[postgis-users] Performance and memory problems

[David Blasby]

strk wrote:
> A quick fix might be removing that pfree() call (memory is freed at
> the end of transaction). Anyway, if you can restrict the number
> of records needed to reproduce the problem I'd take a look at
> them. 
> 
> Dave, looking for possible fixes of the bug reported by Andrea 
> I've incurred of the DirectFunctionCall2 of line  1391
> Is it really needed ? Why don't we make a real direct function call 
> instead ?

Its possible to do it directly, but its much easier to do using 
DirectFunctionCall2 because it handles all the details.  You have to 
remember that postgis function (in fact all postgresql VERSION1 
functions) are actually quite complex.  The details are in the function 
macros (ie. PG_FUNCTION_INFO_V1(...)).

dave