Fix Uint8Array concat crash in Parse IPv4 header

[Zish19]

Fixes #2400

Summary

Fixes a runtime TypeError in the Parse IPv4 header operation caused by calling .concat() on a Uint8Array.

Root Cause

input.slice() returns a Uint8Array, not a standard JavaScript Array.

The operation previously attempted to do:

input.slice(0, 10).concat([0, 0]).concat(input.slice(12, 20));

Since Uint8Array does not implement .concat(), malformed/raw input caused:

TypeError: e.slice(...).concat is not a function

This affects both:

• Raw input mode
• Hex input mode (fromHex() also returns Uint8Array)

Fix

Replaced the unsafe concatenation with array spreading:

[...input.slice(0, 10), 0, 0, ...input.slice(12, 20)]

This preserves the exact byte sequence while producing a standard array compatible with the downstream checksum logic.

Notes

• Minimal isolated fix
• No unrelated refactoring
• No behavioral changes outside checksum header construction
• Prevents crashes on malformed/truncated input
• Valid IPv4 parsing behavior remains unchanged

[CLAassistant]

All committers have signed the CLA.

[Zish19]

CLA issue has been resolved and the branch has been updated successfully.
The fix remains minimal and isolated to the affected checksum header construction logic in ParseIPv4Header.mjs.
Please let me know if any additional changes, tests, or adjustments are needed.

[GCHQDeveloper581]

Please add a test that would fail without this patch and pass with it present.

Please also provide an AI disclosure in the header of your PR stating if AI has been used in creation of the PR and if so what model(s) has(have) been used.

[Zish19]

Thanks for the review.

I’ll add a regression test covering the malformed/raw input case that previously triggered the Uint8Array.concat TypeError, and update the PR description with an AI usage disclosure.