By allowing to specify a separate source bitmap when calling Filter::apply
the same filter can be applied to multiple areas, and also doesn't need
to use a temporary bitmap. This also enables us to apply the filter to
multiple regions properly, even if they are (almost) adjacent.
If no separate source bitmap is supplied then a temporary bitmap is still
necessary.
By moving the Bitmap and Rect out of Filter::Parameters we can re-use
the parameters more efficiently, allowing the filter to be applied
to many bitmaps without having to re-create the filter every time.
Add an overload of GenericConvolutionFilter::apply that can be used
with a GenericConvolutionFilter::ApplyCache instance to avoid having
to allocate a temporary bitmap every time the filter is being applied.
The main inspiration behind this was to have a correct ex CSS unit.
The mean line is based off what it shows in the CSS Values and Units
Level 4 specification, section 6.1.1.
https://www.w3.org/TR/css-values-4/#font-relative-lengths
If we're sharing buffers, we only want to share trivial structures
as anything else could potentially share internal pointers, which
most likely is going to cause problems due to different address
spaces.
Fix the GUI::SystemTheme structure, which was not trivial, which
is now caught at compile time.
Fixes#3650
To make slightly more aesthetically pleasing use of the vertical space,
we now move all vertically centered text lines down by half the amount
of space below the font's baseline.
This is probably not the "correct" way to do this, but it does make
things look nicer with some of our fonts already.
Instead of everyone overriding save_to() and set_property() and doing
a pretty asymmetric job of implementing the various properties, let's
add a bit of structure here.
Object properties are now represented by a Core::Property. Properties
are registered with a getter and setter (optional) in constructors.
I've added some convenience macros for creating and registering
properties, but this does still feel a bit bulky. We'll have to
iterate on this and see where it goes.
Moves Bitmap backing store creation to the static create() methods.
This backing store is then passed into the Bitmap constructor. This
allows us correctly return nullptr from create() in the event that
memory allocation fails.
Errors like this became more likely due to the 'optimized' memory usage.
Also, this prevents the WindowServer from being killed by a goofy program
sharing an incomplete bitmap, and likely some other scenarios.
Indexed bitmaps used to allocate four times the required amount of memory.
Also, we should acknowledge that the underlying data is not always RGBA32,
and instead cast it only when the true type is known.
In all circumstances, this returned exactly the same thing as scanline_u8(),
so let's just remove the silly detour.
This does not add any new dependency on Bitmap-internals, because that already existed.
DC and AC table IDs read in the scan header segment weren't validated
against the IDs of Huffman tables read in the DHT segment. This caused
an OOB read when a Huffman table was accessed using the ID read in the
scan header segment. Furthermore, the decoder now replaces the old DC
or AC table if a redefinition has been found prior to the scan header.
Fixes#3439.
GIFLoader now uses a single frame buffer to cache the last decoded
frame. This drastically reduces memory usage at the small expense of
re-decoding frames on each loop.
RestoreBackground disposal mode is now a transparent fill to allow
background to show through.
RestorePrevious disposal mode now restores the previous frame.
