mirror of
https://github.com/microsoft/vscode.git
synced 2026-10-07 01:08:40 +01:00
Giuseppe Cianci
ad360f0511
Use TypeScript comment ranges so the generator satisfies Unicode hygiene without rewriting string literals, template text, or other code. Cover the Codex 0.155.1 user-verification comments and literal-preservation cases.
61 lines
2.6 KiB
TypeScript
61 lines
2.6 KiB
TypeScript
/*---------------------------------------------------------------------------------------------
|
|
* Copyright (c) Microsoft Corporation. All rights reserved.
|
|
* Licensed under the MIT License. See License.txt in the project root for license information.
|
|
*--------------------------------------------------------------------------------------------*/
|
|
|
|
import assert from 'assert';
|
|
import { suite, test } from 'node:test';
|
|
import { normalizeCommentDashes } from '../generate-protocol.mjs';
|
|
|
|
suite('codex protocol comment normalization', () => {
|
|
test('normalizes the generated user-verification documentation', () => {
|
|
const source = [
|
|
'export type UserVerificationVerifyParams = {',
|
|
'\t/**',
|
|
'\t * Unpadded base64url encoding of 1\u20134096 challenge bytes.',
|
|
'\t */',
|
|
'\tchallenge: string,',
|
|
'\t/**',
|
|
'\t * Display context already approved by the UI; 1\u2013256 UTF-8 bytes.',
|
|
'\t */',
|
|
'\ttitle: string,',
|
|
'};',
|
|
].join('\n');
|
|
assert.strictEqual(normalizeCommentDashes(source), source.replace(/\u2013/g, '-'));
|
|
});
|
|
|
|
test('normalizes line, block, inline and end-of-file comments without changing line endings', () => {
|
|
const source = [
|
|
'// Leading 1\u20132',
|
|
'export type Range = { /* Inline 1\u20132 */',
|
|
'\tvalue: /* Before a type 1\u20132 */ string, /* After a comma 1\u20132 */',
|
|
'}; // Trailing 1\u20132',
|
|
'/* End of file 1\u20132 */',
|
|
].join('\r\n');
|
|
assert.strictEqual(normalizeCommentDashes(source), source.replace(/\u2013/g, '-'));
|
|
});
|
|
|
|
test('preserves quoted literals containing en dashes and comment delimiters', () => {
|
|
const source = [
|
|
'export type Single = \'/* 1\u20132 */\';',
|
|
'export type Double = "// 1\u20132";',
|
|
'export type Escaped = "\\\" /* 1\u20132 */";',
|
|
'export type PlainTemplate = `/* 1\u20132 */`;',
|
|
'export type Template = `${string}/* 1\u20132 */${number}// 1\u20132`;',
|
|
'const pattern = /[/*]\u2013[*/]/;',
|
|
].join('\n');
|
|
assert.strictEqual(normalizeCommentDashes(source), source);
|
|
});
|
|
|
|
test('normalizes actual comments inside template expressions but preserves template text', () => {
|
|
const source = 'const value = `/* 1\u20132 */ ${ /* Comment 1\u20132 */ 1 } // 1\u20132`;';
|
|
assert.strictEqual(normalizeCommentDashes(source), source.replace('Comment 1\u20132', 'Comment 1-2'));
|
|
});
|
|
|
|
test('is idempotent and leaves other punctuation unchanged', () => {
|
|
const source = '// Range 1\u20132; other punctuation \u2014\u2026\nexport type Range = "1\u20132";';
|
|
const expected = source.replace('Range 1\u20132', 'Range 1-2');
|
|
assert.strictEqual(normalizeCommentDashes(normalizeCommentDashes(source)), expected);
|
|
});
|
|
});
|