[postgis-devel] RECHECK not supported in PsotgreSQL HEAD any more?

[Mark Cave-Ayland]

Paul Ramsey wrote:
> Sorry to re-chew the RECHECK stuff, but am I right that our consistent
> test is just checking that SRIDs are the same? I personally think we
> should drop the test for performance reasons and transfer the moral
> hazard to the users: if they want to try do index runs across
> reference systems, go for it. Besides, if they happen to use *any*
> spatial functions in their SQL, they'll get stopped at the point they
> enter the function by an SRID test anyways, so the risk of
> "surprisingly wrong" results is very low.
> 
> P.

Yeah that was the main purpose of the check. I think at some point the 
test should go, it was more a case of fixing SVN trunk before I'm away 
again.

One thing I did think of is that if we remove it, we need to clearly 
document that A && B (along with all of the other operators) will be 
subject to rounding from float8 -> float4. So the warning should be 
don't use the plain operators if you want exact results...


ATB,

Mark.

-- 
Mark Cave-Ayland
Sirius Corporation - The Open Source Experts
http://www.siriusit.co.uk
T: +44 870 608 0063