GitLab

GIT_SILENT

Summary:
This adds a label field, so that devices can provide a label and a more detailled description.
This label should be used as the device label and description used for subText or tooltips for instance.
The description is used as label by default.

CCBUG: 415281

For use in D26113 and D26114

Reviewers: #frameworks, bruns, sitter

Subscribers: ervin, kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D28590

ACK'ed by dfaure
GIT_SILENT

GIT_SILENT

All unit tests pass.

Unit tests still pass.

Unit tests still pass. I had to hack CMakeLists to ensure HAL stuff
are still buildable.

Test Plan: make && ctest

Reviewers: #frameworks, dfaure, bruns, meven

Reviewed By: dfaure

Subscribers: kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D29233

Summary:
As was discussed in https://phabricator.kde.org/D29138 and on IRC (with
frinring and vkrause), add a new enumerator to MediumType with value 0
to indicate no-op. Change one for-loop to use the new enumerator with
QMap::value(key, defaultValue).

Test Plan: make && ctest

Reviewers: #frameworks, dfaure, meven, bruns

Reviewed By: dfaure

Subscribers: kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D29221

Summary:
src/imports/devices.cpp
src/solid/devices/backends/fakehw/*
src/solid/devices/backends/fstab/*

Test Plan: make && ctest

Reviewers: #frameworks, dfaure, apol, meven

Reviewed By: dfaure

Subscribers: bruns, kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D29138

KApiDox & ECMAddQch have been just teached about Q_DECLARE_FLAGS, so the
underlying typedefs are no longer skipped by doxygen, but can be
documented now, allowing links to be generated for these types e.g. when
used as arguments.

The "#" prefix to the enum name in the description text of all the
Q_DECLARE_FLAGS docs seems needed to properly trigger doxygen autolinks.

Wording of pre-existing docs for those flags has been adapted to
pattern used elsewhere in KF.

GIT_SILENT

Summary:
In case the fstab is deleted (caused by editing it), the watch will be
removed. The code already partially dealed with this, i.e. it tried to
readd the watch, but did not check if the operation succeeded and dropped
any further changes if not.

Keep a watch on the containing folder and readd the fstab on change
events. Unfortunately QFilesystemWatcher does not allow fine granular
change events, though as /etc/ is likely mostly silent this should not
matter much.

The bug becomes apparent when e.g. editing the fstab with VIM, which
has a quite broken implementation of creating temporary/backup files
during save: it moves away the original file and creates the new file
directly under the original name, instead of using a atomic rename.

Test Plan:
1. solid-hardware listen
2. open fstab with vi, save a few times
Without the change, solid will loose the watch sooner or later.

Depends on D28779

Reviewers: #frameworks, ngraham, apol

Reviewed By: ngraham, apol

Subscribers: apol, kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D28780

Summary:
The deviceAdded signal was emitted twice, once with the old device list
in m_deviceList, once with the new one.

The first one is actually wrong and was fixed in kdelibs4 by moving the
deviceAdded signal after the m_deviceList update. In KF5, the second
signal was added but the first one never removed.

Move the m_deviceList update to the top (though for deviceRemoved it
does not matter much), and remove the second signal.

Depends on D28779

Test Plan:
1. solid-hardware listen
2. add an entry to the fstab
-> deviceAdded is only emitted once

Reviewers: #frameworks, afiestas, apol

Reviewed By: apol

Subscribers: apol, kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D28782

Summary:
Depends on D28778

GIT_SILENT

Test Plan: No functional changes

Reviewers: #frameworks, apol

Reviewed By: apol

Subscribers: kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D28779

Summary:
QFilesystemWatcher::addPath takes a QString as argument, so avoid
construction of tempary QString on each call.

This also avoids a name clash between the Solid::Backends::Fstab::FSTAB
logging category and the FSTAB macro.

Test Plan: no functional changes

Reviewers: #frameworks, apol

Reviewed By: apol

Subscribers: broulik, kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D28778

Summary:
Commit a99c6136 ("Samba: Ensure to differenciate mounts sharing the
same source") only fixed the multiple mount problem for SMB share. This
can also happen for all other mounts (remounted local file systems,
NFS, ...).

Also clarify the comment.

Reviewers: #frameworks, meven, broulik, sitter

Reviewed By: meven, broulik, sitter

Subscribers: broulik, kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D28488

Summary:
`getmntent` for samba returns fsname of the form "//server/folder" (same as `mount`)
It is not sufficient to use in udi as device name as it does not differenciate mounts sharing the same source but having different mount points.

BUG: 418906
FIXED-IN: 5.69

Test Plan:
With a local cifs mount point:

Before
$ solid-hardware5 list | grep ourfiles
udi = '/org/kde/fstab///meven-synapse/ourfiles'

After
$ solid-hardware5 list | grep ourfiles
udi = '/org/kde/fstab///meven-synapse/ourfiles:/media/samba'

Reviewers: #frameworks, sitter

Reviewed By: sitter

Subscribers: ahmadsamir, anthonyfieroni, sitter, kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D28476

Summary:
Convert license headers to SPDX expressions and add
license files as required by REUSE specification.

Reviewers: cgiboudeaux

Reviewed By: cgiboudeaux

Subscribers: cgiboudeaux, kde-frameworks-devel

Tags: #frameworks

Maniphest Tasks: T11550

Differential Revision: https://phabricator.kde.org/D27742

Summary:
This way the commands syntax is always immediately visible
in the help text.
Help that doesn't help, and
previously the help would essentially just end up telling you to
run --commands to actually get help -.-

The existing syntax argument of addPositionArgument will simply
print the syntax help undearneath the usage line, thus giving
immediate access to information on how to use the application
without first having to go through the --commands argument

(--commands still exists for backwards compat)

Test Plan: -h now useful

Subscribers: broulik, bruns, kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D26841

Summary:
Reading through the code, I realized that the isRemovable check
returned true when the kDADiskDescriptionDeviceInternalKey property
is set to true. But that sounds like the check needs to be inverted:

According to [1] e.g. a disk is non-removable when it is internal.
And kDADiskDescriptionDeviceInternalKey returns whether the disk
is internal, not external.

[1]: https://stackoverflow.com/questions/38499860/thunderbolt-drives-not-marked-as-ejectable-in-disk-arbitration-iokit-although-th#comment64407405_38499860

Output from `solid-hardware5 details` and `... nonportableinfo`
on a MacBook Pro Retina, 13-inch, Mid 2014 with macOS 10.15.2,
before this patch:

```
udi = 'IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/RP06@1C,5/IOPP/SSD0@0/AppleAHCI/PRT0@0/IOAHCIDevice@0/AppleAHCIDiskDriver/IOAHCIBlockStorageDevice/IOBlockStorageDriver/APPLE SSD SM0256F Media'
  parent = 'IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/RP06@1C,5/IOPP/SSD0@0/AppleAHCI/PRT0@0/IOAHCIDevice@0/AppleAHCIDiskDriver/IOAHCIBlockStorageDevice/IOBlockStorageDriver'  (string)
  vendor = ''  (string)
  product = 'APPLE SSD SM0256F                       '  (string)
  description = 'APPLE SSD SM0256F Media'  (string)
  icon = 'drive-removable-media'  (string)
  Block.major = 1  (0x1)  (int)
  Block.minor = 0  (0x0)  (int)
  Block.device = '/dev/disk0'  (string)
  StorageAccess.accessible = false  (bool)
  StorageAccess.filePath = ''  (string)
  StorageAccess.ignored = false  (bool)
  StorageDrive.bus = 'Platform'  (0x5)  (enum)
  StorageDrive.driveType = 'HardDisk'  (0x0)  (enum)
  StorageDrive.removable = true  (bool)
  StorageDrive.hotpluggable = false  (bool)
  StorageDrive.inUse = true  (bool)
  StorageDrive.size = 251000193024  (0x3a70c70000)  (qulonglong)
  StorageVolume.ignored = false  (bool)
  StorageVolume.usage = 'PartitionTable'  (0x3)  (enum)
  StorageVolume.fsType = ''  (string)
  StorageVolume.label = ''  (string)
  StorageVolume.uuid = ''  (string)
  StorageVolume.size = 251000193024  (0x3a70c70000)  (qulonglong)

udi = 'IOService:/AppleACPIPlatformExpert/PCI0@0/AppleACPIPCI/RP06@1C,5/IOPP/SSD0@0/AppleAHCI/PRT0@0/IOAHCIDevice@0/AppleAHCIDiskDriver/IOAHCIBlockStorageDevice/IOBlockStorageDriver/APPLE SSD SM0256F Media'
  BSD Major = 1  (0x1)  (int)
  BSD Minor = 0  (0x0)  (int)
  BSD Name = 'disk0'  (string)
  BSD Unit = 0  (0x0)  (int)
  Content = 'GUID_partition_scheme'  (string)
  Content Hint = ''  (string)
  Ejectable = false  (bool)
  IOBusyInterest = 'IOCommand is not serializable'  (string)
  IOGeneralInterest = 'IOCommand is not serializable'  (string)
  IOMediaIcon = ''  (string)
  IOPolledInterfaceStack = 'IOPolledFilePollers is not serializable'  (string)
  Leaf = false  (bool)
  Open = true  (bool)
  Preferred Block Size = 512  (0x200)  (qlonglong)
  Removable = false  (bool)
  Size = 251000193024  (0x3a70c70000)  (qlonglong)
  Whole = true  (bool)
  Writable = true  (bool)
  className = 'IOMedia'  (string)
```

Note the `Ejectable = false  (bool)` vs.
`StorageDrive.removable = true  (bool)`. The patch here fixes
it to yield `StorageDrive.removable = false  (bool)`

Subscribers: kde-frameworks-devel

Tags: #frameworks

Differential Revision: https://phabricator.kde.org/D27065