[postgis-devel] current_parser_check_flags

strk strk at keybit.net
Tue Dec 7 00:43:25 PST 2010


On Mon, Dec 06, 2010 at 04:08:02PM -0800, Paul Ramsey wrote:

> select astext('0103000000010000000400000000000000000024400000000000003640000000000000244000000000000040400000000000003440000000000000404000000000000034400000000000003640'::geometry);
>                astext
> ------------------------------------
>  POLYGON((10 22,10 32,20 32,20 22))
> (1 row)
> 
> 
> 
> So, decision point: input validity checks or not?

The HEXEWKB you show above might be in dumps.
This was the whole reason for revising the checks in the first place.
You don't want to lock out data.
So, you may want to have _both_ possibilities, so that users may opt
for checked input when willing to, but still can restore data w/out
or get data in for performing invalidity analisys "manually".

Might be another candidate for a GUC, but I really think canonical
input functions (those used on restore) are NOT checking by default.

--strk;

  ()   Free GIS & Flash consultant/developer
  /\   http://strk.keybit.net/services.html



More information about the postgis-devel mailing list