ILNP nonce builders size with floor division: math.ceil(bit_length() // 8) at hopopt.py:1889 and ipv6_opts.py:1892

[JarryShaw]

The ILNP nonce builders size their option with math.ceil(nonce.bit_length() // 8), which is floor division — math.ceil on an int is a no-op — so the declared length is wrong for every nonce whose bit length is not a multiple of 8.

The two sites

pcapkit/protocols/internet/hopopt.py:1889      len=math.ceil(nonce.bit_length() // 8),
pcapkit/protocols/internet/ipv6_opts.py:1892   len=math.ceil(nonce.bit_length() // 8),

Both are the ILNP Nonce option builder.

The codebase already knows the correct form, in eight other places

pcapkit/protocols/internet/hip.py:3156    len=4 + math.ceil(random.bit_length() / 8),
pcapkit/protocols/internet/hip.py:3334    pub_len = math.ceil(pub_val.bit_length() / 8)
pcapkit/protocols/internet/hip.py:4276    len=math.ceil(id.bit_length() / 8),
pcapkit/protocols/internet/hip.py:4302    len=math.ceil(id.bit_length() / 8),
pcapkit/protocols/internet/mh.py:7947     id_len = max(1, math.ceil(identifier.bit_length() / 8))

and pcapkit/protocols/internet/mh.py:7936 carries a comment that states the distinction outright:

id_len = math.ceil(identifier.bit_length() / 8) was the right …

So this is a two-site typo against an established convention, not an unresolved design question.

Measured

math.ceil(n // 8) versus math.ceil(n / 8), executed on origin/main (13a75dfcd):

value=1         bit_length=1   floor=0  correct=1  WRONG
value=2         bit_length=2   floor=0  correct=1  WRONG
value=42        bit_length=6   floor=0  correct=1  WRONG
value=127       bit_length=7   floor=0  correct=1  WRONG
value=255       bit_length=8   floor=1  correct=1
value=256       bit_length=9   floor=1  correct=2  WRONG
value=65536     bit_length=17  floor=2  correct=3  WRONG

Small nonces are the worst case, not boundary values: any nonce below 256 is sized at zero octets. Only a bit length that is an exact multiple of 8 gives the right answer.

Why the test suite does not catch it

There is no ILNP entry in EXPECTED_FAILURES in tests/protocols/test_option_roundtrip_unit.py — confirmed by importing it rather than grepping, since it is built with ** unpacking; 45 entries, none ILNP-related. So the round-trip case passes, which means it is landing on a nonce whose bit length happens to divide by 8.

That is the same blind spot that hid the identical defect in pcapkit/corekit/fields/numbers.py: #591's repair-path suite used 0xFF, 0xFFFF, 0xFFFFFFFF, 0xFFFFFFFFFFFFFFFF and 0x800001 — bit lengths 8, 16, 32, 64 and 24, every one an exact multiple of 8, precisely where floor division and the ceiling agree. A fix here needs a nonce whose bit length is not a multiple of 8, and ideally one below 256.

Notes

• Found by the pre_process sizes a value with floor division disguised as math.ceil, so every octet boundary raises OverflowError #599 worker while fixing the same typo in numbers.py:198; that file's fix is fix(corekit): size the width repair with a real ceiling, not a floored one (#599) #600 and does not touch these two sites, which is why this is filed separately.
• Not the same as pre_process sizes a value with floor division disguised as math.ceil, so every octet boundary raises OverflowError #599: that one is the field-layer repair path, reachable only through the field API. These two are protocol builders on the construction path.
• pcapkit/protocols/internet/hip.py:3200 uses math.ceil(… / 4) rather than / 8. That may be deliberate — a nibble count rather than an octet count — and I have not established which, so it is recorded here as worth a look rather than as a defect.