Check libc code in rtree/core.py about rtree HOT 3 CLOSED

Please make a pull request that fixes this up and I'll be happy to merge it in and roll it into the next release.

On Apr 3, 2014, at 3:39 AM, Ian Kirker [email protected] wrote:

I think this snippet from line 104 of rtree/core.py only works by coincidence.

elif os.name == 'posix':
platform = os.uname()[0]
lib_name = 'libspatialindex_c.so'
if platform == 'Darwin':
lib_name = 'libspatialindex_c.dylib'
free = ctypes.CDLL(find_library('libc')).free
else:
free = ctypes.CDLL(find_library('libc.so.6')).free
If you're actually looking for libc with the find_library function, you'd want find_library('c'), but this code works because:

find_library returns None if it fails to find the library
dlopen (and thus, CDLL) returns a handle that can access libc if you pass None to it as the library parameter.
So, as far as I can tell, this:

if platform == 'Darwin':
    lib_name = 'libspatialindex_c.dylib'
    free = ctypes.CDLL(find_library('libc')).free
else:
    free = ctypes.CDLL(find_library('libc.so.6')).free

could be replaced with:

if platform == 'Darwin':
    lib_name = 'libspatialindex_c.dylib'
free = ctypes.CDLL(None).free

Please do correct me if I'm wrong -- or I guess if this is a deliberate choice for some reason?

—
Reply to this email directly or view it on GitHub.

from rtree.

I will fix this in a new pull request. In short, we should need or use the libc free anyway -- we don't know which .so/.dll it would be coming from. Instead we should use the Index_Free method to make sure that libspatialindex frees its own stuff.

from rtree.

#27 addresses this issue more fully.

from rtree.