On Fri, Feb 22, 2008 at 01:35:33PM -0800, H. Peter Anvin wrote: > > Spinning on about protocol design: > > The boot flag (not *bootable* flag - the boot flag indicates not just > bootability, but selection) really should be orthogonal to partition > type, for obvious reasons.
Please, don't interpret this as a sign that I object to your proposed design, but I think it is departing too much from what GRUB uses to be considered the same thing. You make a valid point about orthogonality IMHO; in our "standarized" scheme, a partition could have the type for data or some filesystem and be bootable at the same time. OTOH, when it comes to GRUB, what its installer wants is a dedicated partition where it can embed its stuff. And it is likely that other bootloaders (specially those that used to rely on the 63-sector gap) will want the same thing. I admit it may have been a mistake in my part to hint that we could standarize based on this partition type -- perhaps the desired standarization requires something completely different. So, what I propose is: let's not mix them up. I have a tight schedule to make GRUB have proper support for GPT in next Debian release; I do also have a desire (which I think is shared) to standarize on an MBR implementation for GPT. With this assumption in mind, may I suggest we split the discussion? I.e. in a separate thread or so; although at this point it isn't that related to Parted, I think. We could even go as far as setting up a mailing list for this. Thoughts? -- Robert Millan <GPLv2> I know my rights; I want my phone call! <DRM> What use is a phone call… if you are unable to speak? (as seen on /.) _______________________________________________ parted-devel mailing list [email protected] http://lists.alioth.debian.org/mailman/listinfo/parted-devel
