Opened 16 years ago
Closed 9 years ago
#2380 closed defect (fixed)
OGRBuildPolygonFromEdges() has not been mapped to SWIG
Reported by: | Even Rouault | Owned by: | hobu |
---|---|---|---|
Priority: | normal | Milestone: | 1.8.1 |
Component: | SWIG (all bindings) | Version: | 1.5.0 |
Severity: | normal | Keywords: | swig OGRBuildPolygonFromEdges |
Cc: | dfuhry@…, warmerdam |
Description
Reported by David Fuhry, see http://lists.osgeo.org/pipermail/gdal-dev/2008-May/017127.html
ogr.BuildPolygonFromEdges as called from python is always throwing an exception for me, and perhaps always segfaulting: $ ./test_build_poly.py BuildPolygonFromEdges failed. Segmentation fault I first experienced the error running swig/python/samples/tigerpoly.py on some TIGER data. I duplicated some of tigerpoly.pl into a reduced python testcase (attached). Tested with both svn head and 1.5.1, on both RHEL4 x86_64 and Debian x86_64. I don't know the debugging procedure for tracing python calls into C++. Suggestions on that or how to fix the bug welcome. Thanks, Dave
Change History (7)
comment:1 by , 16 years ago
Status: | new → assigned |
---|
comment:2 by , 16 years ago
Cc: | added |
---|
Adding Frank in CC to regenerate the SWIG bindings and see if backporting is to be considered.
comment:3 by , 16 years ago
I have regenerated the python bindings. I'm not too keen on backporting new bindings into 1.5.
comment:5 by , 16 years ago
For me, methods not mapped in NG bindings, but present in OG bindings, should be more considered as API breakage between GDAL 1.4 and GDAL 1.5, and I think that users expect they can switch from OG bindings to NG bindings safely.
The ogr.BuildPolygonFromEdges is an example of that, and gdal.EscapeString reported in #2400 is another one.
comment:6 by , 15 years ago
This should be in 1.6. Can you please test and confirm that things work correctly Dave?
comment:7 by , 9 years ago
Resolution: | → fixed |
---|---|
Status: | assigned → closed |
There are now 4 ogr_geom_build_from_edges_ tests in the autotest script /trunk/autotest/ogr/ogr_geom.py. It looks like the initial fix for GDAL 1.6 was good and autotests have been enhanced afterwards.
I have implemented this in r14581 with Dave's test case in r14580, which succeeds. I don't know if this should be backported to 1.5 (too late for 1.5.2 anyway).