* Peter Eisentraut (peter_e(at)gmx(dot)net) wrote:
> - ip4 really only stores a single address, not a netmask, not sometimes
> a netmask, or sometimes a range, or sometimes a network and an address,
> or whatever. That really seems like the most common use case, and no
> matter what you do with the other types, some stupid netmask will appear
> in your output when you least expect it.
This is definitely one of the funny complications with our built-in
types. I don't feel that's a feature either. Nor do I consider it
'worse' that we have a type that actually makes sense. :) Regardless of
who developed it, it's simply trying to do too much in one type. I'm
also not convinced that our built-in types even operate in a completely
sensible way when you consider all the interactions you could have
between the different 'types' of that 'type', but I'll admit that I
haven't got examples or illustrations of that- something better exists
and is what I use and encourage others to use.
In some ways, I would say this is akin to our built-in types vs.
PostGIS. My argument isn't about features or capabilities in either
case (though those are valuable too), it's about what's 'right' and
makes sense, to me anyway.
In response to
pgsql-hackers by date
|Next:||From: Robert Haas||Date: 2011-12-02 13:56:18|
|Subject: Re: Why so few built-in range types?|
|Previous:||From: Peter Geoghegan||Date: 2011-12-02 13:33:30|
|Subject: Re: Inlining comparators as a performance optimisation|