[Vm-dev] Socket aid and advice request

[David T. Lewis]

On Wed, Feb 20, 2013 at 09:56:34PM -0500, John McIntosh wrote:
>  
> On Wed, Feb 20, 2013 at 9:03 PM, David T. Lewis <lewis at mail.msen.com> wrote:
> >
> > On Wed, Feb 20, 2013 at 10:06:21AM -0800, tim Rowledge wrote:
> >>
> >>
> >> On 20-02-2013, at 9:59 AM, Bert Freudenberg <bert at freudenbergs.de> wrote:
> >>
> >> >
> >> >
> >> > On 2013-02-20, at 00:36, tim Rowledge <tim at rowledge.org> wrote:
> >> >
> > There are some problems with using this on Windows related to very
> > slow resolver performance. I don't fully understand the issues, but
> > they are enough of a problem that we made "new network support" a
> > preference in Squeak, and have it turned off for release images.
> 
> There was some problems with resolving the local host name & IP under
> unxi/linux/osx/ios
> as it would get an array of ipv4 & ipv6 names and present something
> unexpected to the caller.
> Oddly an acceptable solution would be to just return 127.0.0.1 and be
> done with it. But no-one
> knew what to do so the easiest thing was to flip it off.

I also want to refer back to a number of points that Andreas made:

  http://lists.squeakfoundation.org/pipermail/squeak-dev/2012-July/164853.html

When we were deciding what to do for the Squeak release, I realized
there were a number of unresolved problems on non-unix platforms, and
also recognized that Andreas was raising valid concerns with respect
to trying to do an add-on approach versus a rewrite (not that there
is anyone out there planning on doing a rewrite, but still...). Under
those circumstances, the only prudent thing to do was to disable the
IPv6 preference for that Squeak release, and come back to the problem
when we had time to think it through properly. When I say "I" here,
I mean it in the sense that I had a role in putting that stuff into the
image, and the rule is "the person who breaks the image has to fix it".

Dave