Enigma M4 gives wrong encryption for some rotor settings

When the Enigma M4 is set to use reflector UKW B thin and 4th rotor Beta in position A, ring A, the encryption should be identical to that of the M3 set to reflector UKW B. See https://www.cryptomuseum.com/crypto/enigma/m4/index.htm in section “Zusatzwalze Beta (β) and Gamma (γ)”.

Cryptii works correctly in this setting for some rotor combinations. By accident, I discovered that it fails to work with certain combinations. I did not notice any particular pattern. In the examples below, I merely changed Rotor 2 from II to I.

Steps to reproduce the bug

1. See correct behavior: https://cryptii.com/pipes/D3cegA
2. See incorrect behavior: https://cryptii.com/pipes/PYzXOQ

Describe what you expected to happen Output text in the 2nd example should be the same as the input.

Version, environment and state Tried with Chrome and Firefox, same behavior on both.

Inside the web app press Ctrl+I, copy the presented alert text and paste it into the block below.

{"version":"4.0.8+master.8ae346e","env":{"name":"firefox","version":"81.0.0","os":"Linux","type":"browser"},"pipe":{"id":1032640313,"url":"https://cryptii.com/pipes/PYzXOQ","title":"The Enigma machine: Encrypt and decrypt online","description":"The Enigma cipher machine is well known for the vital role it played during WWII. Alan Turing and his attempts to crack the Enigma machine code changed history. Nevertheless, many messages could not be decrypted until today.","items":[{"name":"text","title":"Plaintext"},{"name":"enigma","settings":{"model":"M4","reflector":"UKW-B-thin","rotor1":"beta","position1":1,"ring1":1,"rotor2":"I","position2":17,"ring2":3,"rotor3":"III","position3":12,"ring3":5,"rotor4":"I","position4":2,"ring4":2,"plugboard":"bq cr di ej kw mt os px uz gh","includeForeignChars":false}},{"name":"enigma","settings":{"model":"M3","reflector":"UKW-B","rotor1":"I","position1":17,"ring1":3,"rotor2":"III","position2":12,"ring2":5,"rotor3":"I","position3":2,"ring3":2,"plugboard":"bq cr di ej kw mt os px uz gh","includeForeignChars":false}},{"name":"text"}],"content":{"data":"intel ligen cepoi ntsto attac konth eeast wallo fthec astle atdaw n"}}}

[ffraenz]

Again, thank you for taking the time to report this issue. This was a very interesting edge case bringing the Enigma machine simulation closer to its physical counterpart.

[the-lambda-way]

It’s resolved on my end. Thanks.