Opened 7 years ago
Closed 6 years ago
#4040 closed defect (fixed)
Provide opt-out for autofix in overlay functions
Reported by: | dbaston | Owned by: | pramsey |
---|---|---|---|
Priority: | blocker | Milestone: | PostGIS 3.0.0 |
Component: | postgis | Version: | master |
Keywords: | Cc: |
Description (last modified by )
A subset of the overlay functions was recently updated to automatically run ST_MakeValid
on invalid input geometries. If the new behavior is to be the default, it should be documented and an opt-out mechanism made available for users concerned with data integrity.
dev list discusison: https://lists.osgeo.org/pipermail/postgis-devel/2018-February/026885.html
Change History (10)
comment:1 by , 6 years ago
comment:2 by , 6 years ago
Description: | modified (diff) |
---|
comment:3 by , 6 years ago
Milestone: | PostGIS 2.5.0 → PostGIS next |
---|
comment:4 by , 6 years ago
Milestone: | PostGIS next → PostGIS 2.5.0 |
---|
Let's leave blockers with their current release. I have a patch ready to revert the autofix behavior and need to rebase it after merging https://github.com/postgis/postgis/pull/264
comment:6 by , 6 years ago
As Komzzpa noted, this is the wrong pull request.
Should be - https://github.com/postgis/postgis/pull/268
comment:7 by , 6 years ago
Another, 180-degree different, approach to it is in https://github.com/postgis/postgis/pull/275
Similar to legacy.sql, we can provide a file to import to database to revert to "old" behavior. It installs over extension just fine, and provide strong guarantees on preserving data integrity.
comment:9 by , 6 years ago
Milestone: | PostGIS 2.5.0 → PostGIS 3.0.0 |
---|
comment:10 by , 6 years ago
Resolution: | → fixed |
---|---|
Status: | new → closed |
Alternative: since an opt-out method would be complex, revert the changes and add documentation showing how a user can easily implement an exception-handling overlay wrappers at the PL/PGSQL level.