GParted doesn't notice when a file system label is changed to blank.
GParted first calls the file system specific read_label() method. When
the label is blank read_label() correctly sets partition.label to the
zero length string. Second GParted_Core::set_device_partitions() treats
the zero length string to mean that the label is unset and calls
FS_Info::get_label() to retrieve it from the cache of blkid output.
Blkid also doesn't notice when the file system label has been changed to
blank so reports the previous label. Hence GParted displays the
previous file system label.
Fix by making label a private member variable of the class Partition and
providing access methods set_label(), get_label() and label_known()
which track whether the label has been set or not. This only fixes the
fault for file systems which use file system specific commands to read
the label and when these tools are installed. Otherwise GParted uses,
or has to fall back on using, the buggy blkid command to read the file
system label.
NOTE:
Many of the file system specific read_label() methods use a tool which
outputs more than just the label and use Utils::regexp_label() to match
leading text and the label itself. If the surrounding text changes or
disappears altogether to indicated a blank label, regexp_label() doesn't
match anything and returns the zero length string. This is exactly
what is required and is passed to set_label() to set the label to blank.
Bug 685656 - GParted doesn't notice when file system label is changed to
blank
Now that every call to calc_usage_triple() just passes usage figures
returned by get_sectors_*(), remove those parameters, call
get_sectors_*() internally and rename to get_usage_triple().
For specific partition usage values the right hand border of the
partition graphic in the Information dialog would be displayed as grey
rather than the color assigned to the partition.
Steps to reproduce fault:
Create 1024 MiB partition
# lvm pvcreate /dev/sda12
# lvm vgcreate GParted-VG1 /dev/sda12
View partition information
Fragment from Dialog_Partition_Info::init_drawingarea():
139 else if ( partition .sector_usage_known() )
140 {
141 used = Utils::round( ( 400 - BORDER *2 ) / ( dlength / partition .get_sectors_used() ) ) ;
142 unused = Utils::round( ( 400 - BORDER *2 ) / ( dlength / partition .get_sectors_unused() ) ) ;
143 unallocated = 400 - BORDER *2 - used - unused ;
144 }
For this issue the above values are both exactly x.5 and both round
upwards, resulting in unallocated being -1.
used = round((400 - 8*2)/(2097152.0/8192)) = round(1.5)
unused = round((400 - 8*2)/(2097152.0/2088960)) = round(382.5)
unallocated = 400 - 8*2 - 2 - 383 = -1
The simple fix would be to use floor() instead of round() in the
calculation of either used or unused. The same fix would also need to
be applied in Display_Info() for the calculation of the percentage
figures. Unfortunately this simple fix can lead to odd figures when the
used or unused is close to zero and floor() or ceil() is effectively
applied rather than round(). For example:
Size: 227.23 GiB
Used: 28.00 KiB ( 1% )
Unused: 180.00 GiB ( 79% )
Unallocated: 47.23 GiB ( 20% )
Used figure of 28 KiB in 227 GiB partition should be rounded to 0% but
wasn't.
Write Partition::calc_usage_triple() which calculates the "best" figures
by rounding the smaller two figures and subtracts them from the desired
total for the largest figure. Apply to the calculation of the partition
usage percentage figures in the Information dialog and the partition
usage graphic in the same dialog and the main window.
Bug #499202 - gparted does not see the difference if partition size
differs from filesystem size
Most file systems report intrinsic unallocated space using the statvfs()
system call when mounted, but not using their own tools. They are:
ext2/3/4, fat16/32, hfs, nilfs2, reiserfs and xfs. Showing either a
little or no unallocated space, depending on whether a file system is
mounted or not, could be confusing to the user.
When all file systems are created filling their partitions the unused
figure reported by statvfs() and their own tools are the same or very
close. Also the used plus unallocated figure from statvfs() agrees with
the used figure from their own tools.
For all file systems don't display intrinsic unallocated space (that
below the threshold of 2 to 5%), instead include it as used space. As
soon as the amount of unallocated space becomes significant display it
everywhere and also trigger the warning.
For display purposes always use the new Partition methods:
get_sectors_used(), get_sectors_unused(), and get_sectors_unallocated().
When calculating new usage figures during Paste and Resize/Move
operations directly access sectors_used, sectors_unused and
sectors_unallocated members.
Bug #499202 - gparted does not see the difference if partition size
differs from filesystem size
A number of file systems report intrinsic unallocated space even when
they are created filling the partition. As reported using their own
specific tools, they are: jfs, lvm2 pv and ntfs. Therefore when
resizing a partition estimate its minimum size to be used sectors plus
any unallocated sectors up to the significant amount.
Bug #499202 - gparted does not see the difference if partition size
differs from filesystem size
When pasting a copied partition into free space or move/resizing a
partition set its space utilisation so that any unallocated space within
the partition is displayed correctly before the operation is applied.
NOTE:
If the file system does not support file system resizing the Paste and
Move/Resize dialogs don't allow resizing the partition so the preview
will always be correct, unlike the case in the previous patch:
Set unallocated space when performing simple operations (#499202)
Also remove the deprecated and no longer used Partition::Set_Unused()
and Partition::set_used() methods.
Bug #499202 - gparted does not see the difference if partition size
differs from filesystem size
Currently GParted assumes that a file system fills its containing
partition. This is not always true and can occur when resizing is
performed outside of GParted or a resize operation fails. GParted
doesn't display any information about unallocated space to the user
and in most cases it is simply included in used space.
Add partition unallocated space accounting. Make GParted record the
unallocated space for mounted file system and display a warning in the
Partition Information dialog when too much unallocated space is found.
Partition::set_sector_usage( fs_size, fs_unused ), is the new preferred
method of recording file system usage because it allows the unallocated
space in a partition to be calculated. Partition::Set_Unused() and
Partition::set_used() are now deprecated.
NOTES:
1) Set the minimum unallocated space to be 5% before considering it
significant to avoid false reporting. Worst case found was a
mounted xfs file system in a 100MiB partition, which reports as
~4.7% unallocated according to file system size from statvfs().
However, it reports as having no unallocated space using xfs
specific tools.
2) Unallocated space is only a graphical representation for the user.
GParted must still use relevant tools to resize file systems before
shrinking the data and can't assume all unallocated space exists
after the file system at the end of the partition.
Bug #499202 - gparted does not see the difference if partition size
differs from filesystem size
As the Mount Point column is being borrowed to display the PV's VGNAME,
also suppress generation of the "Mount on" submenu for LVM2 PVs.
Bug #160787 - lvm support
Make align to MiB the default setting instead of align to cylinder.
Migrate logic for alignment to cylinder into its own method
snap_to_cylinder, and place common logic in snap_to_alignment.
Add alignment checks for situations where space is needed for Master
Boot Record or Extended Boot Record.
Adjust ranges on spin buttons according to required boot record space.
Copy fix for off by one sector (#596552) from
Dialog_Partition_New::Get_New_Partition to
Dialog_Base_Partition::Get_New_Partition
Enhance resize / move logic for checking locations of nearby logical
partitions to not depend on the partition ordering.
Note: This commit does not include limiting graphic movement according
to required boot record space.
Also add signal handler to alignment menu to update file system
minimum size.
This enhancement is to prepare for adding a third alignment
option to align to MiB.
The device sector size is needed in all calculations that convert
between sectors and bytes. The device sector_size is included in
the partition object because this object is used to contain
operation information in addition to actual partitions and
unallocated space. A second option was considered to pass the
device object to many methods, but this was considered a much
larger task that might not provide significant gains in
maintainability.
Restore copyright entries by original author to those of his last
known repository commit titled "released gparted-0.3.4 on
LarryT's request." on Feb 25, 2007.
Add my own copyright entries for files in which I changed source
code. Files in which I only made spelling changes do not have my
copyright entry added.
* added support for reading volumelabels. Atm we only read ext2/3, but
the infrastructure for adding the other filesystems is in place.
It's simply a matter of finding the right commands and parsing the
output. (see #338528 for more info)
* include/Partition.h,
src/Partition.cc: added test_overlap()
* include/GParted_Core.h,
src/GParted_Core.cc: perform a readonly testrun before the actual
move if destination overlaps source.
* improved errorhandling a bit. At the initialscan we store
errors/warnings now in a list per partition and show the in the
partitioninfo dialog.
While executing an operation we collect all libparted exceptions in
a list and attach this list to the operationdetails when everything
is done.
* include/Partition.h,
src/Partition.cc: added set_used() and operator!=
* src/Dialog_Partition_Copy.cc,
src/GParted_Core.cc,
src/OperationCopy.cc,
src/Win_GParted.cc: it's now possible to copy from partition to
partition. (before this it was only possible to copy to unallocated
space)
* include/DrawingAreaVisualDisk.h,
src/DrawingAreaVisualDisk.cc: another change to highlighting of
selected partition.
* include/Partition.h,
src/Win_GParted.cc: deal properly with just formatted partitions
* src/Operation.cc: added FIXME
* changed the way devices and partitions store their devicepaths.
Instead of holding a 'realpath' and a symbolic path we store paths
in a list. This allows for improved detection of mountpoins, free
space, etc..
Also fixed a nasty bug which showed up when you copy a partition
from one device to another. (thanks to mario for the report)
* in some places i still used MiB's instead of sectors to store sizes.
this has been fixed everywhere. Only the spinbuttons still use
MiB's. I have a few ideas on how to solve this, but i'll take it up
with #usability first.
* include/GParted_Core.h,
include/Partition.h,
src/Dialog_Partition_Info.cc,
src/GParted_Core.cc,
src/Partition.cc,
src/TreeView_Detail.cc: store flags in a list instead of a string.
* Makefile.am,
src/Makefile.am: cleanups
* include/GParted_Core.h,
include/Partition.h,
src/Dialog_Partition_Info.cc,
src/GParted_Core.cc,
src/Partition.cc,
src/TreeView_Detail.cc,
src/Win_GParted.cc: support partitions with multiple mountpoints
(see also #330327)
* include/Device.h,
src/Device.cc: replaced 'device_partitions' with 'partitions'
* src/Win_GParted.cc: use new Device::partitions
* include/Partition.h,
src/Partition.cc: added 'realpath'
* src/Dialog_Partition_Info.cc: use Partition::realpath..
* include/GParted_Core.h,
src/GParted_Core.cc:use ped_partition_get_path() instead of
constructing it ourselves. (#325800)
Also use global maps for mountpoints and shortpaths to prevent
fileaccess from happening more then once per scan.
* include/Partition.h,
src/Partition.cc: added get_length()
* include/Utils.h,
src/Utils.cc,
src/Dialog_Partition_Info.cc,
src/Operation.cc,
src/TreeView_Detail.cc,
src/VBox_VisualDisk.cc,
src/Win_GParted.cc: from now on values >=1024MB are displayed in
GB's (#319840)
* include/Partition.h,
src/Partition.cc: overloaded operator ==
* include/VBox_VisualDisk.h,
src/VBox_VisualDisk.cc: rewrote the graphical partitiondisplay. It
now supports resizing and the extended partition is also selectable.
see also #312656
* include/Win_GParted.h,
src/Win_GParted.cc: use the new graphical partitiondisplay.
* src/TreeView_Detail.cc: minor cleanups
* include/Partition.h,
src/Partition.cc: added 'Glib::ustring mountpoint'
* include/GParted_Core.h,
src/GParted_Core.cc: implemented set_mountpoints() to set mountpoint
in partitions.
* include/Dialog_Partition_Info.h,
src/Dialog_Partition_Info.cc: use Partition::mountpoint instead of
finding it ourselves.
* include/Partition.h,
src/Partition.cc: from now on each partition has a reference to it's
device.
* src/Win_GParted.cc,
src/Dialog_Partition_New.cc: make use of new Set() from partition
* include/GParted_Core.h,
src/GParted_Core.cc: since ped_file_system_clobber() ignored reiser4
i've implemented a custom 'filesystem signature remover'.
Also made use of the fact a partition now knows on which device it
lives.
* src/Operation.cc: the get_string() didn't use
Get_Filesystem_String() and thus returned faulty strings. Also did
some cleanups.
* Again way too many chances to create a detailed entry (i'm glad i'm the only dev atm :P ).
Resizing of ext2/3 works perfect now. I've even tested it on the partition holding my SG seasons =)
Implemented checking of filesystems (only internally used atm).
Done some overall tweaking, finetuning etc.. release 0.0.7 is getting shape.
* instead of listing all partition in one list, logical partitions are now stored in a sublist in extended partition object.
This makes partitionhandling in gparted more natural and transparant. It also allowed me to clean up this ugly Operation class ;)
* configure.in: added checks for libuuid and libdl. Also refined libparted check a bit.
* include/Partition.h,
src/Partition.cc: removed Get_Color()
* include/Utils.h: added inline Glib::ustring Get_Color( const Glib::ustring & filesystem )
* src/Dialog_Partition_New.cc,
src/Win_GParted.cc: make use of Get_Color from Utils.h
* src/Device.cc: fixed a crasher with (at least) pl_PL locale.
* include/Device.h,
src/Device.cc: made 'deep-scan' optional in Read_Disk_Layout. Removed obsolete (already) commented function.
* include/Partition.h,
src/Partition.cc: Took used sectors out of Set() and added a Set_Used() function instead. Also added Reset() .
* src/Dialog_Partition_New.cc: minor change to make use of the new Set() in Partition..
* include/Win_GParted.h,
src/Win_GParted.cc: show pulsing progressbar while refreshing devices.
Mike Fleetwood