dpage@postgresql.org ("Dave Page") writes:
>> ------- Original Message -------
>> From: Euler Taveira de Oliveira <euler@timbira.com>
>> To: Dave Page <dpage@postgresql.org>
>> Sent: 17/11/07, 16:38:10
>> Subject: Re: [pgadmin-hackers] Support for older servers
>>
>> Maybe it's too hard to test a feature in all of those old versions
>
> Yes. It can be extremely time consuming, even to test simple bug
> fixes that *should* be fine on any version. That became apparent
> this cycle when a few version-specific bugs did creep in when I
> thought I could get away with more limited testing, and later got
> found in the field. :-(

I have certainly found this to be troublesome with Slony-I; I have too
frequently been surprised by seeing things break with, say, 8.0, where
features were fine earlier and later, basically because of mishandled
detection of when an interesting new feature (that we intended to use)
actually crept in.

It's fine if you're always looking to simply use "lowest common
denominator" functionality, but if you want to use richer feature
sets, which is *certainly* going to be true for pgadmin, the
variations between Pg versions will matter, and not in a good way :-(.
--
let name="cbbrowne" and tld="linuxfinances.info" in String.concat "@" [name;tld];;
http://linuxfinances.info/info/linuxdistributions.html
"If you give a man a fish, he will eat for a day. If you teach him how
to fish, he will sit in a boat and drink beer all day."