[postgis-devel] [postgis-tickets] r16649 - Committ same changes to concave hull tests to sfcgal as was done for regular

[Regina Obe]

Oh wait :) I see what you mean now about the \i.  The SFCGAL tests often just do \i to call the higher test so they can have different expected.

We should probably think about this more.

For example I don't see why if SFCGAL is enabled it would ever change the output of the geography tests, so I'd like to just chuck the sfcgal one of that.
Same for legacy tests.


> -----Original Message-----
> From: Regina Obe [mailto:lr at pcorp.us]
> Sent: Saturday, July 21, 2018 12:27 AM
> To: 'PostGIS Development Discussion' <postgis-devel at lists.osgeo.org>
> Subject: RE: [postgis-devel] [postgis-tickets] r16649 - Committ same changes
> to concave hull tests to sfcgal as was done for regular
> 
> > > Modified: trunk/regress/sfcgal/concave_hull.sql
> >
> > Does anyone recall why we're using duplicated tests ?
> > It would be useful to drop the duplication somehow (using \i ?)
> >
> > --strk;
> > _______________________________________________
> > postgis-devel mailing list
> > postgis-devel at lists.osgeo.org
> > https://lists.osgeo.org/mailman/listinfo/postgis-devel
> [Regina Obe]
> The thinking was that when SFCGAL is enabled and it is set as backend, then
> there are some functions it overrides so neither GEOS nor core PostGIS might
> be the ones being used.
> 
> e.g. ST_Intersects, ST_3DIntersects, ST_Intersection
> 
> The tests for the most part are duplicated, but I think sometimes the results
> were tweaked to allow valid answers that SFCGAL returns that are different
> from GEOS/PostGIS.
> So thought the tests are the same sometimes the results may be different.
> 
> I'm not clear how you propose the \i to account for sometimes needing to
> use different output files.
> 
> We'd have to look at each one to know.
> 
> For the ones where the output is expected to be the same in both cases, we
> can probably just somehow have sfcgal reference the main test files.
>