Re: cannot boot XP after shrinking partition

[Andrew Clausen]

On Tue, Sep 07, 2004 at 09:57:07AM +0200, Szakacsits Szabolcs wrote:
> > There was never such a bug in Parted. 
> 
> I can't confirm or deny, and currently don't have the time to check it
> out. But it's a fact several tools using libparted has the problem.

I'm only aware of this flaw in QtParted.  I guess I should maintain
a list of tools that use libparted so I can check them.  But, can
you tell me which tools using libparted have this problem?

> > This was only in QtParted - it didn't tell libparted exactly where to
> > put the partition, so Parted aligned it elsewhere.
> 
> At least SUSE's Yast has the same problem.

Yast just invokes parted on the command line.  What makes you think
Yast has this problem?

> The tools don't want to put the partition start anywhere and apparently
> they expect to be unchanged unless told otherwise. But it seems libparted
> also expects to be told that partition start shouldn't be changed
> otherwise it aligns.

Well, if you pass "ped_constraint_any()" to libparted, you're telling
libparted you don't care.  I think this aspect of the libparted API is
straightforward and unambiguous - I don't understand why several
implementors don't understand what a constraint is.  Perhaps if
ped_disk_set_partition_geom() returned the PedGeometry* actually set,
then implementors would get that what goes in isn't necessarily what
comes out.

Cheers,
Andrew