| Age | Commit message (Collapse) | Author |
|---|
|
Have it their way then.
|
|
Previously crossing the 180 degree boundary got treated as a
large rotation. This is of course incorrect.
The error in fact lies only in the filter. Special-case the sign
changes.
We're centering prior to filtering in logic/tracker.cpp so it's
all right in the world.
Issue: #533
Reported-by: @Ounicron
Fixes: #533
|
|
|
|
|
|
Having yaw and pitch as a tuple, let N be the granularity. We're now only
allowing one sample per the granularity level.
Granularity -- "spacing_in_degrees" has a value of 3 degrees. For now the
values must be integral.
Since we're only allowing (yaw, pitch) tuples of given granularity, the
following get treated as distinct:
(0; 0), (0; 3), (0; 6), (1; 42), (3; 3)
The tuple value order can be swapped. There's nothing significant as for what's
pitch and what's yaw.
We drop the remainder between the yaw/pitch value so (0, 0) is index 0, (0; N)
is index 1, (0; 2N) index 3, etc.
This should prevent the calibration function from biasing itself when the user
keeps still during the procedure.
|
|
|
|
- cap->isOpened() got its return value flipped in the conditional.
- the stop(); return false block was in the wrong place
|
|
The setting is saved, but ignored other than when dealing with
the mapping dialog. This breaks the functionality prior to
manual open of the mapping dialog.
This may not be the cause of #568 but is related.
|
|
Sadly the keys won't pass through the application anymore.
Issue: #570
Closes: #570
|
|
"color" is declared in the same block anyway.
Issue: #569
|
|
|
|
Update nl_NL.ts
|
|
Update nl_NL.ts
|
|
Update nl_NL.ts
|
|
Update nl_NL.ts
|
|
Update nl_NL.ts
|
|
|
|
|
|
|
|
|
|
|
|
Helps with choosing automatic threshold slider position.
|
|
|
|
Spotted-by: @mrsanchos
|
|
|
|
|
|
next hotfix or bust
|
|
In case of the options dialog, running close() in doOK() made it go into
closeEvent(), going into doCancel(), and reloading the settings prior to
them getting saved.
It's time for a hotfix2 build.
|
|
|
|
We rolled up our own refcount while using shared_ptr at the same time.
Remove all rolled-up logic and rely on shared_ptr's custom deleter to do
the cleanups.
This greatly simplifies the code here. Unfortunately, _fini ordering
makes Qt crash after the app object runs out of scope:
gui/main.cpp: QApplication app(argc, argv);
Both things considered, use _exit(2) on Unix to avoid running static
destructors.
|
|
|
|
Polish speakers typically know English well enough and translations are
awkward.
|
|
Adjust translations.
|
|
|
|
|
|
It belonged to the "experimental" center method that got removed
already.
|
|
|
|
Use hide() to avoid emitting idempotent events. There's isVisible() but
no isClosed() or equivalent. Worse yet, close() can return true twice in
a row, despite what docs for `bool QWidget::close()' say.
|
|
Add some translation strings
|
|
|
|
|
|
|
|
|
|
|
|
Unprintable characters won't work in addition to not displaying right.
This happens with altgr diacritics on the Polish keyboard on Linux.
Skip all key text with unprintable characters, as if the key was getting
unbound.
|
|
The escape key closed the dialog but prevented save/cancel hooks from
being run. Now we get the hooks running twice but it's better than not
at all.
|
|
- Register an event filter once and keep it there. The refcount goes out
of whack. Will further investigate.
- Use an iterator over all shortcuts registered for given key
combination
- Lessen logspam
Keyboard shortcut support on Linux looks pretty good now.
The issue that blocked keys from being further processed was running
XGrabKey multiple times with various modifiers. XGrabKey will block
unlisted modifiers from being passed to other apps. The solution was to
register once with AnyModifier and care about modifiers only later.
|
|
For Linux.
|
|
Only for Linux. The OSX code should stay intact however.
|
|
|
