.
This can fail to force emoji presentation and can render incorrectly in the interactive menu.

Code

@commitlint/cz-commitlint/src/services/getRuleQuestionConfig.ts[R33-49]

+	const segments = Array.from(segmenter.segment(trimmed));
+
+	if (segments.length === 1) {
+		const char = segments[0].segment;
+
+		switch (true) {
+			case char.includes("\uFE0F"):
+			case char.includes("\uFE0E"):
+				return emoji;
+
+			case isPresentation.test(char):
+				return emoji;
+
+			// Is it a "Text-style" emoji base and not a number? Add VS16!
+			case isEmojiBase.test(char) && !/^[0-9#*]$/.test(char):
+				return trimmed + "\uFE0F" + trailing;
+		}

Evidence

The code uses grapheme segmentation (so ZWJ/skin-tone sequences are often a single segment) and then
appends VS16 to the entire trimmed string. Prompt config emoji is an unconstrained string and is
stored without validation, so multi-codepoint emoji sequences are valid inputs and will flow into
normalizeEmoji.

@commitlint/cz-commitlint/src/services/getRuleQuestionConfig.ts[33-49]
@commitlint/types/src/prompt.ts[18-40]
@commitlint/cz-commitlint/src/store/prompts.ts[14-28]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
`normalizeEmoji()` currently appends U+FE0F to the end of any single *grapheme cluster* that is `\p{Emoji}` but not `\p{Emoji_Presentation}`. Grapheme clusters can be multi-codepoint (ZWJ sequences, skin-tone modifiers), and appending VS16 at the end is the wrong placement.

## Issue Context
`emoji?: string` is user-provided and unconstrained; questions are stored as-is, so multi-codepoint emoji sequences can reach `normalizeEmoji()`.

## Fix Focus Areas
- @commitlint/cz-commitlint/src/services/getRuleQuestionConfig.ts[27-53]
- @commitlint/cz-commitlint/src/store/prompts.ts[14-28]
- @commitlint/types/src/prompt.ts[18-40]

## Suggested fix
Implement one of:
1) **Restrict normalization to single code point** (aligns with PR goal):
  - After trimming trailing whitespace, ensure `Array.from(trimmed).length === 1` before considering adding VS16.
  - Also ensure the string does not contain ZWJ (`\u200D`) or modifiers.

2) **Insert VS16 after the base code point**:
  - `const cps = Array.from(trimmed); return cps[0] + '\uFE0F' + cps.slice(1).join('') + trailing;`
  - Keep existing early-returns for already-present VS15/VS16.

Add a unit test covering an emoji with skin-tone or ZWJ (e.g., `"☝🏽"` or `"👨‍💻"`) asserting it is not modified (option 1) or that VS16 placement is correct (option 2).

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools