layer __getitem__ bug

[timaranz]

I found a bug in the layer.getitem method of the python bindings. It is trying to use the types module but it hasn't been imported (or rather it's been imported and deleted again.), and it doesn't import it explicitly itself.

[timaranz]

patch for swig\python\osgeo\ogr.py

[hobu]

Fixed in r13647 and backported to 1.5 branch in r13648.

[frankie]

klecker:~/debian/debian-gis/gdal/gdal-1.5.2$ ls -l ./swig/python/osgeo/ogr.py -rw-r--r-- 1 frankie 23 95120 2007-11-29 23:52 ./swig/python/osgeo/ogr.py klecker:~/debian/debian-gis/gdal/gdal-1.5.2$ ls -l ./swig/include/python/ogr_python.i -rw-r--r-- 1 frankie 23 8669 2008-02-01 04:53 ./swig/include/python/ogr_python.i

The change done in the .i file is not propagated to the .py file, as you can see from the dates. That still causes  http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=473912 in 1.5.1 and 1.5.2. Isn't that the case to force swing use by make in those cases?

[rouault]

Apparently the swig bindings were not regenerated for 1.5.1 and 1.5.2 releases. We should do that for 1.5.3.

[warmerdam]

Note that our policy is that the bindings maintainer is espected to regenerate and commit modified bindings for languages other than perl (and perhaps c#?) which are generated by mkgdaldist.sh.

So, hobu ought to be regenerated the bindings in 1.5 branch if they are to be updated in 1.5.3.

[hobu]

Bindings have been regenerated since this bug was opened for 1.5 and 1.6. Closing...