When the filter view is created, the first item of the asyncFilters list
would be selected (if the source supports asyncFilters) and have its
properties shown.
This means that when no items were in the asyncFilters list, no
properties would be shown, even if an effect filter was present.
Adds a check to focus the effectFilters list in case there are only
effect filters, which makes the correct properties appear.
This reverts commit f832d14220b572165bf69ad73f30a5b961d02258.
Reverting this until we do more investigation. Currently, any text
that can be typed in a dialog requires the dialog to re-enable hotkeys
on destruction, which is frustrating. There has to be a better way.
Disable hotkeys when a user starts an interaction with the UI where they
are expected to type text and re-enable hotkeys when the interaction is
completed.
The undo/redo functions are setting the current scene. Everytime
it would do this, it would actually transition to the scene.
This forces the current scene, so it fixes a bug where the
transition would be grayed out when undoing/redoing.
Implements undo/redo for both properties and filters. Works by creating
a new callback that gets called to save undo/redo states after a timer
is fired. Also disabled undo/redo until the actions have completed to
prevent a user from being able to disrupt the stack by perfoming actions
before others have finished.
Speeds up loading with large number of filters and prevents deadlock by
not getting the properties of filters during
OBSBasicFilters::UpdateFilters() which uses the source->filter_mutex
Code submissions have continually suffered from formatting
inconsistencies that constantly have to be addressed. Using
clang-format simplifies this by making code formatting more consistent,
and allows automation of the code formatting so that maintainers can
focus more on the code itself instead of code formatting.
In commit 57f8c5e3, a stretch spacer was put in to the filter window UI,
which was meant to be used when there's no video for the source.
However, it was always put in, and would stretch the layout even when
there was video. This fixes that issue by only inserting a stretch
spacer when there's no video for the source.
Contrary to what the name would have you believe, QPointer<> is not used
to delete a pointer when it leaves its specific scope. Instead, it's
used to check to see if the pointer is still valid. For most
QWidget-based objects, this is actually fine because QWidgets that are
assigned to layouts or other widgets will automatically be destroyed --
however, for non-widget objects, this can cause a memory leak.
This patch replaces QPointer with QScopedPointer where applicable to
prevent memory leaks.
Closesobsproject/obs-studio#1367
The OBSBasicFilters window did not register a handler for the
"update_properties" signal. Now it does. Addresses issue #1028.
Closesjp9000/obs-studio#1036
In the filters dialog, when the properties are being updated due to
filters being added/removed/switched, using deleteLater on the previous
view would cause it to linger in the view until the deleteLater is
processed by the Qt queue, displaying two views instead of one.
Additionally because of that, this would also cause the display in the
filters dialog to resize as well, causing the reorder mutex hard lock
which was fixed by the parent commit.
Closesjp9000/obs-studio#714
