Eric Lemoine wrote: > On Thu, Apr 24, 2008 at 2:48 PM, Eric Lemoine <[EMAIL PROTECTED]> wrote: >> Hello >> >> I have a segfault in geos with Shapely (latest pypi version). >> >> Trace: >> >> #0 0x00000000 in ?? () >> #1 0xb195398c in GEOSEquals () from /usr/lib/libgeos_c.so.1 >> #2 0xb6f71b1b in ffi_call_SYSV () at >> /tmp/easy_install-5N8lYe/ctypes-1.0.2/source/libffi/src/x86/sysv.S:60 >> #3 0xb6f71983 in ffi_call (cif=0xb194a548, fn=0xb1953960 >> <GEOSEquals>, rvalue=0x84f3b48, avalue=0xb5992820) >> at /tmp/easy_install-5N8lYe/ctypes-1.0.2/source/libffi/src/x86/ffi.c:221 >> #4 0xb6f6cbf1 in _CallProc (pProc=0xb1953960 <GEOSEquals>, >> argtuple=0xb171b44c, flags=<value optimized out>, argtypes=0x0, >> restype=0x82571ec, checker=0x0) >> at source/callproc.c:668 >> #5 0xb6f67a4c in CFuncPtr_call (self=0xb69cedfc, inargs=0xb171b44c, >> kwds=0x0) at source/_ctypes.c:3362 >> #6 0x080589d7 in PyObject_Call () >> #7 0x080b68c7 in PyEval_EvalFrame () >> #8 0x080ba735 in PyEval_EvalCodeEx () >> ... >> ... >> >> Has anyone seen this? > > I can't figure out what's going on. This happens when I try to save a > feature using SQLAlchemy. Actually I have two tables in my database, > one with polygons and one with points. I have no problem with the > polygon table, only with the point table. At this point, I am > clueless. I've been wondering why SQLAlchemy does a geometry > comparison (geom1 != geom2 apparently) when saving a feature. I don't > even understand to what other geometry it compares the one I'm > attempting to save. > > Thanks, > > -- > Eric
Eric, Try adding a breakpoint at the beginning of BinaryPredicate.__call__ in shapely/predicates.py to see what's up with self.context._geom and other._geom. Sean _______________________________________________ Community mailing list [email protected] http://lists.gispython.org/mailman/listinfo/community
