Support slice operator for array types

[septatrix]

As the title already says, it would be useful to allow code like for element in ringbuffer[head:tail]: do_stuff(element). Currently one has to wrap the object with list(foobar)[start:end] but I see no reason why this shouldn't be natively supported.

If you think this is a great idea and could point me to where this would need to be implemented I would be happy to do so. My initial idea how this could should work was the following

foobar # is Object(prog, 'struct foobar [128]', address=0x100)
foobar[4:8] # is Object(prog, 'struct foobar [4]', address=0x100 + 4*sizeof(foobar))

[osandov]

I love this idea. I'd like for it to also work with pointers. E.g.,

ptr = Object(prog, "int *", 0x1234)
ptr[4:8] # is Object(prog, "int [4]", address=0x1234 + 4 * sizeof(int)

The main implementation of this would be in C in libdrgn. Something like:

struct drgn_error *drgn_object_slice(struct drgn_object *res,
				     const struct drgn_object *obj,
				     int64_t index, uint64_t length);

See the similar functions in libdrgn/object.c, especially drgn_object_subscript().

The second part of this would be the Python bindings. DrgnObject_subscript() would need to be updated to handle a slice argument and call either drgn_object_subscript() or drgn_object_slice() as appropriate. The standard list type is a good example: https://github.com/python/cpython/blob/7c98b0674daa3e4eb3e8f35afb61a0dba61d1780/Objects/listobject.c#L3572-L3598

Let me know if you're still interested in working on this and I'd be happy to provide more details.

[septatrix]

Let me know if you're still interested in working on this and I'd be happy to provide more details.

I did not realize this would require writing CPython extension code which I have never done so far. I'd be up for the challenge but in that case I can provide absolutely no time guarantees when I will get to it as I have more urgent things on my table at the moment

[osandov]

No worries. Feel free to comment here if you start actively working on it. I may get to it at some point otherwise.

[septatrix]

I think this might also help with #377