Opened 12 years ago
Closed 12 years ago
#1543 closed defect (fixed)
ST_GetFaceGeometry exception on invalid edge, null on unexistent face
| Reported by: | strk | Owned by: | strk |
|---|---|---|---|
| Priority: | medium | Milestone: | PostGIS 2.0.0 |
| Component: | topology | Version: | master |
| Keywords: | Cc: |
Description
The SQL/MM document says that an exception should be raised IF the face argument references an unknown face. Doesn't say anything about faces bound by no ring (would be due to invalid topology).
The current implementation results in an exception in the latter case (no ring) but not in the former (unknown face).
In turn this doesn't help validating a topology in that the exception on invalid topology makes the whole validateTopology function bail out.
Change History (3)
comment:1 by , 12 years ago
comment:2 by , 12 years ago
| Summary: | ST_GetFaceGeometry exception on partial face boundary, null on unexistent face → ST_GetFaceGeometry exception on invalid edge, null on unexistent face |
|---|
comment:3 by , 12 years ago
| Resolution: | → fixed |
|---|---|
| Status: | new → closed |
r9184 fixes the exception by making even single-point edges convertible to GEOS. The null-on-unexistent face I'll pretend I didnt see (would only make things slower)
Note:
See TracTickets
for help on using tickets.
This is actually ST_BuildArea's fault, whereas passing 'LINESTRING(0 0, 10 10)' as input returns NULL while 'LINESTRING(0 0)' as input returns an exception (cause the input is invalid…).