[Cc'ing the bug, not the lists.]

On Thu, 2010-04-01 at 02:00 +0100, Justin B Rye wrote:
> Ben Hutchings wrote:
> > On Wed, 2010-03-31 at 11:45 +0100, Justin B Rye wrote:
> >> In my case the name of my IOmega Zip drive changed too.  Yes, I
> >> only had it installed on that machine to see if it would cause
> >> trouble, and it still worked as /dev/sdc1.  Mind you, I imagine it
> >> would be a bit of a pain assigning labels to a pile of 100MB
> >> removable zip-disks if dosfslabel's still buggy (#506786).
> > [...]
> > 
> > Also reported and fixed as #559985.  You're welcome.
> 
> If it's fixed in 3.0.7-1, shouldn't dosfslabel agree with mlabel,
> blkid, mount, etc on what label a vfat file system has?

If the two labels (boot sector and root directory) are inconsistent then
different tools might reasonably disagree.

[...]
> I can use mlabel or mkdosfs to set a label that can then immediately
> be used by mount, but dosfslabel still seems to be modifying some
> ID-string not connected to the file system label or UUID.
> 
> (I should probably be testing this on something less arcane.)

mkdosfs and dosfslabel now write to the same two labels.  I didn't test
mlabel, but a quick inspection suggests that it also sets both.

Note that there is no automatic mechanism to update the link in
/dev/disk/by-label or the blkid cache in /etc/blkid.tab when a disk is
relabelled.  Lookup by label will be unreliable until these are also
updated.

If you can come up with a test case where dosfslabel is still doing the
wrong thing (starting from a blank disk, like my test case in #559985)
then I will try to fix it.

Ben.

-- 
Ben Hutchings
Once a job is fouled up, anything done to improve it makes it worse.

Attachment: signature.asc
Description: This is a digitally signed message part

Reply via email to