This addresses the same issue that
95d416ab77 was intended to fix:
Drawing a drop shadow on a high-res display on Windows could sometimes
produce thin lines around the edge of the shadow. Using the 'clamp'
rather than the 'wrap' clipping mode seems to help.
Previously, plugins and standalone apps could produce slightly different
visual results, particularly anti-aliasing, when displaying on a display
with native scaling applied. The discrepancy was caused by SetDpi being
called with a larger-than-default value (e.g. 192 DPI on a 200% scaled
display) in a standalone app, whereas SetDpi would always be called with
the default value of 96 in a plugin.
Keeping the default value seems to produce better results, so standalone
apps will now retain the default DPI.
This issue could be observed in the GraphicsDemo's SVG pane, when the
"rotation" option was enabled.
Drawables internally enable the unclipped painting flag, which normally
prevents slow clipping when drawing subcomponents of the drawable.
The Direct2D graphics context was using the frame area as the default
area, used to signal that no clipping should be applied. However, when
non-axis-aligned transform was active, this area was incorrectly applied
as a geometric clipping region. D2D geometric clips are relatively slow,
so this caused large slowdowns.
This solution adds a flag that is set whenever a clip is explicitly
requested. If no clip is explicitly requested, then clipping will be
entirely bypassed. This can make rendering of Drawables significantly
faster.
Previously, filling a string containing a space or other non-rendered
character with a gradient would end up filling the entire clip region.
The correct behaviour is to completely skip filling any empty paths.
There were a few "ambiguous operator new/delete" errors that were due to
inheriting from a private base class that used the leak detector. These
errors are resolved by adding the leak detector to the derived classes.
JUCE_API was missing from a few useful types, notably the ARA hosting
types.
This was evident in the FontsDemo when using the CoreGraphics renderer,
selecting a font without an italic face, and enabling the italic style.
The glyph anchor positions were incorrectly transformed by the text
matrix, causing the left margin of the text to become tilted.
We now correct for the slant and scale specified in the text matrix when
computing glyph anchor positions.
This fixes an issue where Direct2D will emit an error when using dirty
rects on the first full frame after resizing.
The issue isn't present on all hardware/drivers, but was observed on a
Windows 11 computer with a 890M iGPU.
Partially reverts a change made in
362a1cc070.
The factory needs to be multithreading-enabled in order to support e.g.
drawing to the screen on the main thread and drawing to an Image on a
background thread, even when no resources other than the graphics device
are shared between the two threads.
Previously, IDXGISwapChain::Present was called on a background thread,
which made it difficult to avoid race conditions. e.g. during a
live-resize of a window, we would occasionally draw old incomplete
frames instead of new frames at the correct size.
The new approach moves the Present call to the main thread via
AsyncUpdater. We attempt to present whenever the swap event wakes, and
whenever a frame is drawn. Only a single Present call may be made after
the swap event wakes. Subsequent Present calls will be ignored until the
next time the swap event wakes.
This allows for styles other than normal/bold/italic/bold-italic to be
selected, and more closely matches the behaviour of font selection on
other platforms.
