On Wed, 2008-06-25 at 08:19 -0700, Linus Torvalds wrote:
quoted text
>=20
> On Wed, 25 Jun 2008, Johannes Berg wrote:
> >=20
> > In networking, we've gone through various incarnations of print_mac()
> > which is similar to the sym() macro Paul proposed, and it turned out to
> > be undesirable because of the way it interacts with static inlines that
> > only optionally contain code at all, the print_mac() function call is
> > still emitted by the compiler. People experimented with marking it
> > __pure but that had other problems.
>=20
> You don't even have to go that esoteric.
>=20
> Just printing things like "sector_t" or "u64" is painful, because the=20
> exact type depends on config options and/or architecture.

Heh, true, but I have the print_mac() disaster firmly imprinted in my
mind ;)

quoted text
> > It would be nice to be able to say
> >=20
> > u8 *eaddr;
> >=20
> > printk(... %M ..., eaddr);
>=20
> For special things, I do think we should extend the format more, and=20
> forget about single-character names. It would be lovely to do them as
> %[mac], %[u64], %[symbol] or similar. Because once you don't rely on gcc=20
> checking the string, you can do it.

True, that does look a lot better and has less potential for confusion.

quoted text
> The problem is that right now we absolutely _do_ rely on gcc checking the=
=20
quoted text
> string, and as such we're forced to use standard patterns, and standard=20
> patterns _only_. And that means that %M isn't an option, but also that if=
=20
quoted text
> we want symbolic names we'd have to use %p, and not some extension.
>=20
> But once you drop the 'standard patterns' requirement, I do think you=20
> should drop it _entirely_, and not just extend it with some pissant=20
> single-character unreadable mess.

Oh yes, I agree. At one point I figured it should be easy enough to
extend gcc with something that allows you to specify the format
character/type to take but alas, such a thing is not possible.

johannes