| Age | Commit message (Collapse) | Author |
|
This commit brings the main view of Dolphin into a usable state
accessibility-wise. Users of screen readers should have a way better
experience while browsing files and folders and navigating along the
file system hierarchy.
This commit fixes most of the remaining already-identified
accessibility issues listed in
https://invent.kde.org/teams/accessibility/collaboration/-/issues/28,
but not all. Namely, these should now be fixed:
1. Orca should read the element type in dolphin (file, folder, device,
link to folder, link to file)
2. Orca should read complete label in icon and compact view mode,
currently it only speaks the name, but there could be additional
information like the number of elements or the file size.
3. Orca is not able to announce Selecting / Unselecting files in
Dolphin. It also never announces how many items are selected in total.
(Announcing the total selection can be done by reading out the view
element or by pressing the Tab key to get to the status bar with the
relevant information.)
4. Dolphin opens on the home directory, but Orca doesn't tell you so.
Consider enclosing the area in a frame/panel which updates its
accessible name each time you modify the current path by entering or
leaving a directory.
5. I don't know what the folder presentation widget is, but it should
be presented as a grid view. Currently, we have a terrible experience
because the entire row of folders is read at once, with no indication
that we can move left and right with the arrows to go between the
elements of a row. When I found that out, however, I discovered that
when you're on the last icon of the first row and press right arrow,
you get to the first icon of the next row, but that's not announced,
instead, the whole row is announced at once
6. Orca should announce the current elements instead of "layered pane"
when the Folder / File view gets the focus in dolphin
7. Orca reads only name in Table View only of Dolphin
8. Items are sometimes confusingly announced as "collapsed" in contexts
in which there is no concept of collapsing/expanding e.g. in icon view
mode.
A lot of code was moved around and renamed. The three accessibility
classes, which all used to be in the same file, are moved into separate
files.
*Acknowledgement*
Thanks to Christian Hempfling and bgt lover for testing as well as
originally identifying a lot of the pain points being addressed here.
This work is part of a my project funded through the NGI0 Entrust Fund,
a fund established by NLnet with financial support from the European
Commission's Next Generation Internet programme, under the aegis of DG
Communications Networks, Content and Technology.
https://kde.org/announcements/2024_ngi_openletter/
|
|
|
|
Before this commit, KItemListViewAccessible would always return
nullptr as its parent. This meant that accessibility software would
have to guess to which window/hierarchy the KItemListView belongs
to. Guessing shouldn't be necessary here.
This commit makes sure that the KItemListView always returns a
sensible parent in the accessible hierarchy. It does so by
explicitly setting the accessible parent for every KItemListView
after construction in the DolphinView contructor. Since
KItemListView now always knows about its accessible parent, the
accessibleInterfaceFactory can always ask the KItemListView for
that information when constructing the QAccessibleInterfaces.
Fixes https://invent.kde.org/system/dolphin/-/issues/47.
|
|
|
|
|
|
Use `value_or(-1)` for those functions that don't use `std::optional`.
|
|
|
|
Unfortunately licensedigger does not strip the trailing * characters.
While at it, use a common style for all source files.
|
|
GIT_SILENT
|
|
Summary:
Fix crash during shutdown. The root cause is that when Dolphin in stopped as
part of an activity, the KItemListViewAccessible destructor is called after
QApplication::exec has returned causing Qt to already having cleaned up the
QAccessibleInterface instances kept in KItemListViewAccessible. Instead of
storing the pointers to QAccessibleInterface we store the QAccessible::Id so
that we can use the QAccessible::deleteAccessibleInterface function for
deleting the instances.
BUG: 402784
Test Plan:
I wasn't able to reproduce the crash in the first place, but I have just
opened and closed Dolphin a few times and verified the the QAccessibleInterface
instances are correctly cleaned up.
Reviewers: #dolphin, elvisangelaccio, ngraham
Reviewed By: #dolphin, elvisangelaccio
Subscribers: kfm-devel
Tags: #dolphin
Differential Revision: https://phabricator.kde.org/D19083
|
|
Summary: The KItemListViewAccessible class has a list of QAccessibleInterface pointers in a member variable m_cells. The problem is that when new entries are created, the newly allocated pointer is not stored in the list, only a nullptr is store, this renders the cleanup code in the destructor useless. This patch simply stores the pointer in the list, causing the destructor to correctly free the memory.
Test Plan: I found this issue using address sanitizer. Simply building Dolphin with -fsanitize=address and opening a window caused the memory leak.
Reviewers: #dolphin, jtamate, elvisangelaccio
Reviewed By: #dolphin, jtamate, elvisangelaccio
Subscribers: elvisangelaccio, kfm-devel
Tags: #dolphin
Differential Revision: https://phabricator.kde.org/D14168
|
|
Summary: I used CLion inspection to hunt all unused #include
Reviewers: #dolphin, elvisangelaccio, markg
Reviewed By: #dolphin, elvisangelaccio, markg
Subscribers: bcooksley, markg, elvisangelaccio, #dolphin
Differential Revision: https://phabricator.kde.org/D10985
|
|
Summary: I used CLion inspection to hunt all unused #include
Reviewers: #dolphin, elvisangelaccio, markg
Reviewed By: #dolphin, elvisangelaccio, markg
Subscribers: markg, elvisangelaccio, #dolphin
Differential Revision: https://phabricator.kde.org/D10985
|
|
|
|
BUG: 359738
FIXED-IN: 16.04.0
REVIEW: 127397
|
|
REVIEW: 126771
|
|
|
|
|
|
|
|
REVIEW: 119428
|
|
Since Dolphin 2.0, we have stored the selected items in a QSet<int>,
which is neither space-efficient nor particularly fast when inserting
many items which are in a consecutive range.
This commit replaces the QSet<int> by a new class "KItemSet", which
stores the items in a sorted list of ranges. For each range, we only
store the first index and the length of the range, so we need a lot
less memory for most common selection patterns, and we also save quite
a few CPU cycles in many situations, because adding an item to the
KItemSet will in many cases not need a memory allocation at all, and
it's particularly easy when inserting sorted items into the KItemSet in
a row.
KItemSet contains a minimal subset of QSet's API which makes it
suitable as a drop-in replacement for our needs. It also has iterators,
such that the items can be iterated through easily, also with foreach.
One advantage of KItemSet compared to QSet<int> is that the items are
always iterated through in ascending order.
REVIEW: 113488
|
|
BUG: 316285
FIXED-IN: 4.10.3
|
|
|
|
Added the license header in kitemlistviewaccessible.*
Accessibility headers not included when QT_NO_ACCESSIBILITY defined
Removed unnecessary debug statement
|
|
CCMAIL: [email protected]
CCMAIL: [email protected]
|
|
REVIEW : 106555
|
|
|
|
|
|
Row and column are 0 based, the cell function
uses 1 based indices.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Added the calls to updateAccessibility, for DragDrop & Focus scenarios
|
|
|
|
Remove comparison of QAccessibleInterfaces - these are pointers that get
instantiated again and again, so they cannot be compared. Use the object()
function in QAI instead for comparison.
|
|
|
|
This leads to crashes since it implies that
casting to QAccessible2Interface is valid.
|
|
|
|
|
|
The base class implementations are actually good and without some problems.
Save one pointer's worth of memory by not having m_container.
rect was returning a relative rectangle always.
State was wrongly returning HasInvokeExtension which can lead to crashes.
|
|
And also bad use of "" instead of QString().
|
|
|
|
|
|
|