[postgis-devel] 1.3.4?

[Mark Cave-Ayland]

Obe, Regina wrote:
> Paul,
> I was hoping to get in the pseudo cascade aggregate union function which 
> is in general 5-10 times faster than current and completes where 
> ST_Union runs out of memory.  Though that problem may have been remedied 
> a bit by your memory patches.
>  
> I haven't had a chance to test with the new 1.3.4 changes and regression 
> tests though so don't want to hold up the train.

I don't feel that we're in a rush to ship 1.3.4 ASAP (see previous post 
relating to existing reports on the bug tracker) but in my mind, given 
the development in trunk I see that we should think about making the 1.3 
branch as stable as possible rather than taking the risk of breaking it. 
  I'd like to guarantee that 1.3.4 is a good quality release by changing 
as little as possible, and focus performance enhancements on SVN trunk.

> Then again would people prefer just waiting till we have GEOS Cascaded 
> Union hopefully in 1.4 and I can just put this in wiki (since its 
> implemented in plpgsql so will work just fine in lower versions of 
> Postgis) as an alternative for people who have those issues until that 
> time or people who just don't want to upgrade to the new GEOS when it 
> first comes out?

Yeah; my vote would be to try and get the functionality in GEOS if 
possible rather than adding extra code into PostGIS to handle this. I 
suspect more will be known after the results of the current memory 
discussions on the GEOS list ;)


ATB,

Mark.

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