Segfault with PyOpenCL 2011.2 on OS X 10.6.8

[praus]

Hello, I'm trying to use poclbm (most recent one from master) for GPU mining on Mac OS X 10.6.8 with HD 5750. I've everything setup correctly, pyopencl is installed using pip and CPU mining works. However when I choose GPU as a mining device I get a segfault. The segfault is caused by a build issue with the kernel file for SHA computation.

I traced the problem to the following line in BitcoinMiner.py:
self.miner = cl.Program(self.context, kernel).build(self.defines)

I also checked with the most recent development version of PyOpenCL from their Git repository and the issue persists. Is there something I can do? I thought about trying to swap the phatk kernel for something different but I haven't explored further yet.

Thanks

Here's a stack trace from gdb:

#1  0x00000001132263d8 in glrCompLoadBinary ()
#2  0x00000001132264db in glrCompLoadBinary ()
#3  0x0000000113226692 in glrCompBuildProgram ()
#4  0x00007fff85027175 in clBuildProgram ()
#5  0x00000001015ff736 in pyopencl::program::build ()
#6  0x00000001015ead40 in pyopenclboost::python::objects::caller_py_function_impl<pyopenclboost::python::detail::caller<void (pyopencl::program::*)(std::string, pyopenclboost::python::api::object), pyopenclboost::python::default_call_policies, pyopenclboost::mpl::vector4<void, pyopencl::program&, std::string, pyopenclboost::python::api::object> > >::operator() ()
#7  0x000000010163aa6a in pyopenclboost::python::objects::function::call ()
#8  0x000000010163ae00 in pyopenclboost::detail::function::void_function_ref_invoker0<pyopenclboost::python::objects::(anonymous namespace)::bind_return, void>::invoke ()
#9  0x0000000101649be1 in pyopenclboost::python::detail::exception_handler::operator() ()
#10 0x0000000101622d02 in pyopenclboost::detail::function::function_obj_invoker2<pyopenclboost::_bi::bind_t<bool, pyopenclboost::python::detail::translate_exception<pyopencl::error, void (*)(pyopencl::error const&)>, pyopenclboost::_bi::list3pyopenclboost::arg<1, pyopenclboost::arg<2>, pyopenclboost::_bi::value<void (*)(pyopencl::error const&)> > >, bool, pyopenclboost::python::detail::exception_handler const&, pyopenclboost::function0<void> const&>::invoke ()
#11 0x000000010164981b in pyopenclboost::python::handle_exception_impl ()
#12 0x00000001016379ce in function_call ()
#13 0x000000010000c122 in PyObject_Call ()
#14 0x00000001000c279a in PyEval_EvalFrameEx ()
#15 0x00000001000c66f6 in PyEval_EvalCodeEx ()
#16 0x00000001000c46d3 in PyEval_EvalFrameEx ()
#17 0x00000001000c66f6 in PyEval_EvalCodeEx ()
#18 0x00000001000c46d3 in PyEval_EvalFrameEx ()
#19 0x00000001000c66f6 in PyEval_EvalCodeEx ()
#20 0x00000001000c46d3 in PyEval_EvalFrameEx ()
#21 0x00000001000c66f6 in PyEval_EvalCodeEx ()
#22 0x00000001000c46d3 in PyEval_EvalFrameEx ()
#23 0x00000001000c5d90 in PyEval_EvalFrameEx ()
#24 0x00000001000c66f6 in PyEval_EvalCodeEx ()
#25 0x000000010003d440 in function_call ()
#26 0x000000010000c122 in PyObject_Call ()
#27 0x00000001000c0ebf in PyEval_EvalFrameEx ()
#28 0x00000001000c5d90 in PyEval_EvalFrameEx ()
#29 0x00000001000c5d90 in PyEval_EvalFrameEx ()
#30 0x00000001000c66f6 in PyEval_EvalCodeEx ()
#31 0x000000010003d440 in function_call ()
#32 0x000000010000c122 in PyObject_Call ()
#33 0x000000010001d60d in instancemethod_call ()
#34 0x000000010000c122 in PyObject_Call ()
#35 0x00000001000be0b7 in PyEval_CallObjectWithKeywords ()
#36 0x00000001001048d3 in t_bootstrap ()
#37 0x00007fff83c02fd6 in _pthread_start ()
#38 0x00007fff83c02e89 in thread_start ()```

[praus]

By the way, I also tried replacing the kernel in current version of poclbm with the kernel from cgminer. That actually compiled after I defined WORKSIZE param but (obviously?) failed on the number of arguments for the search function (21 vs. 24).

[praus]

This issue has been resolved for me by upgrading to 10.7 (10.7.3 to be precise).