On Tue, Mar 11, 2003 at 11:34:57PM +0200, Doytchin Spiridonov wrote:
> Hello Joe,
>
> OK here are some:
>
> try.com
> try.net
> get.com
> get.net
> monitor.com
> car.com
> car.net
>
> etc!
>
> I would say any not so often used word should be already registered at
> the test system but not in active transfer (as people don't test so
> much the transfers :). So it should be easy for you to find several
> thousand domains to test with..
The point isn't that I can randomly find some domains which work
*today*, the point is that domains are needed which provide
predictable results for the life of the API, or there's no point
writing regression tests.
A further example: I ran out of credit in the testing environment,
while writing my test suite - as a result, all my scripts 'broke', and
I spent some fruitless time trying to 'fix' them purely as a result of
test data changing.
OpenSRS seem very keen on Full Access being done properly, with their
RITE test and so on; they need to back this up with some useful test
data in the Test environment.
Chris.
This archive was generated by hypermail 2.1.3 : Tue Oct 19 2004 - 23:37:39 EDT