The Ctrl-Z was interrupting the enveloping, making it finish, but it then got asked to finish again on a mouse up. This change stops the second finish from happening.
Caused by declaring a variable later than a conditionally compiled piece of code.
Net result is that Mac will be less tolerant of a window partly off the top of the screen than Windows will be, and will reset it to the default.
Splash screen still always appears on first screen (intentional).
Also I fixed issue where new project windows would appear on first screen, even if Audacity on second.
Previously Ctrl Click on Wave behaved exactly as Ctrl Click on Track Control Panel. With this change on wave it
(a) No longer toggles. It only adds.
(b) It will extend and can be drag extended.
It is possible to change (a) to be toggle if that is desired.
This should ease the creation and extension of discontinuous track selections.
Fixed by being more tolerant of window being (slightly) off screen. Windows 10 appears to mis-report the window size by 6 pixels. This fix allows the user to restore a window that is (slightly) off screen. That's OK and helps those users who position and size the window very precisely.
It turns out that having fixed the behaviour of Shift-Click on the Track Control Panel, I'd also fixed the behaviour of Shift-Click on the wave. But then I blew it by over riding the Shift-Click behaviour on wave to use the version that does NOT horizontally extend. This change comments out the 'bShift' and so gives us the original Shift-Click behaviour on wave, without undoing the other fixes.
Thanks to Walter Gladwin.
The comment said "Is this correct??" the answer is no. However the problem only showed when other fixes were made elsewhere, and Add Label at Playback Positions then stopped working when in Pause.
I first of all added a warning message if trying to configure while playing.
Then decided that it is more in keeping with current style to grey out the Spectrogram Settings menu item.
So the new warning message should actually never be reached.
Comment more cases of bad sampleCount truncation, and even worse...
Fix progress indicator again: must do a floating-point division...
A little more type agnosticism
Remove more unnecessary casts to sampleCount
Fix more narrowings of sampleCount
Caused by optimisation of NumericTextCtrl::ValueToControls(). It now only updates on a change, and because mValueStr was being updated without calling value to controls, the string was not being identified as having changed.
* sampleCount:
Remove unnecessary casts to sampleCount
Type agnosticism for some other variables that were not sampleCount...
Remove many mentions of sampleCount with auto and decltype...
Use sf_count_t not sampleCount
some uses of size_t
Use long long for argument passed to wxString::ToLongLong
More cautions in SBSMSEffect.cpp...
