Opened 7 years ago
Closed 6 years ago
#807 closed defect (wontfix)
Buffering a linestring whose endpoints match should not be equivalent to buffering a polygon
| Reported by: | MShepherd | Owned by: | strk |
|---|---|---|---|
| Priority: | major | Milestone: | 3.6.3 |
| Component: | Core | Version: | 3.6.0 |
| Severity: | Unassigned | Keywords: | |
| Cc: |
Description
Please see the attached C++ file.
A particular linestring should be buffered into a donut (a polygon with a hole), but is instead buffered into a polygon with no hole. A similar linestring is buffered correctly. Both cases should produce a donut.
Attachments (1)
Change History (8)
by , 7 years ago
| Attachment: | lineBufferBug.cpp added |
|---|
comment:2 by , 7 years ago
I'm not sure if JTS shares the same bug. There is no bug filed in their tracking system that is similar. Does this need to be confirmed in JTS before triage of this ticket can proceed? I have no experience with JTS, but the repro to check would be to buffer the following linestring with no endcaps, round corners, and a distance of 1:
LINESTRING (600003.5632553970200 6000006.802345209800,600000.4312202869900 6000000.3311080700,600016.5353996580200 6000010.459545190400,600007.3861603239800 6000014.701193110100,600003.5632553970200 6000006.802345209800)
For completeness, it would be great to have the bug fixed in JTS and Geos (assuming it's present there as well). For my purposes, I'd be happy if it were only fixed in Geos.
comment:3 by , 7 years ago
I'd also be happy but not enough to support the sadness of figuring out myself how to fix it :)
Checking JTS is part of that process (figuring out). Want to share some of that ? If the bug is not in JTS it would probably be easier to fix.
comment:7 by , 6 years ago
| Resolution: | → wontfix |
|---|---|
| Status: | new → closed |
It seems this will probably never move since no one is interested in it.
A C++ file that shows the bug.