Ticket Hash: | 970ab8322e4995793b6b3acb872d1e6a100af60b | |||
Title: | spatial_ref_sys definitions shouldn't contain TOWGS84[] in WKT definition | |||
Status: | Closed | Type: | Code_Defect | |
Severity: | Critical | Priority: | Immediate | |
Subsystem: | Resolution: | Works_As_Designed | ||
Last Modified: | 2020-08-02 16:10:31 | |||
Version Found In: | master | |||
User Comments: | ||||
anonymous added on 2020-03-12 20:24:35:
(Even Roualt here) I see fossil head contains WKT definitions in srs/srsinit/epsg_* with TOWGS84[] terms in them. Those were useful in a GDAL < 3 & PROJ < 6 world, but they are now a nuisance with GDAL 3 and PROJ >= 6 as can been seen in https://github.com/qgis/QGIS/issues/34992#issuecomment-598392777 I'd suggest that the future of Spatialite no longer contains them since it will likely be more likely deployed with GDAL 3 and PROJ >= 6 sandro added on 2020-08-02 16:10:31: Hi Even, supporting a +towgs84 term in "old style" geodesic strings contained into spatial_ref_sys is absolutely intentional. here is a short rationale: 1. when spatialite is linked against PROJ.6 or PROJ.7 all geodesic definitions are always retrieved directly from the PROJ own sqlite DB 2. so the old geodesic strings in spatial_ref_sys are still there just to support the case when spatialite is linked against some obsolete version of PROJ, and in this case defining +towgs84 is absolutely required in order to preserve a fair historical coherence. future perspectives: when old PROJ versions will be reasonably disappeared, spatial_ref_sys will become just a plain dictionary (SRID, auth_name, auth_srid and ref_sys_name) mainly intended to support Foreign Key constrains in geometry_columns, but will no longer contain any geodesic data at all. |