Re: The netsec thread

[Robert Pluim]

>>>>> On Tue, 30 Jul 2019 17:05:46 +0200, Robert Pluim <address@hidden> said:

>>>>> On Tue, 30 Jul 2019 15:32:13 +0200, Lars Ingebrigtsen <address@hidden> 
>>>>> said:
    Lars> Robert Pluim <address@hidden> writes:
    >>> Yes, once Iʼve finished writing the tests, and undone the port
    >>> stripping in network-lookup-address-info, and made sure it works for
    >>> ipv6. My home internet is down at the moment, so it might be a few
    >>> days.

    Lars> Okidoke.  I could start looking at tweaking some of the other 
behaviour
    Lars> you mentioned (paranoid etc) in the meantime, and then we could
    Lars> tentatively aim at having something ready for merging in...  a couple 
of
    Lars> weeks?

    Robert> Sure. Iʼm on vacation soon, which will give me more time to look at
    Robert> this.

Iʼve just pushed it to the netsec branch, which I didnʼt rebase. The
nsm-should-check stuff will only cause emacs to skip checks to local
IPv4 subnets if itʼs enabled, but thatʼs because
'network-interface-info' only returns the first address it finds for
each interface, which is invariably IPv4. On platforms with
'getifaddrs' or the equivalent, we could extend it to

(network-interface-info ifname &optional addr-type)

Where addr-type is 'ipv4, 'ipv6, 'all

And have 'all mean return a list of address information element lists,
rather than a single one. But I donʼt have time to work on that right
now (and I have no idea what the w32 implementation would look like).

Robert