Re: Extending varlena

Tom Lane wrote:
> Andrew Chernow <ac@esilo.com> writes:
>> Anyways (back on topic), I am in favor of removing limits from any 
>> section of the database ... not just your suggestion.  The end-user 
>> application should impose limits.
> 
> That's nice as an abstract principle, but there are only so many hours
> in the day, so we need to prioritize which limits we're going to get rid
> of.  The 4-byte limit on individual Datum sizes does not strike me as a
> limit that's going to be significant for practical use any time soon.
> (I grant David's premise that people will soon want to work with objects
> that are larger than that --- but not that they'll want to push them
> around as indivisible, store-and-fetch-as-a-unit field values.)
> 
>             regards, tom lane
> 
> 

Yeah, my comments were overly general.  I wasn't suggesting attention be 
put on one limit over another.  I was only saying that the act of 
removing a limit (of which many are arbitrary) is most often a good one.

andrew