valid netmask like 255.0.0.0 255.128.0.0, but 255.254.128.0 is not valid, but ipaddress._BaseV4._is_valid_netmask fails to detect the latter.

Tested in Python 3.6.7, as the code stays the same, affects all after Python 3.6.7.

I couldn't find any use of _is_valid_netmask() in the code, why should this netmask be invalid?

As far as I can tell this should be a valid one, is there a reason to refuse it?

> I couldn't find any use of _is_valid_netmask() in the code, why should this netmask be invalid?

There is an open issue to remove this method which is unused and undocumented issue27860. 

I have less knowledge on the subnet masking area but on searching the web there seems to be an RFC that asks for contiguous ones without intermixing zeroes that I hope OP is referring to.

* https://superuser.com/questions/601252/is-225-225-225-128-a-valid-subnet-mask
* https://superuser.com/questions/1151908/non-contiguous-subnet-mask/1151942#1151942

Still 255.254.128.0 is a valid subnet.

If I understand correctly, rfc1519 relate to how public ip addresses should be attributed. It does not cover what private subnets you use and you still can such submask (as long as you own the whole /16 for example).

rfc950 explicitly allow for such subnets:

> Since the bits that identify the subnet are specified by a bitmask, they need not be adjacent in the address. However, we recommend that the subnet bits be contiguous and located as the most significant bits of the local address.

and I've heard of them being used (hence the complications and the need for rfc1519 for the public address space).

I'm in favor of closing this issue in favor of issue27860.

I opened https://github.com/python/cpython/pull/11591 to remove _valid_mask_octets, _is_valid_netmask and _is_hostmask from ipaddress, the patch proposed in issue27860 has a larger scope so it's probably better to remove those unused methods in this issue.

Does someone know how to set the skip news label on GitHub?

> I opened https://github.com/python/cpython/pull/11591 to remove _valid_mask_octets, _is_valid_netmask and _is_hostmask from ipaddress, the patch proposed in issue27860 has a larger scope so it's probably better to remove those unused methods in this issue.

Thanks, the subject of the issue is misleading and I would propose changing the issue subject or closing this in favor of the other issue and opening a PR against issue27860 stating that this only removes the unused functions and not the complete patch attached. This would require a review from the module maintainer I hope though it's an internal function that is never used.

> Does someone know how to set the skip news label on GitHub?

It needs to be added by a core dev

I changed the subject of this issue as the scope of issue27860 is larger. I will review them and open a new PR for them if appropriate once this one is accepted.

OK, message copied, I'll mark it closed. :)

New changeset e59ec1b05d3e1487ca7754530d3748446c9b7dfd by Inada Naoki (Rémi Lapeyre) in branch 'master':
bpo-35734: ipaddress: remove unused methods (GH-11591)
https://github.com/python/cpython/commit/e59ec1b05d3e1487ca7754530d3748446c9b7dfd