[postgis-users] Stupid Database Mistake?

[Mark Cave-Ayland]

On Wed, 2007-03-07 at 13:44 -0800, Paul Ramsey wrote:
> I just recognized that our definition of GEOMETRY_COLUMNS does not 
> declare that SRID is a foreign key linked to SRID in SPATIAL_REF_SYS. 
> That seems like a potentially dangerous oversight. Does anyone see a 
> problem if that restriction was *added*?
> 
> Paul

Well, I've worked with databases before where I've compiled without
PROJ.4 support and so not bothered loading in spatial_ref_sys.sql. I
can't really see a problem at the moment, since IIRC the PROJ.4 code
throws a reasonably clear error if you try and use transform() with a
non-existent SRID, and to the best of my knowledge, no-one has
complained about it.


Kind regards,

Mark.