Huffman Encoding

This is my attempt to implement the Huffman's data compression algorithm in my programming language, AEC. A few years ago, I implemented it in JavaScript, you can see it here.
This only works in very modern browsers. I'm not too interested in targeting archaic browsers here. This only works in Firefox 62 and newer or Chrome 69 and newer (those with support for WebAssembly.Global). After a few years (if not months), all JavaScript environments will become as capable as they are, rather than like Internet Explorer.




AEC program output:

You can see the source code of the AEC program here.

UPDATE on 13/03/2025: For a reason that escapes me, the Huffman Coding in AEC and the Huffman Coding in JavaScript don't always output the same result. For the input TEO SAMARZIJA, they do, but, for example, for the input HENDRIK WADE BODE, they don't. The Huffman Coding in AEC outputs:
00101100011111010001010110100011110101111101001011000111110
While the Huffman Coding in JavaScript outputs:
01001100101111011001111000001100110101111100011011000111110
I have warned about that problem on Discord. The only potentially relevant difference between the way I implemented the Huffman's Algorithm in AEC and the way I implemented it in JavaScript is this: In AEC, I was trying to save a bit of memory by deleting the tree nodes that have already been used (that is, the two tree nodes with the lowest frequency in the array) from the array, whereas, in JavaScript, I put a boolean in the tree node structure indicating whether the node has already been used in the tree construction. But that shouldn't affect the final results, should it?