IPv6 support for Windows and Mac?

> Michael Rueger wrote:
>> looking at the state of the IPv6 support I couldn't find any  
>> implementations for Windows and Mac.
>> What is the current state of IPv6 support for these platforms?
>
> A while back I looked at what Ian had done in this area and it  
> looked eminently doable. However, I haven't actually done anything  
> and unless there is sudden customer demand for IPv6 (which is not  
> impossible) I wouldn't expect me to get around to this any time soon  
> (meaning: during this year). If you have a real need for IPv6, send  
> some code my way and I'll integrate it.
>
> Cheers,
>  - Andreas

> Andreas Raab wrote:
>
>> this year). If you have a real need for IPv6, send some code my way
>> and I'll integrate it.
>
> OK, here's my first stab at this. It is unfinished, but I've got to run
> for today and thought I could as well post what I have so far.
>
> The files in the zip simply go into the win32 SocketPlugin folder.
>
> I had to add a couple of newer header files and define a few macros to
> make it work on the current mingw version, as the socket header files
> are missing crucial definitions.
>
> The wspiapi.h emulates the newer functions on older Windows versions. If
> we decide to require at least XP we could get rid of it.
>
> Also attached is a partial change set adding image support code. Just
> run "NetNameResolver testIPv6" to run a few sanity checks.
>
> Known problem so far is that certain lookups trigger a getaddrinfo error
> showing up in the error console. I haven't spent much time trying to
> debug this. Hopefully someone with more familiarity with the socket code
> might easily spot what goes wrong.
>
> Probably more problems, haven't tested much yet.
>
> If IPv6 isn't enabled on your windows XP machine you can do so by
>
>     ipv6 install
>
> And uninstall with
>
>     ipv6 uninstall
>
> Comments, suggestions appreciated :-)
>
> Michael
>
>

>
> Well the sqResolverHostNameSize is in the SocketPlugin.c, but the
> required code is not in VMMaker
> Did you want to add the required code to vmmaker?
>
>
> On 30-Mar-09, at 5:51 AM, Michael Rueger wrote:
>
>> John M McIntosh wrote:
>>> Well the mac & iphone VM use the unix socket code so whatever is in
>>> the tree...
>>
>> Hmm, with the latest 3.8.20 VM the new primitives seem to be missing
>> (e.g. NetNameResolver primHostNameSize gives primitive fail).
>>
>> Michael
>
> --
> ===========================================================================
> John M. McIntosh <[hidden email]>
> Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
> ===========================================================================
>
>
>