[postgis-devel] lwgeom: regress tests

strk strk at keybit.net
Tue Jun 8 05:47:23 PDT 2004


On Tue, Jun 08, 2004 at 08:40:18AM -0400, Carl Anderson wrote:
> strk wrote:
> >On Tue, Jun 08, 2004 at 11:13:33AM +0000, Carl Anderson wrote:
> >
> >>On 06/08/2004 05:05:23 AM, strk wrote:
> >>
> >>>All regress tests run with no problems for me.
> >>>
> >>>Just a note. I do not have a 'postgres' user, so I had
> >>>to remove the \c line on top of regres_lots_of_points.sql
> >>>
> >>>If this is not a problem I'd remove the 'You are now connected as...'
> >>>line from the expected result.
> >>>
> >>>--strk;
> >>>
> >>
> >>Another problem for me was on createdb
> >>
> >>createdb -T template0 lwgeom_reg
> >>
> >>will derive from template0 not the usual template1.
> >>
> >>template0 is always clean whilst template1 may have localizations
> >>tsearch, plpgsql, postgis, ...
> >
> >
> >plpgsql and postgis are loaded by regression tests.
> >did localizations or other thikngs make regress fail ?
> >
> >--strk;
> >
> 
> plpgsql is in my template1
> it causes a message to stderr but the normal output still works ok and 
> regression tests pass.
> 
> My point is that unless you start with template0 you are not starting 
> from a known quantity.
> 
> 
> C.

Ooh. Now I got it, I tought you were suggesting to start from template1.
Yes. I agree with you. But template0 was introduced somewhen after pg71
if I'm not wrong... I'd leave it as it is as far as it does not
compromise the test result.
--strk;




More information about the postgis-devel mailing list