Re: PATCH: type safe(r) list_entry repacement: generic_out_cast

From: Linus Torvalds (torvalds@transmeta.com)
Date: Tue Jul 23 2002 - 11:29:27 EST


Hmm, I don't disagree, but "out_cast" reads like "outcast" to me, which
has a totally different meaning than the one you want to imply.

I'd rather call it "container_struct()" or something like that, which at
least to me sounds more natural.

The difference is that in "out_cast()" you try to describe the operation
you're doing. Which is to me not very important or interesting (never mind
the confusion with a real word meaning something totally different).

In contrast, "container_struct()" tries to describe what the _result_ is,
not how we got there. And I think that's the much more important part,
especially when reading the code. You don't care how you get the result,
but you do care abotu what the result actually _means_.

And no, I'm not married to the "container_struct()" name. But any name we
come up with should have that kind of flavor to it, I feel. I think you
can more easily explain something like

        #define to_pci_dev(n) \
                container_struct(struct device, n, struct pci_dev, dev)

by telling somebody "the 'struct device' is contained within the 'struct
pci_dev', and 'to_pci_dev' converts from 'struct device' to the
container". You can explain it at a _conceptual_ level without having to
worry about what the implementation is. And that kind of conceptual notion
is always good.

While in contrast, to explain "out_cast()" you're already starting off at
a low-level compiler implementation level.

Maybe "member_to_container()" would be even better?

                Linus

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Tue Jul 23 2002 - 22:00:43 EST