mirror of
https://github.com/mauriceboe/TREK.git
synced 2026-10-19 05:26:47 +00:00
724c88c201
A mechanical prettier --write over server/src, server/tests, server/scripts and shared/src with the workspaces' .prettierrc; nothing but layout and import order changes, and every file keeps the same set of imports. shared is formatted throughout and its format baseline is empty. The server baseline keeps the 35 files that Prettier would grow past their lint:size limit (they need a split first, which is not a formatting change), and the generated entity files stay as their generator writes them.
47 lines
1.7 KiB
TypeScript
47 lines
1.7 KiB
TypeScript
/**
|
|
* byCodeUnit restates what a bare `.sort()` already did, so the contract is
|
|
* "identical to the default order" — and that is what these check, against the
|
|
* default rather than against hand-written expectations.
|
|
*/
|
|
import { byCodeUnit } from '../../../src/nest/common/compare';
|
|
|
|
import { describe, it, expect } from 'vitest';
|
|
|
|
describe('byCodeUnit', () => {
|
|
it('CMP-001: reports less, greater and equal', () => {
|
|
expect(byCodeUnit('a', 'b')).toBe(-1);
|
|
expect(byCodeUnit('b', 'a')).toBe(1);
|
|
expect(byCodeUnit('a', 'a')).toBe(0);
|
|
});
|
|
|
|
it('CMP-002: orders ISO dates chronologically, which is why the callers can use it', () => {
|
|
expect(['2026-03-01', '2026-01-15', '2026-02-28'].sort(byCodeUnit)).toEqual([
|
|
'2026-01-15',
|
|
'2026-02-28',
|
|
'2026-03-01',
|
|
]);
|
|
});
|
|
|
|
it('CMP-003: reproduces the default sort exactly for the ASCII the callers pass', () => {
|
|
const samples = [
|
|
['EUR', 'usd', 'CHF', 'JPY'],
|
|
['b', 'A', 'a', 'B', ''],
|
|
['plugin:one', 'plugin:One', 'plugin:1', 'plugin:'],
|
|
['2026-01-01', '2026-1-1', '2026-01-1'],
|
|
['x'],
|
|
[],
|
|
];
|
|
for (const sample of samples) {
|
|
expect([...sample].sort(byCodeUnit)).toEqual([...sample].sort());
|
|
}
|
|
});
|
|
|
|
it('CMP-004: is NOT locale-aware — the reason the doc comment warns callers off names', () => {
|
|
// Code-unit order files every accented word behind the whole ASCII range.
|
|
// A list a person reads wants localeCompare instead; this asserts the
|
|
// difference so nobody reaches for byCodeUnit there by accident.
|
|
expect(['Zelt', 'Übernachtung'].sort(byCodeUnit)).toEqual(['Zelt', 'Übernachtung']);
|
|
expect(['Zelt', 'Übernachtung'].sort((a, b) => a.localeCompare(b, 'de'))).toEqual(['Übernachtung', 'Zelt']);
|
|
});
|
|
});
|