> > Well, the reason I was in favor of using Parted was to have a single clean
> > code base for the partition reading, and preferably to use Zthe canonical
> > sanctioned GNU thing if technical concerns permit.
> 
> What is the canonical sanctioned GNU thing?

I was referring to Parted.  Since it has GNU in front of the name and there
is AFAIK nothing else so named that implements partitioning, I take it to
be that.

> > The over-fanciness of Parted's library might be completely appropriate to
> > use in a "partitionfs" as we have discussed before.
> 
> I am not sure that I follow this.  Partitionfs, as currently
> implemented, reads the specified partition table and (effectively)
> generates an array of stores.  The only difference, with respect to the
> partition store type, is that the store proceeds to index the array and
> discard everything else.  Thus, as far as I can see, we want to use to
> the same code as, one, this guarantees that the partition naming scheme
> is the same for both interfaces and, two, we are not actually demanding
> anything additional from partitionfs; or do you have other plans?

My point was that while libstore should never do anything more than what
you just mentioned, it would be fine and dandy to have a much fancier
partitionfs that had did more things (like change the partitions in
response to filesystem operations).  

_______________________________________________
Bug-hurd mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/bug-hurd

Reply via email to