PyLong_AsNativeBytes documentation inaccuracy

[uwu-crj]

Documentation

I am looking at the PyLong_AsNativeBytes section of https://docs.python.org/3/c-api/long.html .

Currently, it says:

All n_bytes of the buffer are written: large buffers are padded with zeroes.

I would expect that in the event that a large buffer is being filled with a negative number it will be padded with 0xFF. I've glanced at the source code and this is indeed the case.

Could the documentation please be amended?

Linked PRs

• gh-143834: Fix PyLong_AsNativeBytes docs for negative number padding  #143840
• [3.14] gh-143834: Fix PyLong_AsNativeBytes docs for negative number padding (GH-143840) #143854
• [3.13] gh-143834: Fix PyLong_AsNativeBytes docs for negative number padding (GH-143840) #143855

[skirpichev]

Indeed, issue seems to be present: https://docs.python.org/3.15/c-api/long.html#c.PyLong_AsNativeBytes

I would say something like this: "All n_bytes of the buffer are written: remaining bytes filled by copies of the sign bit."