Hi all,

I suppose that expecting weave.inline to magically make the following work:

############### from weave import inline class bunch: pass

def oaccess(): x=bunch()

x.a = 1

code = """ // BROKEN! // Try to emulate Python's: print 'x.a',x.a std::cout << "x.a " << x.a << std::endl; """ inline(code,['x'])

oaccess() ###############

is probably a bit too much, right :) ? It would have to extract all

Hey Fernando, the

members of 'x', and make all the right type work in C++. This sounds pretty much near impossible.

Things like these are what makes you really love the convenience of Python vs C/C++ :)

But I wonder what the best way to access object members from within

code is. So far my typical approach is just to make undotted copies of everything I'll need once inside C/C++:

a = x.a b = x.b

inline("do_stuff_with(a,b);", ['a','b'])

But there may be a better way to do it that I've missed. I pored over

Yep. the C the

weave tutorial and the auto-generated C++ code, but I couldn't really find any ideas for a better solution.

The attr method will work: ################################### from weave import inline class bunch: pass def oaccess(): x=bunch() x.a = 1 code = """ // Try to emulate Python's: print 'x.a',x.a std::cout << (int)x.attr("a") << (int)x.attr("a") << std::endl; """ inline(code,['x']) oaccess() The attr method returns a py::object. Notice you have to cast this value to an integer before writing it out. I pretty sure this could actually be fixed by overloading the << operator for py::object -- please do if you feel comfortable trying it. eric