On 20 February 2015 at 16:19, Brian C. Lane <[email protected]> wrote: > What do you expect to happen though? You've explicitly told parted to > remove the 'partition' and that's what's happened. The fact that ext4 > survives is more of an accident than anything else. I also don't really > see the utility of trying to use parted like this -- just make a real > disk label and use that.
I'm coding GParted to support whole disk file systems and as it uses libparted to do the disk querying and manipulation operations, how libparted handles loop partitions mattered. It seems common to put file systems into Linux Software RAID arrays without further partitioning them. Additionally LVM is sometimes put on whole disks without partitioning them first too. These are cases where libparted does (or should, but it doesn't recognise LVM PVs) present loop partition tables. GParted then allows the user to graphically create, resize, reformat, delete those whole disk partitions. Ultimately to make GParted behave consistently over several versions of libparted and provide a full capabilities to the users I wasn't able to rely of libparted to work with loop partition tables other than to recognise them. As such I don't actually need libparted to behave consistently in this case. I saw what I though was a remaining inconsistency after Phillip's work to improve loop partition handling in parted 3.2. That for every file system other than XFS, deletion of the virtual partition on a loop table is a no-operation where as for XFS it damages it beyond recognition. I reasoned that making deletion of virtual partition on a loop table should be made a no-operation. But as I said above it doesn't actually matter to me any more. Mike
