Re: GPT name overflow

[Andrew Clausen]

On Sun, Mar 10, 2002 at 02:40:02PM -0700, Andreas Dilger wrote:
> > So, how does it add "useful" names?  Is it 1-1 with each number?
> > Is there a different name for each partition table type?
> 
> It is _mostly_ a 1-1 mapping of partition type numbers to names, but not
> strictly so.  For example, there is a "linux" type for each disk type
> which maps to the correct numeric value/UUID/name (for Mac).  In some
> cases, there are multiple names which map to the same type (it picks the
> first name matching a number on number->name lookups).  For example, on
> GPT disks the "linux" and "data" types both map to the same UUID.

Ah, ok :)

> There is also the facility to use "names" of the form "type-<value>" in
> case a given value is not supported (I don't have names for _all_ of the
> MSDOS types).  This at least will give you _some_ useful output instead
> of "unknown", and allows you to input unsopported types if you really
> need it.  However, that is a last resort.
> 
> Patch is attached so you can look at it.

Looks good :)

I guess it can't do any harm... I should just discourage people from
using it as the default way of doing things.

So, how should parted (the front end) interface to it?

Perhaps:
* there should be a "print partition" command that prints all kinds of
stuff about the partition, like how much it can be resized, etc.
Included in this should be the partition type.
* a command analogous to "name" for setting the type.

Thanks,
Andrew