みけCAT

In my opinion, having `ARITMETIC_OPTIONS` in the file `src/core/lib/Delim.mjs` looks weird. The file should be for delimiters as its name and comment are suggesting. Embedding the array for choices directly...

Non-printable characters including 0x00 are displayed as `.`. This looks like not a bug but an intended behavior.

This issue is not VBA notation-specific but can happen with simple space-delimited data. [From Hex, To Hex - CyberChef](https://gchq.github.io/CyberChef/#recipe=From_Hex%28%27Space%27%29To_Hex%28%27Space%27,0%29&input=NTUgOEIgRUMgODEgRUMgOTAgMCAwIDAgNTYgNTcgQjkgNEMgNzc) Workaround: split the data and convert each values. [Fork, From Hex,...

Hmmm, it looks like each bytes are added instead of 16-bit or 32-bit blocks as specified in Wikipedia. [From Hex, Fletcher-32 Checksum - CyberChef](https://gchq.github.io/CyberChef/#recipe=From_Hex%28%27Auto%27%29Fletcher-32_Checksum%28%29&input=MTIgMzQ)

Here is more interesting fact: Encoding the decoding result (that seems empty) yields some meaningful data (with the order swapped): [From Hex, CBOR Decode, CBOR Encode, To Hex - CyberChef](https://gchq.github.io/CyberChef/#recipe=From_Hex%28%27Auto%27%29CBOR_Decode%28%29CBOR_Encode%28%29To_Hex%28%27Space%27,0%29&input=YTIgMDQgNDggMzkgMzAgMTcgNjggY2QgZGEgMDUgMTMgMDEgMjY)...

Fmmm, 0x26 doesn't seem causing stop. Example Input: ``` A3 # map(3) 61 # text(1) 26 # "&" 03 # unsigned(3) 61 # text(1) 41 # "A" 26 # negative(6)...

This issue looks like due to a bug in zlib.js. Due to the bug, CRC information is not correctly set to the encryption header. Therefore, tools that check this information...

I found that `zlib.js` has the same problem and submitted an issue. [RawDeflate with FIXED or NONE mode doesn't work well with an 0-byte input · Issue #81 · imaya/zlib.js](https://github.com/imaya/zlib.js/issues/81)

It should be because PKCS is defined so (16-byte padding is added when the length of the plaintext is multiple of 16). [PKCS padding method - IBM Documentation](https://www.ibm.com/docs/en/zos/2.4.0?topic=rules-pkcs-padding-method) I think...

I tried implementing that. My implementation processed 100MB (100,000,000B) input in about 20 seconds on Firefox 106.0.2 and Intel(R) Core(TM) i7-9750H CPU @ 2.60GHz. However, processing larger input looked difficult....