Re: Disk Device API: Some semantical questions

> > On 2003-07-08 at 10:28:25 [-0700], Ingo Weinhold wrote:
> > [...determining BDiskDevice::IsReadOnly()...]
> > > Mmh, our
> > > DriveSetup would have no way of checking whether it can work on 
> > > the
> > > device (i.e. initialize partitions or the device itself) -- other
> > > than
> > > opening the device manually and do a B_GET_DEVICE_GEOMETRY ioctl,
> > > of
> > > course.
> > 
> > Couldn't it just try to open the device read-write?
> 
> I'm not sure, what the policy for opening devices is. E.g. a read-only
> device could allow to be opened for writing as well, but may fail 
> only,
> when one actually tries to write something to it. But anyway, the 
> point
> is, that you would have to use another API to get the information.

Gotcha.

> > On 2003-07-08 at 10:28:25 [-0700], Ingo Weinhold wrote:
> > > I would at least add a BDiskDevice::DeviceFlags() (we already
> > > have BPartition::Flags()), which would comprise this and other
> > > flags.
> > 
> > Which other flags?
> 
> Like whether the media is removable and/or write once.

Okay, my bad, I read that as BDiskSystem::Flags() for some reason. 
BDiskDevice::DeviceFlags() makes sense. :-)

> > On 2003-07-08 at 14:35:43 [-0700], Ingo Weinhold wrote:
> > > On Tue, 08 Jul 2003 22:31:07 +0200 CEST "Axel Dörfler" 
> > > <axeld@pinc-
> > > > BTW my question originally aimed at the "OR" part of the first
> > > > quoted
> > > > paragraph. If it's interesting for the user if the disk system
> > > > has
> > > > write capabilities or not, it should not be ORed with
> > > > IsReadOnlyMedia()
> > > > but be separately accessible.
> > > 
> > > Mmh, I guess, I find it more convenient, if
> > > BPartition::IsReadOnly()
> > > does the OR. The information the caller is interested in, is
> > > probably
> > > whether they can modify the partition's contents. The info whether
> > > the
> > > disk system in general does support writing or not is perhaps
> > > better
> > > provided via something like BDiskSystem::IsReadOnly().
> > 
> > That's why I suggested it 4 replies ago. :-) :-P
> 
> I have the vague memory of indeed having read something like this from
> you. :-)
> 
> > At any rate, I like
> > the OR-semantics for BPartition::ReadOnly() as well.
> 
> Cool. So I won't change it.

Sounds good.

> BTW, what kind of partitions does
> BDiskDeviceRoster::VisitEachInitializablePartition() visit, i.e. what
> is an initializable partition? Something like the device is not read-
> only and there exists a disk system for which the partition's
> CanInitialize() would return true? Sounds a bit complicated.

I think something more along the lines of "the device is not read-only 
and partition->Parent()->DiskSystem()->SupportsChildSystem() returns 
true for some available disk system and there exists an available disk 
system such that 
system->SupportsParentSystem(partition->Parent()->DiskSystem()) returns 
true", again pretending BPartition::DiskSystem() exists instead of 
BPartition::GetDiskSystem(). I guess the question is how useful it is. 
If we want to give a complete overview in DriveSetup of which partitions 
are initializable, we'll have to do something similar anyway. If we want 
to make the user check on a partition by partition basis if a given 
partition is initializable (i.e. by right-clicking and seeing if an 
"initialize" item is avaliable or something similar), then the 
Visit...() function may not be all that necessary. I kind of like the 
first ides better though.

-Tyler