Counterpart to GlyphCacheArrayGL. At first I expected that it'd need
TextureTools::DistanceFieldGL to be expanded with texture array support
but after enough massaging of brain matter I realized that not really,
since the procesing has to be done slice by slice anyway, and having to
upload the whole array just to have a temporary input for processing
would be a waste of memory.
It just binds a layer of it to a framebuffer internally so no shader
changes or extra construction flags are needed. Originally I thought
about making the input an array as well, but ultimately that just
doesn't make sense -- the processing would need to be done slice by
slice anyway and you don't want to allocate the whole excessively sized
texture just for it to be used once, and only a part of it every time.
Wanted to add a proper variant that reads all layers at once into an
Image3D, but that requires the Image APIs to be made less crappy first.
So it's just this for noew because now I badly need it for
Text::DistanceFieldGlyphCacheArrayGL tests.
This was quite nasty, a multi-day effort to trim this down and then
increasingly growing disappointment as I discovered it was affecting
basically any use of the API.
Am I overdoing it? Most unpremultiplication code I found doesn't even
deal with treating alpha being zero correctly, while here I'm handling
also the cases of RGB channels going over alpha, and for packed
formats trying to match precision of the same done with pack()/unpack().
So far it contains just the uniform definitions and utilities related
to line drawing, nothing else, but especially the line utilities were
needed to be able to build MeshTools tests on a GL-less build.
The tests all pass exactly as they did before, apart from slightly
different assertion messages. I moved the guts from Renderer.cpp to
RendererGL.cpp as the code relies on both GL and STL, to not have it
spread over too many places.
I'm going to make a new Renderer class that's unlike the old one, and
isn't templated anymore either, so gotta make room for it first. I assume
that all existing code used the Renderer2D / Renderer3D typedefs so this
should be pretty much painless.
Ultimately this class will be gone, with only (then deprecated)
AbstractRenderer left, and Renderer2D / Renderer3D being aliases to it
-- internally it doesn't actually do anything dfferent, even the
positions are two-component in both cases.
As the time goes, I have less and less patience trying to figure these
things out. It's likely my fault for using the not-well-tested DSA APIs
or something.
For integer attributes, that is. This never worked and the MeshGLTest
even had an explicit test case for this, but it took me until now to
realize that this would be best caught at compile time. So now it is.
But since this functionality dates back to 2010, just removing the
values would likely break a lot of code (well, broken code, but still),
so the disallowed values are now deprecated, giving a compile time
warning when used, and are removed only on non-deprecated builds.
This warning isn't enabled by default by either -Wall or -Wextra, and
for a good reason. When I do that, it fires also for all uses of
Platform::Application, basically any subclass causes that warning to be
printed as well, even if given subclass is used just once, in an
anonymous namespace, and only through the MAGNUM_APPLICATION_MAIN()
macro. The only solution on Clang is to make such subclasses `final`,
but on GCC not even that helps and one is forced to really make the base
destructor virtual. Which doesn't achieve anything, only adds yet
another entry to the vtable.
So, in the SceneGraph it fixes one particular use case where the warning
was triggered, and since the fix isn't so invasive I'm fine with that.
For Platform I have no idea how an acceptable fix would look like, so I
hope nobody needs any of that anytime soon.
See also https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102168, especially
the part saying "This warning should not be used.".
This got deprecated in 069c81b9cb but
without any visible documentation bit, so it was almost impossible to
trace back to a particular version. Or know about it when using the API.
The more I looked at it, the more useless it felt. Also, since it's so
broken, I'm also completely removing it from the test because there's no
point in even testing it.
Unlike the Key enum, which shows what a user would perceive as given key
in a particular layout, the scancode is a layout-independent identifier
for e.g. WASD movement in games.
Unfortunately the API availability is wildly different among the
toolkits -- SDL's is the most complete, GLFW is second, and then there's
Emscripten / HTML5 which provides just string identifiers. I tried to
add these for X11 as well, but quick googling led to a SO question where
it was left unanswered. Not worth my time.
There's the obvious advantage of them now being shorter to type, as one
no longer needs to prefix them with KeyEvent::. But the main reason I
did this was to allow various direct keyboard state queries to be
implemented, such as isKeyPressed(). With them being hidden in the event
class the only way would be to put the query directly there as well,
which isn't nice and is also not very discoverable.
A similar case was with mouse buttons, but that was already fixed with
the PointerEvent rework that happened in earlier commits. There the
additional complication was that MouseEvent::Button and
MouseMoveEvent::Button were incompatible enums. Application::Pointer
fixes that now as well.
Ubuntu 18.04 has 3.2 so this should be safe. Need to use
glfwGetKeyName() to undo keyboard layout weirdness, would be nasty to
have to make a fallback for older versions.
As usual, the most trash fire platform of them all. Ugh. I chose to
ignore certain aspects and suggestions and made it behave more like
Emscripten and SDL2, because that makes more sense to me.
Co-authored-by: nodoteve <nodoteve@yandex.com>
The impossible-to-reliably-disable behavior with compatibility mouse
events is quite a headache. I wish Emscripten implemented pointer events
already so I could ditch this mess -- especially the array of 32 touches
where all of them but one will be unchanged is stupid.
For the internals unfortunately, EmscriptenMouseEvent and
EmscriptenTouchEvent have no common base, so I had to give up on the
current way of querying the event struct directly from event getters, as
that'd be too nasty with the branching and casts. Instead the relevant
fields are put directly into the events themselves.
HTML5 also doesn't provide any relative pointer position. For the mouse
it was rather straightforward, but for the up-to-32 touches I have to
maintain an array of per-finger positions and match them by ID.
Hopefully the linear lookup is fine. I'll probably use the same approach
for the AndroidApplication.
This makes 2.0.6 as the oldest supported because in older versions it's
not possible to disable touch to mouse event translation, and it'd be
too annoying to have it special-cased there. The version bump should be
fine as Ubuntu 18.04 has 2.0.8.
All the new pointer events have float positions, this one was the odd one
out. And I didn't like the name anymore, so I took that as an opportunity
to change the position() data type without introducing a breaking change
for everyone.
Another considered change was adding Z offset to it, since HTML5 APIs
have that. However, all my googling led to just a single SO question from
2015, where someone said it's for trackballs that can navigate in 3D
space. I'm not sure if *scroll* is actually the best way to report those,
and since SDL3 didn't bother adding that and neither Android nor WINAPI
have anything like that, I'm not bothering either.
Unlike the previous commits, this is done for all apps at once, because
it's a comparatively simpler change. The only odd one out is
AbstractXApplication, where I introduced MouseScrollEvent just a few
commits back, so I simply renamed it without leaving a deprecated copy.
Then, ScreenedApplication needed some extra logic to handle the case of
apps not implementing any scroll event at all.
Because this overrides the base pointer*Event() implementations, it
additionally has to call into the parent implementation in order to fall
back to the deprecated mouse event if the new pointer event isn't
handled.
And thus also add actual support for finger and pen input, instead of
reporting them as Button::None. Something that needed fixing ever since
the initial implementation in 2014.
I *really* wanted to make a setup where mouse input would be recognized
as such and reported in the app, but wasted the whole day on that and got
only as far as having it recognized as a stylus input with hover (!!!),
when using some Android Desktop image. Not sure if that's some stupid
mislabeling (because middle and right mouse buttons are reported as such)
or it's just the emulation layer being crap.
While at it, I at least added support for hover events. I still have to
document all the newly found warts and hard-to-remember workflows with
getting a simulator running.
Vladimír Vondruš