Currently:
- load/update/run/etc. take place on the boot.lua coroutine.
- draw happens "asynchronously" on the main thread.
When C needs to throw an error, it doesn't know which thread to
throw the error on. If it throws it on the wrong thread, you get
a crash instead of an error screen.
One way to fix this is to change the error context based on the
thread that's currently running, so that errors in C are thrown
on the correct thread. This is the approach that's taken here.
A potentially better approach would be to run all the code on the
same thread, but I ran into issues when I tried to do this.
It may also be possible to (ab)use the Lua panic handler to catch
errors on one of the threads and somehow forward them to the other.
- lovr.graphics.tock returns the latest value of the timer, or 0.
- Timers are not in the stats table anymore.
This is to prepare for an upcoming internal change that affects timers.
This means Lua print() statements can be uniquely filtered out vs anything else (because internal Lovr logging uses loglevel DEBUG and lovr errors use loglevel WARN).
The "mask" example was failing on Oculus Mobile on the line of glsl:
if (lovrViewID == 1) {
because lovrViewID was unsigned and 1 was signed. One way to fix this would be to replace 1 with 1u as that is unsigned, but this would be the wrong fix because lovrViewID is in fact signed on all platforms other than Oculus Mobile. lovrViewID was depending on platform defined to either gl_ViewportIndex (signed), a signed uniform, or gl_ViewID_OVR (unsigned). The solution is to place an explicit cast in the multiview definition of lovrViewID so that it is signed on all platforms.
lovr.graphics.fill renders a fullscreen quad, it's convenient because
you don't need to set up a mesh and toggle all the pipeline states.
However, if you are dealing with copying/rendering between stereo
textures, you have to write your own shader for that. For now.
Instead of boolean shader flags turning into actual booleans defines
in the shader source, for GLSL they turn into defines. This lets you
use ifdef, which is the more common intended usage.
Also MULTICANVAS is now a boolean shader flag. The old MULTICANVAS
define is deprecated.
Previously, this program function lovr.update(dt) lovr.filesystem.append("/test123", lovr.timer.getTime()) end would fail in lovr because lovrFileWrite required the file to be in write mode (not append)
l_event.c was processing a thread-related event and in the process using a thread struct, even when LOVR_ENABLE_THREAD is undefined and threads do not exist. I caused the thread event type to simply not exist when the thread module is not being built.
Since the event is now only sometimes present, I put it at the end of the enum as slight protection against binary mismatches with dynamically loaded modules.
Because of how and when draws occur in our Oculus Mobile path, during a restart it would attempt to draw a frame after lovrGraphicsDestroy() is called, leading to a crash in lovrGraphicsSetCamera(). This blocks draws until the restart is finished and renderTo() has been called (conveniently detectable using the existing state.renderCallback).
Empty arrays aren't allowed in C++, so a single dummy element has to be added. sds works by scary cast magic so this dummy element is never actually allocated. A #define is used in this patch so in C this compiles exactly the same as before.
Check for GCC version >=4.7 [covers GCC or, theoretically, Intel C Compiler] or __has_builtin(__atomic_add_fetch) [covers Clang]. If either of these are found, prefer the GCC atomic builtins instead of the C11 builtins.
Also: Gates Microsoft atomics on _MSC_VER, not _WIN32. This may help improve compatibility with mingw but has not been tested.
Because the new arr.h contains an array on the stack, we can't
initialize it and then copy it, or the pointer to the stack array
will be pointing to the wrong thing, causing incorrect behavior.
mat4_transform ignores the final row of the matrix, which normally contains scale data. This is fine for modelview matrices, which is all lovr currently uses mat4_transform for. However it fails if you ever use mat4_transform on a projection matrix.
mat4_transform_project takes the final row into account, so you can use it with projection matrices. This was useful in my fork and might be useful to have around in lovr someday later.
- If you have an instanced batch, it will use the instanced mesh. That
has a drawID attribute that uses the identity buffer, which has a vertex
divisor. BUT if you only have one instance, then we won't emit an
instanced draw, and the use of a divisor'd attribute w/ a non-instanced
draw is causing mega problems on macOS.
- This also fixes observed macOS bugs like:
- Needing to have a small UBO
- Flickering at startup
- Flicker when writing to the last byte of a UBO
- etc.
- Also make the generic attribute value for lovrDrawID more correct (scalar instead of vector).
- Slightly dim background color (may change).
- Use a shader for the logo (it's centered now, not quite 100% like the image).
- Adjust text optical weight (I hate it).