Re: Cleaning up the INET/CIDR mess

On Jan 25, 2006, at 10:30 AM, Andrew - Supernews wrote:

> On 2006-01-25, Greg Stark <gsstark(at)mit(dot)edu> wrote:
>> This isn't an obscure old-fashioned thing. People really do use
>> this syntax.
>
> Given how little code now supports 10.1 meaning 10.0.0.1, that seems a
> questionable point.
>

All code that uses gethostbyname() on, at least, Linux, Solaris,
Windows XP,
OS X and (I think) the BSDs and anything else that's even vaguely
posix uses it.

I don't think that's terribly relevant to the PG inet types, though.
Rejecting any
input format that's not a dotted-quad seems the safest thing to do,
and doesn't
lose any useful functionality. Given the number of people in this
thread who
think that the (non-standard, archaic) behaviour of bind is correct
it's clear that
accepting anything other than a real dotted-quad will lead to an
inconsistency
between what the data represents and what the user thinks it
represents, and
that's bound to cause problems.

Cheers,
Steve