test(core): add UTF-8 Base64 regression cases

[1amKhush]

Problem

encodeBase64 is used to construct LNbits Basic Auth credentials and deliberately bridges UTF-8 bytes to the Web Platform btoa API. It currently has no focused regression tests for Unicode handling or Base64 padding.

Scope

Add src/core/utils/base64.test.ts covering:

• empty input;
• ordinary ASCII credentials;
• one-byte and two-byte padding cases;
• non-ASCII/UTF-8 credentials;
• deterministic output matching known Base64 vectors.

Non-goals

• Do not replace the implementation or introduce Node-only globals such as Buffer in production code.
• Do not add LNbits integration or network tests.
• Keep production changes out of scope unless a test exposes a confirmed defect.

Acceptance criteria

• Expected values are independent known vectors rather than values generated by calling the function under test.
• Tests exercise the browser-safe UTF-8 path.
• Existing tests, lint, and type checking continue to pass.

Verification

bun test src/core/utils/base64.test.ts
bun lint
bun typecheck

[jaenil]

Working on this issue .Please assign it to me

[1amKhush]

@jaenil go ahead!

[jaenil]

Hey @ContextVM-org @1amKhush just opened a PR for the test coverage task you assigned me.
PR #99 :
Added unit tests for encodeBase64 — empty string, plain ASCII Basic Auth credentials, 0/1/2 byte padding variants, and UTF-8 multi-byte sequences (accents, umlauts, and 3-byte Japanese hiragana). Test-only change, no production code touched.
Verified locally: bun test src/core/utils/base64.test.ts -> 9 pass / 0 fail, bun run lint and bun run typecheck both clean.