[parted-devel] Patch: Be more conservative about modifying the disk

Joel Granados jgranado at redhat.com
Wed Dec 3 09:26:09 UTC 2008


Hey Mathew:

Yep I get your point.  The solution is to ask the user when in interactive mode.  Default to fixing it (This means when you just press enter), and do nothing when ctrl-c is the input.  Additionally default to fixing the erroneous partition table when in scripting mode (We assume that you know what you are doing when in scripting mode.  yes, this assumes you know about this gpt issue).

I'll see what I can do about my commit access and get back to a whole bunch of issues that have sprung up during the past months.

Regards. 
----- "Matthew S. Harris" <mharris312 at gmail.com> wrote:

> Any more word on this patch?  It's purely a UI issue; you can't
> automatically fix the problem in all cases, so we need to check with
> the user before potentially doing damage.  I know you're bothered
> that
> a user in this situation may have to tell Parted to ignore the
> problem, but certainly that's better than blindly overwriting the
> user's disk even when the user presses ctrl-c.
> 
> If you don't want to take the change, I won't press any more.
> Obviously the situation doesn't happen often in real life or you
> would
> have heard from more people by now.  I just don't want this fix to
> fall through the cracks if it can save someone else from what I went
> through.
> 
> 
> Matthew
> 
> On Wed, Nov 19, 2008 at 09:32, Matthew S. Harris
> <mharris312 at gmail.com> wrote:
> > On Wed, Nov 19, 2008 at 02:47, Joel Granados <jgranado at redhat.com>
> wrote:
> >> You have a point here.  What still bothers me is that the user will
> constantly have to tell parted.  "yes, I know that the gpt table is
> not where its suppose to be, ignore please"
> >
> > I don't have a problem with that at all.  Most users will never see
> > the message.  I was in an apparently unusual situation (since
> nobody
> > else has reported this problem in the past 18 months), and any user
> > would be much happier saying "ignore, please" than digging into the
> > internals of GPT and editing their copy of Parted.
> >
> >> I confess that I could not think of a good solution for this issue
> of the top of my head.  The only thing I could think of is to "fix"
> the position of the table without erasing the previous one.  But then
> you would have problems with the synchronization of the three.  You
> may also have an issue of overlapping lists.
> >
> > Yeah, there's no transparent fix you can apply here; if the other
> OS
> > really sees the partition differently, any writing you do at the
> end
> > of it could corrupt its data.  We just need to fix the UI so that
> the
> > user doesn't get into the situation I was in: Parted says, "the
> > primary GPT table is corrupt," I think "no, it's not, it's a bug in
> > Parted, let me ctrl-c so Parted doesn't try to fix it" ... and then
> > Parted does anyway.
> >
> > The situation is unfortunate, but I think this UI change really
> > doesn't have a downside.
> >
> >
> > Matthew
> >



More information about the parted-devel mailing list