[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: New release

From: Sven Luther
Subject: Re: New release
Date: Tue, 20 Dec 2005 18:13:24 +0100
User-agent: Mutt/1.5.9i

On Tue, Dec 20, 2005 at 01:45:55PM +0100, Patrick Leslie Polzer wrote:
> Hello Sven,
> On Fri, 2 Dec 2005 21:09:42 +0100
> Sven Luther <Sven Luther <address@hidden>> wrote:
>  | I would include there the code for more flexible flags and partition types
>  | and generic handling of values of many kind in partition tables, since this
>  | will break ABI i believe.
> Can you explain your plans in more detail?

Basically, right now we have just flags, which can be compared to a list of
boolean values, which can be set to on and off. 

These flags are abused to set the system type, which behave like mutually
exclusive boolean values, but are in reality just plain values, 8-bit values
for the MBR partition type.

These values can be of many different types, like the amiga/rdb 32bit
partition type, or numeric boot priority, or even the mac string partition
type, or in general, the partition names for those partition tables which
support them.

Right now, we have only boolean flags, and more to the point, we have them
shared by all partition tables, which doesn't necessarily make sense.

So instead of each partition table format exporting a list of boolean flags
supported, it would export a more rich list of supported mode, maybe using a
name/type list or something. Each field would have a name, a type of field
(string, numeric, bitfield, ...) and a size, and there would be functions to
get/set those values.

The actual flags would simply be implemented on top of this as a bitfield
entry, and the system type thingy would also be implemented on top of this as
a field of the right value.

This would have the advantage of doing away with the ugliness of
LVM|RAID|PREP|PALO exclusive flags for example, and allow more easy
manipulation of different partition table attributes.

Another benefit would be for the amiga-filesystems to be able to more easily
handle filesystem information which is found in their partition table entry,
like the block size and such, altough that is another issue.


Sven Luther

reply via email to

[Prev in Thread] Current Thread [Next in Thread]