[postgis-users] ST_Union and 8.3?

[Mark Cave-Ayland]

On Thursday 20 March 2008 13:37:03 Obe, Regina wrote:
> Mark,
>
> Can you think of anything wrong with using ST_MemUnion as a temporary
> fix aside from the fact that I guess for larger unioning jobs it is
> probably less efficient?

I can't say that I've had much to do with that section of code - I think it 
was probably strk that came up with the idea and implementation.

> Has anyone tried the new 8.3.1 to see if it changes the below issue.  I
> don't see anything in the list of changes
> http://www.postgresql.org/docs/8.3/static/release-8-3-1.html that would
> suggest it would, except possibly for
>
> "Fix memory leaks in certain usages of set-returning functions"
>
> I'm planning to try it out in the next day or so.

The discussion on -hackers was that objects created within each SRF were 
living longer than intended. Now in the case of a buggy SRF this could cause 
excess memory usage during execution that would only be freed at the end of 
the SRF. However I have a feeling that most of the memory in PostGIS is being 
held by GEOS so this may/or may not make a difference. But please feel free 
to experiment :D


ATB,

Mark.

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