Why flush() in python's arithmeticcoding's ArithmeticEncoder just need to write out 1? about reference-arithmetic-coding HOT 2 CLOSED

I have understand it, because there are infinite zeros when reading code file is at eof.
and 1000000..... is always in [low, high].
💯

from reference-arithmetic-coding.

You have answered your own question correctly.

Indeed in my arithmetic coding implementation, the termination logic differs from Mark Nelson's logic. This is because my bit input stream generates an infinite sequence of 0 bits after the data terminates, whereas Mark allows the stream to read arbitrary data (such as data from the next segment of unrelated data, not necessarily encoded with arithmetic coding).

Thanks for citing Mark's quality article - I read through it and it was a good refresher for me on the topic.

from reference-arithmetic-coding.