[GRASS-dev] r51975, vlib: improve PostGIS topology support, introduce level 3

Hi Martin,

I find the newly introduced level 3 for topology confusing because 1)
level 3 could sometime in the future mean 3D topology, 2) level 3
applies only to PostGIS. Before that change, the open level had the
same meaning independent of the vector format (GRASS, PGR, PostGIS).
The open level is important for modules, and for modules it does not
matter if topology is pseudo-topology or PostGIS topology or true
GRASS topology. Modules only check if topology is available or not,
but are and should not be interested in what kind of topology is
available. This should be handled by the vector libs. I would
therefore prefer to keep the two open levels and rather have a flag in
struct Format_info_pg indicating whether this map is opened with
pseudo-topology or PostGIS topology.

Markus M

Hi Markus,

2012/6/6 Markus Metz <markus.metz.giswork@googlemail.com>:

first of all, thanks for the quick review!

I find the newly introduced level 3 for topology confusing because 1)
level 3 could sometime in the future mean 3D topology, 2) level 3
applies only to PostGIS. Before that change, the open level had the
same meaning independent of the vector format (GRASS, PGR, PostGIS).

right.

The open level is important for modules, and for modules it does not
matter if topology is pseudo-topology or PostGIS topology or true
GRASS topology. Modules only check if topology is available or not,
but are and should not be interested in what kind of topology is
available. This should be handled by the vector libs. I would

Sounds reasonable to me.

therefore prefer to keep the two open levels and rather have a flag in
struct Format_info_pg indicating whether this map is opened with
pseudo-topology or PostGIS topology.

It's already there (pg_info->toposchema_name).

I was worried about introducing level 3 in this sense, you just made
me sure, that it was not a good idea. I will revert it back to the two
levels (keep level 3 for 3D topology).

Martin

--
Martin Landa <landa.martin gmail.com> * http://geo.fsv.cvut.cz/~landa

2012/6/6 Martin Landa <landa.martin@gmail.com>:

[...]

I was worried about introducing level 3 in this sense, you just made
me sure, that it was not a good idea. I will revert it back to the two
levels (keep level 3 for 3D topology).

applied in r51984. Martin

--
Martin Landa <landa.martin gmail.com> * http://geo.fsv.cvut.cz/~landa

On Wed, Jun 6, 2012 at 3:27 PM, Martin Landa <landa.martin@gmail.com> wrote:

2012/6/6 Martin Landa <landa.martin@gmail.com>:

[...]

I was worried about introducing level 3 in this sense, you just made
me sure, that it was not a good idea. I will revert it back to the two
levels (keep level 3 for 3D topology).

applied in r51984. Martin

Thanks!

BTW, when you're done implementing PostGIS topology, could you provide
a sample script to play around a bit with the new functionality?
Thanks again!

Markus M

Hi,

2012/6/6 Markus Metz <markus.metz.giswork@googlemail.com>:

[...]

BTW, when you're done implementing PostGIS topology, could you provide
a sample script to play around a bit with the new functionality?
Thanks again!

sure, before that I would like to implement also write access
(currently available for feature tables + pseudo-topology). In
progress.

Martin

--
Martin Landa <landa.martin gmail.com> * http://geo.fsv.cvut.cz/~landa