Re: [PATCH] supports IPv6 only remote target

[Pedro Alves <palves@redhat.com>]

On 02/10/2016 07:21 PM, Eli Zaretskii wrote:

> Do we care about W2K support at this point?  AFAIK, quite a few
> MinGW64 headers require and set _WIN32_WINNT to a value that will
> exclude W2K anyway.
> 
> I suggest to wait for Pedro or Joel to chime in on this issue.

I wouldn't care about W2K at this point.

gnulib has a getaddrinfo module, but I think that pulls in gnulib's 
select replacement, which is a big problem for gdb.  So if we cared
about old hosts that don't support getaddrinfo, it'd have to be based
on #ifdef.  But, I believe that the only remaining such systems
would be old Windows hosts, since we dropped support for several old
Unix hosts in 7.10/7.11.  As such, I think we can just use getaddrinfo
unconditionally.

I think this is at least the 3rd IPv6 patch posted over the years.
It's about time we merge some in.  I've added a couple folks to
CC who proposed this before.

It'd be useful to review past discussions, and see if
there's something there we should consider.  See e.g.:

  https://sourceware.org/ml/gdb-patches/2014-02/msg00254.html

AFAICS, Tsutomu's patch doesn't cycle through available
addresses.  Should it?

Tsutomu, do you have an FSF copyright assignment?

Paul, do you have an FSF copyright assignment?

Thanks,
Pedro Alves