Fix: CRC-16 checksum incorrect for non-ASCII characters

[BrookTV]

Issue

From Issue

• CRC-16 checksum incorrect for non-ASCII characters (e.g. Thai merchant name) #7

Fix

Fixes #7

Replace charCodeAt(0) with new TextEncoder().encode(data), which yields the actual UTF-8 byte sequence the EMVCo spec calls for. TextEncoder is a Web API available natively in Node.js 11+, all modern
browsers, Deno, and Bun — no polyfill needed.

tsconfig.json gains "DOM" in lib so TypeScript resolves the TextEncoder type.

Test

Added a test case that parses a PromptPay payload containing non-ascii characters in the merchant name field (tag 59) in strict mode (checksum validation on). The test was failing before this fix.

---

[maythiwat]

Thanks for the PR!

I have two concerns I'd like to get your thoughts on

1. Do Thai banks actually use non-ASCII characters in the Merchant Name (Tag 59) in practice? The EMVCo spec usually limits this to ASCII (using Tag 64 for local languages). Have you encountered real QR payloads from Thai banks failing because of Thai characters in this field?

2. Could replacing charCodeAt with new TextEncoder().encode break older browsers or specific in-app/embedded browsers that might lack native TextEncoder support?

I'll look into this soon. Thanks again for your contribution!