On Thu, Sep 02, 2010 at 12:48:17AM +0100, Mark Cave-Ayland wrote:
> Mark Cave-Ayland wrote:
>
> >Ah wait a second, you're right - I've been looking in the wrong place.
> >It looks like there is still some (old) pgsql2shp code which hasn't been
> >updated to use the new error handling, and hence it doesn't get
> >propagated back up to the main function correctly :( Will fix later, as
> >I have to head out now.
>
> Okay - I've just committed a better fix as r5892 which makes sure that
> in the case of a 0 row result, pgsql2shp will terminate returning an
> error code of 1 (failure) as opposed to 0 (success) when dumping tables
> into shapefiles (serves me right for trying to fix things in a rush when
> I'm just about to head out). Denis, please can you try again after an
> "svn up"?
Just to stress my point out, I was suggesting NOT to look at
number of rows to tell success/failure.
Granted, pgsql2shp is currently unable to produce a 0-row
valid shapefile, but maybe such thing exists.
It is fine for me throwing different non-zero return codes
for different errors. Just I think this error is:
"Can't determine geometry type"
and not:
"There are no rows"
Clearer now ?
--strk;
() Free GIS & Flash consultant/developer
/\ http://strk.keybit.net/services.html
_______________________________________________
postgis-users mailing list
postgis-users@postgis.refractions.net
http://postgis.refractions.net/mailman/listinfo/postgis-users
