... except Audacity.h; and in no others.
Do so even if Experimental.h gets multiply included, as in both the .h and
.cpp files.
This makes it easier to do a text scan to be sure there are no unintended quiet
changes of meaning because of omission of Experimental.h when the flag is
an enabled one.
Also move inclusions of Experimental.h earlier.
Also don't require Experimental.h to be preceded by Audacity.h to define
EXPERIMENTAL_MIDI_OUT correctly.
... Make all line drawing go through our AColor which wraps the wxDC
line-drawing functions.
Despite what wxWidgets documentation says, the functions still don't
consistently include the first point and omit the last point of a line. I
observed inclusion of both on Mac, while on Windows, omission sometimes of the
first point instead of the last.
Discrepancies could be observed in 2.3.0 between Windows and Mac, using the
Window magnifier or command+alt+ +, zooming in closely on the ends of the
shadow lines below and right of tracks, or at the bottom-right corners of the
bevels drawn around vertical rulers.
So where there is an observable one-pixel difference of drawing between
platforms, there is the question, which was the intent when the drawing code
was written? Should the coordinates be corrected by one or not?
I reviewed each case and used my judgment.
Most of the calls are in drawing rulers. Major tick lines were drawn five
pixels long, and minor, three, on Mac. I keep this behavior, but you might
argue for making them four and two.
On the other hand the drawing of ruler grid lines, which you can see in
the equalization and frequency analysis windows, did need a one-pixel correction
to avoid straying out of bounds.
... which will make it easier to change the types of those containers to
std::vectors of other string-like classes
for wxString,
IsEmpty => empty
Clear => clear
Alloc => reserve
for wxArrayString,
Count => size
GetCount => size
IsEmpty => empty
Add => push_back
Clear => clear
Empty => clear
Sort => std::sort (only with default comparator)
SetCount => resize
Last => back
Item => operator []
Alloc => reserve
... Replace some enum constants with the equivalent values from the more
special-purpose enums, as the wx header files recommend, so overloading
selects the ctors with non-int arguments.
In a full rebuild of the debug project on Mac, this reduces the count of
warnings from 264 to 274.
We were for example getting many of these:
\audacity\src\widgets\numerictextctrl.h(171): warning C4458: declaration of 'value' hides class member
MSVC2013 didn't warn about these, but MSVC2017 does.
The main change in wx accessibility is this:
7dab555f71 (diff-04f5191d86f95b1c4d5d9c979da65878)
However wxWindowAccessible has not been updated to take into account of that change. In particular wxWindowAccessible::GetParent() was always wrong, but it was consistent with the rest of the framework. Now it's wrong and inconsistent. This function should return an object with role window, and which has the same name.
The fix is to introduce class WindowAccessible, which is effectively our own version of wxWindowAccessible. This class does not override GetParent(), and so just relies on a standard accessible object to to the right thing in wxIAccessible::get_accParent() (which is does). This class also allows us to have our own version of GetName(), which allows us to set the accessibility names of buttons.
These changes will break the accessibility of Audacity if it is built with wxWidgets 3.0.X. If this is a problem, then there could be some #if stuff in WindowAccessible.h to turn the WindowAccessible class into one which simply inherits from wxWindowAccessible, and doesn't override anything.
... it's either the source of the connection that is being destroyed, or other
object (such as an ancestor window) transitively owning it and so causing it to
be destroyed too;
or, the sink is being destroyed, and that sink is a wxEvtHandler (which is
always so for Disconnect, though not for Unbind in case Bind was passed a
member function of a non-wxEvtHandler).
wxWidgets takes care of erasing the connection in such cases.
This removes most calls to Disconnect and Unbind. Many destructors shrank to
nothing.
Notably, in case of popup menu handling, the call to Disconnect is not removable
because the object being destroyed is neither the source nor the sink.
... Assuming that large unsigned magnitudes with high order bit set are not
the problem, but signed negatives of small magnitude may be:
1) Always cast the unsigned to signed in comparisons, not the other way.
Also:
2) Cast unsigned TERM to signed by itself, before subtracting. Don't cast
the result.
3) Rewrite some comparisons by moving subtracted term to other side.
See commits
d2fe7b1757d77d93d82d53685a118517a4d2e996
f463eda36c059236ecc168919c745eb687170c95
- AutomationCommands replaces GetAllMenuCommands, and can provide
information about menus, buttons and toolbars to a script.
- BatchCommands can now return textual results to a script.
- There's a new GUID for mod-script-pipe and it is included in the .sln.
... and similar wx "variadics," which all treat wxString smartly enough that
you don't need this.
Don't need c_str either to convert wxString to const wxChar * because
wxString has a conversion operator that does the same.
Previously hovering over a down button made no difference.
Also tweaked the appearance of hover-over thumbs on dark theme sliders.
Also tweaked hover images and colours generally.
Classic retains the old style.
Hi Contrast does not distinguish between hover-up and hover-down.
- Dead code from experiments in SelectionBar removed.
- Many warnings about unused parameters fixed with WXUNUSED()
- Many warnings about signed / unsigned comparisons cleaned up.
- Several 'local variable declared but not used' warnings fixed.
... whenever they really describe the size of a buffer that fits in memory, or
of a block file (which is never now more than a megabyte and so could be fit in
memory all at once), or a part thereof.
... And in some places where a library uses signed types, assert that
the reported number is not negative.
What led me to this, is that there are many places where a size_t value for
an allocation is the product of a number of channels and some other number.
