update love-api

15 files changed, 3335 insertions, 322 deletions

diff --git a/meta/3rd/love2d/library/love.graphics.lua b/meta/3rd/love2d/library/love.graphics.lua
index d2f8c7d8..056cfef1 100644
--- a/meta/3rd/love2d/library/love.graphics.lua
+++ b/meta/3rd/love2d/library/love.graphics.lua
@@ -92,7 +92,9 @@ function love.graphics.discard(discardcolor, discardstencil) end
 ---
 ---The right and bottom edges of the object are shifted at an angle defined by the shearing factors.
 ---
----When using the default shader anything drawn with this function will be tinted according to the currently selected color.  Set it to pure white to preserve the object's original colors.
+---When using the default shader anything drawn with this function will be tinted according to the currently selected color.
+---
+---Set it to pure white to preserve the object's original colors.
 ---
 ---@overload fun(texture: love.Texture, quad: love.Quad, x: number, y: number, r: number, sx: number, sy: number, ox: number, oy: number, kx: number, ky: number)
 ---@overload fun(drawable: love.Drawable, transform: love.Transform)
@@ -418,7 +420,9 @@ function love.graphics.getTextureTypes() end
 function love.graphics.getWidth() end
 
 ---
----Sets the scissor to the rectangle created by the intersection of the specified rectangle with the existing scissor.  If no scissor is active yet, it behaves like love.graphics.setScissor.
+---Sets the scissor to the rectangle created by the intersection of the specified rectangle with the existing scissor.
+---
+---If no scissor is active yet, it behaves like love.graphics.setScissor.
 ---
 ---The scissor limits the drawing area to a specified rectangle. This affects all graphics calls, including love.graphics.clear.
 ---
@@ -718,7 +722,9 @@ function love.graphics.present() end
 ---
 ---Draws text on screen. If no Font is set, one will be created and set (once) if needed.
 ---
----As of LOVE 0.7.1, when using translation and scaling functions while drawing text, this function assumes the scale occurs first.  If you don't script with this in mind, the text won't be in the right position, or possibly even on screen.
+---As of LOVE 0.7.1, when using translation and scaling functions while drawing text, this function assumes the scale occurs first.
+---
+---If you don't script with this in mind, the text won't be in the right position, or possibly even on screen.
 ---
 ---love.graphics.print and love.graphics.printf both support UTF-8 encoding. You'll also need a proper Font for special characters.
 ---
diff --git a/meta/3rd/love2d/library/love.lua b/meta/3rd/love2d/library/love.lua
index c1c55c69..8c2e3f18 100644
--- a/meta/3rd/love2d/library/love.lua
+++ b/meta/3rd/love2d/library/love.lua
@@ -1,6 +1,7 @@
 ---@meta
 
 -- version: 11.4
+
 ---@class love
 love = {}
 
diff --git a/meta/3rd/lovr/library/callback.lua b/meta/3rd/lovr/library/callback.lua
index 3fb0a960..13cbef51 100644
--- a/meta/3rd/lovr/library/callback.lua
+++ b/meta/3rd/lovr/library/callback.lua
@@ -1,19 +1,47 @@
 ---@meta
 
 ---
----The `lovr.conf` callback lets you configure default settings for LÖVR.  It is called once right before the game starts.  Make sure you put `lovr.conf` in a file called `conf.lua`, a special file that's loaded before the rest of the framework initializes.
+---The `lovr.conf` callback lets you configure default settings for LÖVR.
+---
+---It is called once right before the game starts.
+---
+---Make sure you put `lovr.conf` in a file called `conf.lua`, a special file that's loaded before the rest of the framework initializes.
+---
+---
+---### NOTE:
+---Disabling the headset module can improve startup time a lot if you aren't intending to use `lovr.headset`.
+---
+---You can set `t.window` to nil to avoid creating the window. You can do it yourself later by using `lovr.graphics.createWindow`.
+---
+---If the `lovr.graphics` module is disabled or the window isn't created, attempting to use any functionality requiring graphics may cause a crash.
+---
+---Enabling the `t.graphics.debug` flag will add additional error checks and will send messages from the GPU driver to the `lovr.log` callback.
+---
+---This will decrease performance but can help provide information on performance problems or other bugs.
+---
+---The `headset.offset` field is a vertical offset applied to the scene for headsets that do not center their tracking origin on the floor.
+---
+---This can be thought of as a "default user height". Setting this offset makes it easier to design experiences that work in both seated and standing VR configurations.
 ---
 ---@type fun(t: table)
 lovr.conf = nil
 
 ---
----This callback is called every frame.  Use it to render the scene.  If a VR headset is connected, anything rendered by this function will appear in the headset display.  The display is cleared to the background color before this function is called.
+---This callback is called every frame.
+---
+---Use it to render the scene.
+---
+---If a VR headset is connected, anything rendered by this function will appear in the headset display.
+---
+---The display is cleared to the background color before this function is called.
 ---
 ---@type fun()
 lovr.draw = nil
 
 ---
----The "lovr.errhand" callback is run whenever an error occurs.  It receives two parameters. The first is a string containing the error message. The second is either nil, or a string containing a traceback (as returned by "debug.traceback()"); if nil, this means "lovr.errhand" is being called in the stack where the error occurred, and it can call "debug.traceback()" itself.
+---The "lovr.errhand" callback is run whenever an error occurs.
+---
+---It receives two parameters. The first is a string containing the error message. The second is either nil, or a string containing a traceback (as returned by "debug.traceback()"); if nil, this means "lovr.errhand" is being called in the stack where the error occurred, and it can call "debug.traceback()" itself.
 ---
 ---"lovr.errhand" should return a handler function to run in a loop to show the error screen. This handler function is of the same type as the one returned by "lovr.run" and has the same requirements (such as pumping events). If an error occurs while this handler is running, the program will terminate immediately-- "lovr.errhand" will not be given a second chance. Errors which occur inside "lovr.errhand" or in the handler it returns may not be cleanly reported, so be careful.
 ---
@@ -23,7 +51,11 @@ lovr.draw = nil
 lovr.errhand = nil
 
 ---
----The `lovr.focus` callback is called whenever the application acquires or loses focus (for example, when opening or closing the Steam dashboard).  The callback receives a single argument, focused, which is a boolean indicating whether or not the application is now focused.  It may make sense to pause the game or reduce visual fidelity when the application loses focus.
+---The `lovr.focus` callback is called whenever the application acquires or loses focus (for example, when opening or closing the Steam dashboard).
+---
+---The callback receives a single argument, focused, which is a boolean indicating whether or not the application is now focused.
+---
+---It may make sense to pause the game or reduce visual fidelity when the application loses focus.
 ---
 ---@type fun(focused: boolean)
 lovr.focus = nil
@@ -41,35 +73,67 @@ lovr.keypressed = nil
 lovr.keyreleased = nil
 
 ---
----This callback is called once when the app starts.  It should be used to perform initial setup work, like loading resources and initializing classes and variables.
+---This callback is called once when the app starts.
+---
+---It should be used to perform initial setup work, like loading resources and initializing classes and variables.
+---
+---
+---### NOTE:
+---If the project was loaded from a restart using `lovr.event.restart`, the return value from the previously-run `lovr.restart` callback will be made available to this callback as the `restart` key in the `args` table.
+---
+---The `args` table follows the [Lua standard](https://en.wikibooks.org/wiki/Lua_Programming/command_line_parameter).
+---
+---The arguments passed in from the shell are put into a global table named `arg` and passed to `lovr.load`, but with indices offset such that the "script" (the project path) is at index 0.
+---
+---So all arguments (if any) intended for the project are at successive indices starting with 1, and the executable and its "internal" arguments are in normal order but stored in negative indices.
 ---
 ---@type fun(args: table)
 lovr.load = nil
 
 ---
----This callback is called when a message is logged.  The default implementation of this callback prints the message to the console using `print`, but it's possible to override this callback to render messages in VR, write them to a file, filter messages, and more.
+---This callback is called when a message is logged.
+---
+---The default implementation of this callback prints the message to the console using `print`, but it's possible to override this callback to render messages in VR, write them to a file, filter messages, and more.
 ---
 ---The message can have a "tag" that is a short string representing the sender, and a "level" indicating how severe the message is.
 ---
----The `t.graphics.debug` flag in `lovr.conf` can be used to get log messages from the GPU driver (tagged as `GL`).  It is also possible to emit your own log messages using `lovr.event.push`.
+---The `t.graphics.debug` flag in `lovr.conf` can be used to get log messages from the GPU driver (tagged as `GL`).
+---
+---It is also possible to emit your own log messages using `lovr.event.push`.
 ---
 ---@type fun(message: string, level: string, tag: string)
 lovr.log = nil
 
 ---
----This callback is called every frame after rendering to the headset and is usually used to render a mirror of the headset display onto the desktop window.  It can be overridden for custom mirroring behavior.  For example, you could render a single eye instead of a stereo view, apply postprocessing effects, add 2D UI, or render the scene from an entirely different viewpoint for a third person camera.
+---This callback is called every frame after rendering to the headset and is usually used to render a mirror of the headset display onto the desktop window.
+---
+---It can be overridden for custom mirroring behavior.
+---
+---For example, you could render a single eye instead of a stereo view, apply postprocessing effects, add 2D UI, or render the scene from an entirely different viewpoint for a third person camera.
+---
+---
+---### NOTE:
+---When this callback is called, the camera is located at `(0, 0, 0)` and is looking down the negative-z axis.
+---
+---Note that the usual graphics state applies while `lovr.mirror` is invoked, so you may need to reset graphics state at the end of `lovr.draw` to get the result you want.
 ---
 ---@type fun()
 lovr.mirror = nil
 
 ---
----This callback contains a permission response previously requested with `lovr.system.requestPermission`.  The callback contains information on whether permission was granted or denied.
+---This callback contains a permission response previously requested with `lovr.system.requestPermission`.
+---
+---The callback contains information on whether permission was granted or denied.
 ---
 ---@type fun(permission: lovr.Permission, granted: boolean)
 lovr.permission = nil
 
 ---
----This callback is called right before the application is about to quit.  Use it to perform any necessary cleanup work.  A truthy value can be returned from this callback to abort quitting.
+---This callback is called right before the application is about to quit.
+---
+---Use it to perform any necessary cleanup work.
+---
+---A truthy value can be returned from this callback to abort quitting.
 ---
 ---@type fun():boolean
 lovr.quit = nil
@@ -81,13 +145,23 @@ lovr.quit = nil
 lovr.resize = nil
 
 ---
----This callback is called when a restart from `lovr.event.restart` is happening.  A value can be returned to send it to the next LÖVR instance, available as the `restart` key in the argument table passed to `lovr.load`.  Object instances can not be used as the restart value, since they are destroyed as part of the cleanup process.
+---This callback is called when a restart from `lovr.event.restart` is happening.
+---
+---A value can be returned to send it to the next LÖVR instance, available as the `restart` key in the argument table passed to `lovr.load`.
+---
+---Object instances can not be used as the restart value, since they are destroyed as part of the cleanup process.
+---
+---
+---### NOTE:
+---Only nil, booleans, numbers, and strings are supported types for the return value.
 ---
 ---@type fun():any
 lovr.restart = nil
 
 ---
----This callback is the main entry point for a LÖVR program.  It is responsible for calling `lovr.load` and returning the main loop function.
+---This callback is the main entry point for a LÖVR program.
+---
+---It is responsible for calling `lovr.load` and returning the main loop function.
 ---
 ---@type fun():function
 lovr.run = nil
@@ -97,11 +171,21 @@ lovr.run = nil
 ---
 ---For example, when `shift + 1` is pressed on an American keyboard, `lovr.textinput` will be called with `!`.
 ---
+---
+---### NOTE:
+---Some characters in UTF-8 unicode take multiple bytes to encode.
+---
+---Due to the way Lua works, the length of these strings will be bigger than 1 even though they are just a single character. `lovr.graphics.print` is compatible with UTF-8 but doing other string processing on these strings may require a library.
+---
+---Lua 5.3+ has support for working with UTF-8 strings.
+---
 ---@type fun(text: string, code: number)
 lovr.textinput = nil
 
 ---
----The `lovr.threaderror` callback is called whenever an error occurs in a Thread.  It receives the Thread object where the error occurred and an error message.
+---The `lovr.threaderror` callback is called whenever an error occurs in a Thread.
+---
+---It receives the Thread object where the error occurred and an error message.
 ---
 ---The default implementation of this callback will call `lovr.errhand` with the error.
 ---
@@ -109,7 +193,11 @@ lovr.textinput = nil
 lovr.threaderror = nil
 
 ---
----The `lovr.update` callback should be used to update your game's logic.  It receives a single parameter, `dt`, which represents the amount of elapsed time between frames.  You can use this value to scale timers, physics, and animations in your game so they play at a smooth, consistent speed.
+---The `lovr.update` callback should be used to update your game's logic.
+---
+---It receives a single parameter, `dt`, which represents the amount of elapsed time between frames.
+---
+---You can use this value to scale timers, physics, and animations in your game so they play at a smooth, consistent speed.
 ---
 ---@type fun(dt: number)
 lovr.update = nil
diff --git a/meta/3rd/lovr/library/lovr.audio.lua b/meta/3rd/lovr/library/lovr.audio.lua
index a768b708..71717874 100644
--- a/meta/3rd/lovr/library/lovr.audio.lua
+++ b/meta/3rd/lovr/library/lovr.audio.lua
@@ -1,13 +1,31 @@
 ---@meta
 
 ---
----The `lovr.audio` module is responsible for playing sound effects and music.  To play a sound, create a `Source` object and call `Source:play` on it.  Currently ogg, wav, and mp3 audio formats are supported.
+---The `lovr.audio` module is responsible for playing sound effects and music.
+---
+---To play a sound, create a `Source` object and call `Source:play` on it.
+---
+---Currently ogg, wav, and mp3 audio formats are supported.
 ---
 ---@class lovr.audio
 lovr.audio = {}
 
 ---
----Returns the global air absorption coefficients for the medium.  This affects Sources that have the `absorption` effect enabled, causing audio volume to drop off with distance as it is absorbed by the medium it's traveling through (air, water, etc.).  The difference between absorption and falloff is that absorption is more subtle and is frequency-dependent, so higher-frequency bands can get absorbed more quickly than lower ones.  This can be used to apply "underwater" effects and stuff.
+---Returns the global air absorption coefficients for the medium.
+---
+---This affects Sources that have the `absorption` effect enabled, causing audio volume to drop off with distance as it is absorbed by the medium it's traveling through (air, water, etc.).
+---
+---The difference between absorption and falloff is that absorption is more subtle and is frequency-dependent, so higher-frequency bands can get absorbed more quickly than lower ones.
+---
+---This can be used to apply "underwater" effects and stuff.
+---
+---
+---### NOTE:
+---Absorption is currently only supported by the phonon spatializer.
+---
+---The frequency bands correspond to `400Hz`, `2.5KHz`, and `15KHz`.
+---
+---The default coefficients are `.0002`, `.0017`, and `.0182` for low, mid, and high.
 ---
 ---@return number low # The absorption coefficient for the low frequency band.
 ---@return number mid # The absorption coefficient for the mid frequency band.
@@ -15,7 +33,9 @@ lovr.audio = {}
 function lovr.audio.getAbsorption() end
 
 ---
----Returns a list of playback or capture devices.  Each device has an `id`, `name`, and a `default` flag indicating whether it's the default device.
+---Returns a list of playback or capture devices.
+---
+---Each device has an `id`, `name`, and a `default` flag indicating whether it's the default device.
 ---
 ---To use a specific device id for playback or capture, pass it to `lovr.audio.setDevice`.
 ---
@@ -55,13 +75,91 @@ function lovr.audio.getPosition() end
 ---
 ---Returns the name of the active spatializer (`simple`, `oculus`, or `phonon`).
 ---
----The `t.audio.spatializer` setting in `lovr.conf` can be used to express a preference for a particular spatializer.  If it's `nil`, all spatializers will be tried in the following order: `phonon`, `oculus`, `simple`.
+---The `t.audio.spatializer` setting in `lovr.conf` can be used to express a preference for a particular spatializer.
+---
+---If it's `nil`, all spatializers will be tried in the following order: `phonon`, `oculus`, `simple`.
+---
+---
+---### NOTE:
+---Using a feature or effect that is not supported by the current spatializer will not error, it just won't do anything.
+---
+---<table>
+---  <thead>
+---    <tr>
+---      <td>Feature</td>
+---      <td>simple</td>
+---      <td>phonon</td>
+---      <td>oculus</td>
+---    </tr>
+---  </thead>
+---  <tbody>
+---    <tr>
+---      <td>Effect: Spatialization</td>
+---      <td>x</td>
+---      <td>x</td>
+---      <td>x</td>
+---    </tr>
+---    <tr>
+---      <td>Effect: Attenuation</td>
+---      <td>x</td>
+---      <td>x</td>
+---      <td></td>
+---    </tr>
+---    <tr>
+---      <td>Effect: Absorption</td>
+---      <td></td>
+---      <td>x</td>
+---      <td></td>
+---    </tr>
+---    <tr>
+---      <td>Effect: Occlusion</td>
+---      <td></td>
+---      <td>x</td>
+---      <td></td>
+---    </tr>
+---    <tr>
+---      <td>Effect: Transmission</td>
+---      <td></td>
+---      <td>x</td>
+---      <td></td>
+---    </tr>
+---    <tr>
+---      <td>Effect: Reverb</td>
+---      <td></td>
+---      <td>x</td>
+---      <td></td>
+---    </tr>
+---    <tr>
+---      <td>lovr.audio.setGeometry</td>
+---      <td></td>
+---      <td>x</td>
+---      <td></td>
+---    </tr>
+---    <tr>
+---      <td>Source:setDirectivity</td>
+---      <td>x</td>
+---      <td>x</td>
+---      <td></td>
+---    </tr>
+---    <tr>
+---      <td>Source:setRadius</td>
+---      <td></td>
+---      <td>x</td>
+---      <td></td>
+---    </tr>
+---  </tbody> </table>
 ---
 ---@return string spatializer # The name of the active spatializer.
 function lovr.audio.getSpatializer() end
 
 ---
----Returns the master volume.  All audio sent to the playback device has its volume multiplied by this factor.
+---Returns the master volume.
+---
+---All audio sent to the playback device has its volume multiplied by this factor.
+---
+---
+---### NOTE:
+---The default volume is 1.0 (0 dB).
 ---
 ---@param units? lovr.VolumeUnit # The units to return (linear or db).
 ---@return number volume # The master volume.
@@ -85,7 +183,21 @@ function lovr.audio.isStarted(type) end
 function lovr.audio.newSource(filename, options) end
 
 ---
----Sets the global air absorption coefficients for the medium.  This affects Sources that have the `absorption` effect enabled, causing audio volume to drop off with distance as it is absorbed by the medium it's traveling through (air, water, etc.).  The difference between absorption and falloff is that absorption is more subtle and is frequency-dependent, so higher-frequency bands can get absorbed more quickly than lower ones.  This can be used to apply "underwater" effects and stuff.
+---Sets the global air absorption coefficients for the medium.
+---
+---This affects Sources that have the `absorption` effect enabled, causing audio volume to drop off with distance as it is absorbed by the medium it's traveling through (air, water, etc.).
+---
+---The difference between absorption and falloff is that absorption is more subtle and is frequency-dependent, so higher-frequency bands can get absorbed more quickly than lower ones.
+---
+---This can be used to apply "underwater" effects and stuff.
+---
+---
+---### NOTE:
+---Absorption is currently only supported by the phonon spatializer.
+---
+---The frequency bands correspond to `400Hz`, `2.5KHz`, and `15KHz`.
+---
+---The default coefficients are `.0002`, `.0017`, and `.0182` for low, mid, and high.
 ---
 ---@param low number # The absorption coefficient for the low frequency band.
 ---@param mid number # The absorption coefficient for the mid frequency band.
@@ -95,13 +207,25 @@ function lovr.audio.setAbsorption(low, mid, high) end
 ---
 ---Switches either the playback or capture device to a new one.
 ---
----If a device for the given type is already active, it will be stopped and destroyed.  The new device will not be started automatically, use `lovr.audio.start` to start it.
+---If a device for the given type is already active, it will be stopped and destroyed.
+---
+---The new device will not be started automatically, use `lovr.audio.start` to start it.
 ---
 ---A device id (previously retrieved using `lovr.audio.getDevices`) can be given to use a specific audio device, or `nil` can be used for the id to use the default audio device.
 ---
----A sink can be also be provided when changing the device.  A sink is an audio stream (`Sound` object with a `stream` type) that will receive all audio samples played (for playback) or all audio samples captured (for capture).  When an audio device with a sink is started, be sure to periodically call `Sound:read` on the sink to read audio samples from it, otherwise it will overflow and discard old data.  The sink can have any format, data will be converted as needed. Using a sink for the playback device will reduce performance, but this isn't the case for capture devices.
+---A sink can be also be provided when changing the device.
+---
+---A sink is an audio stream (`Sound` object with a `stream` type) that will receive all audio samples played (for playback) or all audio samples captured (for capture).
 ---
----Audio devices can be started in `shared` or `exclusive` mode.  Exclusive devices may have lower latency than shared devices, but there's a higher chance that requesting exclusive access to an audio device will fail (either because it isn't supported or allowed).  One strategy is to first try the device in exclusive mode, switching to shared if it doesn't work.
+---When an audio device with a sink is started, be sure to periodically call `Sound:read` on the sink to read audio samples from it, otherwise it will overflow and discard old data.
+---
+---The sink can have any format, data will be converted as needed. Using a sink for the playback device will reduce performance, but this isn't the case for capture devices.
+---
+---Audio devices can be started in `shared` or `exclusive` mode.
+---
+---Exclusive devices may have lower latency than shared devices, but there's a higher chance that requesting exclusive access to an audio device will fail (either because it isn't supported or allowed).
+---
+---One strategy is to first try the device in exclusive mode, switching to shared if it doesn't work.
 ---
 ---@param type? lovr.AudioType # The device to switch.
 ---@param id? userdata # The id of the device to use, or `nil` to use the default device.
@@ -111,10 +235,26 @@ function lovr.audio.setAbsorption(low, mid, high) end
 function lovr.audio.setDevice(type, id, sink, mode) end
 
 ---
----Sets a mesh of triangles to use for modeling audio effects, using a table of vertices or a Model.  When the appropriate effects are enabled, audio from `Source` objects will correctly be occluded by walls and bounce around to create realistic reverb.
+---Sets a mesh of triangles to use for modeling audio effects, using a table of vertices or a Model.
+---
+---When the appropriate effects are enabled, audio from `Source` objects will correctly be occluded by walls and bounce around to create realistic reverb.
 ---
 ---An optional `AudioMaterial` may be provided to specify the acoustic properties of the geometry.
 ---
+---
+---### NOTE:
+---This is currently only supported/used by the `phonon` spatializer.
+---
+---The `Effect`s that use geometry are:
+---
+---- `occlusion`
+---- `reverb`
+---- `transmission`
+---
+---If an existing geometry has been set, this function will replace it.
+---
+---The triangles must use counterclockwise winding.
+---
 ---@overload fun(model: lovr.Model, material: lovr.AudioMaterial):boolean
 ---@param vertices table # A flat table of vertices.  Each vertex is 3 numbers representing its x, y, and z position. The units used for audio coordinates are up to you, but meters are recommended.
 ---@param indices table # A list of indices, indicating how the vertices are connected into triangles.  Indices are 1-indexed and are 32 bits (they can be bigger than 65535).
@@ -152,41 +292,80 @@ function lovr.audio.setPose(x, y, z, angle, ax, ay, az) end
 function lovr.audio.setPosition(x, y, z) end
 
 ---
----Sets the master volume.  All audio sent to the playback device has its volume multiplied by this factor.
+---Sets the master volume.
+---
+---All audio sent to the playback device has its volume multiplied by this factor.
+---
+---
+---### NOTE:
+---The volume will be clamped to a 0-1 range (0 dB).
 ---
 ---@param volume number # The master volume.
 ---@param units? lovr.VolumeUnit # The units of the value.
 function lovr.audio.setVolume(volume, units) end
 
 ---
----Starts the active playback or capture device.  By default the playback device is initialized and started, but this can be controlled using the `t.audio.start` flag in `lovr.conf`.
+---Starts the active playback or capture device.
+---
+---By default the playback device is initialized and started, but this can be controlled using the `t.audio.start` flag in `lovr.conf`.
+---
+---
+---### NOTE:
+---Starting an audio device may fail if:
+---
+---- The device is already started
+---- No device was initialized with `lovr.audio.setDevice`
+---- Lack of `audiocapture` permission on Android (see `lovr.system.requestPermission`)
+---- Some other problem accessing the audio device
 ---
 ---@param type? lovr.AudioType # The type of device to start.
 ---@return boolean started # Whether the device was successfully started.
 function lovr.audio.start(type) end
 
 ---
----Stops the active playback or capture device.  This may fail if:
+---Stops the active playback or capture device.
+---
+---This may fail if:
 ---
 ---- The device is not started
 ---- No device was initialized with `lovr.audio.setDevice`
 ---
+---
+---### NOTE:
+---Switching devices with `lovr.audio.setDevice` will stop the existing one.
+---
 ---@param type? lovr.AudioType # The type of device to stop.
 ---@return boolean stopped # Whether the device was successfully stopped.
 function lovr.audio.stop(type) end
 
 ---
----A Source is an object representing a single sound.  Currently ogg, wav, and mp3 formats are supported.
+---A Source is an object representing a single sound.
+---
+---Currently ogg, wav, and mp3 formats are supported.
+---
+---When a Source is playing, it will send audio to the speakers.
+---
+---Sources do not play automatically when they are created.
 ---
----When a Source is playing, it will send audio to the speakers.  Sources do not play automatically when they are created.  Instead, the `play`, `pause`, and `stop` functions can be used to control when they should play.
+---Instead, the `play`, `pause`, and `stop` functions can be used to control when they should play.
 ---
----`Source:seek` and `Source:tell` can be used to control the playback position of the Source.  A Source can be set to loop when it reaches the end using `Source:setLooping`.
+---`Source:seek` and `Source:tell` can be used to control the playback position of the Source.
+---
+---A Source can be set to loop when it reaches the end using `Source:setLooping`.
 ---
 ---@class lovr.Source
 local Source = {}
 
 ---
----Creates a copy of the Source, referencing the same `Sound` object and inheriting all of the settings of this Source.  However, it will be created in the stopped state and will be rewound to the beginning.
+---Creates a copy of the Source, referencing the same `Sound` object and inheriting all of the settings of this Source.
+---
+---However, it will be created in the stopped state and will be rewound to the beginning.
+---
+---
+---### NOTE:
+---This is a good way to create multiple Sources that play the same sound, since the audio data won't be loaded multiple times and can just be reused.
+---
+---You can also create multiple `Source` objects and pass in the same `Sound` object for each one, which will have the same effect.
 ---
 ---@return lovr.Source source # A genetically identical copy of the Source.
 function Source:clone() end
@@ -196,9 +375,21 @@ function Source:clone() end
 ---
 ---The directivity is controlled by two parameters: the weight and the power.
 ---
----The weight is a number between 0 and 1 controlling the general "shape" of the sound emitted. 0.0 results in a completely omnidirectional sound that can be heard from all directions.  1.0 results in a full dipole shape that can be heard only from the front and back.  0.5 results in a cardioid shape that can only be heard from one direction.  Numbers in between will smoothly transition between these.
+---The weight is a number between 0 and 1 controlling the general "shape" of the sound emitted. 0.0 results in a completely omnidirectional sound that can be heard from all directions.
+---
+---1.0 results in a full dipole shape that can be heard only from the front and back.
+---
+---0.5 results in a cardioid shape that can only be heard from one direction.
+---
+---Numbers in between will smoothly transition between these.
+---
+---The power is a number that controls how "focused" or sharp the shape is.
+---
+---Lower power values can be heard from a wider set of angles.
 ---
----The power is a number that controls how "focused" or sharp the shape is.  Lower power values can be heard from a wider set of angles.  It is an exponent, so it can get arbitrarily large.  Note that a power of zero will still result in an omnidirectional source, regardless of the weight.
+---It is an exponent, so it can get arbitrarily large.
+---
+---Note that a power of zero will still result in an omnidirectional source, regardless of the weight.
 ---
 ---@return number weight # The dipole weight.  0.0 is omnidirectional, 1.0 is a dipole, 0.5 is cardioid.
 ---@return number power # The dipole power, controlling how focused the directivity shape is.
@@ -233,7 +424,9 @@ function Source:getOrientation() end
 function Source:getPose() end
 
 ---
----Returns the position of the Source, in meters.  Setting the position will cause the Source to be distorted and attenuated based on its position relative to the listener.
+---Returns the position of the Source, in meters.
+---
+---Setting the position will cause the Source to be distorted and attenuated based on its position relative to the listener.
 ---
 ---@return number x # The x coordinate.
 ---@return number y # The y coordinate.
@@ -243,13 +436,23 @@ function Source:getPosition() end
 ---
 ---Returns the radius of the Source, in meters.
 ---
----This does not control falloff or attenuation.  It is only used for smoothing out occlusion.  If a Source doesn't have a radius, then when it becomes occluded by a wall its volume will instantly drop.  Giving the Source a radius that approximates its emitter's size will result in a smooth transition between audible and occluded, improving realism.
+---This does not control falloff or attenuation.
+---
+---It is only used for smoothing out occlusion.
+---
+---If a Source doesn't have a radius, then when it becomes occluded by a wall its volume will instantly drop.
+---
+---Giving the Source a radius that approximates its emitter's size will result in a smooth transition between audible and occluded, improving realism.
 ---
 ---@return number radius # The radius of the Source, in meters.
 function Source:getRadius() end
 
 ---
----Returns the `Sound` object backing the Source.  Multiple Sources can share one Sound, allowing its data to only be loaded once.  An easy way to do this sharing is by using `Source:clone`.
+---Returns the `Sound` object backing the Source.
+---
+---Multiple Sources can share one Sound, allowing its data to only be loaded once.
+---
+---An easy way to do this sharing is by using `Source:clone`.
 ---
 ---@return lovr.Sound sound # The Sound object.
 function Source:getSound() end
@@ -264,6 +467,18 @@ function Source:getVolume(units) end
 ---
 ---Returns whether a given `Effect` is enabled for the Source.
 ---
+---
+---### NOTE:
+---The active spatializer will determine which effects are supported.
+---
+---If an unsupported effect is enabled on a Source, no error will be reported.
+---
+---Instead, it will be silently ignored.
+---
+---See `lovr.audio.getSpatializer` for a table showing the effects supported by each spatializer.
+---
+---Calling this function on a Source that was created with `{ effects = false }` will always return false.
+---
 ---@param effect lovr.Effect # The effect.
 ---@return boolean enabled # Whether the effect is enabled.
 function Source:isEffectEnabled(effect) end
@@ -281,12 +496,22 @@ function Source:isLooping() end
 function Source:isPlaying() end
 
 ---
----Pauses the source.  It can be resumed with `Source:resume` or `Source:play`. If a paused source is rewound, it will remain paused.
+---Pauses the source.
+---
+---It can be resumed with `Source:resume` or `Source:play`. If a paused source is rewound, it will remain paused.
 ---
 function Source:pause() end
 
 ---
----Plays the Source.  This doesn't do anything if the Source is already playing.
+---Plays the Source.
+---
+---This doesn't do anything if the Source is already playing.
+---
+---
+---### NOTE:
+---There is a maximum of 64 Sources that can be playing at once.
+---
+---If 64 Sources are already playing, this function will return `false`.
 ---
 ---@return boolean success # Whether the Source successfully started playing.
 function Source:play() end
@@ -294,6 +519,10 @@ function Source:play() end
 ---
 ---Seeks the Source to the specified position.
 ---
+---
+---### NOTE:
+---Seeking a Source backed by a stream `Sound` has no meaningful effect.
+---
 ---@param position number # The position to seek to.
 ---@param unit? lovr.TimeUnit # The units for the seek position.
 function Source:seek(position, unit) end
@@ -303,9 +532,21 @@ function Source:seek(position, unit) end
 ---
 ---The directivity is controlled by two parameters: the weight and the power.
 ---
----The weight is a number between 0 and 1 controlling the general "shape" of the sound emitted. 0.0 results in a completely omnidirectional sound that can be heard from all directions.  1.0 results in a full dipole shape that can be heard only from the front and back.  0.5 results in a cardioid shape that can only be heard from one direction.  Numbers in between will smoothly transition between these.
+---The weight is a number between 0 and 1 controlling the general "shape" of the sound emitted. 0.0 results in a completely omnidirectional sound that can be heard from all directions.
+---
+---1.0 results in a full dipole shape that can be heard only from the front and back.
 ---
----The power is a number that controls how "focused" or sharp the shape is.  Lower power values can be heard from a wider set of angles.  It is an exponent, so it can get arbitrarily large.  Note that a power of zero will still result in an omnidirectional source, regardless of the weight.
+---0.5 results in a cardioid shape that can only be heard from one direction.
+---
+---Numbers in between will smoothly transition between these.
+---
+---The power is a number that controls how "focused" or sharp the shape is.
+---
+---Lower power values can be heard from a wider set of angles.
+---
+---It is an exponent, so it can get arbitrarily large.
+---
+---Note that a power of zero will still result in an omnidirectional source, regardless of the weight.
 ---
 ---@param weight number # The dipole weight.  0.0 is omnidirectional, 1.0 is a dipole, 0.5 is cardioid.
 ---@param power number # The dipole power, controlling how focused the directivity shape is.
@@ -314,6 +555,18 @@ function Source:setDirectivity(weight, power) end
 ---
 ---Enables or disables an effect on the Source.
 ---
+---
+---### NOTE:
+---The active spatializer will determine which effects are supported.
+---
+---If an unsupported effect is enabled on a Source, no error will be reported.
+---
+---Instead, it will be silently ignored.
+---
+---See `lovr.audio.getSpatializer` for a table showing the effects supported by each spatializer.
+---
+---Calling this function on a Source that was created with `{ effects = false }` will throw an error.
+---
 ---@param effect lovr.Effect # The effect.
 ---@param enable boolean # Whether the effect should be enabled.
 function Source:setEffectEnabled(effect, enable) end
@@ -321,6 +574,10 @@ function Source:setEffectEnabled(effect, enable) end
 ---
 ---Sets whether or not the Source loops.
 ---
+---
+---### NOTE:
+---Attempting to loop a Source backed by a stream `Sound` will cause an error.
+---
 ---@param loop boolean # Whether or not the Source will loop.
 function Source:setLooping(loop) end
 
@@ -346,9 +603,13 @@ function Source:setOrientation(angle, ax, ay, az) end
 function Source:setPose(x, y, z, angle, ax, ay, az) end
 
 ---
----Sets the position of the Source, in meters.  Setting the position will cause the Source to be distorted and attenuated based on its position relative to the listener.
+---Sets the position of the Source, in meters.
+---
+---Setting the position will cause the Source to be distorted and attenuated based on its position relative to the listener.
 ---
----Only mono sources can be positioned.  Setting the position of a stereo Source will cause an error.
+---Only mono sources can be positioned.
+---
+---Setting the position of a stereo Source will cause an error.
 ---
 ---@param x number # The x coordinate.
 ---@param y number # The y coordinate.
@@ -358,7 +619,13 @@ function Source:setPosition(x, y, z) end
 ---
 ---Sets the radius of the Source, in meters.
 ---
----This does not control falloff or attenuation.  It is only used for smoothing out occlusion.  If a Source doesn't have a radius, then when it becomes occluded by a wall its volume will instantly drop.  Giving the Source a radius that approximates its emitter's size will result in a smooth transition between audible and occluded, improving realism.
+---This does not control falloff or attenuation.
+---
+---It is only used for smoothing out occlusion.
+---
+---If a Source doesn't have a radius, then when it becomes occluded by a wall its volume will instantly drop.
+---
+---Giving the Source a radius that approximates its emitter's size will result in a smooth transition between audible and occluded, improving realism.
 ---
 ---@param radius number # The new radius of the Source, in meters.
 function Source:setRadius(radius) end
@@ -366,6 +633,10 @@ function Source:setRadius(radius) end
 ---
 ---Sets the current volume factor for the Source.
 ---
+---
+---### NOTE:
+---The volume will be clamped to a 0-1 range (0 dB).
+---
 ---@param volume number # The new volume.
 ---@param units? lovr.VolumeUnit # The units of the value.
 function Source:setVolume(volume, units) end
@@ -378,6 +649,10 @@ function Source:stop() end
 ---
 ---Returns the current playback position of the Source.
 ---
+---
+---### NOTE:
+---The return value for Sources backed by a stream `Sound` has no meaning.
+---
 ---@param unit? lovr.TimeUnit # The unit to return.
 ---@return number position # The current playback position.
 function Source:tell(unit) end
@@ -406,15 +681,23 @@ function Source:tell(unit) end
 ---Concrete.
 ---
 ---| '"concrete"'
+
 ---| '"glass"'
+
 ---| '"gravel"'
+
 ---| '"metal"'
+
 ---| '"plaster"'
+
 ---| '"rock"'
+
 ---| '"wood"'
 
 ---
----Audio devices can be created in shared mode or exclusive mode.  In exclusive mode, the audio device is the only one active on the system, which gives better performance and lower latency. However, exclusive devices aren't always supported and might not be allowed, so there is a higher chance that creating one will fail.
+---Audio devices can be created in shared mode or exclusive mode.
+---
+---In exclusive mode, the audio device is the only one active on the system, which gives better performance and lower latency. However, exclusive devices aren't always supported and might not be allowed, so there is a higher chance that creating one will fail.
 ---
 ---@alias lovr.AudioShareMode
 ---
@@ -442,6 +725,16 @@ function Source:tell(unit) end
 ---
 ---Different types of effects that can be applied with `Source:setEffectEnabled`.
 ---
+---
+---### NOTE:
+---The active spatializer will determine which effects are supported.
+---
+---If an unsupported effect is enabled on a Source, no error will be reported.
+---
+---Instead, it will be silently ignored.
+---
+---TODO: expose a table of supported effects for spatializers in docs or from Lua.
+---
 ---@alias lovr.Effect
 ---
 ---Models absorption as sound travels through the air, water, etc.
@@ -469,7 +762,9 @@ function Source:tell(unit) end
 ---| '"transmission"'
 
 ---
----When figuring out how long a Source is or seeking to a specific position in the sound file, units can be expressed in terms of seconds or in terms of frames.  A frame is one set of samples for each channel (one sample for mono, two samples for stereo).
+---When figuring out how long a Source is or seeking to a specific position in the sound file, units can be expressed in terms of seconds or in terms of frames.
+---
+---A frame is one set of samples for each channel (one sample for mono, two samples for stereo).
 ---
 ---@alias lovr.TimeUnit
 ---
diff --git a/meta/3rd/lovr/library/lovr.data.lua b/meta/3rd/lovr/library/lovr.data.lua
index 9c34b189..a699a2f3 100644
--- a/meta/3rd/lovr/library/lovr.data.lua
+++ b/meta/3rd/lovr/library/lovr.data.lua
@@ -17,7 +17,17 @@ lovr.data = {}
 function lovr.data.newBlob(size, name) end
 
 ---
----Creates a new Image.  Image data can be loaded and decoded from an image file, or a raw block of pixels with a specified width, height, and format can be created.
+---Creates a new Image.
+---
+---Image data can be loaded and decoded from an image file, or a raw block of pixels with a specified width, height, and format can be created.
+---
+---
+---### NOTE:
+---The supported image file formats are png, jpg, hdr, dds (DXT1, DXT3, DXT5), ktx, and astc.
+---
+---Only 2D textures are supported for DXT/ASTC.
+---
+---Currently textures loaded as KTX need to be in DXT/ASTC formats.
 ---
 ---@overload fun(width: number, height: number, format: lovr.TextureFormat, data: lovr.Blob):lovr.Image
 ---@overload fun(source: lovr.Image):lovr.Image
@@ -28,7 +38,9 @@ function lovr.data.newBlob(size, name) end
 function lovr.data.newImage(filename, flip) end
 
 ---
----Loads a 3D model from a file.  The supported 3D file formats are OBJ and glTF.
+---Loads a 3D model from a file.
+---
+---The supported 3D file formats are OBJ and glTF.
 ---
 ---@overload fun(blob: lovr.Blob):lovr.ModelData
 ---@param filename string # The filename of the model to load.
@@ -45,11 +57,43 @@ function lovr.data.newModelData(filename) end
 function lovr.data.newRasterizer(size) end
 
 ---
----Creates a new Sound.  A sound can be loaded from an audio file, or it can be created empty with capacity for a certain number of audio frames.
+---Creates a new Sound.
+---
+---A sound can be loaded from an audio file, or it can be created empty with capacity for a certain number of audio frames.
 ---
 ---When loading audio from a file, use the `decode` option to control whether compressed audio should remain compressed or immediately get decoded to raw samples.
 ---
----When creating an empty sound, the `contents` parameter can be set to `'stream'` to create an audio stream.  On streams, `Sound:setFrames` will always write to the end of the stream, and `Sound:getFrames` will always read the oldest samples from the beginning.  The number of frames in the sound is the total capacity of the stream's buffer.
+---When creating an empty sound, the `contents` parameter can be set to `'stream'` to create an audio stream.
+---
+---On streams, `Sound:setFrames` will always write to the end of the stream, and `Sound:getFrames` will always read the oldest samples from the beginning.
+---
+---The number of frames in the sound is the total capacity of the stream's buffer.
+---
+---
+---### NOTE:
+---It is highly recommended to use an audio format that matches the format of the audio module: `f32` sample formats at a sample rate of 48000, with 1 channel for spatialized sources or 2 channels for unspatialized sources.
+---
+---This will avoid the need to convert audio during playback, which boosts performance of the audio thread.
+---
+---The WAV importer supports 16, 24, and 32 bit integer data and 32 bit floating point data.
+---
+---The data must be mono, stereo, or 4-channel full-sphere ambisonic.
+---
+---The `WAVE_FORMAT_EXTENSIBLE` extension is supported.
+---
+---Ambisonic channel layouts are supported for import (but not yet for playback).
+---
+---Ambisonic data can be loaded from WAV files.
+---
+---It must be first-order full-sphere ambisonic data with 4 channels.
+---
+---If the WAV has a `WAVE_FORMAT_EXTENSIBLE` chunk with an `AMBISONIC_B_FORMAT` format GUID, then the data is understood as using the AMB format with Furse-Malham channel ordering and normalization.
+---
+---*All other* 4-channel files are assumed to be using the AmbiX format with ACN channel ordering and SN3D normalization.
+---
+---AMB files will get automatically converted to AmbiX on import, so ambisonic Sounds will always be in a consistent format.
+---
+---OGG and MP3 files will always have the `f32` format when loaded.
 ---
 ---@overload fun(filename: string, decode: boolean):lovr.Sound
 ---@overload fun(blob: lovr.Blob, decode: boolean):lovr.Sound
@@ -62,7 +106,11 @@ function lovr.data.newRasterizer(size) end
 function lovr.data.newSound(frames, format, channels, sampleRate, contents) end
 
 ---
----A Blob is an object that holds binary data.  It can be passed to most functions that take filename arguments, like `lovr.graphics.newModel` or `lovr.audio.newSource`.  Blobs aren't usually necessary for simple projects, but they can be really helpful if:
+---A Blob is an object that holds binary data.
+---
+---It can be passed to most functions that take filename arguments, like `lovr.graphics.newModel` or `lovr.audio.newSource`.
+---
+---Blobs aren't usually necessary for simple projects, but they can be really helpful if:
 ---
 ---- You need to work with low level binary data, potentially using the LuaJIT FFI for increased
 ---  performance.
@@ -76,13 +124,19 @@ function lovr.data.newSound(frames, format, channels, sampleRate, contents) end
 local Blob = {}
 
 ---
----Returns the filename the Blob was loaded from, or the custom name given to it when it was created.  This label is also used in error messages.
+---Returns the filename the Blob was loaded from, or the custom name given to it when it was created.
+---
+---This label is also used in error messages.
 ---
 ---@return string name # The name of the Blob.
 function Blob:getName() end
 
 ---
----Returns a raw pointer to the Blob's data.  This can be used to interface with other C libraries using the LuaJIT FFI.  Use this only if you know what you're doing!
+---Returns a raw pointer to the Blob's data.
+---
+---This can be used to interface with other C libraries using the LuaJIT FFI.
+---
+---Use this only if you know what you're doing!
 ---
 ---@return userdata pointer # A pointer to the data.
 function Blob:getPointer() end
@@ -100,7 +154,13 @@ function Blob:getSize() end
 function Blob:getString() end
 
 ---
----An Image stores raw 2D pixel info for `Texture`s.  It has a width, height, and format.  The Image can be initialized with the contents of an image file or it can be created with uninitialized contents.  The supported image formats are `png`, `jpg`, `hdr`, `dds`, `ktx`, and `astc`.
+---An Image stores raw 2D pixel info for `Texture`s.
+---
+---It has a width, height, and format.
+---
+---The Image can be initialized with the contents of an image file or it can be created with uninitialized contents.
+---
+---The supported image formats are `png`, `jpg`, `hdr`, `dds`, `ktx`, and `astc`.
 ---
 ---Usually you can just use Textures, but Image can be useful if you want to manipulate individual pixels, load Textures in a background thread, or use the FFI to efficiently access the raw image data.
 ---
@@ -108,7 +168,9 @@ function Blob:getString() end
 local Image = {}
 
 ---
----Encodes the Image to an uncompressed png.  This intended mainly for debugging.
+---Encodes the Image to an uncompressed png.
+---
+---This intended mainly for debugging.
 ---
 ---@return lovr.Blob blob # A new Blob containing the PNG image data.
 function Image:encode() end
@@ -141,6 +203,10 @@ function Image:getHeight() end
 ---
 ---Returns the value of a pixel of the Image.
 ---
+---
+---### NOTE:
+---The following texture formats are supported: `rgba`, `rgb`, `r32f`, `rg32f`, and `rgba32f`.
+---
 ---@param x number # The x coordinate of the pixel to get (0-indexed).
 ---@param y number # The y coordinate of the pixel to get (0-indexed).
 ---@return number r # The red component of the pixel, from 0.0 to 1.0.
@@ -158,6 +224,14 @@ function Image:getWidth() end
 ---
 ---Copies a rectangle of pixels from one Image to this one.
 ---
+---
+---### NOTE:
+---The two Images must have the same pixel format.
+---
+---Compressed images cannot be copied.
+---
+---The rectangle cannot go outside the dimensions of the source or destination textures.
+---
 ---@param source lovr.Image # The Image to copy pixels from.
 ---@param x? number # The x coordinate to paste to (0-indexed).
 ---@param y? number # The y coordinate to paste to (0-indexed).
@@ -170,6 +244,10 @@ function Image:paste(source, x, y, fromX, fromY, width, height) end
 ---
 ---Sets the value of a pixel of the Image.
 ---
+---
+---### NOTE:
+---The following texture formats are supported: `rgba`, `rgb`, `r32f`, `rg32f`, and `rgba32f`.
+---
 ---@param x number # The x coordinate of the pixel to set (0-indexed).
 ---@param y number # The y coordinate of the pixel to set (0-indexed).
 ---@param r number # The red component of the pixel, from 0.0 to 1.0.
@@ -179,7 +257,9 @@ function Image:paste(source, x, y, fromX, fromY, width, height) end
 function Image:setPixel(x, y, r, g, b, a) end
 
 ---
----A ModelData is a container object that loads and holds data contained in 3D model files.  This can include a variety of things like the node structure of the asset, the vertex data it contains, contains, the `Image` and `Material` properties, and any included animations.
+---A ModelData is a container object that loads and holds data contained in 3D model files.
+---
+---This can include a variety of things like the node structure of the asset, the vertex data it contains, contains, the `Image` and `Material` properties, and any included animations.
 ---
 ---The current supported formats are OBJ, glTF, and STL.
 ---
@@ -197,19 +277,27 @@ local ModelData = {}
 local Rasterizer = {}
 
 ---
----Returns the advance metric of the font, in pixels.  The advance is how many pixels the font advances horizontally after each glyph is rendered.  This does not include kerning.
+---Returns the advance metric of the font, in pixels.
+---
+---The advance is how many pixels the font advances horizontally after each glyph is rendered.
+---
+---This does not include kerning.
 ---
 ---@return number advance # The advance of the font, in pixels.
 function Rasterizer:getAdvance() end
 
 ---
----Returns the ascent metric of the font, in pixels.  The ascent represents how far any glyph of the font ascends above the baseline.
+---Returns the ascent metric of the font, in pixels.
+---
+---The ascent represents how far any glyph of the font ascends above the baseline.
 ---
 ---@return number ascent # The ascent of the font, in pixels.
 function Rasterizer:getAscent() end
 
 ---
----Returns the descent metric of the font, in pixels.  The descent represents how far any glyph of the font descends below the baseline.
+---Returns the descent metric of the font, in pixels.
+---
+---The descent represents how far any glyph of the font descends below the baseline.
 ---
 ---@return number descent # The descent of the font, in pixels.
 function Rasterizer:getDescent() end
@@ -227,7 +315,9 @@ function Rasterizer:getGlyphCount() end
 function Rasterizer:getHeight() end
 
 ---
----Returns the line height metric of the font, in pixels.  This is how far apart lines are.
+---Returns the line height metric of the font, in pixels.
+---
+---This is how far apart lines are.
 ---
 ---@return number height # The line height of the font, in pixels.
 function Rasterizer:getLineHeight() end
@@ -239,31 +329,75 @@ function Rasterizer:getLineHeight() end
 function Rasterizer:hasGlyphs() end
 
 ---
----A Sound stores the data for a sound.  The supported sound formats are OGG, WAV, and MP3.  Sounds cannot be played directly.  Instead, there are `Source` objects in `lovr.audio` that are used for audio playback.  All Source objects are backed by one of these Sounds, and multiple Sources can share a single Sound to reduce memory usage.
+---A Sound stores the data for a sound.
+---
+---The supported sound formats are OGG, WAV, and MP3.
+---
+---Sounds cannot be played directly.
+---
+---Instead, there are `Source` objects in `lovr.audio` that are used for audio playback.
+---
+---All Source objects are backed by one of these Sounds, and multiple Sources can share a single Sound to reduce memory usage.
 ---
 ---Metadata
 ------
 ---
----Sounds hold a fixed number of frames.  Each frame contains one audio sample for each channel. The `SampleFormat` of the Sound is the data type used for each sample (floating point, integer, etc.).  The Sound has a `ChannelLayout`, representing the number of audio channels and how they map to speakers (mono, stereo, etc.).  The sample rate of the Sound indicates how many frames should be played per second.  The duration of the sound (in seconds) is the number of frames divided by the sample rate.
+---Sounds hold a fixed number of frames.
+---
+---Each frame contains one audio sample for each channel. The `SampleFormat` of the Sound is the data type used for each sample (floating point, integer, etc.).
+---
+---The Sound has a `ChannelLayout`, representing the number of audio channels and how they map to speakers (mono, stereo, etc.).
+---
+---The sample rate of the Sound indicates how many frames should be played per second.
+---
+---The duration of the sound (in seconds) is the number of frames divided by the sample rate.
 ---
 ---Compression
 ------
 ---
----Sounds can be compressed.  Compressed sounds are stored compressed in memory and are decoded as they are played.  This uses a lot less memory but increases CPU usage during playback.  OGG and MP3 are compressed audio formats.  When creating a sound from a compressed format, there is an option to immediately decode it, storing it uncompressed in memory.  It can be a good idea to decode short sound effects, since they won't use very much memory even when uncompressed and it will improve CPU usage.  Compressed sounds can not be written to using `Sound:setFrames`.
+---Sounds can be compressed.
+---
+---Compressed sounds are stored compressed in memory and are decoded as they are played.
+---
+---This uses a lot less memory but increases CPU usage during playback.
+---
+---OGG and MP3 are compressed audio formats.
+---
+---When creating a sound from a compressed format, there is an option to immediately decode it, storing it uncompressed in memory.
+---
+---It can be a good idea to decode short sound effects, since they won't use very much memory even when uncompressed and it will improve CPU usage.
+---
+---Compressed sounds can not be written to using `Sound:setFrames`.
 ---
 ---Streams
 ------
 ---
----Sounds can be created as a stream by passing `'stream'` as their contents when creating them. Audio frames can be written to the end of the stream, and read from the beginning.  This works well for situations where data is being generated in real time or streamed in from some other data source.
+---Sounds can be created as a stream by passing `'stream'` as their contents when creating them. Audio frames can be written to the end of the stream, and read from the beginning.
+---
+---This works well for situations where data is being generated in real time or streamed in from some other data source.
 ---
----Sources can be backed by a stream and they'll just play whatever audio is pushed to the stream. The audio module also lets you use a stream as a "sink" for an audio device.  For playback devices, this works like loopback, so the mixed audio from all playing Sources will get written to the stream.  For capture devices, all the microphone input will get written to the stream. Conversion between sample formats, channel layouts, and sample rates will happen automatically.
+---Sources can be backed by a stream and they'll just play whatever audio is pushed to the stream. The audio module also lets you use a stream as a "sink" for an audio device.
 ---
----Keep in mind that streams can still only hold a fixed number of frames.  If too much data is written before it is read, older frames will start to get overwritten.  Similary, it's possible to read too much data without writing fast enough.
+---For playback devices, this works like loopback, so the mixed audio from all playing Sources will get written to the stream.
+---
+---For capture devices, all the microphone input will get written to the stream. Conversion between sample formats, channel layouts, and sample rates will happen automatically.
+---
+---Keep in mind that streams can still only hold a fixed number of frames.
+---
+---If too much data is written before it is read, older frames will start to get overwritten.
+---
+---Similary, it's possible to read too much data without writing fast enough.
 ---
 ---Ambisonics
 ------
 ---
----Ambisonic sounds can be imported from WAVs, but can not yet be played.  Sounds with a `ChannelLayout` of `ambisonic` are stored as first-order full-sphere ambisonics using the AmbiX format (ACN channel ordering and SN3D channel normalization).  The AMB format is supported for import and will automatically get converted to AmbiX.  See `lovr.data.newSound` for more info.
+---Ambisonic sounds can be imported from WAVs, but can not yet be played.
+---
+---Sounds with a `ChannelLayout` of `ambisonic` are stored as first-order full-sphere ambisonics using the AmbiX format (ACN channel ordering and SN3D channel normalization).
+---
+---The AMB format is supported for import and will automatically get converted to AmbiX.
+---
+---See `lovr.data.newSound` for more info.
 ---
 ---@class lovr.Sound
 local Sound = {}
@@ -271,11 +405,19 @@ local Sound = {}
 ---
 ---Returns a Blob containing the raw bytes of the Sound.
 ---
+---
+---### NOTE:
+---Samples for each channel are stored interleaved.
+---
+---The data type of each sample is given by `Sound:getFormat`.
+---
 ---@return lovr.Blob blob # The Blob instance containing the bytes for the `Sound`.
 function Sound:getBlob() end
 
 ---
----Returns the number of channels in the Sound.  Mono sounds have 1 channel, stereo sounds have 2 channels, and ambisonic sounds have 4 channels.
+---Returns the number of channels in the Sound.
+---
+---Mono sounds have 1 channel, stereo sounds have 2 channels, and ambisonic sounds have 4 channels.
 ---
 ---@return number channels # The number of channels in the sound.
 function Sound:getChannelCount() end
@@ -289,6 +431,10 @@ function Sound:getChannelLayout() end
 ---
 ---Returns the duration of the Sound, in seconds.
 ---
+---
+---### NOTE:
+---This can be computed as `(frameCount / sampleRate)`.
+---
 ---@return number duration # The duration of the Sound, in seconds.
 function Sound:getDuration() end
 
@@ -299,7 +445,13 @@ function Sound:getDuration() end
 function Sound:getFormat() end
 
 ---
----Returns the number of frames in the Sound.  A frame stores one sample for each channel.
+---Returns the number of frames in the Sound.
+---
+---A frame stores one sample for each channel.
+---
+---
+---### NOTE:
+---For streams, this returns the number of frames in the stream's buffer.
 ---
 ---@return number frames # The number of frames in the Sound.
 function Sound:getFrameCount() end
@@ -323,17 +475,31 @@ function Sound:getFrames(count, srcOffset) end
 ---
 ---Returns the total number of samples in the Sound.
 ---
+---
+---### NOTE:
+---For streams, this returns the number of samples in the stream's buffer.
+---
 ---@return number samples # The total number of samples in the Sound.
 function Sound:getSampleCount() end
 
 ---
----Returns the sample rate of the Sound, in Hz.  This is the number of frames that are played every second.  It's usually a high number like 48000.
+---Returns the sample rate of the Sound, in Hz.
+---
+---This is the number of frames that are played every second.
+---
+---It's usually a high number like 48000.
 ---
 ---@return number frequency # The number of frames per second in the Sound.
 function Sound:getSampleRate() end
 
 ---
----Returns whether the Sound is compressed.  Compressed sounds are loaded from compressed audio formats like MP3 and OGG.  They use a lot less memory but require some extra CPU work during playback.  Compressed sounds can not be modified using `Sound:setFrames`.
+---Returns whether the Sound is compressed.
+---
+---Compressed sounds are loaded from compressed audio formats like MP3 and OGG.
+---
+---They use a lot less memory but require some extra CPU work during playback.
+---
+---Compressed sounds can not be modified using `Sound:setFrames`.
 ---
 ---@return boolean compressed # Whether the Sound is compressed.
 function Sound:isCompressed() end
@@ -365,11 +531,17 @@ function Sound:setFrames(t, count, dstOffset, srcOffset) end
 ---
 ---| '"mono"'
 ---
----2 channels.  The first channel is for the left speaker and the second is for the right.
+---2 channels.
+---
+---The first channel is for the left speaker and the second is for the right.
 ---
 ---| '"stereo"'
 ---
----4 channels.  Ambisonic channels don't map directly to speakers but instead represent directions in 3D space, sort of like the images of a skybox.  Currently, ambisonic sounds can only be loaded, not played.
+---4 channels.
+---
+---Ambisonic channels don't map directly to speakers but instead represent directions in 3D space, sort of like the images of a skybox.
+---
+---Currently, ambisonic sounds can only be loaded, not played.
 ---
 ---| '"ambisonic"'
 
diff --git a/meta/3rd/lovr/library/lovr.event.lua b/meta/3rd/lovr/library/lovr.event.lua
index 5beeb5a1..f2e9509d 100644
--- a/meta/3rd/lovr/library/lovr.event.lua
+++ b/meta/3rd/lovr/library/lovr.event.lua
@@ -5,6 +5,12 @@
 ---
 ---Due to its low-level nature, it's rare to use `lovr.event` in simple projects.
 ---
+---
+---### NOTE:
+---You can define your own custom events by adding a function to the `lovr.handlers` table with a key of the name of the event you want to add.
+---
+---Then, push the event using `lovr.event.push`.
+---
 ---@class lovr.event
 lovr.event = {}
 
@@ -14,24 +20,52 @@ lovr.event = {}
 function lovr.event.clear() end
 
 ---
----This function returns a Lua iterator for all of the unprocessed items in the event queue.  Each event consists of a name as a string, followed by event-specific arguments.  This function is called in the default implementation of `lovr.run`, so it is normally not necessary to poll for events yourself.
+---This function returns a Lua iterator for all of the unprocessed items in the event queue.
+---
+---Each event consists of a name as a string, followed by event-specific arguments.
+---
+---This function is called in the default implementation of `lovr.run`, so it is normally not necessary to poll for events yourself.
 ---
 ---@return function iterator # The iterator function, usable in a for loop.
 function lovr.event.poll() end
 
 ---
----Fills the event queue with unprocessed events from the operating system.  This function should be called often, otherwise the operating system will consider the application unresponsive. This function is called in the default implementation of `lovr.run`.
+---Fills the event queue with unprocessed events from the operating system.
+---
+---This function should be called often, otherwise the operating system will consider the application unresponsive. This function is called in the default implementation of `lovr.run`.
 ---
 function lovr.event.pump() end
 
 ---
----Pushes an event onto the event queue.  It will be processed the next time `lovr.event.poll` is called.  For an event to be processed properly, there needs to be a function in the `lovr.handlers` table with a key that's the same as the event name.
+---Pushes an event onto the event queue.
+---
+---It will be processed the next time `lovr.event.poll` is called.
+---
+---For an event to be processed properly, there needs to be a function in the `lovr.handlers` table with a key that's the same as the event name.
+---
+---
+---### NOTE:
+---Only nil, booleans, numbers, strings, and LÖVR objects are supported types for event data.
 ---
 ---@param name string # The name of the event.
 function lovr.event.push(name) end
 
 ---
----Pushes an event to quit.  An optional number can be passed to set the exit code for the application.  An exit code of zero indicates normal termination, whereas a nonzero exit code indicates that an error occurred.
+---Pushes an event to quit.
+---
+---An optional number can be passed to set the exit code for the application.
+---
+---An exit code of zero indicates normal termination, whereas a nonzero exit code indicates that an error occurred.
+---
+---
+---### NOTE:
+---This function is equivalent to calling `lovr.event.push('quit', <args>)`.
+---
+---The event won't be processed until the next time `lovr.event.poll` is called.
+---
+---The `lovr.quit` callback will be called when the event is processed, which can be used to do any cleanup work.
+---
+---The callback can also return `false` to abort the quitting process.
 ---
 ---@param code? number # The exit code of the program.
 function lovr.event.quit(code) end
@@ -39,10 +73,18 @@ function lovr.event.quit(code) end
 ---
 ---Pushes an event to restart the framework.
 ---
+---
+---### NOTE:
+---The event won't be processed until the next time `lovr.event.poll` is called.
+---
+---The `lovr.restart` callback can be used to persist a value between restarts.
+---
 function lovr.event.restart() end
 
 ---
----Keys that can be pressed on a keyboard.  Notably, numpad keys are missing right now.
+---Keys that can be pressed on a keyboard.
+---
+---Notably, numpad keys are missing right now.
 ---
 ---@alias lovr.KeyCode
 ---
diff --git a/meta/3rd/lovr/library/lovr.filesystem.lua b/meta/3rd/lovr/library/lovr.filesystem.lua
index 37fe8e0e..4f8214db 100644
--- a/meta/3rd/lovr/library/lovr.filesystem.lua
+++ b/meta/3rd/lovr/library/lovr.filesystem.lua
@@ -3,12 +3,40 @@
 ---
 ---The `lovr.filesystem` module provides access to the filesystem.
 ---
+---
+---### NOTE:
+---LÖVR programs can only write to a single directory, called the save directory.
+---
+---The location of the save directory is platform-specific:
+---
+---<table>
+---  <tr>
+---    <td>Windows</td>
+---    <td><code>C:\Users\&lt;user&gt;\AppData\Roaming\LOVR\&lt;identity&gt;</code></td>
+---  </tr>
+---  <tr>
+---    <td>macOS</td>
+---    <td><code>/Users/&lt;user&gt;/Library/Application Support/LOVR/&lt;identity&gt;</code></td>
+---  </tr> </table>
+---
+---`<identity>` should be a unique identifier for your app.
+---
+---It can be set either in `lovr.conf` or by using `lovr.filesystem.setIdentity`.
+---
+---All filenames are relative to either the save directory or the directory containing the project source.
+---
+---Files in the save directory take precedence over files in the project.
+---
 ---@class lovr.filesystem
 lovr.filesystem = {}
 
 ---
 ---Appends content to the end of a file.
 ---
+---
+---### NOTE:
+---If the file does not exist, it is created.
+---
 ---@overload fun(filename: string, blob: lovr.Blob):number
 ---@param filename string # The file to append to.
 ---@param content string # A string to write to the end of the file.
@@ -16,14 +44,18 @@ lovr.filesystem = {}
 function lovr.filesystem.append(filename, content) end
 
 ---
----Creates a directory in the save directory.  Any parent directories that don't exist will also be created.
+---Creates a directory in the save directory.
+---
+---Any parent directories that don't exist will also be created.
 ---
 ---@param path string # The directory to create, recursively.
 ---@return boolean success # Whether the directory was created.
 function lovr.filesystem.createDirectory(path) end
 
 ---
----Returns the application data directory.  This will be something like:
+---Returns the application data directory.
+---
+---This will be something like:
 ---
 ---- `C:\Users\user\AppData\Roaming` on Windows.
 ---- `/home/user/.config` on Linux.
@@ -35,6 +67,10 @@ function lovr.filesystem.getAppdataDirectory() end
 ---
 ---Returns a sorted table containing all files and folders in a single directory.
 ---
+---
+---### NOTE:
+---This function calls `table.sort` to sort the results, so if `table.sort` is not available in the global scope the results are not guaranteed to be sorted.
+---
 ---@param path string # The directory.
 ---@return lovr.items table # A table with a string for each file and subfolder in the directory.
 function lovr.filesystem.getDirectoryItems(path) end
@@ -46,7 +82,15 @@ function lovr.filesystem.getDirectoryItems(path) end
 function lovr.filesystem.getExecutablePath() end
 
 ---
----Returns the identity of the game, which is used as the name of the save directory.  The default is `default`.  It can be changed using `t.identity` in `lovr.conf`.
+---Returns the identity of the game, which is used as the name of the save directory.
+---
+---The default is `default`.
+---
+---It can be changed using `t.identity` in `lovr.conf`.
+---
+---
+---### NOTE:
+---On Android, this is always the package id (like `org.lovr.app`).
 ---
 ---@return string identity # The name of the save directory, or `nil` if it isn't set.
 function lovr.filesystem.getIdentity() end
@@ -59,14 +103,26 @@ function lovr.filesystem.getIdentity() end
 function lovr.filesystem.getLastModified(path) end
 
 ---
----Get the absolute path of the mounted archive containing a path in the virtual filesystem.  This can be used to determine if a file is in the game's source directory or the save directory.
+---Get the absolute path of the mounted archive containing a path in the virtual filesystem.
+---
+---This can be used to determine if a file is in the game's source directory or the save directory.
 ---
 ---@param path string # The path to check.
 ---@return string realpath # The absolute path of the mounted archive containing `path`.
 function lovr.filesystem.getRealDirectory(path) end
 
 ---
----Returns the require path.  The require path is a semicolon-separated list of patterns that LÖVR will use to search for files when they are `require`d.  Any question marks in the pattern will be replaced with the module that is being required.  It is similar to Lua\'s `package.path` variable, but the main difference is that the patterns are relative to the virtual filesystem.
+---Returns the require path.
+---
+---The require path is a semicolon-separated list of patterns that LÖVR will use to search for files when they are `require`d.
+---
+---Any question marks in the pattern will be replaced with the module that is being required.
+---
+---It is similar to Lua\'s `package.path` variable, but the main difference is that the patterns are relative to the virtual filesystem.
+---
+---
+---### NOTE:
+---The default reqiure path is '?.lua;?/init.lua'.
 ---
 ---@return string path # The semicolon separated list of search patterns.
 function lovr.filesystem.getRequirePath() end
@@ -74,12 +130,24 @@ function lovr.filesystem.getRequirePath() end
 ---
 ---Returns the absolute path to the save directory.
 ---
+---
+---### NOTE:
+---The save directory takes the following form:
+---
+---``` <appdata>/LOVR/<identity> ```
+---
+---Where `<appdata>` is `lovr.filesystem.getAppdataDirectory` and `<identity>` is `lovr.filesystem.getIdentity` and can be customized using `lovr.conf`.
+---
 ---@return string path # The absolute path to the save directory.
 function lovr.filesystem.getSaveDirectory() end
 
 ---
 ---Returns the size of a file, in bytes.
 ---
+---
+---### NOTE:
+---If the file does not exist, an error is thrown.
+---
 ---@param file string # The file.
 ---@return number size # The size of the file, in bytes.
 function lovr.filesystem.getSize(file) end
@@ -97,7 +165,9 @@ function lovr.filesystem.getSource() end
 function lovr.filesystem.getUserDirectory() end
 
 ---
----Returns the absolute path of the working directory.  Usually this is where the executable was started from.
+---Returns the absolute path of the working directory.
+---
+---Usually this is where the executable was started from.
 ---
 ---@return string path # The current working directory, or `nil` if it's unknown.
 function lovr.filesystem.getWorkingDirectory() end
@@ -125,12 +195,26 @@ function lovr.filesystem.isFused() end
 ---
 ---Load a file containing Lua code, returning a Lua chunk that can be run.
 ---
+---
+---### NOTE:
+---An error is thrown if the file contains syntax errors.
+---
 ---@param filename string # The file to load.
 ---@return function chunk # The runnable chunk.
 function lovr.filesystem.load(filename) end
 
 ---
----Mounts a directory or `.zip` archive, adding it to the virtual filesystem.  This allows you to read files from it.
+---Mounts a directory or `.zip` archive, adding it to the virtual filesystem.
+---
+---This allows you to read files from it.
+---
+---
+---### NOTE:
+---The `append` option lets you control the priority of the archive's files in the event of naming collisions.
+---
+---This function is not thread safe.
+---
+---Mounting or unmounting an archive while other threads call lovr.filesystem functions is not supported.
 ---
 ---@param path string # The path to mount.
 ---@param mountpoint? string # The path in the virtual filesystem to mount to.
@@ -149,6 +233,10 @@ function lovr.filesystem.newBlob(filename) end
 ---
 ---Read the contents of a file.
 ---
+---
+---### NOTE:
+---If the file does not exist or cannot be read, nil is returned.
+---
 ---@param filename string # The name of the file to read.
 ---@param bytes? number # The number of bytes to read (if -1, all bytes will be read).
 ---@return string contents # The contents of the file.
@@ -158,6 +246,12 @@ function lovr.filesystem.read(filename, bytes) end
 ---
 ---Remove a file or directory in the save directory.
 ---
+---
+---### NOTE:
+---A directory can only be removed if it is empty.
+---
+---To recursively remove a folder, use this function with `lovr.filesystem.getDirectoryItems`.
+---
 ---@param path string # The file or directory to remove.
 ---@return boolean success # Whether the path was removed.
 function lovr.filesystem.remove(path) end
@@ -169,13 +263,25 @@ function lovr.filesystem.remove(path) end
 function lovr.filesystem.setIdentity(identity) end
 
 ---
----Sets the require path.  The require path is a semicolon-separated list of patterns that LÖVR will use to search for files when they are `require`d.  Any question marks in the pattern will be replaced with the module that is being required.  It is similar to Lua\'s `package.path` variable, but the main difference is that the patterns are relative to the save directory and the project directory.
+---Sets the require path.
+---
+---The require path is a semicolon-separated list of patterns that LÖVR will use to search for files when they are `require`d.
+---
+---Any question marks in the pattern will be replaced with the module that is being required.
+---
+---It is similar to Lua\'s `package.path` variable, but the main difference is that the patterns are relative to the save directory and the project directory.
+---
+---
+---### NOTE:
+---The default reqiure path is '?.lua;?/init.lua'.
 ---
 ---@param path? string # An optional semicolon separated list of search patterns.
 function lovr.filesystem.setRequirePath(path) end
 
 ---
----Sets the location of the project's source.  This can only be done once, and is usually done internally.
+---Sets the location of the project's source.
+---
+---This can only be done once, and is usually done internally.
 ---
 ---@param identity string # The path containing the project's source.
 function lovr.filesystem.setSource(identity) end
@@ -183,6 +289,12 @@ function lovr.filesystem.setSource(identity) end
 ---
 ---Unmounts a directory or archive previously mounted with `lovr.filesystem.mount`.
 ---
+---
+---### NOTE:
+---This function is not thread safe.
+---
+---Mounting or unmounting an archive while other threads call lovr.filesystem functions is not supported.
+---
 ---@param path string # The path to unmount.
 ---@return boolean success # Whether the archive was unmounted.
 function lovr.filesystem.unmount(path) end
@@ -190,6 +302,12 @@ function lovr.filesystem.unmount(path) end
 ---
 ---Write to a file.
 ---
+---
+---### NOTE:
+---If the file does not exist, it is created.
+---
+---If the file already has data in it, it will be replaced with the new content.
+---
 ---@overload fun(filename: string, blob: lovr.Blob):number
 ---@param filename string # The file to write to.
 ---@param content string # A string to write to the file.
diff --git a/meta/3rd/lovr/library/lovr.graphics.lua b/meta/3rd/lovr/library/lovr.graphics.lua
index 788126f7..f0ccb310 100644
--- a/meta/3rd/lovr/library/lovr.graphics.lua
+++ b/meta/3rd/lovr/library/lovr.graphics.lua
@@ -1,7 +1,9 @@
 ---@meta
 
 ---
----The `lovr.graphics` module renders graphics to displays.  Anything rendered using this module will automatically show up in the VR headset if one is connected, otherwise it will just show up in a window on the desktop.
+---The `lovr.graphics` module renders graphics to displays.
+---
+---Anything rendered using this module will automatically show up in the VR headset if one is connected, otherwise it will just show up in a window on the desktop.
 ---
 ---@class lovr.graphics
 lovr.graphics = {}
@@ -9,6 +11,10 @@ lovr.graphics = {}
 ---
 ---Draws an arc.
 ---
+---
+---### NOTE:
+---The local normal vector of the circle is `(0, 0, 1)`.
+---
 ---@overload fun(material: lovr.Material, x: number, y: number, z: number, radius: number, angle: number, ax: number, ay: number, az: number, start: number, end: number, segments: number)
 ---@overload fun(mode: lovr.DrawStyle, transform: lovr.mat4, start: number, end: number, segments: number)
 ---@overload fun(material: lovr.Material, transform: lovr.mat4, start: number, end: number, segments: number)
@@ -31,7 +37,9 @@ lovr.graphics = {}
 function lovr.graphics.arc(mode, x, y, z, radius, angle, ax, ay, az, start, end, segments) end
 
 ---
----Draws a box.  This is similar to `lovr.graphics.cube` except you can have different values for the width, height, and depth of the box.
+---Draws a box.
+---
+---This is similar to `lovr.graphics.cube` except you can have different values for the width, height, and depth of the box.
 ---
 ---@overload fun(material: lovr.Material, x: number, y: number, z: number, width: number, height: number, depth: number, angle: number, ax: number, ay: number, az: number)
 ---@overload fun(mode: lovr.DrawStyle, transform: lovr.mat4)
@@ -52,6 +60,10 @@ function lovr.graphics.box(mode, x, y, z, width, height, depth, angle, ax, ay, a
 ---
 ---Draws a 2D circle.
 ---
+---
+---### NOTE:
+---The local normal vector of the circle is `(0, 0, 1)`.
+---
 ---@overload fun(material: lovr.Material, x: number, y: number, z: number, radius: number, angle: number, ax: number, ay: number, az: number, segments: number)
 ---@overload fun(mode: lovr.DrawStyle, transform: lovr.mat4, segments: number)
 ---@overload fun(material: lovr.Material, transform: lovr.mat4, segments: number)
@@ -68,7 +80,15 @@ function lovr.graphics.box(mode, x, y, z, width, height, depth, angle, ax, ay, a
 function lovr.graphics.circle(mode, x, y, z, radius, angle, ax, ay, az, segments) end
 
 ---
----Clears the screen, resetting the color, depth, and stencil information to default values.  This function is called automatically by `lovr.run` at the beginning of each frame to clear out the data from the previous frame.
+---Clears the screen, resetting the color, depth, and stencil information to default values.
+---
+---This function is called automatically by `lovr.run` at the beginning of each frame to clear out the data from the previous frame.
+---
+---
+---### NOTE:
+---The first two variants of this function can be mixed and matched, meaning you can use booleans for some of the values and numeric values for others.
+---
+---If you are using `lovr.graphics.setStencilTest`, it will not affect how the screen gets cleared. Instead, you can use `lovr.graphics.fill` to draw a fullscreen quad, which will get masked by the active stencil.
 ---
 ---@overload fun(r: number, g: number, b: number, a: number, z: number, s: number)
 ---@overload fun(hex: number)
@@ -78,10 +98,18 @@ function lovr.graphics.circle(mode, x, y, z, radius, angle, ax, ay, az, segments
 function lovr.graphics.clear(color, depth, stencil) end
 
 ---
----This function runs a compute shader on the GPU.  Compute shaders must be created with `lovr.graphics.newComputeShader` and they should implement the `void compute();` GLSL function. Running a compute shader doesn't actually do anything, but the Shader can modify data stored in `Texture`s or `ShaderBlock`s to get interesting things to happen.
+---This function runs a compute shader on the GPU.
+---
+---Compute shaders must be created with `lovr.graphics.newComputeShader` and they should implement the `void compute();` GLSL function. Running a compute shader doesn't actually do anything, but the Shader can modify data stored in `Texture`s or `ShaderBlock`s to get interesting things to happen.
 ---
 ---When running the compute shader, you can specify the number of times to run it in 3 dimensions, which is useful to iterate over large numbers of elements like pixels or array elements.
 ---
+---
+---### NOTE:
+---Only compute shaders created with `lovr.graphics.newComputeShader` can be used here.
+---
+---There are GPU-specific limits on the `x`, `y`, and `z` values which can be queried in the `compute` entry of `lovr.graphics.getLimits`.
+---
 ---@param shader lovr.Shader # The compute shader to run.
 ---@param x? number # The amount of times to run in the x direction.
 ---@param y? number # The amount of times to run in the y direction.
@@ -91,6 +119,16 @@ function lovr.graphics.compute(shader, x, y, z) end
 ---
 ---Create the desktop window, usually used to mirror the headset display.
 ---
+---
+---### NOTE:
+---This function can only be called once.
+---
+---It is normally called internally, but you can override this by setting window to `nil` in conf.lua.
+---
+---See `lovr.conf` for more information.
+---
+---The window must be created before any `lovr.graphics` functions can be used.
+---
 ---@param flags {width: number, height: number, fullscreen: boolean, resizable: boolean, msaa: number, title: string, icon: string, vsync: number} # Flags to customize the window's appearance and behavior.
 function lovr.graphics.createWindow(flags) end
 
@@ -114,6 +152,10 @@ function lovr.graphics.cube(mode, x, y, z, size, angle, ax, ay, az) end
 ---
 ---Draws a cylinder.
 ---
+---
+---### NOTE:
+---Currently, cylinders don't have UVs.
+---
 ---@overload fun(material: lovr.Material, x: number, y: number, z: number, length: number, angle: number, ax: number, ay: number, az: number, r1: number, r2: number, capped: boolean, segments: number)
 ---@param x? number # The x coordinate of the center of the cylinder.
 ---@param y? number # The y coordinate of the center of the cylinder.
@@ -130,7 +172,9 @@ function lovr.graphics.cube(mode, x, y, z, size, angle, ax, ay, az) end
 function lovr.graphics.cylinder(x, y, z, length, angle, ax, ay, az, r1, r2, capped, segments) end
 
 ---
----Discards pixel information in the active Canvas or display.  This is mostly used as an optimization hint for the GPU, and is usually most helpful on mobile devices.
+---Discards pixel information in the active Canvas or display.
+---
+---This is mostly used as an optimization hint for the GPU, and is usually most helpful on mobile devices.
 ---
 ---@param color? boolean # Whether or not to discard color information.
 ---@param depth? boolean # Whether or not to discard depth information.
@@ -140,6 +184,12 @@ function lovr.graphics.discard(color, depth, stencil) end
 ---
 ---Draws a fullscreen textured quad.
 ---
+---
+---### NOTE:
+---This function ignores stereo rendering, so it will stretch the input across the entire Canvas if it's stereo.
+---
+---Special shaders are currently required for correct stereo fills.
+---
 ---@overload fun()
 ---@param texture lovr.Texture # The texture to use.
 ---@param u? number # The x component of the uv offset.
@@ -149,18 +199,37 @@ function lovr.graphics.discard(color, depth, stencil) end
 function lovr.graphics.fill(texture, u, v, w, h) end
 
 ---
----Flushes the internal queue of draw batches.  Under normal circumstances this is done automatically when needed, but the ability to flush manually may be helpful if you're integrating a LÖVR project with some external rendering code.
+---Flushes the internal queue of draw batches.
+---
+---Under normal circumstances this is done automatically when needed, but the ability to flush manually may be helpful if you're integrating a LÖVR project with some external rendering code.
 ---
 function lovr.graphics.flush() end
 
 ---
----Returns whether or not alpha sampling is enabled.  Alpha sampling is also known as alpha-to-coverage.  When it is enabled, the alpha channel of a pixel is factored into how antialiasing is computed, so the edges of a transparent texture will be correctly antialiased.
+---Returns whether or not alpha sampling is enabled.
+---
+---Alpha sampling is also known as alpha-to-coverage.
+---
+---When it is enabled, the alpha channel of a pixel is factored into how antialiasing is computed, so the edges of a transparent texture will be correctly antialiased.
+---
+---
+---### NOTE:
+---- Alpha sampling is disabled by default.
+---- This feature can be used for a simple transparency effect, pixels with an alpha of zero will
+---  have their depth value discarded, allowing things behind them to show through (normally you
+---  have to sort objects or write a shader for this).
 ---
 ---@return boolean enabled # Whether or not alpha sampling is enabled.
 function lovr.graphics.getAlphaSampling() end
 
 ---
----Returns the current background color.  Color components are from 0.0 to 1.0.
+---Returns the current background color.
+---
+---Color components are from 0.0 to 1.0.
+---
+---
+---### NOTE:
+---The default background color is `(0.0, 0.0, 0.0, 1.0)`.
 ---
 ---@return number r # The red component of the background color.
 ---@return number g # The green component of the background color.
@@ -169,22 +238,40 @@ function lovr.graphics.getAlphaSampling() end
 function lovr.graphics.getBackgroundColor() end
 
 ---
----Returns the current blend mode.  The blend mode controls how each pixel's color is blended with the previous pixel's color when drawn.
+---Returns the current blend mode.
+---
+---The blend mode controls how each pixel's color is blended with the previous pixel's color when drawn.
 ---
 ---If blending is disabled, `nil` will be returned.
 ---
+---
+---### NOTE:
+---The default blend mode is `alpha` and `alphamultiply`.
+---
 ---@return lovr.BlendMode blend # The current blend mode.
 ---@return lovr.BlendAlphaMode alphaBlend # The current alpha blend mode.
 function lovr.graphics.getBlendMode() end
 
 ---
----Returns the active Canvas.  Usually when you render something it will render directly to the headset.  If a Canvas object is active, things will be rendered to the textures attached to the Canvas instead.
+---Returns the active Canvas.
+---
+---Usually when you render something it will render directly to the headset.
+---
+---If a Canvas object is active, things will be rendered to the textures attached to the Canvas instead.
 ---
 ---@return lovr.Canvas canvas # The active Canvas, or `nil` if no canvas is set.
 function lovr.graphics.getCanvas() end
 
 ---
----Returns the current global color factor.  Color components are from 0.0 to 1.0.  Every pixel drawn will be multiplied (i.e. tinted) by this color.
+---Returns the current global color factor.
+---
+---Color components are from 0.0 to 1.0.
+---
+---Every pixel drawn will be multiplied (i.e. tinted) by this color.
+---
+---
+---### NOTE:
+---The default color is `(1.0, 1.0, 1.0, 1.0)`.
 ---
 ---@return number r # The red component of the color.
 ---@return number g # The green component of the color.
@@ -193,12 +280,26 @@ function lovr.graphics.getCanvas() end
 function lovr.graphics.getColor() end
 
 ---
----Returns a boolean for each color channel (red, green, blue, alpha) indicating whether it is enabled.  When a color channel is enabled, it will be affected by drawing commands and clear commands.
+---Returns a boolean for each color channel (red, green, blue, alpha) indicating whether it is enabled.
+---
+---When a color channel is enabled, it will be affected by drawing commands and clear commands.
+---
+---
+---### NOTE:
+---By default, all color channels are enabled.
+---
+---Disabling all of the color channels can be useful if you only want to write to the depth buffer or the stencil buffer.
 ---
 function lovr.graphics.getColorMask() end
 
 ---
----Returns the default filter mode for new Textures.  This controls how textures are sampled when they are minified, magnified, or stretched.
+---Returns the default filter mode for new Textures.
+---
+---This controls how textures are sampled when they are minified, magnified, or stretched.
+---
+---
+---### NOTE:
+---The default filter is `trilinear`.
 ---
 ---@return lovr.FilterMode mode # The filter mode.
 ---@return number anisotropy # The level of anisotropy.
@@ -207,6 +308,28 @@ function lovr.graphics.getDefaultFilter() end
 ---
 ---Returns the current depth test settings.
 ---
+---
+---### NOTE:
+---The depth test is an advanced technique to control how 3D objects overlap each other when they are rendered.
+---
+---It works as follows:
+---
+---- Each pixel keeps track of its z value as well as its color.
+---- If `write` is enabled when something is drawn, then any pixels that are drawn will have their
+---  z values updated.
+---- Additionally, anything drawn will first compare the existing z value of a pixel with the new z
+---  value.
+---
+---The `compareMode` setting determines how this comparison is performed.
+---
+---If the
+---  comparison succeeds, the new pixel will overwrite the previous one, otherwise that pixel won't
+---  be rendered to.
+---
+---Smaller z values are closer to the camera.
+---
+---The default compare mode is `lequal`, which usually gives good results for normal 3D rendering.
+---
 ---@return lovr.CompareMode compareMode # The current comparison method for depth testing.
 ---@return boolean write # Whether pixels will have their z value updated when rendered to.
 function lovr.graphics.getDepthTest() end
@@ -245,11 +368,23 @@ function lovr.graphics.getLimits() end
 ---
 ---Returns the current line width.
 ---
+---
+---### NOTE:
+---The default line width is `1`.
+---
 ---@return number width # The current line width, in pixels.
 function lovr.graphics.getLineWidth() end
 
 ---
----Returns the pixel density of the window.  On "high-dpi" displays, this will be `2.0`, indicating that there are 2 pixels for every window coordinate.  On a normal display it will be `1.0`, meaning that the pixel to window-coordinate ratio is 1:1.
+---Returns the pixel density of the window.
+---
+---On "high-dpi" displays, this will be `2.0`, indicating that there are 2 pixels for every window coordinate.
+---
+---On a normal display it will be `1.0`, meaning that the pixel to window-coordinate ratio is 1:1.
+---
+---
+---### NOTE:
+---If the window isn't created yet, this function will return 0.
 ---
 ---@return number density # The pixel density of the window.
 function lovr.graphics.getPixelDensity() end
@@ -257,12 +392,24 @@ function lovr.graphics.getPixelDensity() end
 ---
 ---Returns the current point size.
 ---
+---
+---### NOTE:
+---The default point size is `1.0`.
+---
 ---@return number size # The current point size, in pixels.
 function lovr.graphics.getPointSize() end
 
 ---
 ---Returns the projection for a single view.
 ---
+---
+---### NOTE:
+---Non-stereo rendering will only use the first view.
+---
+---The projection matrices are available as the `mat4 lovrProjections[2]` variable in shaders.
+---
+---The current projection matrix is available as `lovrProjection`.
+---
 ---@overload fun(view: number, matrix: lovr.Mat4):lovr.Mat4
 ---@param view number # The view index.
 ---@return number left # The left field of view angle, in radians.
@@ -284,7 +431,19 @@ function lovr.graphics.getShader() end
 function lovr.graphics.getStats() end
 
 ---
----Returns the current stencil test.  The stencil test lets you mask out pixels that meet certain criteria, based on the contents of the stencil buffer.  The stencil buffer can be modified using `lovr.graphics.stencil`.  After rendering to the stencil buffer, the stencil test can be set to control how subsequent drawing functions are masked by the stencil buffer.
+---Returns the current stencil test.
+---
+---The stencil test lets you mask out pixels that meet certain criteria, based on the contents of the stencil buffer.
+---
+---The stencil buffer can be modified using `lovr.graphics.stencil`.
+---
+---After rendering to the stencil buffer, the stencil test can be set to control how subsequent drawing functions are masked by the stencil buffer.
+---
+---
+---### NOTE:
+---Stencil values are between 0 and 255.
+---
+---By default, the stencil test is disabled.
 ---
 ---@return lovr.CompareMode compareMode # The comparison method used to decide if a pixel should be visible, or nil if the stencil test is disabled.
 ---@return number compareValue # The value stencil values are compared against, or nil if the stencil test is disabled.
@@ -311,7 +470,17 @@ function lovr.graphics.getViewPose(view) end
 function lovr.graphics.getWidth() end
 
 ---
----Returns the current polygon winding.  The winding direction determines which face of a triangle is the front face and which is the back face.  This lets the graphics engine cull the back faces of polygons, improving performance.
+---Returns the current polygon winding.
+---
+---The winding direction determines which face of a triangle is the front face and which is the back face.
+---
+---This lets the graphics engine cull the back faces of polygons, improving performance.
+---
+---
+---### NOTE:
+---Culling is initially disabled and must be enabled using `lovr.graphics.setCullingEnabled`.
+---
+---The default winding direction is counterclockwise.
 ---
 ---@return lovr.Winding winding # The current winding direction.
 function lovr.graphics.getWinding() end
@@ -319,11 +488,23 @@ function lovr.graphics.getWinding() end
 ---
 ---Returns whether the desktop window is currently created.
 ---
+---
+---### NOTE:
+---Most of the `lovr.graphics` functionality will only work if a window is created.
+---
 ---@return boolean present # Whether a window is created.
 function lovr.graphics.hasWindow() end
 
 ---
----Returns whether or not culling is active.  Culling is an optimization that avoids rendering the back face of polygons.  This improves performance by reducing the number of polygons drawn, but requires that the vertices in triangles are specified in a consistent clockwise or counter clockwise order.
+---Returns whether or not culling is active.
+---
+---Culling is an optimization that avoids rendering the back face of polygons.
+---
+---This improves performance by reducing the number of polygons drawn, but requires that the vertices in triangles are specified in a consistent clockwise or counter clockwise order.
+---
+---
+---### NOTE:
+---Culling is disabled by default.
 ---
 ---@return boolean isEnabled # Whether or not culling is enabled.
 function lovr.graphics.isCullingEnabled() end
@@ -331,11 +512,19 @@ function lovr.graphics.isCullingEnabled() end
 ---
 ---Returns a boolean indicating whether or not wireframe rendering is enabled.
 ---
+---
+---### NOTE:
+---Wireframe rendering is initially disabled.
+---
+---Wireframe rendering is only supported on desktop systems.
+---
 ---@return boolean isWireframe # Whether or not wireframe rendering is enabled.
 function lovr.graphics.isWireframe() end
 
 ---
----Draws lines between points.  Each point will be connected to the previous point in the list.
+---Draws lines between points.
+---
+---Each point will be connected to the previous point in the list.
 ---
 ---@overload fun(points: table)
 ---@param x1 number # The x coordinate of the first point.
@@ -347,10 +536,18 @@ function lovr.graphics.isWireframe() end
 function lovr.graphics.line(x1, y1, z1, x2, y2, z2) end
 
 ---
----Creates a new Canvas.  You can specify Textures to attach to it, or just specify a width and height and attach textures later using `Canvas:setTexture`.
+---Creates a new Canvas.
+---
+---You can specify Textures to attach to it, or just specify a width and height and attach textures later using `Canvas:setTexture`.
 ---
 ---Once created, you can render to the Canvas using `Canvas:renderTo`, or `lovr.graphics.setCanvas`.
 ---
+---
+---### NOTE:
+---Textures created by this function will have `clamp` as their `WrapMode`.
+---
+---Stereo Canvases will either have their width doubled or use array textures for their attachments, depending on their implementation.
+---
 ---@overload fun(..., flags: table):lovr.Canvas
 ---@overload fun(attachments: table, flags: table):lovr.Canvas
 ---@param width number # The width of the canvas, in pixels.
@@ -362,16 +559,32 @@ function lovr.graphics.newCanvas(width, height, flags) end
 ---
 ---Creates a new compute Shader, used for running generic compute operations on the GPU.
 ---
+---
+---### NOTE:
+---Compute shaders are not supported on all hardware, use `lovr.graphics.getFeatures` to check if they're available on the current system.
+---
+---The source code for a compute shader needs to implement the `void compute();` GLSL function. This function doesn't return anything, but the compute shader is able to write data out to `Texture`s or `ShaderBlock`s.
+---
+---The GLSL version used for compute shaders is GLSL 430.
+---
+---Currently, up to 32 shader flags are supported.
+---
 ---@param source string # The code or filename of the compute shader.
 ---@param options? {flags: table} # Optional settings for the Shader.
 ---@return lovr.Shader shader # The new compute Shader.
 function lovr.graphics.newComputeShader(source, options) end
 
 ---
----Creates a new Font.  It can be used to render text with `lovr.graphics.print`.
+---Creates a new Font.
+---
+---It can be used to render text with `lovr.graphics.print`.
 ---
 ---Currently, the only supported font format is TTF.
 ---
+---
+---### NOTE:
+---Larger font sizes will lead to more detailed curves at the cost of performance.
+---
 ---@overload fun(size: number, padding: number, spread: number):lovr.Font
 ---@overload fun(rasterizer: lovr.Rasterizer, padding: number, spread: number):lovr.Font
 ---@param filename string # The filename of the font file.
@@ -382,7 +595,18 @@ function lovr.graphics.newComputeShader(source, options) end
 function lovr.graphics.newFont(filename, size, padding, spread) end
 
 ---
----Creates a new Material.  Materials are sets of colors, textures, and other parameters that affect the appearance of objects.  They can be applied to `Model`s, `Mesh`es, and most graphics primitives accept a Material as an optional first argument.
+---Creates a new Material.
+---
+---Materials are sets of colors, textures, and other parameters that affect the appearance of objects.
+---
+---They can be applied to `Model`s, `Mesh`es, and most graphics primitives accept a Material as an optional first argument.
+---
+---
+---### NOTE:
+---- Scalar properties will default to `1.0`.
+---- Color properties will default to `(1.0, 1.0, 1.0, 1.0)`, except for `emissive` which will
+---  default to `(0.0, 0.0, 0.0, 0.0)`.
+---- Textures will default to `nil` (a single 1x1 white pixel will be used for them).
 ---
 ---@overload fun(texture: lovr.Texture, r: number, g: number, b: number, a: number):lovr.Material
 ---@overload fun(canvas: lovr.Canvas, r: number, g: number, b: number, a: number):lovr.Material
@@ -392,10 +616,28 @@ function lovr.graphics.newFont(filename, size, padding, spread) end
 function lovr.graphics.newMaterial() end
 
 ---
----Creates a new Mesh.  Meshes contain the data for an arbitrary set of vertices, and can be drawn. You must specify either the capacity for the Mesh or an initial set of vertex data.  Optionally, a custom format table can be used to specify the set of vertex attributes the mesh will provide to the active shader.  The draw mode and usage hint can also optionally be specified.
+---Creates a new Mesh.
+---
+---Meshes contain the data for an arbitrary set of vertices, and can be drawn. You must specify either the capacity for the Mesh or an initial set of vertex data.
+---
+---Optionally, a custom format table can be used to specify the set of vertex attributes the mesh will provide to the active shader.
+---
+---The draw mode and usage hint can also optionally be specified.
 ---
 ---The default data type for an attribute is `float`, and the default component count is 1.
 ---
+---
+---### NOTE:
+---Once created, the size and format of the Mesh cannot be changed.'
+---
+---The default mesh format is:
+---
+---    {
+---      { 'lovrPosition',    'float', 3 },
+---      { 'lovrNormal',      'float', 3 },
+---      { 'lovrTexCoord',    'float', 2 }
+---    }
+---
 ---@overload fun(vertices: table, mode: lovr.DrawMode, usage: lovr.MeshUsage, readable: boolean):lovr.Mesh
 ---@overload fun(blob: lovr.Blob, mode: lovr.DrawMode, usage: lovr.MeshUsage, readable: boolean):lovr.Mesh
 ---@overload fun(format: table, size: number, mode: lovr.DrawMode, usage: lovr.MeshUsage, readable: boolean):lovr.Mesh
@@ -409,7 +651,23 @@ function lovr.graphics.newMaterial() end
 function lovr.graphics.newMesh(size, mode, usage, readable) end
 
 ---
----Creates a new Model from a file.  The supported 3D file formats are OBJ, glTF, and STL.
+---Creates a new Model from a file.
+---
+---The supported 3D file formats are OBJ, glTF, and STL.
+---
+---
+---### NOTE:
+---Diffuse and emissive textures will be loaded in the sRGB encoding, all other textures will be loaded as linear.
+---
+---Currently, the following features are not supported by the model importer:
+---
+---- OBJ: Quads are not supported (only triangles).
+---- glTF: Sparse accessors are not supported.
+---- glTF: Morph targets are not supported.
+---- glTF: base64 images are not supported (base64 buffer data works though).
+---- glTF: Only the default scene is loaded.
+---- glTF: Currently, each skin in a Model can have up to 48 joints.
+---- STL: ASCII STL files are not supported.
 ---
 ---@overload fun(modelData: lovr.ModelData):lovr.Model
 ---@param filename string # The filename of the model to load.
@@ -419,6 +677,87 @@ function lovr.graphics.newModel(filename) end
 ---
 ---Creates a new Shader.
 ---
+---
+---### NOTE:
+---The `flags` table should contain string keys, with boolean or numeric values.
+---
+---These flags can be used to customize the behavior of Shaders from Lua, by using the flags in the shader source code.
+---
+---Numeric flags will be available as constants named `FLAG_<flagName>`.
+---
+---Boolean flags can be used with `#ifdef` and will only be defined if the value in the Lua table was `true`.
+---
+---The following flags are used by shaders provided by LÖVR:
+---
+---- `animated` is a boolean flag that will cause the shader to position vertices based on the pose
+---  of an animated skeleton.
+---
+---This should usually only be used for animated `Model`s, since it
+---  needs a skeleton to work properly and is slower than normal rendering.
+---- `alphaCutoff` is a numeric flag that can be used to implement simple "cutout" style
+---  transparency, where pixels with alpha below a certain threshold will be discarded.
+---
+---The value
+---  of the flag should be a number between 0.0 and 1.0, representing the alpha threshold.
+---- `uniformScale` is a boolean flag used for optimization.
+---
+---If the Shader is only going to be
+---  used with objects that have a *uniform* scale (i.e. the x, y, and z components of the scale
+---  are all the same number), then this flag can be set to use a faster method to compute the
+---  `lovrNormalMatrix` uniform variable.
+---- `multicanvas` is a boolean flag that should be set when rendering to multiple Textures
+---  attached to a `Canvas`.
+---
+---When set, the fragment shader should implement the `colors` function
+---  instead of the `color` function, and can write color values to the `lovrCanvas` array instead
+---  of returning a single color.
+---
+---Each color in the array gets written to the corresponding
+---  texture attached to the canvas.
+---- `highp` is a boolean flag specific to mobile GPUs that changes the default precision for
+---  fragment shaders to use high precision instead of the default medium precision.
+---
+---This can fix
+---  visual issues caused by a lack of precision, but isn't guaranteed to be supported on some
+---  lower-end systems.
+---- The following flags are used only by the `standard` PBR shader:
+---  - `normalMap` should be set to `true` to render objects with a normal map, providing a more
+---  detailed, bumpy appearance.
+---
+---Currently, this requires the model to have vertex tangents.
+---  - `emissive` should be set to `true` to apply emissive maps to rendered objects.
+---
+---This is
+---    usually used to apply glowing lights or screens to objects, since the emissive texture is
+---    not affected at all by lighting.
+---  - `indirectLighting` is an *awesome* boolean flag that will apply realistic reflections and
+---    lighting to the surface of an object, based on a specially-created skybox.
+---
+---See the
+---    `Standard Shader` guide for more information.
+---  - `occlusion` is a boolean flag that uses the ambient occlusion texture in the model.
+---
+---It only
+---    affects indirect lighting, so it will only have an effect if the `indirectLighting` flag is
+---    also enabled.
+---  - `skipTonemap` is a flag that will skip the tonemapping process.
+---
+---Tonemapping is an important
+---    process that maps the high definition physical color values down to a 0 - 1 range for
+---    display.
+---
+---There are lots of different tonemapping algorithms that give different artistic
+---    effects.
+---
+---The default tonemapping in the standard shader is the ACES algorithm, but you can
+---    use this flag to turn off ACES and use your own tonemapping.
+---
+---Currently, up to 32 shader flags are supported.
+---
+---The `stereo` option is only necessary for Android.
+---
+---Currently on Android, only stereo shaders can be used with stereo Canvases, and mono Shaders can only be used with mono Canvases.
+---
 ---@overload fun(default: lovr.DefaultShader, options: table):lovr.Shader
 ---@param vertex string # The code or filename of the vertex shader.  If nil, the default vertex shader is used.
 ---@param fragment string # The code or filename of the fragment shader.  If nil, the default fragment shader is used.
@@ -429,6 +768,14 @@ function lovr.graphics.newShader(vertex, fragment, options) end
 ---
 ---Creates a new ShaderBlock from a table of variable definitions with their names and types.
 ---
+---
+---### NOTE:
+---`compute` blocks can only be true if compute shaders are supported, see `lovr.graphics.getFeatures`.
+---
+---Compute blocks may be slightly slower than uniform blocks, but they can also be much, much larger.
+---
+---Uniform blocks are usually limited to around 16 kilobytes in size, depending on hardware.
+---
 ---@param type lovr.BlockType # Whether the block will be used for read-only uniform data or compute shaders.
 ---@param uniforms table # A table where the keys are uniform names and the values are uniform types.  Uniform arrays can be specified by supplying a table as the uniform's value containing the type and the array size.
 ---@param flags? {usage: lovr.BufferUsage, readable: boolean} # Optional settings.
@@ -438,6 +785,12 @@ function lovr.graphics.newShaderBlock(type, uniforms, flags) end
 ---
 ---Creates a new Texture from an image file.
 ---
+---
+---### NOTE:
+---The "linear" flag should be set to true for textures that don't contain color information, such as normal maps.
+---
+---Right now the supported image file formats are png, jpg, hdr, dds (DXT1, DXT3, DXT5), ktx, and astc.
+---
 ---@overload fun(images: table, flags: table):lovr.Texture
 ---@overload fun(width: number, height: number, depth: number, flags: table):lovr.Texture
 ---@overload fun(blob: lovr.Blob, flags: table):lovr.Texture
@@ -450,11 +803,21 @@ function lovr.graphics.newTexture(filename, flags) end
 ---
 ---Resets the transformation to the origin.
 ---
+---
+---### NOTE:
+---This is called at the beginning of every frame to reset the coordinate space.
+---
 function lovr.graphics.origin() end
 
 ---
 ---Draws a plane with a given position, size, and orientation.
 ---
+---
+---### NOTE:
+---The `u`, `v`, `w`, `h` arguments can be used to select a subregion of the diffuse texture to apply to the plane.
+---
+---One efficient technique for rendering many planes with different textures is to pack all of the textures into a single image, and then use the uv arguments to select a sub-rectangle to use for each plane.
+---
 ---@overload fun(material: lovr.Material, x: number, y: number, z: number, width: number, height: number, angle: number, ax: number, ay: number, az: number, u: number, v: number, w: number, h: number)
 ---@param mode lovr.DrawStyle # How to draw the plane.
 ---@param x? number # The x coordinate of the center of the plane.
@@ -484,16 +847,40 @@ function lovr.graphics.points(x, y, z) end
 ---
 ---Pops the current transform from the stack, returning to the transformation that was applied before `lovr.graphics.push` was called.
 ---
+---
+---### NOTE:
+---An error is thrown if there isn't a transform to pop.
+---
+---This can happen if you forget to call push before calling pop, or if you have an unbalanced sequence of pushes and pops.
+---
 function lovr.graphics.pop() end
 
 ---
----Presents the results of pending drawing operations to the window.  This is automatically called after `lovr.draw` by the default `lovr.run` function.
+---Presents the results of pending drawing operations to the window.
+---
+---This is automatically called after `lovr.draw` by the default `lovr.run` function.
 ---
 function lovr.graphics.present() end
 
 ---
 ---Draws text in 3D space using the active font.
 ---
+---
+---### NOTE:
+---Unicode text is supported.
+---
+---Use `\n` to add line breaks.
+---
+---`\t` will be rendered as four spaces.
+---
+---LÖVR uses a fancy technique for font rendering called multichannel signed distance fields.
+---
+---This leads to crisp text in VR, but requires a special shader to use.
+---
+---LÖVR internally switches to the appropriate shader, but only when the default shader is already set.
+---
+---If you see strange font artifacts, make sure you switch back to the default shader by using `lovr.graphics.setShader()` before you draw text.
+---
 ---@param str string # The text to render.
 ---@param x? number # The x coordinate of the center of the text.
 ---@param y? number # The y coordinate of the center of the text.
@@ -509,7 +896,15 @@ function lovr.graphics.present() end
 function lovr.graphics.print(str, x, y, z, scale, angle, ax, ay, az, wrap, halign, valign) end
 
 ---
----Pushes a copy of the current transform onto the transformation stack.  After changing the transform using `lovr.graphics.translate`, `lovr.graphics.rotate`, and `lovr.graphics.scale`, the original state can be restored using `lovr.graphics.pop`.
+---Pushes a copy of the current transform onto the transformation stack.
+---
+---After changing the transform using `lovr.graphics.translate`, `lovr.graphics.rotate`, and `lovr.graphics.scale`, the original state can be restored using `lovr.graphics.pop`.
+---
+---
+---### NOTE:
+---An error is thrown if more than 64 matrices are pushed.
+---
+---This can happen accidentally if a push isn't followed by a corresponding pop.
 ---
 function lovr.graphics.push() end
 
@@ -523,6 +918,10 @@ function lovr.graphics.reset() end
 ---
 ---The rotation will last until `lovr.draw` returns or the transformation is popped off the transformation stack.
 ---
+---
+---### NOTE:
+---Order matters when scaling, translating, and rotating the coordinate system.
+---
 ---@param angle? number # The amount to rotate the coordinate system by, in radians.
 ---@param ax? number # The x component of the axis of rotation.
 ---@param ay? number # The y component of the axis of rotation.
@@ -530,23 +929,46 @@ function lovr.graphics.reset() end
 function lovr.graphics.rotate(angle, ax, ay, az) end
 
 ---
----Scales the coordinate system in 3 dimensions.  This will cause objects to appear bigger or smaller.
+---Scales the coordinate system in 3 dimensions.
+---
+---This will cause objects to appear bigger or smaller.
 ---
 ---The scaling will last until `lovr.draw` returns or the transformation is popped off the transformation stack.
 ---
+---
+---### NOTE:
+---Order matters when scaling, translating, and rotating the coordinate system.
+---
 ---@param x? number # The amount to scale on the x axis.
 ---@param y? number # The amount to scale on the y axis.
 ---@param z? number # The amount to scale on the z axis.
 function lovr.graphics.scale(x, y, z) end
 
 ---
----Enables or disables alpha sampling.  Alpha sampling is also known as alpha-to-coverage.  When it is enabled, the alpha channel of a pixel is factored into how antialiasing is computed, so the edges of a transparent texture will be correctly antialiased.
+---Enables or disables alpha sampling.
+---
+---Alpha sampling is also known as alpha-to-coverage.
+---
+---When it is enabled, the alpha channel of a pixel is factored into how antialiasing is computed, so the edges of a transparent texture will be correctly antialiased.
+---
+---
+---### NOTE:
+---- Alpha sampling is disabled by default.
+---- This feature can be used for a simple transparency effect, pixels with an alpha of zero will
+---  have their depth value discarded, allowing things behind them to show through (normally you
+---  have to sort objects or write a shader for this).
 ---
 ---@param enabled boolean # Whether or not alpha sampling is enabled.
 function lovr.graphics.setAlphaSampling(enabled) end
 
 ---
----Sets the background color used to clear the screen.  Color components are from 0.0 to 1.0.
+---Sets the background color used to clear the screen.
+---
+---Color components are from 0.0 to 1.0.
+---
+---
+---### NOTE:
+---The default background color is `(0.0, 0.0, 0.0, 1.0)`.
 ---
 ---@overload fun(hex: number, a: number)
 ---@overload fun(color: table)
@@ -557,7 +979,13 @@ function lovr.graphics.setAlphaSampling(enabled) end
 function lovr.graphics.setBackgroundColor(r, g, b, a) end
 
 ---
----Sets the blend mode.  The blend mode controls how each pixel's color is blended with the previous pixel's color when drawn.
+---Sets the blend mode.
+---
+---The blend mode controls how each pixel's color is blended with the previous pixel's color when drawn.
+---
+---
+---### NOTE:
+---The default blend mode is `alpha` and `alphamultiply`.
 ---
 ---@overload fun()
 ---@param blend lovr.BlendMode # The blend mode.
@@ -565,13 +993,25 @@ function lovr.graphics.setBackgroundColor(r, g, b, a) end
 function lovr.graphics.setBlendMode(blend, alphaBlend) end
 
 ---
----Sets or disables the active Canvas object.  If there is an active Canvas, things will be rendered to the Textures attached to that Canvas instead of to the headset.
+---Sets or disables the active Canvas object.
+---
+---If there is an active Canvas, things will be rendered to the Textures attached to that Canvas instead of to the headset.
 ---
 ---@param canvas? lovr.Canvas # The new active Canvas object, or `nil` to just render to the headset.
 function lovr.graphics.setCanvas(canvas) end
 
 ---
----Sets the color used for drawing objects.  Color components are from 0.0 to 1.0.  Every pixel drawn will be multiplied (i.e. tinted) by this color.  This is a global setting, so it will affect all subsequent drawing operations.
+---Sets the color used for drawing objects.
+---
+---Color components are from 0.0 to 1.0.
+---
+---Every pixel drawn will be multiplied (i.e. tinted) by this color.
+---
+---This is a global setting, so it will affect all subsequent drawing operations.
+---
+---
+---### NOTE:
+---The default color is `(1.0, 1.0, 1.0, 1.0)`.
 ---
 ---@overload fun(hex: number, a: number)
 ---@overload fun(color: table)
@@ -582,7 +1022,15 @@ function lovr.graphics.setCanvas(canvas) end
 function lovr.graphics.setColor(r, g, b, a) end
 
 ---
----Enables and disables individual color channels.  When a color channel is enabled, it will be affected by drawing commands and clear commands.
+---Enables and disables individual color channels.
+---
+---When a color channel is enabled, it will be affected by drawing commands and clear commands.
+---
+---
+---### NOTE:
+---By default, all color channels are enabled.
+---
+---Disabling all of the color channels can be useful if you only want to write to the depth buffer or the stencil buffer.
 ---
 ---@param r boolean # Whether the red color channel should be enabled.
 ---@param g boolean # Whether the green color channel should be enabled.
@@ -591,20 +1039,60 @@ function lovr.graphics.setColor(r, g, b, a) end
 function lovr.graphics.setColorMask(r, g, b, a) end
 
 ---
----Enables or disables culling.  Culling is an optimization that avoids rendering the back face of polygons.  This improves performance by reducing the number of polygons drawn, but requires that the vertices in triangles are specified in a consistent clockwise or counter clockwise order.
+---Enables or disables culling.
+---
+---Culling is an optimization that avoids rendering the back face of polygons.
+---
+---This improves performance by reducing the number of polygons drawn, but requires that the vertices in triangles are specified in a consistent clockwise or counter clockwise order.
+---
+---
+---### NOTE:
+---Culling is disabled by default.
 ---
 ---@param isEnabled boolean # Whether or not culling should be enabled.
 function lovr.graphics.setCullingEnabled(isEnabled) end
 
 ---
----Sets the default filter mode for new Textures.  This controls how textures are sampled when they are minified, magnified, or stretched.
+---Sets the default filter mode for new Textures.
+---
+---This controls how textures are sampled when they are minified, magnified, or stretched.
+---
+---
+---### NOTE:
+---The default filter is `trilinear`.
+---
+---The maximum supported anisotropy level can be queried using `lovr.graphics.getLimits`.
 ---
 ---@param mode lovr.FilterMode # The filter mode.
 ---@param anisotropy number # The level of anisotropy to use.
 function lovr.graphics.setDefaultFilter(mode, anisotropy) end
 
 ---
----Sets the current depth test.  The depth test controls how overlapping objects are rendered.
+---Sets the current depth test.
+---
+---The depth test controls how overlapping objects are rendered.
+---
+---
+---### NOTE:
+---The depth test is an advanced technique to control how 3D objects overlap each other when they are rendered.
+---
+---It works as follows:
+---
+---- Each pixel keeps track of its z value as well as its color.
+---- If `write` is enabled when something is drawn, then any pixels that are drawn will have their
+---  z values updated.
+---- Additionally, anything drawn will first compare the existing z value of a pixel with the new z
+---  value.
+---
+---The `compareMode` setting determines how this comparison is performed.
+---
+---If the
+---  comparison succeeds, the new pixel will overwrite the previous one, otherwise that pixel won't
+---  be rendered to.
+---
+---Smaller z values are closer to the camera.
+---
+---The default compare mode is `lequal`, which usually gives good results for normal 3D rendering.
 ---
 ---@param compareMode? lovr.CompareMode # The new depth test.  Use `nil` to disable the depth test.
 ---@param write? boolean # Whether pixels will have their z value updated when rendered to.
@@ -619,19 +1107,49 @@ function lovr.graphics.setFont(font) end
 ---
 ---Sets the width of lines rendered using `lovr.graphics.line`.
 ---
+---
+---### NOTE:
+---The default line width is `1`.
+---
+---GPU driver support for line widths is poor.
+---
+---The actual width of lines may be different from what is set here.
+---
+---In particular, some graphics drivers only support a line width of `1`.
+---
+---Currently this function only supports integer values from 1 to 255.
+---
 ---@param width? number # The new line width, in pixels.
 function lovr.graphics.setLineWidth(width) end
 
 ---
 ---Sets the width of drawn points, in pixels.
 ---
+---
+---### NOTE:
+---The default point size is `1.0`.
+---
 ---@param size? number # The new point size.
 function lovr.graphics.setPointSize(size) end
 
 ---
----Sets the projection for a single view.  4 field of view angles can be used, similar to the field of view returned by `lovr.headset.getViewAngles`.  Alternatively, a projection matrix can be used for other types of projections like orthographic, oblique, etc.
+---Sets the projection for a single view.
+---
+---4 field of view angles can be used, similar to the field of view returned by `lovr.headset.getViewAngles`.
+---
+---Alternatively, a projection matrix can be used for other types of projections like orthographic, oblique, etc.
+---
+---Two views are supported, one for each eye.
+---
+---When rendering to the headset, both projections are changed to match the ones used by the headset.
+---
+---
+---### NOTE:
+---Non-stereo rendering will only use the first view.
 ---
----Two views are supported, one for each eye.  When rendering to the headset, both projections are changed to match the ones used by the headset.
+---The projection matrices are available as the `mat4 lovrProjections[2]` variable in shaders.
+---
+---The current projection matrix is available as `lovrProjection`.
 ---
 ---@overload fun(view: number, matrix: lovr.Mat4)
 ---@param view number # The index of the view to update.
@@ -649,7 +1167,19 @@ function lovr.graphics.setProjection(view, left, right, up, down) end
 function lovr.graphics.setShader(shader) end
 
 ---
----Sets the stencil test.  The stencil test lets you mask out pixels that meet certain criteria, based on the contents of the stencil buffer.  The stencil buffer can be modified using `lovr.graphics.stencil`.  After rendering to the stencil buffer, the stencil test can be set to control how subsequent drawing functions are masked by the stencil buffer.
+---Sets the stencil test.
+---
+---The stencil test lets you mask out pixels that meet certain criteria, based on the contents of the stencil buffer.
+---
+---The stencil buffer can be modified using `lovr.graphics.stencil`.
+---
+---After rendering to the stencil buffer, the stencil test can be set to control how subsequent drawing functions are masked by the stencil buffer.
+---
+---
+---### NOTE:
+---Stencil values are between 0 and 255.
+---
+---By default, the stencil test is disabled.
 ---
 ---@overload fun()
 ---@param compareMode lovr.CompareMode # The comparison method used to decide if a pixel should be visible, or nil if the stencil test is disabled.
@@ -657,9 +1187,23 @@ function lovr.graphics.setShader(shader) end
 function lovr.graphics.setStencilTest(compareMode, compareValue) end
 
 ---
----Sets the pose for a single view.  Objects rendered in this view will appear as though the camera is positioned using the given pose.
+---Sets the pose for a single view.
 ---
----Two views are supported, one for each eye.  When rendering to the headset, both views are changed to match the estimated eye positions.  These view poses are also available using `lovr.headset.getViewPose`.
+---Objects rendered in this view will appear as though the camera is positioned using the given pose.
+---
+---Two views are supported, one for each eye.
+---
+---When rendering to the headset, both views are changed to match the estimated eye positions.
+---
+---These view poses are also available using `lovr.headset.getViewPose`.
+---
+---
+---### NOTE:
+---Non-stereo rendering will only use the first view.
+---
+---The inverted view poses (view matrices) are available as the `mat4 lovrViews[2]` variable in shaders.
+---
+---The current view matrix is available as `lovrView`.
 ---
 ---@overload fun(view: number, matrix: lovr.Mat4, inverted: boolean)
 ---@param view number # The index of the view to update.
@@ -673,19 +1217,41 @@ function lovr.graphics.setStencilTest(compareMode, compareValue) end
 function lovr.graphics.setViewPose(view, x, y, z, angle, ax, ay, az) end
 
 ---
----Sets the polygon winding.  The winding direction determines which face of a triangle is the front face and which is the back face.  This lets the graphics engine cull the back faces of polygons, improving performance.  The default is counterclockwise.
+---Sets the polygon winding.
+---
+---The winding direction determines which face of a triangle is the front face and which is the back face.
+---
+---This lets the graphics engine cull the back faces of polygons, improving performance.
+---
+---The default is counterclockwise.
+---
+---
+---### NOTE:
+---Culling is initially disabled and must be enabled using `lovr.graphics.setCullingEnabled`.
+---
+---The default winding direction is counterclockwise.
 ---
 ---@param winding lovr.Winding # The new winding direction.
 function lovr.graphics.setWinding(winding) end
 
 ---
----Enables or disables wireframe rendering.  This is meant to be used as a debugging tool.
+---Enables or disables wireframe rendering.
+---
+---This is meant to be used as a debugging tool.
+---
+---
+---### NOTE:
+---Wireframe rendering is initially disabled.
+---
+---Wireframe rendering is only supported on desktop systems.
 ---
 ---@param wireframe boolean # Whether or not wireframe rendering should be enabled.
 function lovr.graphics.setWireframe(wireframe) end
 
 ---
----Render a skybox from a texture.  Two common kinds of skybox textures are supported: A 2D equirectangular texture with a spherical coordinates can be used, or a "cubemap" texture created from 6 images.
+---Render a skybox from a texture.
+---
+---Two common kinds of skybox textures are supported: A 2D equirectangular texture with a spherical coordinates can be used, or a "cubemap" texture created from 6 images.
 ---
 ---@param texture lovr.Texture # The texture to use.
 function lovr.graphics.skybox(texture) end
@@ -707,6 +1273,10 @@ function lovr.graphics.sphere(x, y, z, radius, angle, ax, ay, az) end
 ---
 ---Renders to the stencil buffer using a function.
 ---
+---
+---### NOTE:
+---Stencil values are between 0 and 255.
+---
 ---@overload fun(callback: function, action: lovr.StencilAction, value: number, initial: number)
 ---@param callback function # The function that will be called to render to the stencil buffer.
 ---@param action? lovr.StencilAction # How to modify the stencil value of pixels that are rendered to.
@@ -717,18 +1287,46 @@ function lovr.graphics.stencil(callback, action, value, keep) end
 ---
 ---Starts a named timer on the GPU, which can be stopped using `lovr.graphics.tock`.
 ---
+---
+---### NOTE:
+---The timer can be stopped by calling `lovr.graphics.tock` using the same name.
+---
+---All drawing commands between the tick and the tock will be timed.
+---
+---It is not possible to nest calls to tick and tock.
+---
+---GPU timers are not supported on all systems.
+---
+---Check the `timers` feature using `lovr.graphics.getFeatures` to see if it is supported on the current system.
+---
 ---@param label string # The name of the timer.
 function lovr.graphics.tick(label) end
 
 ---
 ---Stops a named timer on the GPU, previously started with `lovr.graphics.tick`.
 ---
+---
+---### NOTE:
+---All drawing commands between tick and tock will be timed.
+---
+---It is not possible to nest calls to tick and tock.
+---
+---The results are delayed, and might be `nil` for the first few frames.
+---
+---This function returns the most recent available timer value.
+---
+---GPU timers are not supported on all systems.
+---
+---Check the `timers` feature using `lovr.graphics.getFeatures` to see if it is supported on the current system.
+---
 ---@param label string # The name of the timer.
 ---@return number time # The number of seconds elapsed, or `nil` if the data isn't ready yet.
 function lovr.graphics.tock(label) end
 
 ---
----Apply a transform to the coordinate system, changing its translation, rotation, and scale using a single function.  A `mat4` can also be used.
+---Apply a transform to the coordinate system, changing its translation, rotation, and scale using a single function.
+---
+---A `mat4` can also be used.
 ---
 ---The transformation will last until `lovr.draw` returns or the transformation is popped off the transformation stack.
 ---
@@ -746,25 +1344,45 @@ function lovr.graphics.tock(label) end
 function lovr.graphics.transform(x, y, z, sx, sy, sz, angle, ax, ay, az) end
 
 ---
----Translates the coordinate system in three dimensions.  All graphics operations that use coordinates will behave as if they are offset by the translation value.
+---Translates the coordinate system in three dimensions.
+---
+---All graphics operations that use coordinates will behave as if they are offset by the translation value.
 ---
 ---The translation will last until `lovr.draw` returns or the transformation is popped off the transformation stack.
 ---
+---
+---### NOTE:
+---Order matters when scaling, translating, and rotating the coordinate system.
+---
 ---@param x? number # The amount to translate on the x axis.
 ---@param y? number # The amount to translate on the y axis.
 ---@param z? number # The amount to translate on the z axis.
 function lovr.graphics.translate(x, y, z) end
 
 ---
----A Canvas is also known as a framebuffer or render-to-texture.  It allows you to render to a texture instead of directly to the screen.  This lets you postprocess or transform the results later before finally rendering them to the screen.
+---A Canvas is also known as a framebuffer or render-to-texture.
+---
+---It allows you to render to a texture instead of directly to the screen.
+---
+---This lets you postprocess or transform the results later before finally rendering them to the screen.
 ---
 ---After creating a Canvas, you can attach Textures to it using `Canvas:setTexture`.
 ---
+---
+---### NOTE:
+---Up to four textures can be attached to a Canvas and anything rendered to the Canvas will be broadcast to all attached Textures.
+---
+---If you want to do render different things to different textures, use the `multicanvas` shader flag when creating your shader and implement the `void colors` function instead of the usual `vec4 color` function.
+---
+---You can then assign different output colors to `lovrCanvas[0]`, `lovrCanvas[1]`, etc. instead of returning a single color. Each color written to the array will end up in the corresponding texture attached to the Canvas.
+---
 ---@class lovr.Canvas
 local Canvas = {}
 
 ---
----Returns the depth buffer used by the Canvas as a Texture.  If the Canvas was not created with a readable depth buffer (the `depth.readable` flag in `lovr.graphics.newCanvas`), then this function will return `nil`.
+---Returns the depth buffer used by the Canvas as a Texture.
+---
+---If the Canvas was not created with a readable depth buffer (the `depth.readable` flag in `lovr.graphics.newCanvas`), then this function will return `nil`.
 ---
 ---@return lovr.Texture texture # The depth Texture of the Canvas.
 function Canvas:getDepthTexture() end
@@ -772,6 +1390,10 @@ function Canvas:getDepthTexture() end
 ---
 ---Returns the dimensions of the Canvas, its Textures, and its depth buffer.
 ---
+---
+---### NOTE:
+---The dimensions of a Canvas can not be changed after it is created.
+---
 ---@return number width # The width of the Canvas, in pixels.
 ---@return number height # The height of the Canvas, in pixels.
 function Canvas:getDimensions() end
@@ -779,11 +1401,21 @@ function Canvas:getDimensions() end
 ---
 ---Returns the height of the Canvas, its Textures, and its depth buffer.
 ---
+---
+---### NOTE:
+---The height of a Canvas can not be changed after it is created.
+---
 ---@return number height # The height of the Canvas, in pixels.
 function Canvas:getHeight() end
 
 ---
----Returns the number of multisample antialiasing samples to use when rendering to the Canvas. Increasing this number will make the contents of the Canvas appear more smooth at the cost of performance.  It is common to use powers of 2 for this value.
+---Returns the number of multisample antialiasing samples to use when rendering to the Canvas. Increasing this number will make the contents of the Canvas appear more smooth at the cost of performance.
+---
+---It is common to use powers of 2 for this value.
+---
+---
+---### NOTE:
+---All textures attached to the Canvas must be created with this MSAA value.
 ---
 ---@return number samples # The number of MSAA samples.
 function Canvas:getMSAA() end
@@ -791,16 +1423,26 @@ function Canvas:getMSAA() end
 ---
 ---Returns the set of Textures currently attached to the Canvas.
 ---
+---
+---### NOTE:
+---Up to 4 Textures can be attached at once.
+---
 function Canvas:getTexture() end
 
 ---
 ---Returns the width of the Canvas, its Textures, and its depth buffer.
 ---
+---
+---### NOTE:
+---The width of a Canvas can not be changed after it is created.
+---
 ---@return number width # The width of the Canvas, in pixels.
 function Canvas:getWidth() end
 
 ---
----Returns whether the Canvas was created with the `stereo` flag.  Drawing something to a stereo Canvas will draw it to two viewports in the left and right half of the Canvas, using transform information from two different eyes.
+---Returns whether the Canvas was created with the `stereo` flag.
+---
+---Drawing something to a stereo Canvas will draw it to two viewports in the left and right half of the Canvas, using transform information from two different eyes.
 ---
 ---@return boolean stereo # Whether the Canvas is stereo.
 function Canvas:isStereo() end
@@ -813,54 +1455,103 @@ function Canvas:isStereo() end
 function Canvas:newImage(index) end
 
 ---
----Renders to the Canvas using a function.  All graphics functions inside the callback will affect the Canvas contents instead of directly rendering to the headset.  This can be used in `lovr.update`.
+---Renders to the Canvas using a function.
+---
+---All graphics functions inside the callback will affect the Canvas contents instead of directly rendering to the headset.
+---
+---This can be used in `lovr.update`.
+---
+---
+---### NOTE:
+---Make sure you clear the contents of the canvas before rendering by using `lovr.graphics.clear`. Otherwise there might be data in the canvas left over from a previous frame.
+---
+---Also note that the transform stack is not modified by this function.
+---
+---If you plan on modifying the transform stack inside your callback it may be a good idea to use `lovr.graphics.push` and `lovr.graphics.pop` so you can revert to the previous transform afterwards.
 ---
 ---@param callback function # The function to use to render to the Canvas.
 function Canvas:renderTo(callback) end
 
 ---
----Attaches one or more Textures to the Canvas.  When rendering to the Canvas, everything will be drawn to all attached Textures.  You can attach different layers of an array, cubemap, or volume texture, and also attach different mipmap levels of Textures.
+---Attaches one or more Textures to the Canvas.
+---
+---When rendering to the Canvas, everything will be drawn to all attached Textures.
+---
+---You can attach different layers of an array, cubemap, or volume texture, and also attach different mipmap levels of Textures.
+---
+---
+---### NOTE:
+---There are some restrictions on how textures can be attached:
+---
+---- Up to 4 textures can be attached at once.
+---- Textures must have the same dimensions and multisample settings as the Canvas.
+---
+---To specify layers and mipmaps to attach, specify them after the Texture.
+---
+---You can also optionally wrap them in a table.
 ---
 function Canvas:setTexture() end
 
 ---
----A Font is an object created from a TTF file.  It can be used to render text with `lovr.graphics.print`.
+---A Font is an object created from a TTF file.
+---
+---It can be used to render text with `lovr.graphics.print`.
 ---
 ---@class lovr.Font
 local Font = {}
 
 ---
----Returns the maximum distance that any glyph will extend above the Font's baseline.  Units are generally in meters, see `Font:getPixelDensity`.
+---Returns the maximum distance that any glyph will extend above the Font's baseline.
+---
+---Units are generally in meters, see `Font:getPixelDensity`.
 ---
 ---@return number ascent # The ascent of the Font.
 function Font:getAscent() end
 
 ---
----Returns the baseline of the Font.  This is where the characters "rest on", relative to the y coordinate of the drawn text.  Units are generally in meters, see `Font:setPixelDensity`.
+---Returns the baseline of the Font.
+---
+---This is where the characters "rest on", relative to the y coordinate of the drawn text.
+---
+---Units are generally in meters, see `Font:setPixelDensity`.
 ---
 ---@return number baseline # The baseline of the Font.
 function Font:getBaseline() end
 
 ---
----Returns the maximum distance that any glyph will extend below the Font's baseline.  Units are generally in meters, see `Font:getPixelDensity` for more information.  Note that due to the coordinate system for fonts, this is a negative value.
+---Returns the maximum distance that any glyph will extend below the Font's baseline.
+---
+---Units are generally in meters, see `Font:getPixelDensity` for more information.
+---
+---Note that due to the coordinate system for fonts, this is a negative value.
 ---
 ---@return number descent # The descent of the Font.
 function Font:getDescent() end
 
 ---
----Returns the height of a line of text.  Units are in meters, see `Font:setPixelDensity`.
+---Returns the height of a line of text.
+---
+---Units are in meters, see `Font:setPixelDensity`.
 ---
 ---@return number height # The height of a rendered line of text.
 function Font:getHeight() end
 
 ---
----Returns the current line height multiplier of the Font.  The default is 1.0.
+---Returns the current line height multiplier of the Font.
+---
+---The default is 1.0.
 ---
 ---@return number lineHeight # The line height.
 function Font:getLineHeight() end
 
 ---
----Returns the current pixel density for the Font.  The default is 1.0.  Normally, this is in pixels per meter.  When rendering to a 2D texture, the units are pixels.
+---Returns the current pixel density for the Font.
+---
+---The default is 1.0.
+---
+---Normally, this is in pixels per meter.
+---
+---When rendering to a 2D texture, the units are pixels.
 ---
 ---@return number pixelDensity # The current pixel density.
 function Font:getPixelDensity() end
@@ -874,6 +1565,11 @@ function Font:getRasterizer() end
 ---
 ---Returns the width and line count of a string when rendered using the font, taking into account an optional wrap limit.
 ---
+---
+---### NOTE:
+---To get the correct units returned, make sure the pixel density is set with
+---    `Font:setPixelDensity`.
+---
 ---@param text string # The text to get the width of.
 ---@param wrap? number # The width at which to wrap lines, or 0 for no wrap.
 ---@return number width # The maximum width of any line in the text.
@@ -881,32 +1577,52 @@ function Font:getRasterizer() end
 function Font:getWidth(text, wrap) end
 
 ---
----Returns whether the Font has a set of glyphs.  Any combination of strings and numbers (corresponding to character codes) can be specified.  This function will return true if the Font is able to render *all* of the glyphs.
+---Returns whether the Font has a set of glyphs.
+---
+---Any combination of strings and numbers (corresponding to character codes) can be specified.
+---
+---This function will return true if the Font is able to render *all* of the glyphs.
+---
+---
+---### NOTE:
+---It is a good idea to use this function when you're rendering an unknown or user-supplied string to avoid utterly embarrassing crashes.
 ---
 ---@return boolean has # Whether the Font has the glyphs.
 function Font:hasGlyphs() end
 
 ---
----Sets the line height of the Font, which controls how far lines apart lines are vertically separated.  This value is a ratio and the default is 1.0.
+---Sets the line height of the Font, which controls how far lines apart lines are vertically separated.
+---
+---This value is a ratio and the default is 1.0.
 ---
 ---@param lineHeight number # The new line height.
 function Font:setLineHeight(lineHeight) end
 
 ---
----Sets the pixel density for the Font.  Normally, this is in pixels per meter.  When rendering to a 2D texture, the units are pixels.
+---Sets the pixel density for the Font.
+---
+---Normally, this is in pixels per meter.
+---
+---When rendering to a 2D texture, the units are pixels.
 ---
 ---@overload fun(self: lovr.Font)
 ---@param pixelDensity number # The new pixel density.
 function Font:setPixelDensity(pixelDensity) end
 
 ---
----A Material is an object used to control how objects appear, through coloring, texturing, and shading.  The Material itself holds sets of colors, textures, and other parameters that are made available to Shaders.
+---A Material is an object used to control how objects appear, through coloring, texturing, and shading.
+---
+---The Material itself holds sets of colors, textures, and other parameters that are made available to Shaders.
 ---
 ---@class lovr.Material
 local Material = {}
 
 ---
----Returns a color property for a Material.  Different types of colors are supported for different lighting parameters.  Colors default to `(1.0, 1.0, 1.0, 1.0)` and are gamma corrected.
+---Returns a color property for a Material.
+---
+---Different types of colors are supported for different lighting parameters.
+---
+---Colors default to `(1.0, 1.0, 1.0, 1.0)` and are gamma corrected.
 ---
 ---@param colorType? lovr.MaterialColor # The type of color to get.
 ---@return number r # The red component of the color.
@@ -916,14 +1632,20 @@ local Material = {}
 function Material:getColor(colorType) end
 
 ---
----Returns a numeric property of a Material.  Scalar properties default to 1.0.
+---Returns a numeric property of a Material.
+---
+---Scalar properties default to 1.0.
 ---
 ---@param scalarType lovr.MaterialScalar # The type of property to get.
 ---@return number x # The value of the property.
 function Material:getScalar(scalarType) end
 
 ---
----Returns a texture for a Material.  Several predefined `MaterialTexture`s are supported.  Any texture that is `nil` will use a single white pixel as a fallback.
+---Returns a texture for a Material.
+---
+---Several predefined `MaterialTexture`s are supported.
+---
+---Any texture that is `nil` will use a single white pixel as a fallback.
 ---
 ---@param textureType? lovr.MaterialTexture # The type of texture to get.
 ---@return lovr.Texture texture # The texture that is set, or `nil` if no texture is set.
@@ -932,6 +1654,10 @@ function Material:getTexture(textureType) end
 ---
 ---Returns the transformation applied to texture coordinates of the Material.
 ---
+---
+---### NOTE:
+---Although texture coordinates will automatically be transformed by the Material's transform, the material transform is exposed as the `mat3 lovrMaterialTransform` uniform variable in shaders, allowing it to be used for other purposes.
+---
 ---@return number ox # The texture coordinate x offset.
 ---@return number oy # The texture coordinate y offset.
 ---@return number sx # The texture coordinate x scale.
@@ -940,7 +1666,11 @@ function Material:getTexture(textureType) end
 function Material:getTransform() end
 
 ---
----Sets a color property for a Material.  Different types of colors are supported for different lighting parameters.  Colors default to `(1.0, 1.0, 1.0, 1.0)` and are gamma corrected.
+---Sets a color property for a Material.
+---
+---Different types of colors are supported for different lighting parameters.
+---
+---Colors default to `(1.0, 1.0, 1.0, 1.0)` and are gamma corrected.
 ---
 ---@overload fun(self: lovr.Material, r: number, g: number, b: number, a: number)
 ---@overload fun(self: lovr.Material, colorType: lovr.MaterialColor, hex: number, a: number)
@@ -953,14 +1683,20 @@ function Material:getTransform() end
 function Material:setColor(colorType, r, g, b, a) end
 
 ---
----Sets a numeric property of a Material.  Scalar properties default to 1.0.
+---Sets a numeric property of a Material.
+---
+---Scalar properties default to 1.0.
 ---
 ---@param scalarType lovr.MaterialScalar # The type of property to set.
 ---@param x number # The value of the property.
 function Material:setScalar(scalarType, x) end
 
 ---
----Sets a texture for a Material.  Several predefined `MaterialTexture`s are supported.  Any texture that is `nil` will use a single white pixel as a fallback.
+---Sets a texture for a Material.
+---
+---Several predefined `MaterialTexture`s are supported.
+---
+---Any texture that is `nil` will use a single white pixel as a fallback.
 ---
 ---@overload fun(self: lovr.Material, texture: lovr.Texture)
 ---@param textureType? lovr.MaterialTexture # The type of texture to set.
@@ -968,7 +1704,13 @@ function Material:setScalar(scalarType, x) end
 function Material:setTexture(textureType, texture) end
 
 ---
----Sets the transformation applied to texture coordinates of the Material.  This lets you offset, scale, or rotate textures as they are applied to geometry.
+---Sets the transformation applied to texture coordinates of the Material.
+---
+---This lets you offset, scale, or rotate textures as they are applied to geometry.
+---
+---
+---### NOTE:
+---Although texture coordinates will automatically be transformed by the Material's transform, the material transform is exposed as the `mat3 lovrMaterialTransform` uniform variable in shaders, allowing it to be used for other purposes.
 ---
 ---@param ox number # The texture coordinate x offset.
 ---@param oy number # The texture coordinate y offset.
@@ -980,17 +1722,85 @@ function Material:setTransform(ox, oy, sx, sy, angle) end
 ---
 ---A Mesh is a low-level graphics object that stores and renders a list of vertices.
 ---
----Meshes are really flexible since you can pack pretty much whatever you want in them.  This makes them great for rendering arbitrary geometry, but it also makes them kinda difficult to use since you have to place each vertex yourself.
+---Meshes are really flexible since you can pack pretty much whatever you want in them.
+---
+---This makes them great for rendering arbitrary geometry, but it also makes them kinda difficult to use since you have to place each vertex yourself.
+---
+---It's possible to batch geometry with Meshes too.
+---
+---Instead of drawing a shape 100 times, it's much faster to pack 100 copies of the shape into a Mesh and draw the Mesh once.
 ---
----It's possible to batch geometry with Meshes too.  Instead of drawing a shape 100 times, it's much faster to pack 100 copies of the shape into a Mesh and draw the Mesh once.  Even storing just one copy in the Mesh and drawing that 100 times is usually faster.
+---Even storing just one copy in the Mesh and drawing that 100 times is usually faster.
 ---
 ---Meshes are also a good choice if you have an object that changes its shape over time.
 ---
+---
+---### NOTE:
+---Each vertex in a Mesh can hold several pieces of data.
+---
+---For example, you might want a vertex to keep track of its position, color, and a weight.
+---
+---Each one of these pieces of information is called a vertex **attribute**.
+---
+---A vertex attribute must have a name, a type, and a size.
+---
+---Here's what the "position" attribute would look like as a Lua table:
+---
+---    { 'vPosition', 'float', 3 } -- 3 floats, one for x, y, and z
+---
+---Every vertex in a Mesh must have the same set of attributes.
+---
+---We call this set of attributes the **format** of the Mesh, and it's specified as a simple table of attributes.
+---
+---For example, we could represent the format described above as:
+---
+---    {
+---      { 'vPosition', 'float', 3 },
+---      { 'vColor',    'byte',  4 },
+---      { 'vWeight',   'int',   1 }
+---    }
+---
+---When creating a Mesh, you can give it any format you want, or use the default.
+---
+---The default Mesh format looks like this:
+---
+---    {
+---      { 'lovrPosition',    'float', 3 },
+---      { 'lovrNormal',      'float', 3 },
+---      { 'lovrTexCoord',    'float', 2 }
+---    }
+---
+---Great, so why do we go through the trouble of naming everything in our vertex and saying what type and size it is?  The cool part is that we can access this data in a Shader.
+---
+---We can write a vertex Shader that has `in` variables for every vertex attribute in our Mesh:
+---
+---    in vec3 vPosition;
+---    in vec4 vColor;
+---    in int vWeight;
+---
+---    vec4 position(mat4 projection, mat4 transform, vec4 vertex) {
+---      // Here we can access the vPosition, vColor, and vWeight of each vertex in the Mesh!
+---    }
+---
+---Specifying custom vertex data is really powerful and is often used for lighting, animation, and more!
+---
+---For more on the different data types available for the attributes, see `AttributeType`.
+---
 ---@class lovr.Mesh
 local Mesh = {}
 
 ---
----Attaches attributes from another Mesh onto this one.  This can be used to share vertex data across multiple meshes without duplicating the data, and can also be used for instanced rendering by using the `divisor` parameter.
+---Attaches attributes from another Mesh onto this one.
+---
+---This can be used to share vertex data across multiple meshes without duplicating the data, and can also be used for instanced rendering by using the `divisor` parameter.
+---
+---
+---### NOTE:
+---The attribute divisor is a  number used to control how the attribute data relates to instancing. If 0, then the attribute data is considered "per vertex", and each vertex will get the next element of the attribute's data.
+---
+---If the divisor 1 or more, then the attribute data is considered "per instance", and every N instances will get the next element of the attribute data.
+---
+---To prevent cycles, it is not possible to attach attributes onto a Mesh that already has attributes attached to a different Mesh.
 ---
 ---@overload fun(self: lovr.Mesh, mesh: lovr.Mesh, divisor: number, ...)
 ---@overload fun(self: lovr.Mesh, mesh: lovr.Mesh, divisor: number, attributes: table)
@@ -1028,7 +1838,9 @@ function Mesh:draw(x, y, z, scale, angle, ax, ay, az, instances) end
 function Mesh:getDrawMode() end
 
 ---
----Retrieve the current draw range for the Mesh.  The draw range is a subset of the vertices of the Mesh that will be drawn.
+---Retrieve the current draw range for the Mesh.
+---
+---The draw range is a subset of the vertices of the Mesh that will be drawn.
 ---
 ---@return number start # The index of the first vertex that will be drawn, or nil if no draw range is set.
 ---@return number count # The number of vertices that will be drawn, or nil if no draw range is set.
@@ -1041,7 +1853,11 @@ function Mesh:getDrawRange() end
 function Mesh:getMaterial() end
 
 ---
----Gets the data for a single vertex in the Mesh.  The set of data returned depends on the Mesh's vertex format.  The default vertex format consists of 8 floating point numbers: the vertex position, the vertex normal, and the texture coordinates.
+---Gets the data for a single vertex in the Mesh.
+---
+---The set of data returned depends on the Mesh's vertex format.
+---
+---The default vertex format consists of 8 floating point numbers: the vertex position, the vertex normal, and the texture coordinates.
 ---
 ---@param index number # The index of the vertex to retrieve.
 function Mesh:getVertex(index) end
@@ -1049,6 +1865,10 @@ function Mesh:getVertex(index) end
 ---
 ---Returns the components of a specific attribute of a single vertex in the Mesh.
 ---
+---
+---### NOTE:
+---Meshes without a custom format have the vertex position as their first attribute, the normal vector as the second attribute, and the texture coordinate as the third attribute.
+---
 ---@param index number # The index of the vertex to retrieve the attribute of.
 ---@param attribute number # The index of the attribute to retrieve the components of.
 function Mesh:getVertexAttribute(index, attribute) end
@@ -1056,17 +1876,27 @@ function Mesh:getVertexAttribute(index, attribute) end
 ---
 ---Returns the maximum number of vertices the Mesh can hold.
 ---
+---
+---### NOTE:
+---The size can only be set when creating the Mesh, and cannot be changed afterwards.
+---
+---A subset of the Mesh's vertices can be rendered, see `Mesh:setDrawRange`.
+---
 ---@return number size # The number of vertices the Mesh can hold.
 function Mesh:getVertexCount() end
 
 ---
----Get the format table of the Mesh's vertices.  The format table describes the set of data that each vertex contains.
+---Get the format table of the Mesh's vertices.
+---
+---The format table describes the set of data that each vertex contains.
 ---
 ---@return table format # The table of vertex attributes.  Each attribute is a table containing the name of the attribute, the `AttributeType`, and the number of components.
 function Mesh:getVertexFormat() end
 
 ---
----Returns the current vertex map for the Mesh.  The vertex map is a list of indices in the Mesh, allowing the reordering or reuse of vertices.
+---Returns the current vertex map for the Mesh.
+---
+---The vertex map is a list of indices in the Mesh, allowing the reordering or reuse of vertices.
 ---
 ---@overload fun(self: lovr.Mesh, t: table):table
 ---@overload fun(self: lovr.Mesh, blob: lovr.Blob)
@@ -1074,14 +1904,18 @@ function Mesh:getVertexFormat() end
 function Mesh:getVertexMap() end
 
 ---
----Returns whether or not a vertex attribute is enabled.  Disabled attributes won't be sent to shaders.
+---Returns whether or not a vertex attribute is enabled.
+---
+---Disabled attributes won't be sent to shaders.
 ---
 ---@param attribute string # The name of the attribute.
 ---@return boolean enabled # Whether or not the attribute is enabled when drawing the Mesh.
 function Mesh:isAttributeEnabled(attribute) end
 
 ---
----Sets whether a vertex attribute is enabled.  Disabled attributes won't be sent to shaders.
+---Sets whether a vertex attribute is enabled.
+---
+---Disabled attributes won't be sent to shaders.
 ---
 ---@param attribute string # The name of the attribute.
 ---@param enabled boolean # Whether or not the attribute is enabled when drawing the Mesh.
@@ -1094,7 +1928,9 @@ function Mesh:setAttributeEnabled(attribute, enabled) end
 function Mesh:setDrawMode(mode) end
 
 ---
----Set the draw range for the Mesh.  The draw range is a subset of the vertices of the Mesh that will be drawn.
+---Set the draw range for the Mesh.
+---
+---The draw range is a subset of the vertices of the Mesh that will be drawn.
 ---
 ---@overload fun(self: lovr.Mesh)
 ---@param start number # The first vertex that will be drawn.
@@ -1102,7 +1938,9 @@ function Mesh:setDrawMode(mode) end
 function Mesh:setDrawRange(start, count) end
 
 ---
----Applies a Material to the Mesh.  This will cause it to use the Material's properties whenever it is rendered.
+---Applies a Material to the Mesh.
+---
+---This will cause it to use the Material's properties whenever it is rendered.
 ---
 ---@param material lovr.Material # The Material to apply.
 function Mesh:setMaterial(material) end
@@ -1110,18 +1948,28 @@ function Mesh:setMaterial(material) end
 ---
 ---Update a single vertex in the Mesh.
 ---
+---
+---### NOTE:
+---Any unspecified components will be set to 0.
+---
 ---@param index number # The index of the vertex to set.
 function Mesh:setVertex(index) end
 
 ---
 ---Set the components of a specific attribute of a vertex in the Mesh.
 ---
+---
+---### NOTE:
+---Meshes without a custom format have the vertex position as their first attribute, the normal vector as the second attribute, and the texture coordinate as the third attribute.
+---
 ---@param index number # The index of the vertex to update.
 ---@param attribute number # The index of the attribute to update.
 function Mesh:setVertexAttribute(index, attribute) end
 
 ---
----Sets the vertex map.  The vertex map is a list of indices in the Mesh, allowing the reordering or reuse of vertices.
+---Sets the vertex map.
+---
+---The vertex map is a list of indices in the Mesh, allowing the reordering or reuse of vertices.
 ---
 ---Often, a vertex map is used to improve performance, since it usually requires less data to specify the index of a vertex than it does to specify all of the data for a vertex.
 ---
@@ -1132,6 +1980,10 @@ function Mesh:setVertexMap(map) end
 ---
 ---Updates multiple vertices in the Mesh.
 ---
+---
+---### NOTE:
+---The start index plus the number of vertices in the table should not exceed the maximum size of the Mesh.
+---
 ---@overload fun(self: lovr.Mesh, blob: lovr.Blob, start: number, count: number)
 ---@param vertices table # The new set of vertices.
 ---@param start? number # The index of the vertex to start replacing at.
@@ -1139,7 +1991,9 @@ function Mesh:setVertexMap(map) end
 function Mesh:setVertices(vertices, start, count) end
 
 ---
----A Model is a drawable object loaded from a 3D file format.  The supported 3D file formats are OBJ, glTF, and STL.
+---A Model is a drawable object loaded from a 3D file format.
+---
+---The supported 3D file formats are OBJ, glTF, and STL.
 ---
 ---@class lovr.Model
 local Model = {}
@@ -1147,7 +2001,17 @@ local Model = {}
 ---
 ---Applies an animation to the current pose of the Model.
 ---
----The animation is evaluated at the specified timestamp, and mixed with the current pose of the Model using the alpha value.  An alpha value of 1.0 will completely override the pose of the Model with the animation's pose.
+---The animation is evaluated at the specified timestamp, and mixed with the current pose of the Model using the alpha value.
+---
+---An alpha value of 1.0 will completely override the pose of the Model with the animation's pose.
+---
+---
+---### NOTE:
+---For animations to properly show up, use a Shader created with the `animated` flag set to `true`. See `lovr.graphics.newShader` for more.
+---
+---Animations are always mixed in with the current pose, and the pose only ever changes by calling `Model:animate` and `Model:pose`.
+---
+---To clear the pose of a Model to the default, use `Model:pose(nil)`.
 ---
 ---@overload fun(self: lovr.Model, index: number, time: number, alpha: number)
 ---@param name string # The name of an animation.
@@ -1241,6 +2105,12 @@ function Model:getNodeName(index) end
 ---
 ---Returns the pose of a single node in the Model in a given `CoordinateSpace`.
 ---
+---
+---### NOTE:
+---For skinned nodes to render correctly, use a Shader created with the `animated` flag set to `true`.
+---
+---See `lovr.graphics.newShader` for more.
+---
 ---@overload fun(self: lovr.Model, index: number, space: lovr.CoordinateSpace):number, number, number, number, number, number, number
 ---@param name string # The name of the node.
 ---@param space? lovr.CoordinateSpace # Whether the pose should be returned relative to the node's parent or relative to the root node of the Model.
@@ -1256,7 +2126,9 @@ function Model:getNodePose(name, space) end
 ---
 ---Returns 2 tables containing mesh data for the Model.
 ---
----The first table is a list of vertex positions and contains 3 numbers for the x, y, and z coordinate of each vertex.  The second table is a list of triangles and contains 1-based indices into the first table representing the first, second, and third vertices that make up each triangle.
+---The first table is a list of vertex positions and contains 3 numbers for the x, y, and z coordinate of each vertex.
+---
+---The second table is a list of triangles and contains 1-based indices into the first table representing the first, second, and third vertices that make up each triangle.
 ---
 ---The vertex positions will be affected by node transforms.
 ---
@@ -1265,16 +2137,34 @@ function Model:getNodePose(name, space) end
 function Model:getTriangles() end
 
 ---
----Returns whether the Model has any nodes associated with animated joints.  This can be used to approximately determine whether an animated shader needs to be used with an arbitrary Model.
+---Returns whether the Model has any nodes associated with animated joints.
+---
+---This can be used to approximately determine whether an animated shader needs to be used with an arbitrary Model.
+---
+---
+---### NOTE:
+---A model can still be animated even if this function returns false, since node transforms can still be animated with keyframes without skinning.
+---
+---These types of animations don't need to use a Shader with the `animated = true` flag, though.
 ---
 ---@return boolean skeletal # Whether the Model has any nodes that use skeletal animation.
 function Model:hasJoints() end
 
 ---
----Applies a pose to a single node of the Model.  The input pose is assumed to be relative to the pose of the node's parent.  This is useful for applying inverse kinematics (IK) to a chain of bones in a skeleton.
+---Applies a pose to a single node of the Model.
+---
+---The input pose is assumed to be relative to the pose of the node's parent.
+---
+---This is useful for applying inverse kinematics (IK) to a chain of bones in a skeleton.
 ---
 ---The alpha parameter can be used to mix between the node's current pose and the input pose.
 ---
+---
+---### NOTE:
+---For skinned nodes to render correctly, use a Shader created with the `animated` flag set to `true`.
+---
+---See `lovr.graphics.newShader` for more.
+---
 ---@overload fun(self: lovr.Model, index: number, x: number, y: number, z: number, angle: number, ax: number, ay: number, az: number, alpha: number)
 ---@overload fun(self: lovr.Model)
 ---@param name string # The name of the node.
@@ -1289,7 +2179,100 @@ function Model:hasJoints() end
 function Model:pose(name, x, y, z, angle, ax, ay, az, alpha) end
 
 ---
----Shaders are GLSL programs that transform the way vertices and pixels show up on the screen. They can be used for lighting, postprocessing, particles, animation, and much more.  You can use `lovr.graphics.setShader` to change the active Shader.
+---Shaders are GLSL programs that transform the way vertices and pixels show up on the screen. They can be used for lighting, postprocessing, particles, animation, and much more.
+---
+---You can use `lovr.graphics.setShader` to change the active Shader.
+---
+---
+---### NOTE:
+---GLSL version `330` is used on desktop systems, and `300 es` on WebGL/Android.
+---
+---The default vertex shader:
+---
+---    vec4 position(mat4 projection, mat4 transform, vec4 vertex) {
+---      return projection * transform * vertex;
+---    }
+---
+---The default fragment shader:
+---
+---    vec4 color(vec4 graphicsColor, sampler2D image, vec2 uv) {
+---      return graphicsColor * lovrDiffuseColor * lovrVertexColor * texture(image, uv);
+---    }
+---
+---Additionally, the following headers are prepended to the shader source, giving you convenient access to a default set of uniform variables and vertex attributes.
+---
+---Vertex shader header:
+---
+---    in vec3 lovrPosition; // The vertex position
+---    in vec3 lovrNormal; // The vertex normal vector
+---    in vec2 lovrTexCoord;
+---    in vec4 lovrVertexColor;
+---    in vec3 lovrTangent;
+---    in uvec4 lovrBones;
+---    in vec4 lovrBoneWeights;
+---    in uint lovrDrawID;
+---    out vec4 lovrGraphicsColor;
+---    uniform mat4 lovrModel;
+---    uniform mat4 lovrView;
+---    uniform mat4 lovrProjection;
+---    uniform mat4 lovrTransform; // Model-View matrix
+---    uniform mat3 lovrNormalMatrix; // Inverse-transpose of lovrModel
+---    uniform mat3 lovrMaterialTransform;
+---    uniform float lovrPointSize;
+---    uniform mat4 lovrPose[48];
+---    uniform int lovrViewportCount;
+---    uniform int lovrViewID;
+---    const mat4 lovrPoseMatrix; // Bone-weighted pose
+---    const int lovrInstanceID; // Current instance ID
+---
+---Fragment shader header:
+---
+---    in vec2 lovrTexCoord;
+---    in vec4 lovrVertexColor;
+---    in vec4 lovrGraphicsColor;
+---    out vec4 lovrCanvas[gl_MaxDrawBuffers];
+---    uniform float lovrMetalness;
+---    uniform float lovrRoughness;
+---    uniform vec4 lovrDiffuseColor;
+---    uniform vec4 lovrEmissiveColor;
+---    uniform sampler2D lovrDiffuseTexture;
+---    uniform sampler2D lovrEmissiveTexture;
+---    uniform sampler2D lovrMetalnessTexture;
+---    uniform sampler2D lovrRoughnessTexture;
+---    uniform sampler2D lovrOcclusionTexture;
+---    uniform sampler2D lovrNormalTexture;
+---    uniform samplerCube lovrEnvironmentTexture;
+---    uniform int lovrViewportCount;
+---    uniform int lovrViewID;
+---
+---### Compute Shaders
+---
+---Compute shaders can be created with `lovr.graphics.newComputeShader` and run with `lovr.graphics.compute`.
+---
+---Currently, compute shaders are written with raw GLSL.
+---
+---There is no default compute shader, instead the `void compute();` function must be implemented.
+---
+---You can use the `layout` qualifier to specify a local work group size:
+---
+---    layout(local_size_x = X, local_size_y = Y, local_size_z = Z) in;
+---
+---And the following built in variables can be used:
+---
+---    in uvec3 gl_NumWorkGroups;       // The size passed to lovr.graphics.compute
+---    in uvec3 gl_WorkGroupSize;       // The local work group size
+---    in uvec3 gl_WorkGroupID;         // The current global work group
+---    in uvec3 gl_LocalInvocationID;   // The current local work group
+---    in uvec3 gl_GlobalInvocationID;  // A unique ID combining the global and local IDs
+---    in uint gl_LocalInvocationIndex; // A 1D index of the LocalInvocationID
+---
+---Compute shaders don't return anything but they can write data to `Texture`s or `ShaderBlock`s. To bind a texture in a way that can be written to a compute shader, declare the uniforms with a type of `image2D`, `imageCube`, etc. instead of the usual `sampler2D` or `samplerCube`.
+---
+---Once a texture is bound to an image uniform, you can use the `imageLoad` and `imageStore` GLSL functions to read and write pixels in the image.
+---
+---Variables in `ShaderBlock`s can be written to using assignment syntax.
+---
+---LÖVR handles synchronization of textures and shader blocks so there is no need to use manual memory barriers to synchronize writes to resources from compute shaders.
 ---
 ---@class lovr.Shader
 local Shader = {}
@@ -1297,7 +2280,9 @@ local Shader = {}
 ---
 ---Returns the type of the Shader, which will be "graphics" or "compute".
 ---
----Graphics shaders are created with `lovr.graphics.newShader` and can be used for rendering with `lovr.graphics.setShader`.  Compute shaders are created with `lovr.graphics.newComputeShader` and can be run using `lovr.graphics.compute`.
+---Graphics shaders are created with `lovr.graphics.newShader` and can be used for rendering with `lovr.graphics.setShader`.
+---
+---Compute shaders are created with `lovr.graphics.newComputeShader` and can be run using `lovr.graphics.compute`.
 ---
 ---@return lovr.ShaderType type # The type of the Shader.
 function Shader:getType() end
@@ -1305,7 +2290,11 @@ function Shader:getType() end
 ---
 ---Returns whether a Shader has a block.
 ---
----A block is added to the Shader code at creation time using `ShaderBlock:getShaderCode`.  The block name (not the namespace) is used to link up the ShaderBlock object to the Shader.  This function can be used to check if a Shader was created with a block using the given name.
+---A block is added to the Shader code at creation time using `ShaderBlock:getShaderCode`.
+---
+---The block name (not the namespace) is used to link up the ShaderBlock object to the Shader.
+---
+---This function can be used to check if a Shader was created with a block using the given name.
 ---
 ---@param block string # The name of the block.
 ---@return boolean present # Whether the shader has the specified block.
@@ -1314,6 +2303,10 @@ function Shader:hasBlock(block) end
 ---
 ---Returns whether a Shader has a particular uniform variable.
 ---
+---
+---### NOTE:
+---If a uniform variable is defined but unused in the shader, the shader compiler will optimize it out and the uniform will not report itself as present.
+---
 ---@param uniform string # The name of the uniform variable.
 ---@return boolean present # Whether the shader has the specified uniform.
 function Shader:hasUniform(uniform) end
@@ -1321,13 +2314,106 @@ function Shader:hasUniform(uniform) end
 ---
 ---Updates a uniform variable in the Shader.
 ---
+---
+---### NOTE:
+---The shader does not need to be active to update its uniforms.
+---
+---The following type combinations are supported:
+---
+---<table>
+---  <thead>
+---    <tr>
+---      <td>Uniform type</td>
+---      <td>LÖVR type</td>
+---    </tr>
+---  </thead>
+---  <tbody>
+---    <tr>
+---      <td>float</td>
+---      <td>number</td>
+---    </tr>
+---    <tr>
+---      <td>int</td>
+---      <td>number</td>
+---    </tr>
+---    <tr>
+---      <td>vec2</td>
+---      <td>{ x, y }</td>
+---    </tr>
+---    <tr>
+---      <td>vec3</td>
+---      <td>{ x, y, z } or vec3</td>
+---    </tr>
+---    <tr>
+---      <td>vec4</td>
+---      <td>{ x, y, z, w }</td>
+---    </tr>
+---    <tr>
+---      <td>ivec2</td>
+---      <td>{ x, y }</td>
+---    </tr>
+---    <tr>
+---      <td>ivec3</td>
+---      <td>{ x, y, z }</td>
+---    </tr>
+---    <tr>
+---      <td>ivec4</td>
+---      <td>{ x, y, z, w }</td>
+---    </tr>
+---    <tr>
+---      <td>mat2</td>
+---      <td>{ x, ... }</td>
+---    </tr>
+---    <tr>
+---      <td>mat3</td>
+---      <td>{ x, ... }</td>
+---    </tr>
+---    <tr>
+---      <td>mat4</td>
+---      <td>{ x, ... } or mat4</td>
+---    </tr>
+---    <tr>
+---      <td>sampler</td>
+---      <td>Texture</td>
+---    </tr>
+---    <tr>
+---      <td>image</td>
+---      <td>Texture</td>
+---    </tr>
+---  </tbody> </table>
+---
+---Uniform arrays can be wrapped in tables or passed as multiple arguments.
+---
+---Textures must match the type of sampler or image they are being sent to.
+---
+---The following sampler (and image) types are currently supported:
+---
+---- `sampler2D`
+---- `sampler3D`
+---- `samplerCube`
+---- `sampler2DArray`
+---
+---`Blob`s can be used to pass arbitrary binary data to Shader variables.
+---
 ---@param uniform string # The name of the uniform to update.
 ---@param value any # The new value of the uniform.
 ---@return boolean success # Whether the uniform exists and was updated.
 function Shader:send(uniform, value) end
 
 ---
----Sends a ShaderBlock to a Shader.  After the block is sent, you can update the data in the block without needing to resend the block.  The block can be sent to multiple shaders and they will all see the same data from the block.
+---Sends a ShaderBlock to a Shader.
+---
+---After the block is sent, you can update the data in the block without needing to resend the block.
+---
+---The block can be sent to multiple shaders and they will all see the same data from the block.
+---
+---
+---### NOTE:
+---The Shader does not need to be active to send it a block.
+---
+---Make sure the ShaderBlock's variables line up with the block variables declared in the shader code, otherwise you'll get garbage data in the block.
+---
+---An easy way to do this is to use `ShaderBlock:getShaderCode` to get a GLSL snippet that is compatible with the block.
 ---
 ---@param name string # The name of the block to send to.
 ---@param block lovr.ShaderBlock # The ShaderBlock to associate with the specified block.
@@ -1335,7 +2421,11 @@ function Shader:send(uniform, value) end
 function Shader:sendBlock(name, block, access) end
 
 ---
----Sends a Texture to a Shader for writing.  This is meant to be used with compute shaders and only works with uniforms declared as `image2D`, `imageCube`, `image2DArray`, and `image3D`.  The normal `Shader:send` function accepts Textures and should be used most of the time.
+---Sends a Texture to a Shader for writing.
+---
+---This is meant to be used with compute shaders and only works with uniforms declared as `image2D`, `imageCube`, `image2DArray`, and `image3D`.
+---
+---The normal `Shader:send` function accepts Textures and should be used most of the time.
 ---
 ---@overload fun(self: lovr.Shader, name: string, index: number, texture: lovr.Texture, slice: number, mipmap: number, access: lovr.UniformAccess)
 ---@param name string # The name of the image uniform.
@@ -1346,27 +2436,48 @@ function Shader:sendBlock(name, block, access) end
 function Shader:sendImage(name, texture, slice, mipmap, access) end
 
 ---
----ShaderBlocks are objects that can hold large amounts of data and can be sent to Shaders.  It is common to use "uniform" variables to send data to shaders, but uniforms are usually limited to a few kilobytes in size.  ShaderBlocks are useful for a few reasons:
+---ShaderBlocks are objects that can hold large amounts of data and can be sent to Shaders.
+---
+---It is common to use "uniform" variables to send data to shaders, but uniforms are usually limited to a few kilobytes in size.
+---
+---ShaderBlocks are useful for a few reasons:
 ---
 ---- They can hold a lot more data.
 ---- Shaders can modify the data in them, which is really useful for compute shaders.
 ---- Setting the data in a ShaderBlock updates the data for all Shaders using the block, so you
 ---  don't need to go around setting the same uniforms in lots of different shaders.
 ---
----On systems that support compute shaders, ShaderBlocks can optionally be "writable".  A writable ShaderBlock is a little bit slower than a non-writable one, but shaders can modify its contents and it can be much, much larger than a non-writable ShaderBlock.
+---On systems that support compute shaders, ShaderBlocks can optionally be "writable".
+---
+---A writable ShaderBlock is a little bit slower than a non-writable one, but shaders can modify its contents and it can be much, much larger than a non-writable ShaderBlock.
+---
+---
+---### NOTE:
+---- A Shader can use up to 8 ShaderBlocks.
+---- ShaderBlocks can not contain textures.
+---- Some systems have bugs with `vec3` variables in ShaderBlocks.
+---
+---If you run into strange bugs,
+---  try switching to a `vec4` for the variable.
 ---
 ---@class lovr.ShaderBlock
 local ShaderBlock = {}
 
 ---
----Returns the byte offset of a variable in a ShaderBlock.  This is useful if you want to manually send binary data to the ShaderBlock using a `Blob` in `ShaderBlock:send`.
+---Returns the byte offset of a variable in a ShaderBlock.
+---
+---This is useful if you want to manually send binary data to the ShaderBlock using a `Blob` in `ShaderBlock:send`.
 ---
 ---@param field string # The name of the variable to get the offset of.
 ---@return number offset # The byte offset of the variable.
 function ShaderBlock:getOffset(field) end
 
 ---
----Before a ShaderBlock can be used in a Shader, the Shader has to have the block's variables defined in its source code.  This can be a tedious process, so you can call this function to return a GLSL string that contains this definition.  Roughly, it will look something like this:
+---Before a ShaderBlock can be used in a Shader, the Shader has to have the block's variables defined in its source code.
+---
+---This can be a tedious process, so you can call this function to return a GLSL string that contains this definition.
+---
+---Roughly, it will look something like this:
 ---
 ---    layout(std140) uniform <label> {
 ---      <type> <name>[<count>];
@@ -1392,6 +2503,12 @@ function ShaderBlock:getType() end
 ---
 ---Returns a variable in the ShaderBlock.
 ---
+---
+---### NOTE:
+---This function is really slow!  Only read back values when you need to.
+---
+---Vectors and matrices will be returned as (flat) tables.
+---
 ---@param name string # The name of the variable to read.
 ---@return any value # The value of the variable.
 function ShaderBlock:read(name) end
@@ -1399,13 +2516,25 @@ function ShaderBlock:read(name) end
 ---
 ---Updates a variable in the ShaderBlock.
 ---
+---
+---### NOTE:
+---For scalar or vector types, use tables of numbers or `vec3`s for each vector.
+---
+---For matrix types, use tables of numbers or `mat4` objects.
+---
+---`Blob`s can also be used to pass arbitrary binary data to individual variables.
+---
 ---@overload fun(self: lovr.ShaderBlock, blob: lovr.Blob, offset: number, extent: number):number
 ---@param variable string # The name of the variable to update.
 ---@param value any # The new value of the uniform.
 function ShaderBlock:send(variable, value) end
 
 ---
----A Texture is an image that can be applied to `Material`s.  The supported file formats are `.png`, `.jpg`, `.hdr`, `.dds`, `.ktx`, and `.astc`.  DDS and ASTC are compressed formats, which are recommended because they're smaller and faster.
+---A Texture is an image that can be applied to `Material`s.
+---
+---The supported file formats are `.png`, `.jpg`, `.hdr`, `.dds`, `.ktx`, and `.astc`.
+---
+---DDS and ASTC are compressed formats, which are recommended because they're smaller and faster.
 ---
 ---@class lovr.Texture
 local Texture = {}
@@ -1440,7 +2569,13 @@ function Texture:getDimensions(mipmap) end
 function Texture:getFilter() end
 
 ---
----Returns the format of the Texture.  This describes how many color channels are in the texture as well as the size of each one.  The most common format used is `rgba`, which contains red, green, blue, and alpha color channels.  See `TextureFormat` for all of the possible formats.
+---Returns the format of the Texture.
+---
+---This describes how many color channels are in the texture as well as the size of each one.
+---
+---The most common format used is `rgba`, which contains red, green, blue, and alpha color channels.
+---
+---See `TextureFormat` for all of the possible formats.
 ---
 ---@return lovr.TextureFormat format # The format of the Texture.
 function Texture:getFormat() end
@@ -1489,7 +2624,11 @@ function Texture:getWrap() end
 function Texture:replacePixels(image, x, y, slice, mipmap) end
 
 ---
----Sets the compare mode for a texture.  This is only used for "shadow samplers", which are uniform variables in shaders with type `sampler2DShadow`.  Sampling a shadow sampler uses a sort of virtual depth test, and the compare mode of the texture is used to control how the depth test is performed.
+---Sets the compare mode for a texture.
+---
+---This is only used for "shadow samplers", which are uniform variables in shaders with type `sampler2DShadow`.
+---
+---Sampling a shadow sampler uses a sort of virtual depth test, and the compare mode of the texture is used to control how the depth test is performed.
 ---
 ---@param compareMode? lovr.CompareMode # The new compare mode.  Use `nil` to disable the compare mode.
 function Texture:setCompareMode(compareMode) end
@@ -1497,12 +2636,22 @@ function Texture:setCompareMode(compareMode) end
 ---
 ---Sets the `FilterMode` used by the texture.
 ---
+---
+---### NOTE:
+---The default setting for new textures can be set with `lovr.graphics.setDefaultFilter`.
+---
+---The maximum supported anisotropy level can be queried using `lovr.graphics.getLimits`.
+---
 ---@param mode lovr.FilterMode # The filter mode.
 ---@param anisotropy number # The level of anisotropy to use.
 function Texture:setFilter(mode, anisotropy) end
 
 ---
----Sets the wrap mode of a texture.  The wrap mode controls how the texture is sampled when texture coordinates lie outside the usual 0 - 1 range.  The default for both directions is `repeat`.
+---Sets the wrap mode of a texture.
+---
+---The wrap mode controls how the texture is sampled when texture coordinates lie outside the usual 0 - 1 range.
+---
+---The default for both directions is `repeat`.
 ---
 ---@param horizontal lovr.WrapMode # How the texture should wrap horizontally.
 ---@param vertical? lovr.WrapMode # How the texture should wrap vertically.
@@ -1526,7 +2675,11 @@ function Texture:setWrap(horizontal, vertical) end
 ---| '"closed"'
 
 ---
----Here are the different data types available for vertex attributes in a Mesh.  The ones that have a smaller range take up less memory, which improves performance a bit.  The "u" stands for "unsigned", which means it can't hold negative values but instead has a larger positive range.
+---Here are the different data types available for vertex attributes in a Mesh.
+---
+---The ones that have a smaller range take up less memory, which improves performance a bit.
+---
+---The "u" stands for "unsigned", which means it can't hold negative values but instead has a larger positive range.
 ---
 ---@alias lovr.AttributeType
 ---
@@ -1561,13 +2714,21 @@ function Texture:setWrap(horizontal, vertical) end
 ---
 ---Different ways the alpha channel of pixels affects blending.
 ---
+---
+---### NOTE:
+---The premultiplied mode should be used if pixels being drawn have already been blended, or "pre-multiplied", by the alpha channel.
+---
+---This happens when rendering a framebuffer that contains pixels with transparent alpha values, since the stored color values have already been faded by alpha and don't need to be faded a second time with the alphamultiply blend mode.
+---
 ---@alias lovr.BlendAlphaMode
 ---
 ---Color channel values are multiplied by the alpha channel during blending.
 ---
 ---| '"alphamultiply"'
 ---
----Color channels are not multiplied by the alpha channel.  This should be used if the pixels being drawn have already been blended, or "pre-multiplied", by the alpha channel.
+---Color channels are not multiplied by the alpha channel.
+---
+---This should be used if the pixels being drawn have already been blended, or "pre-multiplied", by the alpha channel.
 ---
 ---| '"premultiplied"'
 
@@ -1639,7 +2800,13 @@ function Texture:setWrap(horizontal, vertical) end
 ---| '"stream"'
 
 ---
----The method used to compare z values when deciding how to overlap rendered objects.  This is called the "depth test", and it happens on a pixel-by-pixel basis every time new objects are drawn.  If the depth test "passes" for a pixel, then the pixel color will be replaced by the new color and the depth value in the depth buffer will be updated.  Otherwise, the pixel will not be changed and the depth value will not be updated.
+---The method used to compare z values when deciding how to overlap rendered objects.
+---
+---This is called the "depth test", and it happens on a pixel-by-pixel basis every time new objects are drawn.
+---
+---If the depth test "passes" for a pixel, then the pixel color will be replaced by the new color and the depth value in the depth buffer will be updated.
+---
+---Otherwise, the pixel will not be changed and the depth value will not be updated.
 ---
 ---@alias lovr.CompareMode
 ---
@@ -1681,7 +2848,11 @@ function Texture:setWrap(horizontal, vertical) end
 ---| '"global"'
 
 ---
----The following shaders are built in to LÖVR, and can be used as an argument to `lovr.graphics.newShader` instead of providing raw GLSL shader code.  The shaders can be further customized by using the `flags` argument.  If you pass in `nil` to `lovr.graphics.setShader`, LÖVR will automatically pick a DefaultShader to use based on whatever is being drawn.
+---The following shaders are built in to LÖVR, and can be used as an argument to `lovr.graphics.newShader` instead of providing raw GLSL shader code.
+---
+---The shaders can be further customized by using the `flags` argument.
+---
+---If you pass in `nil` to `lovr.graphics.setShader`, LÖVR will automatically pick a DefaultShader to use based on whatever is being drawn.
 ---
 ---@alias lovr.DefaultShader
 ---
@@ -1710,7 +2881,9 @@ function Texture:setWrap(horizontal, vertical) end
 ---| '"fill"'
 
 ---
----Meshes are lists of arbitrary vertices.  These vertices can be connected in different ways, leading to different shapes like lines and triangles.
+---Meshes are lists of arbitrary vertices.
+---
+---These vertices can be connected in different ways, leading to different shapes like lines and triangles.
 ---
 ---@alias lovr.DrawMode
 ---
@@ -1730,7 +2903,9 @@ function Texture:setWrap(horizontal, vertical) end
 ---
 ---| '"lineloop"'
 ---
----The first three vertices define a triangle.  Each vertex after that creates a triangle using the new vertex and last two vertices.
+---The first three vertices define a triangle.
+---
+---Each vertex after that creates a triangle using the new vertex and last two vertices.
 ---
 ---| '"strip"'
 ---
@@ -1738,7 +2913,9 @@ function Texture:setWrap(horizontal, vertical) end
 ---
 ---| '"triangles"'
 ---
----Draws a set of triangles.  Each one shares the first vertex as a common point, leading to a fan-like shape.
+---Draws a set of triangles.
+---
+---Each one shares the first vertex as a common point, leading to a fan-like shape.
 ---
 ---| '"fan"'
 
@@ -1756,11 +2933,17 @@ function Texture:setWrap(horizontal, vertical) end
 ---| '"line"'
 
 ---
----The method used to downsample (or upsample) a texture.  "nearest" can be used for a pixelated effect, whereas "linear" leads to more smooth results.  Nearest is slightly faster than linear.
+---The method used to downsample (or upsample) a texture.
+---
+---"nearest" can be used for a pixelated effect, whereas "linear" leads to more smooth results.
+---
+---Nearest is slightly faster than linear.
 ---
 ---@alias lovr.FilterMode
 ---
----Fast nearest-neighbor sampling.  Leads to a pixelated style.
+---Fast nearest-neighbor sampling.
+---
+---Leads to a pixelated style.
 ---
 ---| '"nearest"'
 ---
@@ -1849,7 +3032,9 @@ function Texture:setWrap(horizontal, vertical) end
 ---| '"environment"'
 
 ---
----Meshes can have a usage hint, describing how they are planning on being updated.  Setting the usage hint allows the graphics driver optimize how it handles the data in the Mesh.
+---Meshes can have a usage hint, describing how they are planning on being updated.
+---
+---Setting the usage hint allows the graphics driver optimize how it handles the data in the Mesh.
 ---
 ---@alias lovr.MeshUsage
 ---
@@ -1866,7 +3051,11 @@ function Texture:setWrap(horizontal, vertical) end
 ---| '"stream"'
 
 ---
----Shaders can be used for either rendering operations or generic compute tasks.  Graphics shaders are created with `lovr.graphics.newShader` and compute shaders are created with `lovr.graphics.newComputeShader`.  `Shader:getType` can be used on an existing Shader to figure out what type it is.
+---Shaders can be used for either rendering operations or generic compute tasks.
+---
+---Graphics shaders are created with `lovr.graphics.newShader` and compute shaders are created with `lovr.graphics.newComputeShader`.
+---
+---`Shader:getType` can be used on an existing Shader to figure out what type it is.
 ---
 ---@alias lovr.ShaderType
 ---
@@ -1908,7 +3097,9 @@ function Texture:setWrap(horizontal, vertical) end
 ---| '"invert"'
 
 ---
----Textures can store their pixels in different formats.  The set of color channels and the number of bits stored for each channel can differ, allowing Textures to optimize their storage for certain kinds of image formats or rendering techniques.
+---Textures can store their pixels in different formats.
+---
+---The set of color channels and the number of bits stored for each channel can differ, allowing Textures to optimize their storage for certain kinds of image formats or rendering techniques.
 ---
 ---@alias lovr.TextureFormat
 ---
@@ -1994,7 +3185,9 @@ function Texture:setWrap(horizontal, vertical) end
 ---| '"volume"'
 
 ---
----When binding writable resources to shaders using `Shader:sendBlock` and `Shader:sendImage`, an access pattern can be specified as a hint that says whether you plan to read or write to the resource (or both).  Sometimes, LÖVR or the GPU driver can use this hint to get better performance or avoid stalling.
+---When binding writable resources to shaders using `Shader:sendBlock` and `Shader:sendImage`, an access pattern can be specified as a hint that says whether you plan to read or write to the resource (or both).
+---
+---Sometimes, LÖVR or the GPU driver can use this hint to get better performance or avoid stalling.
 ---
 ---@alias lovr.UniformAccess
 ---
diff --git a/meta/3rd/lovr/library/lovr.headset.lua b/meta/3rd/lovr/library/lovr.headset.lua
index e45d3009..218b70b9 100644
--- a/meta/3rd/lovr/library/lovr.headset.lua
+++ b/meta/3rd/lovr/library/lovr.headset.lua
@@ -1,16 +1,34 @@
 ---@meta
 
 ---
----The `lovr.headset` module is where all the magical VR functionality is.  With it, you can access connected VR hardware and get information about the available space the player has.  Note that all units are reported in meters.  Position `(0, 0, 0)` is the center of the play area.
+---The `lovr.headset` module is where all the magical VR functionality is.
+---
+---With it, you can access connected VR hardware and get information about the available space the player has.
+---
+---Note that all units are reported in meters.
+---
+---Position `(0, 0, 0)` is the center of the play area.
 ---
 ---@class lovr.headset
 lovr.headset = {}
 
 ---
----Animates a device model to match its current input state.  The buttons and joysticks on a controller will move as they're pressed/moved and hand models will move to match skeletal input.
+---Animates a device model to match its current input state.
+---
+---The buttons and joysticks on a controller will move as they're pressed/moved and hand models will move to match skeletal input.
 ---
 ---The model should have been created using `lovr.headset.newModel` with the `animated` flag set to `true`.
 ---
+---
+---### NOTE:
+---Currently this function is supported for OpenVR controller models and Oculus hand models.
+---
+---This function may animate using node-based animation or skeletal animation.
+---
+---`Model:hasJoints` can be used on a Model so you know if a Shader with the `animated` ShaderFlag needs to be used to render the results properly.
+---
+---It's possible to use models that weren't created with `lovr.headset.newModel` but they need to be set up carefully to have the same structure as the models provided by the headset SDK.
+---
 ---@param device? lovr.Device # The device to use for the animation data.
 ---@param model lovr.Model # The model to animate.
 ---@return boolean success # Whether the animation was applied successfully to the Model.  If the Model was not compatible or animation data for the device was not available, this will be `false`.
@@ -26,7 +44,17 @@ function lovr.headset.animate(device, model) end
 function lovr.headset.getAngularVelocity(device) end
 
 ---
----Get the current state of an analog axis on a device.  Some axes are multidimensional, for example a 2D touchpad or thumbstick with x/y axes.  For multidimensional axes, this function will return multiple values, one number for each axis.  In these cases, it can be useful to use the `select` function built in to Lua to select a particular axis component.
+---Get the current state of an analog axis on a device.
+---
+---Some axes are multidimensional, for example a 2D touchpad or thumbstick with x/y axes.
+---
+---For multidimensional axes, this function will return multiple values, one number for each axis.
+---
+---In these cases, it can be useful to use the `select` function built in to Lua to select a particular axis component.
+---
+---
+---### NOTE:
+---The axis values will be between 0 and 1 for 1D axes, and between -1 and 1 for each component of a multidimensional axis.
 ---
 ---@param device lovr.Device # The device.
 ---@param axis lovr.DeviceAxis # The axis.
@@ -35,12 +63,20 @@ function lovr.headset.getAxis(device, axis) end
 ---
 ---Returns the depth of the play area, in meters.
 ---
+---
+---### NOTE:
+---This currently returns 0 on the Quest.
+---
 ---@return number depth # The depth of the play area, in meters.
 function lovr.headset.getBoundsDepth() end
 
 ---
 ---Returns the size of the play area, in meters.
 ---
+---
+---### NOTE:
+---This currently returns 0 on the Quest.
+---
 ---@return number width # The width of the play area, in meters.
 ---@return number depth # The depth of the play area, in meters.
 function lovr.headset.getBoundsDimensions() end
@@ -55,11 +91,23 @@ function lovr.headset.getBoundsGeometry(t) end
 ---
 ---Returns the width of the play area, in meters.
 ---
+---
+---### NOTE:
+---This currently returns 0 on the Quest.
+---
 ---@return number width # The width of the play area, in meters.
 function lovr.headset.getBoundsWidth() end
 
 ---
----Returns the near and far clipping planes used to render to the headset.  Objects closer than the near clipping plane or further than the far clipping plane will be clipped out of view.
+---Returns the near and far clipping planes used to render to the headset.
+---
+---Objects closer than the near clipping plane or further than the far clipping plane will be clipped out of view.
+---
+---
+---### NOTE:
+---The default near and far clipping planes are 0.1 meters and 100.0 meters.
+---
+---This is not currently supported by the `vrapi` headset driver.
 ---
 ---@return number near # The distance to the near clipping plane, in meters.
 ---@return number far # The distance to the far clipping plane, in meters.
@@ -85,7 +133,11 @@ function lovr.headset.getDisplayFrequency() end
 function lovr.headset.getDisplayHeight() end
 
 ---
----Returns 2D triangle vertices that represent areas of the headset display that will never be seen by the user (due to the circular lenses).  This area can be masked out by rendering it to the depth buffer or stencil buffer.  Then, further drawing operations can skip rendering those pixels using the depth test (`lovr.graphics.setDepthTest`) or stencil test (`lovr.graphics.setStencilTest`), which improves performance.
+---Returns 2D triangle vertices that represent areas of the headset display that will never be seen by the user (due to the circular lenses).
+---
+---This area can be masked out by rendering it to the depth buffer or stencil buffer.
+---
+---Then, further drawing operations can skip rendering those pixels using the depth test (`lovr.graphics.setDepthTest`) or stencil test (`lovr.graphics.setStencilTest`), which improves performance.
 ---
 ---@return table points # A table of points.  Each point is a table with two numbers between 0 and 1.
 function lovr.headset.getDisplayMask() end
@@ -97,7 +149,9 @@ function lovr.headset.getDisplayMask() end
 function lovr.headset.getDisplayWidth() end
 
 ---
----Returns the `HeadsetDriver` that is currently in use, optionally for a specific device.  The order of headset drivers can be changed using `lovr.conf` to prefer or exclude specific VR APIs.
+---Returns the `HeadsetDriver` that is currently in use, optionally for a specific device.
+---
+---The order of headset drivers can be changed using `lovr.conf` to prefer or exclude specific VR APIs.
 ---
 ---@overload fun(device: lovr.Device):lovr.HeadsetDriver
 ---@return lovr.HeadsetDriver driver # The driver of the headset in use, e.g. "OpenVR".
@@ -106,21 +160,71 @@ function lovr.headset.getDriver() end
 ---
 ---Returns a table with all of the currently tracked hand devices.
 ---
+---
+---### NOTE:
+---The hand paths will *always* be either `hand/left` or `hand/right`.
+---
 ---@return table hands # The currently tracked hand devices.
 function lovr.headset.getHands() end
 
 ---
 ---Returns a Texture that contains whatever is currently rendered to the headset.
 ---
----Sometimes this can be `nil` if the current headset driver doesn't have a mirror texture, which can happen if the driver renders directly to the display.  Currently the `desktop`, `webxr`, and `vrapi` drivers do not have a mirror texture.
+---Sometimes this can be `nil` if the current headset driver doesn't have a mirror texture, which can happen if the driver renders directly to the display.
 ---
----It also isn't guaranteed that the same Texture will be returned by subsequent calls to this function.  Currently, the `oculus` driver exhibits this behavior.
+---Currently the `desktop`, `webxr`, and `vrapi` drivers do not have a mirror texture.
+---
+---It also isn't guaranteed that the same Texture will be returned by subsequent calls to this function.
+---
+---Currently, the `oculus` driver exhibits this behavior.
 ---
 ---@return lovr.Texture mirror # The mirror texture.
 function lovr.headset.getMirrorTexture() end
 
 ---
----Returns the name of the headset as a string.  The exact string that is returned depends on the hardware and VR SDK that is currently in use.
+---Returns the name of the headset as a string.
+---
+---The exact string that is returned depends on the hardware and VR SDK that is currently in use.
+---
+---
+---### NOTE:
+---<table>
+---  <thead>
+---    <tr>
+---      <td>driver</td>
+---      <td>name</td>
+---    </tr>
+---  </thead>
+---  <tbody>
+---    <tr>
+---      <td>desktop</td>
+---      <td><code>Simulator</code></td>
+---    </tr>
+---    <tr>
+---      <td>openvr</td>
+---      <td>varies</td>
+---    </tr>
+---    <tr>
+---      <td>openxr</td>
+---      <td>varies</td>
+---    </tr>
+---    <tr>
+---      <td>vrapi</td>
+---      <td><code>Oculus Quest</code> or <code>Oculus Quest 2</code></td>
+---    </tr>
+---    <tr>
+---      <td>webxr</td>
+---      <td>always nil</td>
+---    </tr>
+---    <tr>
+---      <td>oculus</td>
+---      <td>varies</td>
+---    </tr>
+---    <tr>
+---      <td>pico</td>
+---      <td><code>Pico</code></td>
+---    </tr>
+---  </tbody> </table>
 ---
 ---@return string name # The name of the headset as a string.
 function lovr.headset.getName() end
@@ -128,6 +232,10 @@ function lovr.headset.getName() end
 ---
 ---Returns the current orientation of a device, in angle/axis form.
 ---
+---
+---### NOTE:
+---If the device isn't tracked, all zeroes will be returned.
+---
 ---@param device? lovr.Device # The device to get the orientation of.
 ---@return number angle # The amount of rotation around the axis of rotation, in radians.
 ---@return number ax # The x component of the axis of rotation.
@@ -136,7 +244,9 @@ function lovr.headset.getName() end
 function lovr.headset.getOrientation(device) end
 
 ---
----Returns the type of origin used for the tracking volume.  The different types of origins are explained on the `HeadsetOrigin` page.
+---Returns the type of origin used for the tracking volume.
+---
+---The different types of origins are explained on the `HeadsetOrigin` page.
 ---
 ---@return lovr.HeadsetOrigin origin # The type of origin.
 function lovr.headset.getOriginType() end
@@ -144,6 +254,12 @@ function lovr.headset.getOriginType() end
 ---
 ---Returns the current position and orientation of a device.
 ---
+---
+---### NOTE:
+---Units are in meters.
+---
+---If the device isn't tracked, all zeroes will be returned.
+---
 ---@param device? lovr.Device # The device to get the pose of.
 ---@return number x # The x position.
 ---@return number y # The y position.
@@ -157,6 +273,10 @@ function lovr.headset.getPose(device) end
 ---
 ---Returns the current position of a device, in meters, relative to the play area.
 ---
+---
+---### NOTE:
+---If the device isn't tracked, all zeroes will be returned.
+---
 ---@param device? lovr.Device # The device to get the position of.
 ---@return number x # The x position of the device.
 ---@return number y # The y position of the device.
@@ -164,7 +284,140 @@ function lovr.headset.getPose(device) end
 function lovr.headset.getPosition(device) end
 
 ---
----Returns a list of joint poses tracked by a device.  Currently, only hand devices are able to track joints.
+---Returns a list of joint poses tracked by a device.
+---
+---Currently, only hand devices are able to track joints.
+---
+---
+---### NOTE:
+---If the Device does not support tracking joints or the poses are unavailable, `nil` is returned.
+---
+---The joint orientation is similar to the graphics coordinate system: -Z is the forwards direction, pointing towards the fingertips.
+---
+---The +Y direction is "up", pointing out of the back of the hand.
+---
+---The +X direction is to the right, perpendicular to X and Z.
+---
+---Hand joints are returned in the following order:
+---
+---<table>
+---  <thead>
+---    <tr>
+---      <td colspan="2">Joint</td>
+---      <td>Index</td>
+---    </tr>
+---  </thead>
+---  <tbody>
+---    <tr>
+---      <td colspan="2">Palm</td>
+---      <td>1</td>
+---    </tr>
+---    <tr>
+---      <td colspan="2">Wrist</td>
+---      <td>2</td>
+---    </tr>
+---    <tr>
+---      <td rowspan="4">Thumb</td>
+---      <td>Metacarpal</td>
+---      <td>3</td>
+---    </tr>
+---    <tr>
+---      <td>Proximal</td>
+---      <td>4</td>
+---    </tr>
+---    <tr>
+---      <td>Distal</td>
+---      <td>5</td>
+---    </tr>
+---    <tr>
+---      <td>Tip</td>
+---      <td>6</td>
+---    </tr>
+---    <tr>
+---      <td rowspan="5">Index</td>
+---      <td>Metacarpal</td>
+---      <td>7</td>
+---    </tr>
+---    <tr>
+---      <td>Proximal</td>
+---      <td>8</td>
+---    </tr>
+---    <tr>
+---      <td>Intermediate</td>
+---      <td>9</td>
+---    </tr>
+---    <tr>
+---      <td>Distal</td>
+---      <td>10</td>
+---    </tr>
+---    <tr>
+---      <td>Tip</td>
+---      <td>11</td>
+---    </tr>
+---    <tr>
+---      <td rowspan="5">Middle</td>
+---      <td>Metacarpal</td>
+---      <td>12</td>
+---    </tr>
+---    <tr>
+---      <td>Proximal</td>
+---      <td>13</td>
+---    </tr>
+---    <tr>
+---      <td>Intermediate</td>
+---      <td>14</td>
+---    </tr>
+---    <tr>
+---      <td>Distal</td>
+---      <td>15</td>
+---    </tr>
+---    <tr>
+---      <td>Tip</td>
+---      <td>16</td>
+---    </tr>
+---    <tr>
+---      <td rowspan="5">Ring</td>
+---      <td>Metacarpal</td>
+---      <td>17</td>
+---    </tr>
+---    <tr>
+---      <td>Proximal</td>
+---      <td>18</td>
+---    </tr>
+---    <tr>
+---      <td>Intermediate</td>
+---      <td>19</td>
+---    </tr>
+---    <tr>
+---      <td>Distal</td>
+---      <td>20</td>
+---    </tr>
+---    <tr>
+---      <td>Tip</td>
+---      <td>21</td>
+---    </tr>
+---    <tr>
+---      <td rowspan="5">Pinky</td>
+---      <td>Metacarpal</td>
+---      <td>22</td>
+---    </tr>
+---    <tr>
+---      <td>Proximal</td>
+---      <td>23</td>
+---    </tr>
+---    <tr>
+---      <td>Intermediate</td>
+---      <td>24</td>
+---    </tr>
+---    <tr>
+---      <td>Distal</td>
+---      <td>25</td>
+---    </tr>
+---    <tr>
+---      <td>Tip</td>
+---      <td>26</td>
+---    </tr>
+---  </tbody> </table>
 ---
 ---@overload fun(device: lovr.Device, t: table):table
 ---@param device lovr.Device # The Device to query.
@@ -176,6 +429,10 @@ function lovr.headset.getSkeleton(device) end
 ---
 ---This can be used as a replacement for `lovr.timer.getTime` for timestamps that are used for rendering to get a smoother result that is synchronized with the display of the headset.
 ---
+---
+---### NOTE:
+---This has a different epoch than `lovr.timer.getTime`, so it is not guaranteed to be close to that value.
+---
 ---@return number time # The predicted display time, in seconds.
 function lovr.headset.getTime() end
 
@@ -203,15 +460,21 @@ function lovr.headset.getVelocity(device) end
 function lovr.headset.getViewAngles(view) end
 
 ---
----Returns the number of views used for rendering.  Each view consists of a pose in space and a set of angle values that determine the field of view.
+---Returns the number of views used for rendering.
+---
+---Each view consists of a pose in space and a set of angle values that determine the field of view.
+---
+---This is usually 2 for stereo rendering configurations, but it can also be different.
 ---
----This is usually 2 for stereo rendering configurations, but it can also be different.  For example, one way of doing foveated rendering uses 2 views for each eye -- one low quality view with a wider field of view, and a high quality view with a narrower field of view.
+---For example, one way of doing foveated rendering uses 2 views for each eye -- one low quality view with a wider field of view, and a high quality view with a narrower field of view.
 ---
 ---@return number count # The number of views.
 function lovr.headset.getViewCount() end
 
 ---
----Returns the pose of one of the headset views.  This info can be used to create view matrices or do other eye-dependent calculations.
+---Returns the pose of one of the headset views.
+---
+---This info can be used to create view matrices or do other eye-dependent calculations.
 ---
 ---If tracking data is unavailable for the view or the index is invalid, `nil` is returned.
 ---
@@ -244,6 +507,10 @@ function lovr.headset.isTouched(device, button) end
 ---
 ---Returns whether any active headset driver is currently returning pose information for a device.
 ---
+---
+---### NOTE:
+---If a device is tracked, it is guaranteed to return a valid pose until the next call to `lovr.headset.update`.
+---
 ---@param device? lovr.Device # The device to get the pose of.
 ---@return boolean tracked # Whether the device is currently tracked.
 function lovr.headset.isTracked(device) end
@@ -251,6 +518,10 @@ function lovr.headset.isTracked(device) end
 ---
 ---Returns a new Model for the specified device.
 ---
+---
+---### NOTE:
+---This is only supported on the `openvr` and `vrapi` drivers right now.
+---
 ---@param device? lovr.Device # The device to load a model for.
 ---@param options? {animated: boolean} # Options for loading the model.
 ---@return lovr.Model model # The new Model, or `nil` if a model could not be loaded.
@@ -259,15 +530,33 @@ function lovr.headset.newModel(device, options) end
 ---
 ---Renders to each eye of the headset using a function.
 ---
----This function takes care of setting the appropriate graphics transformations to ensure that the scene is rendered as though it is being viewed through each eye of the player.  It also takes care of setting the correct projection for the headset lenses.
+---This function takes care of setting the appropriate graphics transformations to ensure that the scene is rendered as though it is being viewed through each eye of the player.
+---
+---It also takes care of setting the correct projection for the headset lenses.
 ---
 ---If the headset module is enabled, this function is called automatically by `lovr.run` with `lovr.draw` as the callback.
 ---
+---
+---### NOTE:
+---When using the `pico` headset driver, headset rendering is asynchronous and the callback passed to `lovr.headset.renderTo` will not be called immediately.
+---
+---At the beginning of the callback, the display is cleared to the background color.
+---
+---The background color can be changed using `lovr.graphics.setBackgroundColor`.
+---
+---If the callback is `nil`, an empty frame cleared to current graphics background color will be submitted to the headset.
+---
 ---@param callback function # The function used to render.  Any functions called will render to the headset instead of to the window.
 function lovr.headset.renderTo(callback) end
 
 ---
----Sets the near and far clipping planes used to render to the headset.  Objects closer than the near clipping plane or further than the far clipping plane will be clipped out of view.
+---Sets the near and far clipping planes used to render to the headset.
+---
+---Objects closer than the near clipping plane or further than the far clipping plane will be clipped out of view.
+---
+---
+---### NOTE:
+---The default clip distances are 0.1 and 100.0.
 ---
 ---@param near number # The distance to the near clipping plane, in meters.
 ---@param far number # The distance to the far clipping plane, in meters.
@@ -276,6 +565,16 @@ function lovr.headset.setClipDistance(near, far) end
 ---
 ---Causes the device to vibrate with a custom strength, duration, and frequency, if possible.
 ---
+---
+---### NOTE:
+---When using the `openvr` headset driver on an HTC Vive, the value for the `duration` currently must be less than .004 seconds.
+---
+---Call this function several frames in a row for stronger or prolonged vibration patterns.
+---
+---On the Oculus Quest, devices can only be vibrated once per frame.
+---
+---Any attempts after the first will return `false`.
+---
 ---@param device? lovr.Device # The device to vibrate.
 ---@param strength? number # The strength of the vibration (amplitude), between 0 and 1.
 ---@param duration? number # The duration of the vibration, in seconds.
@@ -286,6 +585,14 @@ function lovr.headset.vibrate(device, strength, duration, frequency) end
 ---
 ---Returns whether a button on a device was pressed this frame.
 ---
+---
+---### NOTE:
+---Some headset backends are not able to return pressed/released information.
+---
+---These drivers will always return false for `lovr.headset.wasPressed` and `lovr.headset.wasReleased`.
+---
+---Typically the internal `lovr.headset.update` function will update pressed/released status.
+---
 ---@param device lovr.Device # The device.
 ---@param button lovr.DeviceButton # The button to check.
 ---@return boolean pressed # Whether the button on the device was pressed this frame.
@@ -294,6 +601,14 @@ function lovr.headset.wasPressed(device, button) end
 ---
 ---Returns whether a button on a device was released this frame.
 ---
+---
+---### NOTE:
+---Some headset backends are not able to return pressed/released information.
+---
+---These drivers will always return false for `lovr.headset.wasPressed` and `lovr.headset.wasReleased`.
+---
+---Typically the internal `lovr.headset.update` function will update pressed/released status.
+---
 ---@param device lovr.Device # The device.
 ---@param button lovr.DeviceButton # The button to check.
 ---@return boolean released # Whether the button on the device was released this frame.
@@ -463,9 +778,17 @@ function lovr.headset.wasReleased(device, button) end
 ---| '"proximity"'
 
 ---
----These are all of the supported VR APIs that LÖVR can use to power the lovr.headset module.  You can change the order of headset drivers using `lovr.conf` to prefer or exclude specific VR APIs.
+---These are all of the supported VR APIs that LÖVR can use to power the lovr.headset module.
+---
+---You can change the order of headset drivers using `lovr.conf` to prefer or exclude specific VR APIs.
+---
+---At startup, LÖVR searches through the list of drivers in order.
 ---
----At startup, LÖVR searches through the list of drivers in order.  One headset driver will be used for rendering to the VR display, and all supported headset drivers will be used for device input.  The way this works is that when poses or button input is requested, the input drivers are queried (in the order they appear in `conf.lua`) to see if any of them currently have data for the specified device.  The first one that returns data will be used to provide the result. This allows projects to support multiple types of hardware devices.
+---One headset driver will be used for rendering to the VR display, and all supported headset drivers will be used for device input.
+---
+---The way this works is that when poses or button input is requested, the input drivers are queried (in the order they appear in `conf.lua`) to see if any of them currently have data for the specified device.
+---
+---The first one that returns data will be used to provide the result. This allows projects to support multiple types of hardware devices.
 ---
 ---@alias lovr.HeadsetDriver
 ---
@@ -498,7 +821,11 @@ function lovr.headset.wasReleased(device, button) end
 ---| '"webxr"'
 
 ---
----Represents the different types of origins for coordinate spaces.  An origin of "floor" means that the origin is on the floor in the middle of a room-scale play area.  An origin of "head" means that no positional tracking is available, and consequently the origin is always at the position of the headset.
+---Represents the different types of origins for coordinate spaces.
+---
+---An origin of "floor" means that the origin is on the floor in the middle of a room-scale play area.
+---
+---An origin of "head" means that no positional tracking is available, and consequently the origin is always at the position of the headset.
 ---
 ---@alias lovr.HeadsetOrigin
 ---
diff --git a/meta/3rd/lovr/library/lovr.lua b/meta/3rd/lovr/library/lovr.lua
index 00a8f4b7..83acdec7 100644
--- a/meta/3rd/lovr/library/lovr.lua
+++ b/meta/3rd/lovr/library/lovr.lua
@@ -15,14 +15,30 @@ lovr = {}
 function lovr.getVersion() end
 
 ---
----This is not a real object, but describes the behavior shared by all objects.  Think of it as the superclass of all LÖVR objects.
+---This is not a real object, but describes the behavior shared by all objects.
 ---
----In addition to the methods here, all objects have a `__tostring` metamethod that returns the name of the object's type.  So to check if a LÖVR object is an instance of "Blob", you can do `tostring(object) == 'Blob'`.
+---Think of it as the superclass of all LÖVR objects.
+---
+---In addition to the methods here, all objects have a `__tostring` metamethod that returns the name of the object's type.
+---
+---So to check if a LÖVR object is an instance of "Blob", you can do `tostring(object) == 'Blob'`.
+---
+---
+---### NOTE:
+---Note that the functions here don't apply to any vector objects, see `Vectors`.
 ---
 ---@class lovr.Object
 local Object = {}
 
 ---
----Immediately destroys Lua's reference to the object it's called on.  After calling this function on an object, it is an error to do anything with the object from Lua (call methods on it, pass it to other functions, etc.).  If nothing else is using the object, it will be destroyed immediately, which can be used to destroy something earlier than it would normally be garbage collected in order to reduce memory.
+---Immediately destroys Lua's reference to the object it's called on.
+---
+---After calling this function on an object, it is an error to do anything with the object from Lua (call methods on it, pass it to other functions, etc.).
+---
+---If nothing else is using the object, it will be destroyed immediately, which can be used to destroy something earlier than it would normally be garbage collected in order to reduce memory.
+---
+---
+---### NOTE:
+---The object may not be destroyed immediately if something else is referring to it (e.g. it is pushed to a Channel or exists in the payload of a pending event).
 ---
 function Object:release() end
diff --git a/meta/3rd/lovr/library/lovr.math.lua b/meta/3rd/lovr/library/lovr.math.lua
index 9c47ec48..7cf348c6 100644
--- a/meta/3rd/lovr/library/lovr.math.lua
+++ b/meta/3rd/lovr/library/lovr.math.lua
@@ -46,7 +46,9 @@ function lovr.math.getRandomSeed() end
 function lovr.math.linearToGamma(lr, lg, lb) end
 
 ---
----Creates a temporary 4D matrix.  This function takes the same arguments as `Mat4:set`.
+---Creates a temporary 4D matrix.
+---
+---This function takes the same arguments as `Mat4:set`.
 ---
 function lovr.math.mat4() end
 
@@ -62,12 +64,16 @@ function lovr.math.mat4() end
 function lovr.math.newCurve(x, y, z) end
 
 ---
----Creates a new 4D matrix.  This function takes the same arguments as `Mat4:set`.
+---Creates a new 4D matrix.
+---
+---This function takes the same arguments as `Mat4:set`.
 ---
 function lovr.math.newMat4() end
 
 ---
----Creates a new quaternion.  This function takes the same arguments as `Quat:set`.
+---Creates a new quaternion.
+---
+---This function takes the same arguments as `Quat:set`.
 ---
 function lovr.math.newQuat() end
 
@@ -80,22 +86,30 @@ function lovr.math.newQuat() end
 function lovr.math.newRandomGenerator() end
 
 ---
----Creates a new 2D vector.  This function takes the same arguments as `Vec2:set`.
+---Creates a new 2D vector.
+---
+---This function takes the same arguments as `Vec2:set`.
 ---
 function lovr.math.newVec2() end
 
 ---
----Creates a new 3D vector.  This function takes the same arguments as `Vec3:set`.
+---Creates a new 3D vector.
+---
+---This function takes the same arguments as `Vec3:set`.
 ---
 function lovr.math.newVec3() end
 
 ---
----Creates a new 4D vector.  This function takes the same arguments as `Vec4:set`.
+---Creates a new 4D vector.
+---
+---This function takes the same arguments as `Vec4:set`.
 ---
 function lovr.math.newVec4() end
 
 ---
----Returns a 1D, 2D, 3D, or 4D perlin noise value.  The number will be between 0 and 1, and it will always be 0.5 when the inputs are integers.
+---Returns a 1D, 2D, 3D, or 4D perlin noise value.
+---
+---The number will be between 0 and 1, and it will always be 0.5 when the inputs are integers.
 ---
 ---@overload fun(x: number, y: number):number
 ---@overload fun(x: number, y: number, z: number):number
@@ -105,12 +119,20 @@ function lovr.math.newVec4() end
 function lovr.math.noise(x) end
 
 ---
----Creates a temporary quaternion.  This function takes the same arguments as `Quat:set`.
+---Creates a temporary quaternion.
+---
+---This function takes the same arguments as `Quat:set`.
 ---
 function lovr.math.quat() end
 
 ---
----Returns a uniformly distributed pseudo-random number.  This function has improved randomness over Lua's `math.random` and also guarantees that the sequence of random numbers will be the same on all platforms (given the same seed).
+---Returns a uniformly distributed pseudo-random number.
+---
+---This function has improved randomness over Lua's `math.random` and also guarantees that the sequence of random numbers will be the same on all platforms (given the same seed).
+---
+---
+---### NOTE:
+---You can set the random seed using `lovr.math.setRandomSeed`.
 ---
 ---@overload fun(high: number):number
 ---@overload fun(low: number, high: number):number
@@ -118,7 +140,9 @@ function lovr.math.quat() end
 function lovr.math.random() end
 
 ---
----Returns a pseudo-random number from a normal distribution (a bell curve).  You can control the center of the bell curve (the mean value) and the overall width (sigma, or standard deviation).
+---Returns a pseudo-random number from a normal distribution (a bell curve).
+---
+---You can control the center of the bell curve (the mean value) and the overall width (sigma, or standard deviation).
 ---
 ---@param sigma? number # The standard deviation of the distribution.  This can be thought of how "wide" the range of numbers is or how much variability there is.
 ---@param mu? number # The average value returned.
@@ -126,30 +150,44 @@ function lovr.math.random() end
 function lovr.math.randomNormal(sigma, mu) end
 
 ---
----Seed the random generator with a new seed.  Each seed will cause `lovr.math.random` and `lovr.math.randomNormal` to produce a unique sequence of random numbers.  This is done once automatically at startup by `lovr.run`.
+---Seed the random generator with a new seed.
+---
+---Each seed will cause `lovr.math.random` and `lovr.math.randomNormal` to produce a unique sequence of random numbers.
+---
+---This is done once automatically at startup by `lovr.run`.
 ---
 ---@param seed number # The new seed.
 function lovr.math.setRandomSeed(seed) end
 
 ---
----Creates a temporary 2D vector.  This function takes the same arguments as `Vec2:set`.
+---Creates a temporary 2D vector.
+---
+---This function takes the same arguments as `Vec2:set`.
 ---
 function lovr.math.vec2() end
 
 ---
----Creates a temporary 3D vector.  This function takes the same arguments as `Vec3:set`.
+---Creates a temporary 3D vector.
+---
+---This function takes the same arguments as `Vec3:set`.
 ---
 function lovr.math.vec3() end
 
 ---
----Creates a temporary 4D vector.  This function takes the same arguments as `Vec4:set`.
+---Creates a temporary 4D vector.
+---
+---This function takes the same arguments as `Vec4:set`.
 ---
 function lovr.math.vec4() end
 
 ---
----A Curve is an object that represents a Bézier curve in three dimensions.  Curves are defined by an arbitrary number of control points (note that the curve only passes through the first and last control point).
+---A Curve is an object that represents a Bézier curve in three dimensions.
 ---
----Once a Curve is created with `lovr.math.newCurve`, you can use `Curve:evaluate` to get a point on the curve or `Curve:render` to get a list of all of the points on the curve.  These points can be passed directly to `lovr.graphics.points` or `lovr.graphics.line` to render the curve.
+---Curves are defined by an arbitrary number of control points (note that the curve only passes through the first and last control point).
+---
+---Once a Curve is created with `lovr.math.newCurve`, you can use `Curve:evaluate` to get a point on the curve or `Curve:render` to get a list of all of the points on the curve.
+---
+---These points can be passed directly to `lovr.graphics.points` or `lovr.graphics.line` to render the curve.
 ---
 ---Note that for longer or more complicated curves (like in a drawing application) it can be easier to store the path as several Curve objects.
 ---
@@ -159,6 +197,10 @@ local Curve = {}
 ---
 ---Inserts a new control point into the Curve at the specified index.
 ---
+---
+---### NOTE:
+---An error will be thrown if the index is less than one or more than the number of control points.
+---
 ---@param x number # The x coordinate of the control point.
 ---@param y number # The y coordinate of the control point.
 ---@param z number # The z coordinate of the control point.
@@ -166,7 +208,13 @@ local Curve = {}
 function Curve:addPoint(x, y, z, index) end
 
 ---
----Returns a point on the Curve given a parameter `t` from 0 to 1.  0 will return the first control point, 1 will return the last point, .5 will return a point in the "middle" of the Curve, etc.
+---Returns a point on the Curve given a parameter `t` from 0 to 1.
+---
+---0 will return the first control point, 1 will return the last point, .5 will return a point in the "middle" of the Curve, etc.
+---
+---
+---### NOTE:
+---An error will be thrown if `t` is not between 0 and 1, or if the Curve has less than two points.
 ---
 ---@param t number # The parameter to evaluate the Curve at.
 ---@return number x # The x position of the point.
@@ -177,6 +225,10 @@ function Curve:evaluate(t) end
 ---
 ---Returns a control point of the Curve.
 ---
+---
+---### NOTE:
+---An error will be thrown if the index is less than one or more than the number of control points.
+---
 ---@param index number # The index to retrieve.
 ---@return number x # The x coordinate of the control point.
 ---@return number y # The y coordinate of the control point.
@@ -190,7 +242,13 @@ function Curve:getPoint(index) end
 function Curve:getPointCount() end
 
 ---
----Returns a direction vector for the Curve given a parameter `t` from 0 to 1.  0 will return the direction at the first control point, 1 will return the direction at the last point, .5 will return the direction at the "middle" of the Curve, etc.
+---Returns a direction vector for the Curve given a parameter `t` from 0 to 1.
+---
+---0 will return the direction at the first control point, 1 will return the direction at the last point, .5 will return the direction at the "middle" of the Curve, etc.
+---
+---
+---### NOTE:
+---The direction vector returned by this function will have a length of one.
 ---
 ---@param t number # Where on the Curve to compute the direction.
 ---@return number x # The x position of the point.
@@ -201,11 +259,21 @@ function Curve:getTangent(t) end
 ---
 ---Removes a control point from the Curve.
 ---
+---
+---### NOTE:
+---An error will be thrown if the index is less than one or more than the number of control points.
+---
 ---@param index number # The index of the control point to remove.
 function Curve:removePoint(index) end
 
 ---
----Returns a list of points on the Curve.  The number of points can be specified to get a more or less detailed representation, and it is also possible to render a subsection of the Curve.
+---Returns a list of points on the Curve.
+---
+---The number of points can be specified to get a more or less detailed representation, and it is also possible to render a subsection of the Curve.
+---
+---
+---### NOTE:
+---This function will always return 2 points if the Curve is a line with only 2 control points.
 ---
 ---@param n? number # The number of points to use.
 ---@param t1? number # How far along the curve to start rendering.
@@ -216,6 +284,10 @@ function Curve:render(n, t1, t2) end
 ---
 ---Changes the position of a control point on the Curve.
 ---
+---
+---### NOTE:
+---An error will be thrown if the index is less than one or more than the number of control points.
+---
 ---@param index number # The index to modify.
 ---@param x number # The new x coordinate.
 ---@param y number # The new y coordinate.
@@ -225,6 +297,12 @@ function Curve:setPoint(index, x, y, z) end
 ---
 ---Returns a new Curve created by slicing the Curve at the specified start and end points.
 ---
+---
+---### NOTE:
+---The new Curve will have the same number of control points as the existing curve.
+---
+---An error will be thrown if t1 or t2 are not between 0 and 1, or if the Curve has less than two points.
+---
 ---@param t1 number # The starting point to slice at.
 ---@param t2 number # The ending point to slice at.
 ---@return lovr.Curve curve # A new Curve.
@@ -276,7 +354,15 @@ function Mat4:invert() end
 function Mat4:lookAt(from, to, up) end
 
 ---
----Multiplies this matrix by another value.  Multiplying by a matrix combines their two transforms together.  Multiplying by a vector applies the transformation from the matrix to the vector and returns the vector.
+---Multiplies this matrix by another value.
+---
+---Multiplying by a matrix combines their two transforms together.
+---
+---Multiplying by a vector applies the transformation from the matrix to the vector and returns the vector.
+---
+---
+---### NOTE:
+---When multiplying by a vec4, the vector is treated as either a point if its w component is 1, or a direction vector if the w is 0 (the matrix translation won't be applied).
 ---
 ---@overload fun(self: lovr.Mat4, v3: lovr.Vec3):lovr.Vec3
 ---@overload fun(self: lovr.Mat4, v4: lovr.Vec4):lovr.Vec4
@@ -377,7 +463,11 @@ function Mat4:unpack(raw) end
 local Quat = {}
 
 ---
----Conjugates the input quaternion in place, returning the input.  If the quaternion is normalized, this is the same as inverting it.  It negates the (x, y, z) components of the quaternion.
+---Conjugates the input quaternion in place, returning the input.
+---
+---If the quaternion is normalized, this is the same as inverting it.
+---
+---It negates the (x, y, z) components of the quaternion.
 ---
 ---@return lovr.Quat q # The original quaternion.
 function Quat:conjugate() end
@@ -395,7 +485,11 @@ function Quat:direction() end
 function Quat:length() end
 
 ---
----Multiplies this quaternion by another value.  If the value is a quaternion, the rotations in the two quaternions are applied sequentially and the result is stored in the first quaternion.  If the value is a vector, then the input vector is rotated by the quaternion and returned.
+---Multiplies this quaternion by another value.
+---
+---If the value is a quaternion, the rotations in the two quaternions are applied sequentially and the result is stored in the first quaternion.
+---
+---If the value is a vector, then the input vector is rotated by the quaternion and returned.
 ---
 ---@overload fun(self: lovr.Quat, v3: lovr.vec3):lovr.vec3
 ---@param r lovr.quat # A quaternion to combine with the original.
@@ -405,11 +499,19 @@ function Quat:mul(r) end
 ---
 ---Adjusts the values in the quaternion so that its length becomes 1.
 ---
+---
+---### NOTE:
+---A common source of bugs with quaternions is to forget to normalize them after performing a series of operations on them.
+---
+---Try normalizing a quaternion if some of the calculations aren't working quite right!
+---
 ---@return lovr.Quat q # The original quaternion.
 function Quat:normalize() end
 
 ---
----Sets the components of the quaternion.  There are lots of different ways to specify the new components, the summary is:
+---Sets the components of the quaternion.
+---
+---There are lots of different ways to specify the new components, the summary is:
 ---
 ---- Four numbers can be used to specify an angle/axis rotation, similar to other LÖVR functions.
 ---- Four numbers plus the fifth `raw` flag can be used to set the raw values of the quaternion.
@@ -436,7 +538,11 @@ function Quat:set(angle, ax, ay, az, raw) end
 ---
 ---Performs a spherical linear interpolation between this quaternion and another one, which can be used for smoothly animating between two rotations.
 ---
----The amount of interpolation is controlled by a parameter `t`.  A `t` value of zero leaves the original quaternion unchanged, whereas a `t` of one sets the original quaternion exactly equal to the target.  A value between `0` and `1` returns a rotation between the two based on the value.
+---The amount of interpolation is controlled by a parameter `t`.
+---
+---A `t` value of zero leaves the original quaternion unchanged, whereas a `t` of one sets the original quaternion exactly equal to the target.
+---
+---A value between `0` and `1` returns a rotation between the two based on the value.
 ---
 ---@param r lovr.Quat # The quaternion to slerp towards.
 ---@param t number # The lerping parameter.
@@ -462,12 +568,22 @@ local RandomGenerator = {}
 ---
 ---Returns the seed used to initialize the RandomGenerator.
 ---
+---
+---### NOTE:
+---Since the seed is a 64 bit integer, each 32 bits of the seed are returned separately to avoid precision issues.
+---
 ---@return number low # The lower 32 bits of the seed.
 ---@return number high # The upper 32 bits of the seed.
 function RandomGenerator:getSeed() end
 
 ---
----Returns the current state of the RandomGenerator.  This can be used with `RandomGenerator:setState` to reliably restore a previous state of the generator.
+---Returns the current state of the RandomGenerator.
+---
+---This can be used with `RandomGenerator:setState` to reliably restore a previous state of the generator.
+---
+---
+---### NOTE:
+---The seed represents the starting state of the RandomGenerator, whereas the state represents the current state of the generator.
 ---
 ---@return string state # The serialized state.
 function RandomGenerator:getState() end
@@ -481,7 +597,9 @@ function RandomGenerator:getState() end
 function RandomGenerator:random() end
 
 ---
----Returns a pseudo-random number from a normal distribution (a bell curve).  You can control the center of the bell curve (the mean value) and the overall width (sigma, or standard deviation).
+---Returns a pseudo-random number from a normal distribution (a bell curve).
+---
+---You can control the center of the bell curve (the mean value) and the overall width (sigma, or standard deviation).
 ---
 ---@param sigma? number # The standard deviation of the distribution.  This can be thought of how "wide" the range of numbers is or how much variability there is.
 ---@param mu? number # The average value returned.
@@ -489,7 +607,13 @@ function RandomGenerator:random() end
 function RandomGenerator:randomNormal(sigma, mu) end
 
 ---
----Seed the RandomGenerator with a new seed.  Each seed will cause the RandomGenerator to produce a unique sequence of random numbers.
+---Seed the RandomGenerator with a new seed.
+---
+---Each seed will cause the RandomGenerator to produce a unique sequence of random numbers.
+---
+---
+---### NOTE:
+---For precise 64 bit seeds, you should specify the lower and upper 32 bits of the seed separately. Otherwise, seeds larger than 2^53 will start to lose precision.
 ---
 ---@overload fun(self: lovr.RandomGenerator, low: number, high: number)
 ---@param seed number # The random seed.
@@ -498,6 +622,10 @@ function RandomGenerator:setSeed(seed) end
 ---
 ---Sets the state of the RandomGenerator, as previously obtained using `RandomGenerator:getState`. This can be used to reliably restore a previous state of the generator.
 ---
+---
+---### NOTE:
+---The seed represents the starting state of the RandomGenerator, whereas the state represents the current state of the generator.
+---
 ---@param state string # The serialized state.
 function RandomGenerator:setState(state) end
 
@@ -534,6 +662,14 @@ function Vec2:div(u) end
 ---
 ---Returns the dot product between this vector and another one.
 ---
+---
+---### NOTE:
+---This is computed as:
+---
+---    dot = v.x * u.x + v.y * u.y
+---
+---The vectors are not normalized before computing the dot product.
+---
 ---@overload fun(self: lovr.Vec2, x: number, y: number):number
 ---@param u lovr.Vec2 # The vector to compute the dot product with.
 ---@return number dot # The dot product between `v` and `u`.
@@ -542,11 +678,19 @@ function Vec2:dot(u) end
 ---
 ---Returns the length of the vector.
 ---
+---
+---### NOTE:
+---The length is equivalent to this:
+---
+---    math.sqrt(v.x * v.x + v.y * v.y)
+---
 ---@return number length # The length of the vector.
 function Vec2:length() end
 
 ---
----Performs a linear interpolation between this vector and another one, which can be used to smoothly animate between two vectors, based on a parameter value.  A parameter value of `0` will leave the vector unchanged, a parameter value of `1` will set the vector to be equal to the input vector, and a value of `.5` will set the components to be halfway between the two vectors.
+---Performs a linear interpolation between this vector and another one, which can be used to smoothly animate between two vectors, based on a parameter value.
+---
+---A parameter value of `0` will leave the vector unchanged, a parameter value of `1` will set the vector to be equal to the input vector, and a value of `.5` will set the components to be halfway between the two vectors.
 ---
 ---@overload fun(self: lovr.Vec2, x: number, y: number):lovr.Vec2
 ---@return lovr.Vec2 v # The original vector, containing the new lerped values.
@@ -605,7 +749,13 @@ local Vec3 = {}
 function Vec3:add(u) end
 
 ---
----Sets this vector to be equal to the cross product between this vector and another one.  The new `v` will be perpendicular to both the old `v` and `u`.
+---Sets this vector to be equal to the cross product between this vector and another one.
+---
+---The new `v` will be perpendicular to both the old `v` and `u`.
+---
+---
+---### NOTE:
+---The vectors are not normalized before or after computing the cross product.
 ---
 ---@overload fun(self: lovr.Vec3, x: number, y: number, z: number):lovr.Vec3
 ---@param u lovr.Vec3 # The vector to compute the cross product with.
@@ -631,6 +781,14 @@ function Vec3:div(u) end
 ---
 ---Returns the dot product between this vector and another one.
 ---
+---
+---### NOTE:
+---This is computed as:
+---
+---    dot = v.x * u.x + v.y * u.y + v.z * u.z
+---
+---The vectors are not normalized before computing the dot product.
+---
 ---@overload fun(self: lovr.Vec3, x: number, y: number, z: number):number
 ---@param u lovr.Vec3 # The vector to compute the dot product with.
 ---@return number dot # The dot product between `v` and `u`.
@@ -639,11 +797,19 @@ function Vec3:dot(u) end
 ---
 ---Returns the length of the vector.
 ---
+---
+---### NOTE:
+---The length is equivalent to this:
+---
+---    math.sqrt(v.x * v.x + v.y * v.y + v.z * v.z)
+---
 ---@return number length # The length of the vector.
 function Vec3:length() end
 
 ---
----Performs a linear interpolation between this vector and another one, which can be used to smoothly animate between two vectors, based on a parameter value.  A parameter value of `0` will leave the vector unchanged, a parameter value of `1` will set the vector to be equal to the input vector, and a value of `.5` will set the components to be halfway between the two vectors.
+---Performs a linear interpolation between this vector and another one, which can be used to smoothly animate between two vectors, based on a parameter value.
+---
+---A parameter value of `0` will leave the vector unchanged, a parameter value of `1` will set the vector to be equal to the input vector, and a value of `.5` will set the components to be halfway between the two vectors.
 ---
 ---@overload fun(self: lovr.Vec3, x: number, y: number, z: number, t: number):lovr.Vec3
 ---@param u lovr.Vec3 # The vector to lerp towards.
@@ -725,6 +891,14 @@ function Vec4:div(u) end
 ---
 ---Returns the dot product between this vector and another one.
 ---
+---
+---### NOTE:
+---This is computed as:
+---
+---    dot = v.x * u.x + v.y * u.y + v.z * u.z + v.w * u.w
+---
+---The vectors are not normalized before computing the dot product.
+---
 ---@overload fun(self: lovr.Vec4, x: number, y: number, z: number, w: number):number
 ---@param u lovr.Vec4 # The vector to compute the dot product with.
 ---@return number dot # The dot product between `v` and `u`.
@@ -733,11 +907,19 @@ function Vec4:dot(u) end
 ---
 ---Returns the length of the vector.
 ---
+---
+---### NOTE:
+---The length is equivalent to this:
+---
+---    math.sqrt(v.x * v.x + v.y * v.y * v.z + v.z + v.w * v.w)
+---
 ---@return number length # The length of the vector.
 function Vec4:length() end
 
 ---
----Performs a linear interpolation between this vector and another one, which can be used to smoothly animate between two vectors, based on a parameter value.  A parameter value of `0` will leave the vector unchanged, a parameter value of `1` will set the vector to be equal to the input vector, and a value of `.5` will set the components to be halfway between the two vectors.
+---Performs a linear interpolation between this vector and another one, which can be used to smoothly animate between two vectors, based on a parameter value.
+---
+---A parameter value of `0` will leave the vector unchanged, a parameter value of `1` will set the vector to be equal to the input vector, and a value of `.5` will set the components to be halfway between the two vectors.
 ---
 ---@overload fun(self: lovr.Vec4, x: number, y: number, z: number, w: number, t: number):lovr.Vec4
 ---@param u lovr.Vec4 # The vector to lerp towards.
@@ -788,7 +970,102 @@ function Vec4:sub(u) end
 function Vec4:unpack() end
 
 ---
----LÖVR has math objects for vectors, matrices, and quaternions, collectively called "vector objects".  Vectors are useful because they can represent a multidimensional quantity (like a 3D position) using just a single value.
+---LÖVR has math objects for vectors, matrices, and quaternions, collectively called "vector objects".
+---
+---Vectors are useful because they can represent a multidimensional quantity (like a 3D position) using just a single value.
+---
+---
+---### NOTE:
+---Most LÖVR functions that accept positions, orientations, transforms, velocities, etc. also accept vector objects, so they can be used interchangeably with numbers:
+---
+---    function lovr.draw()
+---      -- position and size are vec3's, rotation is a quat
+---      lovr.graphics.box('fill', position, size, rotation)
+---    end
+---
+---### Temporary vs. Permanent
+---
+---Vectors can be created in two different ways: **permanent** and **temporary**.
+---
+---**Permanent** vectors behave like normal Lua values.
+---
+---They are individual objects that are garbage collected when no longer needed.
+---
+---They're created using the usual `lovr.math.new<Type>` syntax:
+---
+---    self.position = lovr.math.newVec3(x, y, z)
+---
+---**Temporary** vectors are created from a shared pool of vector objects.
+---
+---This makes them faster because they use temporary memory and do not need to be garbage collected.
+---
+---To make a temporary vector, leave off the `new` prefix:
+---
+---    local position = lovr.math.vec3(x, y, z)
+---
+---As a further shorthand, these vector constructors are placed on the global scope.
+---
+---If you prefer to keep the global scope clean, this can be configured using the `t.math.globals` flag in `lovr.conf`.
+---
+---    local position = vec3(x1, y1, z1) + vec3(x2, y2, z2)
+---
+---Temporary vectors, with all their speed, come with an important restriction: they can only be used during the frame in which they were created.
+---
+---Saving them into variables and using them later on will throw an error:
+---
+---    local position = vec3(1, 2, 3)
+---
+---    function lovr.update(dt)
+---      -- Reusing a temporary vector across frames will error:
+---      position:add(vec3(dt))
+---    end
+---
+---It's possible to overflow the temporary vector pool.
+---
+---If that happens, `lovr.math.drain` can be used to periodically drain the pool, invalidating any existing temporary vectors.
+---
+---### Metamethods
+---
+---Vectors have metamethods, allowing them to be used using the normal math operators like `+`, `-`, `*`, `/`, etc.
+---
+---    print(vec3(2, 4, 6) * .5 + vec3(10, 20, 30))
+---
+---These metamethods will create new temporary vectors.
+---
+---### Components and Swizzles
+---
+---The raw components of a vector can be accessed like normal fields:
+---
+---    print(vec3(1, 2, 3).z) --> 3
+---    print(mat4()[16]) --> 1
+---
+---Also, multiple fields can be accessed and combined into a new (temporary) vector, called swizzling:
+---
+---    local position = vec3(10, 5, 1)
+---    print(position.xy) --> vec2(10, 5)
+---    print(position.xyy) --> vec3(10, 5, 5)
+---    print(position.zyxz) --> vec4(1, 5, 10, 1)
+---
+---The following fields are supported for vectors:
+---
+---- `x`, `y`, `z`, `w`
+---- `r`, `g`, `b`, `a`
+---- `s`, `t`, `p`, `q`
+---
+---Quaternions support `x`, `y`, `z`, and `w`.
+---
+---Matrices use numbers for accessing individual components in "column-major" order.
+---
+---All fields can also be assigned to.
+---
+---    -- Swap the components of a 2D vector
+---    v.xy = v.yx
+---
+---The `unpack` function can be used (on any vector type) to access all of the individual components of a vector object.
+---
+---For quaternions you can choose whether you want to unpack the angle/axis representation or the raw quaternion components.
+---
+---Similarly, matrices support raw unpacking as well as decomposition into translation/scale/rotation values.
 ---
 ---@class lovr.Vectors
 local Vectors = {}
diff --git a/meta/3rd/lovr/library/lovr.physics.lua b/meta/3rd/lovr/library/lovr.physics.lua
index 7fef948b..fa65009b 100644
--- a/meta/3rd/lovr/library/lovr.physics.lua
+++ b/meta/3rd/lovr/library/lovr.physics.lua
@@ -9,6 +9,12 @@ lovr.physics = {}
 ---
 ---Creates a new BallJoint.
 ---
+---
+---### NOTE:
+---A ball joint is like a ball and socket between the two colliders.
+---
+---It tries to keep the distance between the colliders and the anchor position the same, but does not constrain the angle between them.
+---
 ---@param colliderA lovr.Collider # The first collider to attach the Joint to.
 ---@param colliderB lovr.Collider # The second collider to attach the Joint to.
 ---@param x number # The x position of the joint anchor point, in world coordinates.
@@ -20,6 +26,10 @@ function lovr.physics.newBallJoint(colliderA, colliderB, x, y, z) end
 ---
 ---Creates a new BoxShape.
 ---
+---
+---### NOTE:
+---A Shape can be attached to a Collider using `Collider:addShape`.
+---
 ---@param width? number # The width of the box, in meters.
 ---@param height? number # The height of the box, in meters.
 ---@param depth? number # The depth of the box, in meters.
@@ -27,7 +37,13 @@ function lovr.physics.newBallJoint(colliderA, colliderB, x, y, z) end
 function lovr.physics.newBoxShape(width, height, depth) end
 
 ---
----Creates a new CapsuleShape.  Capsules are cylinders with hemispheres on each end.
+---Creates a new CapsuleShape.
+---
+---Capsules are cylinders with hemispheres on each end.
+---
+---
+---### NOTE:
+---A Shape can be attached to a Collider using `Collider:addShape`.
 ---
 ---@param radius? number # The radius of the capsule, in meters.
 ---@param length? number # The length of the capsule, not including the caps, in meters.
@@ -37,6 +53,10 @@ function lovr.physics.newCapsuleShape(radius, length) end
 ---
 ---Creates a new CylinderShape.
 ---
+---
+---### NOTE:
+---A Shape can be attached to a Collider using `Collider:addShape`.
+---
 ---@param radius? number # The radius of the cylinder, in meters.
 ---@param length? number # The length of the cylinder, in meters.
 ---@return lovr.CylinderShape cylinder # The new CylinderShape.
@@ -45,6 +65,14 @@ function lovr.physics.newCylinderShape(radius, length) end
 ---
 ---Creates a new DistanceJoint.
 ---
+---
+---### NOTE:
+---A distance joint tries to keep the two colliders a fixed distance apart.
+---
+---The distance is determined by the initial distance between the anchor points.
+---
+---The joint allows for rotation on the anchor points.
+---
 ---@param colliderA lovr.Collider # The first collider to attach the Joint to.
 ---@param colliderB lovr.Collider # The second collider to attach the Joint to.
 ---@param x1 number # The x position of the first anchor point, in world coordinates.
@@ -59,6 +87,10 @@ function lovr.physics.newDistanceJoint(colliderA, colliderB, x1, y1, z1, x2, y2,
 ---
 ---Creates a new HingeJoint.
 ---
+---
+---### NOTE:
+---A hinge joint constrains two colliders to allow rotation only around the hinge's axis.
+---
 ---@param colliderA lovr.Collider # The first collider to attach the Joint to.
 ---@param colliderB lovr.Collider # The second collider to attach the Joint to.
 ---@param x number # The x position of the hinge anchor, in world coordinates.
@@ -73,6 +105,10 @@ function lovr.physics.newHingeJoint(colliderA, colliderB, x, y, z, ax, ay, az) e
 ---
 ---Creates a new SliderJoint.
 ---
+---
+---### NOTE:
+---A slider joint constrains two colliders to only allow movement along the slider's axis.
+---
 ---@param colliderA lovr.Collider # The first collider to attach the Joint to.
 ---@param colliderB lovr.Collider # The second collider to attach the Joint to.
 ---@param ax number # The x component of the slider axis.
@@ -84,6 +120,10 @@ function lovr.physics.newSliderJoint(colliderA, colliderB, ax, ay, az) end
 ---
 ---Creates a new SphereShape.
 ---
+---
+---### NOTE:
+---A Shape can be attached to a Collider using `Collider:addShape`.
+---
 ---@param radius? number # The radius of the sphere, in meters.
 ---@return lovr.SphereShape sphere # The new SphereShape.
 function lovr.physics.newSphereShape(radius) end
@@ -91,6 +131,10 @@ function lovr.physics.newSphereShape(radius) end
 ---
 ---Creates a new physics World, which tracks the overall physics simulation, holds collider objects, and resolves collisions between them.
 ---
+---
+---### NOTE:
+---A World must be updated with `World:update` in `lovr.update` for the physics simulation to advance.
+---
 ---@param xg? number # The x component of the gravity force.
 ---@param yg? number # The y component of the gravity force.
 ---@param zg? number # The z component of the gravity force.
@@ -100,13 +144,19 @@ function lovr.physics.newSphereShape(radius) end
 function lovr.physics.newWorld(xg, yg, zg, allowSleep, tags) end
 
 ---
----A BallJoint is a type of `Joint` that acts like a ball and socket between two colliders.  It allows the colliders to rotate freely around an anchor point, but does not allow the colliders' distance from the anchor point to change.
+---A BallJoint is a type of `Joint` that acts like a ball and socket between two colliders.
+---
+---It allows the colliders to rotate freely around an anchor point, but does not allow the colliders' distance from the anchor point to change.
 ---
 ---@class lovr.BallJoint
 local BallJoint = {}
 
 ---
----Returns the anchor points of the BallJoint, in world coordinates.  The first point is the anchor on the first collider, and the second point is on the second collider.  The joint tries to keep these points the same, but they may be different if the joint is forced apart by some other means.
+---Returns the anchor points of the BallJoint, in world coordinates.
+---
+---The first point is the anchor on the first collider, and the second point is on the second collider.
+---
+---The joint tries to keep these points the same, but they may be different if the joint is forced apart by some other means.
 ---
 ---@return number x1 # The x coordinate of the first anchor point, in world coordinates.
 ---@return number y1 # The y coordinate of the first anchor point, in world coordinates.
@@ -117,13 +167,17 @@ local BallJoint = {}
 function BallJoint:getAnchors() end
 
 ---
----Returns the response time of the joint.  See `World:setResponseTime` for more info.
+---Returns the response time of the joint.
+---
+---See `World:setResponseTime` for more info.
 ---
 ---@return number responseTime # The response time setting for the joint.
 function BallJoint:getResponseTime() end
 
 ---
----Returns the tightness of the joint.  See `World:setTightness` for how this affects the joint.
+---Returns the tightness of the joint.
+---
+---See `World:setTightness` for how this affects the joint.
 ---
 ---@return number tightness # The tightness of the joint.
 function BallJoint:getTightness() end
@@ -137,13 +191,17 @@ function BallJoint:getTightness() end
 function BallJoint:setAnchor(x, y, z) end
 
 ---
----Sets the response time of the joint.  See `World:setResponseTime` for more info.
+---Sets the response time of the joint.
+---
+---See `World:setResponseTime` for more info.
 ---
 ---@param responseTime number # The new response time setting for the joint.
 function BallJoint:setResponseTime(responseTime) end
 
 ---
----Sets the tightness of the joint.  See `World:setTightness` for how this affects the joint.
+---Sets the tightness of the joint.
+---
+---See `World:setTightness` for how this affects the joint.
 ---
 ---@param tightness number # The tightness of the joint.
 function BallJoint:setTightness(tightness) end
@@ -201,13 +259,17 @@ function CapsuleShape:setLength(length) end
 function CapsuleShape:setRadius(radius) end
 
 ---
----Colliders are objects that represent a single rigid body in a physics simulation.  They can have forces applied to them and collide with other colliders.
+---Colliders are objects that represent a single rigid body in a physics simulation.
+---
+---They can have forces applied to them and collide with other colliders.
 ---
 ---@class lovr.Collider
 local Collider = {}
 
 ---
----Attaches a Shape to the collider.  Attached shapes will collide with other shapes in the world.
+---Attaches a Shape to the collider.
+---
+---Attached shapes will collide with other shapes in the world.
 ---
 ---@param shape lovr.Shape # The Shape to attach.
 function Collider:addShape(shape) end
@@ -215,6 +277,10 @@ function Collider:addShape(shape) end
 ---
 ---Applies a force to the Collider.
 ---
+---
+---### NOTE:
+---If the Collider is asleep, it will need to be woken up with `Collider:setAwake` for this function to have any affect.
+---
 ---@overload fun(self: lovr.Collider, x: number, y: number, z: number, px: number, py: number, pz: number)
 ---@param x number # The x component of the force to apply.
 ---@param y number # The y component of the force to apply.
@@ -224,6 +290,10 @@ function Collider:applyForce(x, y, z) end
 ---
 ---Applies torque to the Collider.
 ---
+---
+---### NOTE:
+---If the Collider is asleep, it will need to be woken up with `Collider:setAwake` for this function to have any affect.
+---
 ---@param x number # The x component of the torque.
 ---@param y number # The y component of the torque.
 ---@param z number # The z component of the torque.
@@ -232,6 +302,10 @@ function Collider:applyTorque(x, y, z) end
 ---
 ---Destroy the Collider, removing it from the World.
 ---
+---
+---### NOTE:
+---Calling functions on the collider after destroying it is a bad idea.
+---
 function Collider:destroy() end
 
 ---
@@ -246,7 +320,13 @@ function Collider:destroy() end
 function Collider:getAABB() end
 
 ---
----Returns the angular damping parameters of the Collider.  Angular damping makes things less "spinny", making them slow down their angular velocity over time.
+---Returns the angular damping parameters of the Collider.
+---
+---Angular damping makes things less "spinny", making them slow down their angular velocity over time.
+---
+---
+---### NOTE:
+---Angular damping can also be set on the World.
 ---
 ---@return number damping # The angular damping.
 ---@return number threshold # Velocity limit below which the damping is not applied.
@@ -261,7 +341,11 @@ function Collider:getAngularDamping() end
 function Collider:getAngularVelocity() end
 
 ---
----Returns the friction of the Collider.  By default, the friction of two Colliders is combined (multiplied) when they collide to generate a friction force.  The initial friction is 0.
+---Returns the friction of the Collider.
+---
+---By default, the friction of two Colliders is combined (multiplied) when they collide to generate a friction force.
+---
+---The initial friction is 0.
 ---
 ---@return number friction # The friction of the Collider.
 function Collider:getFriction() end
@@ -273,14 +357,28 @@ function Collider:getFriction() end
 function Collider:getJoints() end
 
 ---
----Returns the Collider's linear damping parameters.  Linear damping is similar to drag or air resistance, slowing the Collider down over time.
+---Returns the Collider's linear damping parameters.
+---
+---Linear damping is similar to drag or air resistance, slowing the Collider down over time.
+---
+---
+---### NOTE:
+---A linear damping of 0 means the Collider won't slow down over time.
+---
+---This is the default.
+---
+---Linear damping can also be set on the World using `World:setLinearDamping`, which will affect all new colliders.
 ---
 ---@return number damping # The linear damping.
 ---@return number threshold # Velocity limit below which the damping is not applied.
 function Collider:getLinearDamping() end
 
 ---
----Returns the linear velocity of the Collider.  This is how fast the Collider is moving.  There is also angular velocity, which is how fast the Collider is spinning.
+---Returns the linear velocity of the Collider.
+---
+---This is how fast the Collider is moving.
+---
+---There is also angular velocity, which is how fast the Collider is spinning.
 ---
 ---@return number vx # The x velocity of the Collider, in meters per second.
 ---@return number vy # The y velocity of the Collider, in meters per second.
@@ -340,7 +438,9 @@ function Collider:getLocalPoint(wx, wy, wz) end
 function Collider:getLocalVector(wx, wy, wz) end
 
 ---
----Returns the total mass of the Collider.  The mass of a Collider depends on its attached shapes.
+---Returns the total mass of the Collider.
+---
+---The mass of a Collider depends on its attached shapes.
 ---
 ---@return number mass # The mass of the Collider, in kilograms.
 function Collider:getMass() end
@@ -385,7 +485,11 @@ function Collider:getPose() end
 function Collider:getPosition() end
 
 ---
----Returns the restitution (bounciness) of the Collider.  By default, the restitution of two Colliders is combined (the max is used) when they collide to cause them to bounce away from each other.  The initial restitution is 0.
+---Returns the restitution (bounciness) of the Collider.
+---
+---By default, the restitution of two Colliders is combined (the max is used) when they collide to cause them to bounce away from each other.
+---
+---The initial restitution is 0.
 ---
 ---@return number restitution # The restitution of the Collider.
 function Collider:getRestitution() end
@@ -399,18 +503,30 @@ function Collider:getShapes() end
 ---
 ---Returns the Collider's tag.
 ---
+---
+---### NOTE:
+---Collision between tags can be enabled and disabled using `World:enableCollisionBetween` and `World:disableCollisionBetween`.
+---
 ---@return string tag # The Collider's collision tag.
 function Collider:getTag() end
 
 ---
 ---Returns the user data associated with the Collider.
 ---
+---
+---### NOTE:
+---User data can be useful to identify the Collider in callbacks.
+---
 ---@return any data # The custom value associated with the Collider.
 function Collider:getUserData() end
 
 ---
 ---Returns the World the Collider is in.
 ---
+---
+---### NOTE:
+---Colliders can only be in one World at a time.
+---
 ---@return lovr.World world # The World the Collider is in.
 function Collider:getWorld() end
 
@@ -451,23 +567,47 @@ function Collider:isGravityIgnored() end
 ---
 ---Returns whether the Collider is kinematic.
 ---
+---
+---### NOTE:
+---Kinematic colliders behave as though they have infinite mass, ignoring external forces like gravity, joints, or collisions (though non-kinematic colliders will collide with them). They can be useful for static objects like floors or walls.
+---
 ---@return boolean kinematic # Whether the Collider is kinematic.
 function Collider:isKinematic() end
 
 ---
 ---Returns whether the Collider is allowed to sleep.
 ---
+---
+---### NOTE:
+---If sleeping is enabled, the simulation will put the Collider to sleep if it hasn't moved in a while. Sleeping colliders don't impact the physics simulation, which makes updates more efficient and improves physics performance.
+---
+---However, the physics engine isn't perfect at waking up sleeping colliders and this can lead to bugs where colliders don't react to forces or collisions properly.
+---
+---It is possible to set the default value for new colliders using `World:setSleepingAllowed`.
+---
+---Colliders can be manually put to sleep or woken up using `Collider:setAwake`.
+---
 ---@return boolean allowed # Whether the Collider can go to sleep.
 function Collider:isSleepingAllowed() end
 
 ---
 ---Removes a Shape from the Collider.
 ---
+---
+---### NOTE:
+---Colliders without any shapes won't collide with anything.
+---
 ---@param shape lovr.Shape # The Shape to remove.
 function Collider:removeShape(shape) end
 
 ---
----Sets the angular damping of the Collider.  Angular damping makes things less "spinny", causing them to slow down their angular velocity over time. Damping is only applied when angular velocity is over the threshold value.
+---Sets the angular damping of the Collider.
+---
+---Angular damping makes things less "spinny", causing them to slow down their angular velocity over time. Damping is only applied when angular velocity is over the threshold value.
+---
+---
+---### NOTE:
+---Angular damping can also be set on the World.
 ---
 ---@param damping number # The angular damping.
 ---@param threshold? number # Velocity limit below which the damping is not applied.
@@ -482,13 +622,19 @@ function Collider:setAngularDamping(damping, threshold) end
 function Collider:setAngularVelocity(vx, vy, vz) end
 
 ---
----Manually puts the Collider to sleep or wakes it up.  You can do this if you know a Collider won't be touched for a while or if you need to it be active.
+---Manually puts the Collider to sleep or wakes it up.
+---
+---You can do this if you know a Collider won't be touched for a while or if you need to it be active.
 ---
 ---@param awake boolean # Whether the Collider should be awake.
 function Collider:setAwake(awake) end
 
 ---
----Sets the friction of the Collider.  By default, the friction of two Colliders is combined (multiplied) when they collide to generate a friction force.  The initial friction is 0.
+---Sets the friction of the Collider.
+---
+---By default, the friction of two Colliders is combined (multiplied) when they collide to generate a friction force.
+---
+---The initial friction is 0.
 ---
 ---@param friction number # The new friction.
 function Collider:setFriction(friction) end
@@ -502,18 +648,34 @@ function Collider:setGravityIgnored(ignored) end
 ---
 ---Sets whether the Collider is kinematic.
 ---
+---
+---### NOTE:
+---Kinematic colliders behave as though they have infinite mass, ignoring external forces like gravity, joints, or collisions (though non-kinematic colliders will collide with them). They can be useful for static objects like floors or walls.
+---
 ---@param kinematic boolean # Whether the Collider is kinematic.
 function Collider:setKinematic(kinematic) end
 
 ---
----Sets the Collider's linear damping parameter.  Linear damping is similar to drag or air resistance, slowing the Collider down over time. Damping is only applied when linear velocity is over the threshold value.
+---Sets the Collider's linear damping parameter.
+---
+---Linear damping is similar to drag or air resistance, slowing the Collider down over time. Damping is only applied when linear velocity is over the threshold value.
+---
+---
+---### NOTE:
+---A linear damping of 0 means the Collider won't slow down over time.
+---
+---This is the default.
+---
+---Linear damping can also be set on the World using `World:setLinearDamping`, which will affect all new colliders.
 ---
 ---@param damping number # The linear damping.
 ---@param threshold? number # Velocity limit below which the damping is not applied.
 function Collider:setLinearDamping(damping, threshold) end
 
 ---
----Sets the linear velocity of the Collider directly.  Usually it's preferred to use `Collider:applyForce` to change velocity since instantaneous velocity changes can lead to weird glitches.
+---Sets the linear velocity of the Collider directly.
+---
+---Usually it's preferred to use `Collider:applyForce` to change velocity since instantaneous velocity changes can lead to weird glitches.
 ---
 ---@param vx number # The x velocity of the Collider, in meters per second.
 ---@param vy number # The y velocity of the Collider, in meters per second.
@@ -566,7 +728,9 @@ function Collider:setPose(x, y, z, angle, ax, ay, az) end
 function Collider:setPosition(x, y, z) end
 
 ---
----Sets the restitution (bounciness) of the Collider.  By default, the restitution of two Colliders is combined (the max is used) when they collide to cause them to bounce away from each other. The initial restitution is 0.
+---Sets the restitution (bounciness) of the Collider.
+---
+---By default, the restitution of two Colliders is combined (the max is used) when they collide to cause them to bounce away from each other. The initial restitution is 0.
 ---
 ---@param restitution number # The new restitution.
 function Collider:setRestitution(restitution) end
@@ -574,18 +738,36 @@ function Collider:setRestitution(restitution) end
 ---
 ---Sets whether the Collider is allowed to sleep.
 ---
+---
+---### NOTE:
+---If sleeping is enabled, the simulation will put the Collider to sleep if it hasn't moved in a while. Sleeping colliders don't impact the physics simulation, which makes updates more efficient and improves physics performance.
+---
+---However, the physics engine isn't perfect at waking up sleeping colliders and this can lead to bugs where colliders don't react to forces or collisions properly.
+---
+---It is possible to set the default value for new colliders using `World:setSleepingAllowed`.
+---
+---Colliders can be manually put to sleep or woken up using `Collider:setAwake`.
+---
 ---@param allowed boolean # Whether the Collider can go to sleep.
 function Collider:setSleepingAllowed(allowed) end
 
 ---
 ---Sets the Collider's tag.
 ---
+---
+---### NOTE:
+---Collision between tags can be enabled and disabled using `World:enableCollisionBetween` and `World:disableCollisionBetween`.
+---
 ---@param tag string # The Collider's collision tag.
 function Collider:setTag(tag) end
 
 ---
 ---Associates a custom value with the Collider.
 ---
+---
+---### NOTE:
+---User data can be useful to identify the Collider in callbacks.
+---
 ---@param data any # The custom value to associate with the Collider.
 function Collider:setUserData(data) end
 
@@ -620,7 +802,9 @@ function CylinderShape:setLength(length) end
 function CylinderShape:setRadius(radius) end
 
 ---
----A DistanceJoint is a type of `Joint` that tries to keep two colliders a fixed distance apart. The distance is determined by the initial distance between the anchor points.  The joint allows for rotation on the anchor points.
+---A DistanceJoint is a type of `Joint` that tries to keep two colliders a fixed distance apart. The distance is determined by the initial distance between the anchor points.
+---
+---The joint allows for rotation on the anchor points.
 ---
 ---@class lovr.DistanceJoint
 local DistanceJoint = {}
@@ -637,19 +821,25 @@ local DistanceJoint = {}
 function DistanceJoint:getAnchors() end
 
 ---
----Returns the target distance for the DistanceJoint.  The joint tries to keep the Colliders this far apart.
+---Returns the target distance for the DistanceJoint.
+---
+---The joint tries to keep the Colliders this far apart.
 ---
 ---@return number distance # The target distance.
 function DistanceJoint:getDistance() end
 
 ---
----Returns the response time of the joint.  See `World:setResponseTime` for more info.
+---Returns the response time of the joint.
+---
+---See `World:setResponseTime` for more info.
 ---
 ---@return number responseTime # The response time setting for the joint.
 function DistanceJoint:getResponseTime() end
 
 ---
----Returns the tightness of the joint.  See `World:setTightness` for how this affects the joint.
+---Returns the tightness of the joint.
+---
+---See `World:setTightness` for how this affects the joint.
 ---
 ---@return number tightness # The tightness of the joint.
 function DistanceJoint:getTightness() end
@@ -666,19 +856,25 @@ function DistanceJoint:getTightness() end
 function DistanceJoint:setAnchors(x1, y1, z1, x2, y2, z2) end
 
 ---
----Sets the target distance for the DistanceJoint.  The joint tries to keep the Colliders this far apart.
+---Sets the target distance for the DistanceJoint.
+---
+---The joint tries to keep the Colliders this far apart.
 ---
 ---@param distance number # The new target distance.
 function DistanceJoint:setDistance(distance) end
 
 ---
----Sets the response time of the joint.  See `World:setResponseTime` for more info.
+---Sets the response time of the joint.
+---
+---See `World:setResponseTime` for more info.
 ---
 ---@param responseTime number # The new response time setting for the joint.
 function DistanceJoint:setResponseTime(responseTime) end
 
 ---
----Sets the tightness of the joint.  See `World:setTightness` for how this affects the joint.
+---Sets the tightness of the joint.
+---
+---See `World:setTightness` for how this affects the joint.
 ---
 ---@param tightness number # The tightness of the joint.
 function DistanceJoint:setTightness(tightness) end
@@ -701,7 +897,9 @@ local HingeJoint = {}
 function HingeJoint:getAnchors() end
 
 ---
----Get the angle between the two colliders attached to the HingeJoint.  When the joint is created or when the anchor or axis is set, the current angle is the new "zero" angle.
+---Get the angle between the two colliders attached to the HingeJoint.
+---
+---When the joint is created or when the anchor or axis is set, the current angle is the new "zero" angle.
 ---
 ---@return number angle # The hinge angle, in radians.
 function HingeJoint:getAngle() end
@@ -715,20 +913,26 @@ function HingeJoint:getAngle() end
 function HingeJoint:getAxis() end
 
 ---
----Returns the upper and lower limits of the hinge angle.  These will be between -π and π.
+---Returns the upper and lower limits of the hinge angle.
+---
+---These will be between -π and π.
 ---
 ---@return number lower # The lower limit, in radians.
 ---@return number upper # The upper limit, in radians.
 function HingeJoint:getLimits() end
 
 ---
----Returns the lower limit of the hinge angle.  This will be greater than -π.
+---Returns the lower limit of the hinge angle.
+---
+---This will be greater than -π.
 ---
 ---@return number limit # The lower limit, in radians.
 function HingeJoint:getLowerLimit() end
 
 ---
----Returns the upper limit of the hinge angle.  This will be less than π.
+---Returns the upper limit of the hinge angle.
+---
+---This will be less than π.
 ---
 ---@return number limit # The upper limit, in radians.
 function HingeJoint:getUpperLimit() end
@@ -750,20 +954,26 @@ function HingeJoint:setAnchor(x, y, z) end
 function HingeJoint:setAxis(x, y, z) end
 
 ---
----Sets the upper and lower limits of the hinge angle.  These should be between -π and π.
+---Sets the upper and lower limits of the hinge angle.
+---
+---These should be between -π and π.
 ---
 ---@param lower number # The lower limit, in radians.
 ---@param upper number # The upper limit, in radians.
 function HingeJoint:setLimits(lower, upper) end
 
 ---
----Sets the lower limit of the hinge angle.  This should be greater than -π.
+---Sets the lower limit of the hinge angle.
+---
+---This should be greater than -π.
 ---
 ---@param limit number # The lower limit, in radians.
 function HingeJoint:setLowerLimit(limit) end
 
 ---
----Sets the upper limit of the hinge angle.  This should be less than π.
+---Sets the upper limit of the hinge angle.
+---
+---This should be less than π.
 ---
 ---@param limit number # The upper limit, in radians.
 function HingeJoint:setUpperLimit(limit) end
@@ -777,10 +987,16 @@ local Joint = {}
 ---
 ---Destroy the Joint, removing it from Colliders it's attached to.
 ---
+---
+---### NOTE:
+---Calling functions on the Joint after destroying it is a bad idea.
+---
 function Joint:destroy() end
 
 ---
----Returns the Colliders the Joint is attached to.  All Joints are attached to two colliders.
+---Returns the Colliders the Joint is attached to.
+---
+---All Joints are attached to two colliders.
 ---
 ---@return lovr.Collider colliderA # The first Collider.
 ---@return lovr.Collider colliderB # The second Collider.
@@ -825,6 +1041,10 @@ local Shape = {}
 ---
 ---Destroy the Shape, removing it from Colliders it's attached to.
 ---
+---
+---### NOTE:
+---Calling functions on the Shape after destroying it is a bad idea.
+---
 function Shape:destroy() end
 
 ---
@@ -841,6 +1061,10 @@ function Shape:getAABB() end
 ---
 ---Returns the Collider the Shape is attached to.
 ---
+---
+---### NOTE:
+---A Shape can only be attached to one Collider at a time.
+---
 ---@return lovr.Collider collider # The Collider the Shape is attached to.
 function Shape:getCollider() end
 
@@ -881,17 +1105,27 @@ function Shape:getType() end
 ---
 ---Returns the user data associated with the Shape.
 ---
+---
+---### NOTE:
+---User data can be useful to identify the Shape in callbacks.
+---
 ---@return any data # The custom value associated with the Shape.
 function Shape:getUserData() end
 
 ---
 ---Returns whether the Shape is enabled.
 ---
+---
+---### NOTE:
+---Disabled shapes won't collide with anything.
+---
 ---@return boolean enabled # Whether the Shape is enabled.
 function Shape:isEnabled() end
 
 ---
----Returns whether the Shape is a sensor.  Sensors do not trigger any collision response, but they still report collisions in `World:collide`.
+---Returns whether the Shape is a sensor.
+---
+---Sensors do not trigger any collision response, but they still report collisions in `World:collide`.
 ---
 ---@return boolean sensor # Whether the Shape is a sensor.
 function Shape:isSensor() end
@@ -899,12 +1133,20 @@ function Shape:isSensor() end
 ---
 ---Enable or disable the Shape.
 ---
+---
+---### NOTE:
+---Disabled shapes won't collide with anything.
+---
 ---@param enabled boolean # Whether the Shape should be enabled.
 function Shape:setEnabled(enabled) end
 
 ---
 ---Set the orientation of the Shape relative to its Collider.
 ---
+---
+---### NOTE:
+---If the Shape isn't attached to a Collider, this will error.
+---
 ---@param angle number # The number of radians the Shape is rotated.
 ---@param ax number # The x component of the rotation axis.
 ---@param ay number # The y component of the rotation axis.
@@ -914,13 +1156,19 @@ function Shape:setOrientation(angle, ax, ay, az) end
 ---
 ---Set the position of the Shape relative to its Collider.
 ---
+---
+---### NOTE:
+---If the Shape isn't attached to a Collider, this will error.
+---
 ---@param x number # The x offset.
 ---@param y number # The y offset.
 ---@param z number # The z offset.
 function Shape:setPosition(x, y, z) end
 
 ---
----Sets whether this Shape is a sensor.  Sensors do not trigger any collision response, but they still report collisions in `World:collide`.
+---Sets whether this Shape is a sensor.
+---
+---Sensors do not trigger any collision response, but they still report collisions in `World:collide`.
 ---
 ---@param sensor boolean # Whether the Shape should be a sensor.
 function Shape:setSensor(sensor) end
@@ -928,6 +1176,10 @@ function Shape:setSensor(sensor) end
 ---
 ---Sets the user data associated with the Shape.
 ---
+---
+---### NOTE:
+---User data can be useful to identify the Shape in callbacks.
+---
 ---@param data any # The custom value associated with the Shape.
 function Shape:setUserData(data) end
 
@@ -1012,11 +1264,29 @@ function SphereShape:setDimensions(radius) end
 ---
 ---A World is an object that holds the colliders, joints, and shapes in a physics simulation.
 ---
+---
+---### NOTE:
+---Be sure to update the World in `lovr.update` using `World:update`, otherwise everything will stand still.
+---
 ---@class lovr.World
 local World = {}
 
 ---
----Attempt to collide two shapes.  Internally this uses joints and forces to ensure the colliders attached to the shapes do not pass through each other.  Collisions can be customized using friction and restitution (bounciness) parameters, and default to using a mix of the colliders' friction and restitution parameters.  Usually this is called automatically by `World:update`.
+---Attempt to collide two shapes.
+---
+---Internally this uses joints and forces to ensure the colliders attached to the shapes do not pass through each other.
+---
+---Collisions can be customized using friction and restitution (bounciness) parameters, and default to using a mix of the colliders' friction and restitution parameters.
+---
+---Usually this is called automatically by `World:update`.
+---
+---
+---### NOTE:
+---For friction, numbers in the range of 0-1 are common, but larger numbers can also be used.
+---
+---For restitution, numbers in the range 0-1 should be used.
+---
+---This function respects collision tags, so using `World:disableCollisionBetween` and `World:enableCollisionBetween` will change the behavior of this function.
 ---
 ---@param shapeA lovr.Shape # The first shape.
 ---@param shapeB lovr.Shape # The second shape.
@@ -1026,18 +1296,30 @@ local World = {}
 function World:collide(shapeA, shapeB, friction, restitution) end
 
 ---
----Detects which pairs of shapes in the world are near each other and could be colliding.  After calling this function, the `World:overlaps` iterator can be used to iterate over the overlaps, and `World:collide` can be used to resolve a collision for the shapes (if any). Usually this is called automatically by `World:update`.
+---Detects which pairs of shapes in the world are near each other and could be colliding.
+---
+---After calling this function, the `World:overlaps` iterator can be used to iterate over the overlaps, and `World:collide` can be used to resolve a collision for the shapes (if any). Usually this is called automatically by `World:update`.
 ---
 function World:computeOverlaps() end
 
 ---
 ---Destroy the World!
 ---
+---
+---### NOTE:
+---Bad things will happen if you destroy the world and then try to access it or anything that was in it.
+---
 function World:destroy() end
 
 ---
 ---Disables collision between two collision tags.
 ---
+---
+---### NOTE:
+---Tags must be set up when creating the World, see `lovr.physics.newWorld`.
+---
+---By default, collision is enabled between all tags.
+---
 ---@param tag1 string # The first tag.
 ---@param tag2 string # The second tag.
 function World:disableCollisionBetween(tag1, tag2) end
@@ -1045,12 +1327,24 @@ function World:disableCollisionBetween(tag1, tag2) end
 ---
 ---Enables collision between two collision tags.
 ---
+---
+---### NOTE:
+---Tags must be set up when creating the World, see `lovr.physics.newWorld`.
+---
+---By default, collision is enabled between all tags.
+---
 ---@param tag1 string # The first tag.
 ---@param tag2 string # The second tag.
 function World:enableCollisionBetween(tag1, tag2) end
 
 ---
----Returns the angular damping parameters of the World.  Angular damping makes things less "spinny", making them slow down their angular velocity over time.
+---Returns the angular damping parameters of the World.
+---
+---Angular damping makes things less "spinny", making them slow down their angular velocity over time.
+---
+---
+---### NOTE:
+---Angular damping can also be set on individual colliders.
 ---
 ---@return number damping # The angular damping.
 ---@return number threshold # Velocity limit below which the damping is not applied.
@@ -1072,7 +1366,17 @@ function World:getColliders() end
 function World:getGravity() end
 
 ---
----Returns the linear damping parameters of the World.  Linear damping is similar to drag or air resistance, slowing down colliders over time as they move.
+---Returns the linear damping parameters of the World.
+---
+---Linear damping is similar to drag or air resistance, slowing down colliders over time as they move.
+---
+---
+---### NOTE:
+---A linear damping of 0 means colliders won't slow down over time.
+---
+---This is the default.
+---
+---Linear damping can also be set on individual colliders.
 ---
 ---@return number damping # The linear damping.
 ---@return number threshold # Velocity limit below which the damping is not applied.
@@ -1081,15 +1385,21 @@ function World:getLinearDamping() end
 ---
 ---Returns the response time factor of the World.
 ---
----The response time controls how relaxed collisions and joints are in the physics simulation, and functions similar to inertia.  A low response time means collisions are resolved quickly, and higher values make objects more spongy and soft.
+---The response time controls how relaxed collisions and joints are in the physics simulation, and functions similar to inertia.
 ---
----The value can be any positive number.  It can be changed on a per-joint basis for `DistanceJoint` and `BallJoint` objects.
+---A low response time means collisions are resolved quickly, and higher values make objects more spongy and soft.
+---
+---The value can be any positive number.
+---
+---It can be changed on a per-joint basis for `DistanceJoint` and `BallJoint` objects.
 ---
 ---@return number responseTime # The response time setting for the World.
 function World:getResponseTime() end
 
 ---
----Returns the tightness of the joint.  See `World:setTightness` for how this affects the joint.
+---Returns the tightness of the joint.
+---
+---See `World:setTightness` for how this affects the joint.
 ---
 ---@return number tightness # The tightness of the joint.
 function World:getTightness() end
@@ -1097,6 +1407,12 @@ function World:getTightness() end
 ---
 ---Returns whether collisions are currently enabled between two tags.
 ---
+---
+---### NOTE:
+---Tags must be set up when creating the World, see `lovr.physics.newWorld`.
+---
+---By default, collision is enabled between all tags.
+---
 ---@param tag1 string # The first tag.
 ---@param tag2 string # The second tag.
 ---@return boolean enabled # Whether or not two colliders with the specified tags will collide.
@@ -1105,6 +1421,18 @@ function World:isCollisionEnabledBetween(tag1, tag2) end
 ---
 ---Returns whether colliders can go to sleep in the World.
 ---
+---
+---### NOTE:
+---If sleeping is enabled, the World will try to detect colliders that haven't moved for a while and put them to sleep.
+---
+---Sleeping colliders don't impact the physics simulation, which makes updates more efficient and improves physics performance.
+---
+---However, the physics engine isn't perfect at waking up sleeping colliders and this can lead to bugs where colliders don't react to forces or collisions properly.
+---
+---This can be set on individual colliders.
+---
+---Colliders can be manually put to sleep or woken up using `Collider:setAwake`.
+---
 ---@return boolean allowed # Whether colliders can sleep.
 function World:isSleepingAllowed() end
 
@@ -1134,6 +1462,17 @@ function World:newCapsuleCollider(x, y, z, radius, length) end
 ---
 ---Adds a new Collider to the World.
 ---
+---
+---### NOTE:
+---This function creates a collider without any shapes attached to it, which means it won't collide with anything.
+---
+---To add a shape to the collider, use `Collider:addShape`, or use one of the following functions to create the collider:
+---
+---- `World:newBoxCollider`
+---- `World:newCapsuleCollider`
+---- `World:newCylinderCollider`
+---- `World:newSphereCollider`
+---
 ---@param x? number # The x position of the Collider.
 ---@param y? number # The y position of the Collider.
 ---@param z? number # The z position of the Collider.
@@ -1171,7 +1510,9 @@ function World:newMeshCollider(vertices, indices) end
 function World:newSphereCollider(x, y, z, radius) end
 
 ---
----Returns an iterator that can be used to iterate over "overlaps", or potential collisions between pairs of shapes in the World.  This should be called after using `World:detectOverlaps` to compute the list of overlaps. Usually this is called automatically by `World:update`.
+---Returns an iterator that can be used to iterate over "overlaps", or potential collisions between pairs of shapes in the World.
+---
+---This should be called after using `World:detectOverlaps` to compute the list of overlaps. Usually this is called automatically by `World:update`.
 ---
 ---@return function iterator # A Lua iterator, usable in a for loop.
 function World:overlaps() end
@@ -1179,6 +1520,10 @@ function World:overlaps() end
 ---
 ---Casts a ray through the World, calling a function every time the ray intersects with a Shape.
 ---
+---
+---### NOTE:
+---The callback is passed the shape that was hit, the hit position (in world coordinates), and the normal vector of the hit.
+---
 ---@param x1 number # The x coordinate of the starting position of the ray.
 ---@param y1 number # The y coordinate of the starting position of the ray.
 ---@param z1 number # The z coordinate of the starting position of the ray.
@@ -1189,7 +1534,13 @@ function World:overlaps() end
 function World:raycast(x1, y1, z1, x2, y2, z2, callback) end
 
 ---
----Sets the angular damping of the World.  Angular damping makes things less "spinny", making them slow down their angular velocity over time. Damping is only applied when angular velocity is over the threshold value.
+---Sets the angular damping of the World.
+---
+---Angular damping makes things less "spinny", making them slow down their angular velocity over time. Damping is only applied when angular velocity is over the threshold value.
+---
+---
+---### NOTE:
+---Angular damping can also be set on individual colliders.
 ---
 ---@param damping number # The angular damping.
 ---@param threshold? number # Velocity limit below which the damping is not applied.
@@ -1204,7 +1555,17 @@ function World:setAngularDamping(damping, threshold) end
 function World:setGravity(xg, yg, zg) end
 
 ---
----Sets the linear damping of the World.  Linear damping is similar to drag or air resistance, slowing down colliders over time as they move. Damping is only applied when linear velocity is over the threshold value.
+---Sets the linear damping of the World.
+---
+---Linear damping is similar to drag or air resistance, slowing down colliders over time as they move. Damping is only applied when linear velocity is over the threshold value.
+---
+---
+---### NOTE:
+---A linear damping of 0 means colliders won't slow down over time.
+---
+---This is the default.
+---
+---Linear damping can also be set on individual colliders.
 ---
 ---@param damping number # The linear damping.
 ---@param threshold? number # Velocity limit below which the damping is not applied.
@@ -1213,9 +1574,13 @@ function World:setLinearDamping(damping, threshold) end
 ---
 ---Sets the response time factor of the World.
 ---
----The response time controls how relaxed collisions and joints are in the physics simulation, and functions similar to inertia.  A low response time means collisions are resolved quickly, and higher values make objects more spongy and soft.
+---The response time controls how relaxed collisions and joints are in the physics simulation, and functions similar to inertia.
 ---
----The value can be any positive number.  It can be changed on a per-joint basis for `DistanceJoint` and `BallJoint` objects.
+---A low response time means collisions are resolved quickly, and higher values make objects more spongy and soft.
+---
+---The value can be any positive number.
+---
+---It can be changed on a per-joint basis for `DistanceJoint` and `BallJoint` objects.
 ---
 ---@param responseTime number # The new response time setting for the World.
 function World:setResponseTime(responseTime) end
@@ -1223,13 +1588,33 @@ function World:setResponseTime(responseTime) end
 ---
 ---Sets whether colliders can go to sleep in the World.
 ---
+---
+---### NOTE:
+---If sleeping is enabled, the World will try to detect colliders that haven't moved for a while and put them to sleep.
+---
+---Sleeping colliders don't impact the physics simulation, which makes updates more efficient and improves physics performance.
+---
+---However, the physics engine isn't perfect at waking up sleeping colliders and this can lead to bugs where colliders don't react to forces or collisions properly.
+---
+---This can be set on individual colliders.
+---
+---Colliders can be manually put to sleep or woken up using `Collider:setAwake`.
+---
 ---@param allowed boolean # Whether colliders can sleep.
 function World:setSleepingAllowed(allowed) end
 
 ---
 ---Sets the tightness of joints in the World.
 ---
----The tightness controls how much force is applied to colliders connected by joints.  With a value of 0, no force will be applied and joints won't have any effect.  With a tightness of 1, a strong force will be used to try to keep the Colliders constrained.  A tightness larger than 1 will overcorrect the joints, which can sometimes be desirable.  Negative tightness values are not supported.
+---The tightness controls how much force is applied to colliders connected by joints.
+---
+---With a value of 0, no force will be applied and joints won't have any effect.
+---
+---With a tightness of 1, a strong force will be used to try to keep the Colliders constrained.
+---
+---A tightness larger than 1 will overcorrect the joints, which can sometimes be desirable.
+---
+---Negative tightness values are not supported.
 ---
 ---@param tightness number # The new tightness for the World.
 function World:setTightness(tightness) end
@@ -1237,6 +1622,25 @@ function World:setTightness(tightness) end
 ---
 ---Updates the World, advancing the physics simulation forward in time and resolving collisions between colliders in the World.
 ---
+---
+---### NOTE:
+---It is common to pass the `dt` variable from `lovr.update` into this function.
+---
+---The default collision resolver function is:
+---
+---    function defaultResolver(world)
+---      world:computeOverlaps()
+---      for shapeA, shapeB in world:overlaps() do
+---        world:collide(shapeA, shapeB)
+---      end
+---    end
+---
+---Additional logic could be introduced to the collision resolver function to add custom collision behavior or to change the collision parameters (like friction and restitution) on a per-collision basis.
+---
+---> If possible, use a fixed timestep value for updating the World. It will greatly improve the
+---> accuracy of the simulation and reduce bugs. For more information on implementing a fixed
+---> timestep loop, see [this article](http://gafferongames.com/game-physics/fix-your-timestep/).
+---
 ---@param dt number # The amount of time to advance the simulation forward.
 ---@param resolver? function # The collision resolver function to use.  This will be called before updating to allow for custom collision processing.  If absent, a default will be used.
 function World:update(dt, resolver) end
diff --git a/meta/3rd/lovr/library/lovr.system.lua b/meta/3rd/lovr/library/lovr.system.lua
index d3255f75..35161332 100644
--- a/meta/3rd/lovr/library/lovr.system.lua
+++ b/meta/3rd/lovr/library/lovr.system.lua
@@ -19,7 +19,13 @@ function lovr.system.getCoreCount() end
 function lovr.system.getOS() end
 
 ---
----Requests permission to use a feature.  Usually this will pop up a dialog box that the user needs to confirm.  Once the permission request has been acknowledged, the `lovr.permission` callback will be called with the result.  Currently, this is only used for requesting microphone access on Android devices.
+---Requests permission to use a feature.
+---
+---Usually this will pop up a dialog box that the user needs to confirm.
+---
+---Once the permission request has been acknowledged, the `lovr.permission` callback will be called with the result.
+---
+---Currently, this is only used for requesting microphone access on Android devices.
 ---
 ---@param permission lovr.Permission # The permission to request.
 function lovr.system.requestPermission(permission) end
diff --git a/meta/3rd/lovr/library/lovr.thread.lua b/meta/3rd/lovr/library/lovr.thread.lua
index ca329a6b..08619426 100644
--- a/meta/3rd/lovr/library/lovr.thread.lua
+++ b/meta/3rd/lovr/library/lovr.thread.lua
@@ -5,18 +5,26 @@
 ---
 ---These are operating system level threads, which are different from Lua coroutines.
 ---
----Threads are useful for performing expensive background computation without affecting the framerate or performance of the main thread.  Some examples of this include asset loading, networking and network requests, and physics simulation.
+---Threads are useful for performing expensive background computation without affecting the framerate or performance of the main thread.
+---
+---Some examples of this include asset loading, networking and network requests, and physics simulation.
 ---
 ---Threads come with some caveats:
 ---
----- Threads run in a bare Lua environment.  The `lovr` module (and any of lovr's modules) need to
+---- Threads run in a bare Lua environment.
+---
+---The `lovr` module (and any of lovr's modules) need to
 ---  be required before they can be used.
 ---  - To get `require` to work properly, add `require 'lovr.filesystem'` to the thread code.
----- Threads are completely isolated from other threads.  They do not have access to the variables
+---- Threads are completely isolated from other threads.
+---
+---They do not have access to the variables
 ---  or functions of other threads, and communication between threads must be coordinated through
 ---  `Channel` objects.
 ---- The graphics module (or any functions that perform rendering) cannot be used in a thread.
----  Note that this includes creating graphics objects like Models and Textures.  There are "data"
+---  Note that this includes creating graphics objects like Models and Textures.
+---
+---There are "data"
 ---  equivalent `ModelData` and `Image` objects that can be used in threads though.
 ---- `lovr.event.pump` cannot be called from a thread.
 ---- Crashes or problems can happen if two threads access the same object at the same time, so
@@ -35,6 +43,16 @@ function lovr.thread.getChannel(name) end
 ---
 ---Creates a new Thread from Lua code.
 ---
+---
+---### NOTE:
+---The Thread won\'t start running immediately.
+---
+---Use `Thread:start` to start it.
+---
+---The string argument is assumed to be a filename if there isn't a newline in the first 1024 characters.
+---
+---For really short thread code, an extra newline can be added to trick LÖVR into loading it properly.
+---
 ---@overload fun(filename: string):lovr.Thread
 ---@overload fun(blob: lovr.Blob):lovr.Thread
 ---@param code string # The code to run in the Thread.
@@ -42,7 +60,13 @@ function lovr.thread.getChannel(name) end
 function lovr.thread.newThread(code) end
 
 ---
----A Channel is an object used to communicate between `Thread` objects.  Channels are obtained by name using `lovr.thread.getChannel`.  Different threads can send messages on the same Channel to communicate with each other.  Messages can be sent and received on a Channel using `Channel:push` and `Channel:pop`, and are received in a first-in-first-out fashion. The following types of data can be passed through Channels: nil, boolean, number, string, and any LÖVR object.
+---A Channel is an object used to communicate between `Thread` objects.
+---
+---Channels are obtained by name using `lovr.thread.getChannel`.
+---
+---Different threads can send messages on the same Channel to communicate with each other.
+---
+---Messages can be sent and received on a Channel using `Channel:push` and `Channel:pop`, and are received in a first-in-first-out fashion. The following types of data can be passed through Channels: nil, boolean, number, string, and any LÖVR object.
 ---
 ---@class lovr.Channel
 local Channel = {}
@@ -53,28 +77,52 @@ local Channel = {}
 function Channel:clear() end
 
 ---
----Returns whether or not the message with the given ID has been read.  Every call to `Channel:push` returns a message ID.
+---Returns whether or not the message with the given ID has been read.
+---
+---Every call to `Channel:push` returns a message ID.
 ---
 ---@param id number # The ID of the message to check.
 ---@return boolean read # Whether the message has been read.
 function Channel:hasRead(id) end
 
 ---
----Returns a message from the Channel without popping it from the queue.  If the Channel is empty, `nil` is returned.  This can be useful to determine if the Channel is empty.
+---Returns a message from the Channel without popping it from the queue.
+---
+---If the Channel is empty, `nil` is returned.
+---
+---This can be useful to determine if the Channel is empty.
+---
+---
+---### NOTE:
+---The second return value can be used to detect if a `nil` message is in the queue.
 ---
 ---@return any message # The message, or `nil` if there is no message.
 ---@return boolean present # Whether a message was returned (use to detect nil).
 function Channel:peek() end
 
 ---
----Pops a message from the Channel.  If the Channel is empty, an optional timeout argument can be used to wait for a message, otherwise `nil` is returned.
+---Pops a message from the Channel.
+---
+---If the Channel is empty, an optional timeout argument can be used to wait for a message, otherwise `nil` is returned.
+---
+---
+---### NOTE:
+---Threads can get stuck forever waiting on Channel messages, so be careful.
 ---
 ---@param wait? number # How long to wait for a message to be popped, in seconds.  `true` can be used to wait forever and `false` can be used to avoid waiting.
 ---@return any message # The received message, or `nil` if nothing was received.
 function Channel:pop(wait) end
 
 ---
----Pushes a message onto the Channel.  The following types of data can be pushed: nil, boolean, number, string, and userdata.  Tables should be serialized to strings.
+---Pushes a message onto the Channel.
+---
+---The following types of data can be pushed: nil, boolean, number, string, and userdata.
+---
+---Tables should be serialized to strings.
+---
+---
+---### NOTE:
+---Threads can get stuck forever waiting on Channel messages, so be careful.
 ---
 ---@param message any # The message to push.
 ---@param wait? number # How long to wait for the message to be popped, in seconds.  `true` can be used to wait forever and `false` can be used to avoid waiting.
@@ -83,7 +131,11 @@ function Channel:pop(wait) end
 function Channel:push(message, wait) end
 
 ---
----A Thread is an object that runs a chunk of Lua code in the background.  Threads are completely isolated from other threads, meaning they have their own Lua context and can't access the variables and functions of other threads.  Communication between threads is limited and is accomplished by using `Channel` objects.
+---A Thread is an object that runs a chunk of Lua code in the background.
+---
+---Threads are completely isolated from other threads, meaning they have their own Lua context and can't access the variables and functions of other threads.
+---
+---Communication between threads is limited and is accomplished by using `Channel` objects.
 ---
 ---To get `require` to work properly, add `require 'lovr.filesystem'` to the thread code.
 ---
diff --git a/meta/3rd/lovr/library/lovr.timer.lua b/meta/3rd/lovr/library/lovr.timer.lua
index b4693e4c..4bcb02f1 100644
--- a/meta/3rd/lovr/library/lovr.timer.lua
+++ b/meta/3rd/lovr/library/lovr.timer.lua
@@ -1,7 +1,9 @@
 ---@meta
 
 ---
----The `lovr.timer` module provides a few functions that deal with time.  All times are measured in seconds.
+---The `lovr.timer` module provides a few functions that deal with time.
+---
+---All times are measured in seconds.
 ---
 ---@class lovr.timer
 lovr.timer = {}
@@ -13,7 +15,15 @@ lovr.timer = {}
 function lovr.timer.getAverageDelta() end
 
 ---
----Returns the time between the last two frames.  This is the same value as the `dt` argument provided to `lovr.update`.
+---Returns the time between the last two frames.
+---
+---This is the same value as the `dt` argument provided to `lovr.update`.
+---
+---
+---### NOTE:
+---The return value of this function will remain the same until `lovr.timer.step` is called.
+---
+---This function should not be used to measure times for game behavior or benchmarking, use `lovr.timer.getTime` for that.
 ---
 ---@return number dt # The delta time, in seconds.
 function lovr.timer.getDelta() end
@@ -25,19 +35,25 @@ function lovr.timer.getDelta() end
 function lovr.timer.getFPS() end
 
 ---
----Returns the time since some time in the past.  This can be used to measure the difference between two points in time.
+---Returns the time since some time in the past.
+---
+---This can be used to measure the difference between two points in time.
 ---
 ---@return number time # The current time, in seconds.
 function lovr.timer.getTime() end
 
 ---
----Sleeps the application for a specified number of seconds.  While the game is asleep, no code will be run, no graphics will be drawn, and the window will be unresponsive.
+---Sleeps the application for a specified number of seconds.
+---
+---While the game is asleep, no code will be run, no graphics will be drawn, and the window will be unresponsive.
 ---
 ---@param duration number # The number of seconds to sleep for.
 function lovr.timer.sleep(duration) end
 
 ---
----Steps the timer, returning the new delta time.  This is called automatically in `lovr.run` and it's used to calculate the new `dt` to pass to `lovr.update`.
+---Steps the timer, returning the new delta time.
+---
+---This is called automatically in `lovr.run` and it's used to calculate the new `dt` to pass to `lovr.update`.
 ---
 ---@return number delta # The amount of time since the last call to this function, in seconds.
 function lovr.timer.step() end