[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: bugs

>>>>> "k" == kiewicz  <Arkadiusz> writes:

    k> Hi, rlinetd-0.4 still doesn't work properly when family
    k> ipv{4,6} is specified.  Maybe this is getaddrinfo() problem -
    k> who knows ;-) but even if getaddrinfo will be ok rlinetd won't
    k> work properly !

  I think this is indeed a bug in getaddrinfo() - I'll have a look at
the specs to check (tomorrow! :).

    k> so it seems that my teory is true (bind() on wildcard causes
    k> binding to Ipv4 interfaces, too)... If it is realy true this
    k> must be workarounded in rlinetd or RFC must be changed :-)

  And this is also, I think, working as specified. An IPv6-bound
socket will accept IPv4 connections. This is the entire purpose of the 
IPv4-mapped space (addresses like ::FFFF:a.b.c.d).

  I'm pretty sure about the second point, and I'll try to dig up some
specs on the first to see if the getaddrinfo specification says what I 
think it does.