QGIS » QGIS Application

Giovanni,
if the priority is global ok, I agree there are issues more relevant then this.
I thought priority was scoped to the category the issue was assigned to. In that case it's high because the current GeoNode provider is unasable with the current release of GeoNode.

Giovanni Allegri wrote:

Giovanni,
if the priority is global ok, I agree there are issues more relevant then this.
I thought priority was scoped to the category the issue was assigned to. In that case it's high because the current GeoNode provider is unasable with the current release of GeoNode.

The rationale we use is: "high" is for issues causing crashes or data corruption or that are regressions. Does this issue fit it?

ok then this is not the case. This bug simply makes the provider unusable.

This bug is a bit of a red herring, yes it is a problem however if the code is made to correctly parse the minor version then no GeoNode layers are returned because the wrong code is executed to parse the json response.

The following is what is returned from http://master.demo.geonode.org/api/layers/

{"geonode_version": "2.10rc4", "meta": {"limit": 1000, "next": null, "offset": 0, "previous": null, "total_count": 1}, "objects": [{"abstract": "No abstract provided", "alternate":
 "geonode:ukcs_coast", "bbox_x0": "-22.756981842778200", "bbox_x1": "11.999999997221800", "bbox_y0": "47.999999998915300", "bbox_y1": "64.000000008915300", "constraints_other": null,
 "csw_type": "dataset", "csw_wkt_geometry": "POLYGON((-22.7595163995 47.9987371688,-22.7595163995 63.9997273802,11.9989578796 63.9997273802,11.9989578796 47.9987371688,-22.7595163995 
47.9987371688))", "data_quality_statement": null, "date": "2019-01-28T02:39:50.810622", "date_type": "publication", "default_style": "/api/styles/1/", "detail_url": 
"/layers/geonode:ukcs_coast", "dirty_state": false, "edition": null, "geogig_link": null, "gtype": "MultiPolygon", "has_time": false, "id": 1, "is_approved": true, "is_published": true,
 "keywords": ["features", "ukcs_coast"], "language": "eng", "license": 1, "maintenance_frequency": null, "name": "ukcs_coast", "online": true, "owner__username": "ward_r", "owner_name": 
"ward_r", "popular_count": 0, "purpose": null, "rating": 0, "regions": ["Global"], "resource_uri": "/api/layers/1/", "restriction_code_type": null, "share_count": 0, "site_url": 
"http://master.demo.geonode.org/", "spatial_representation_type": null, "srid": "EPSG:4230", "store_type": "dataStore", "supplemental_information": "No information provided", 
"temporal_extent_end": null, "temporal_extent_start": null, "thumbnail_url": "http://master.demo.geonode.org/uploaded/thumbs/layer-46d3505a-22d8-11e9-922b-0242ac120006-thumb.png", 
"title": "ukcs_coast", "typename": "geonode:ukcs_coast", "uuid": "46d3505a-22d8-11e9-922b-0242ac120006"}]}

Based on a version number of 2.10 the code executed expects to see values for the key 'links' which is not present in the above response. If the code is forced to execute the other code block where the expected version is 2.6 then the layers are correctly parsed. I have a GeoNode 2.8 instance that follows the same pattern and so GeoNode layers never appear in QGIS as they should do.

I have a working patch that parses the GeoNode json response based on the contents rather than the version numbers. It appears the GeoNode documentation does not state what the responses format should be let alone how it has changed over the versions. If somebody could point me at a GeoNode instance to provide the alternative json response I can create a unit test and create a pull request to GeoNode functionality working again in QGIS.

The issue of missing links has been addressed recently (https://github.com/GeoNode/geonode/commit/704b8d55363a8f1489fdb62d2501f7c85dc922e8) but master demo is not updated yet.
You can test it using our own dev deployment with the fix applied: http://dev.geonode.geo-solutions.it/api/layers/

Alexander Bruy wrote:

Applied in changeset qgis|440f8d4cbb655942f868dde9ab23b0cc5d69c3b2.

Thanks very much @Alex Bruy!

Thanks also from me ;)