* Resolve symlinks to avoid including the same thing twice (like
version-numered dylib symlinks)
* Look in all Qt prefix paths for plugins - the package may not
necessarily have the same path
* reduce install_name_tool invocations to make it a bit faster
* change dylib IDs to remove original source path
Adds a Nix flake, allowing easy building and running of melonDS using the Nix package manager, as well as potentially very stable and reproducible CI in the future.
* 0 dot disp defaults to the 24 bit integer limit
* useless correction
it goes through the reset function to set the variable on boot anyway but why not have the initialized state be correct too
* Make `EmuInstance::cheatFile` use a `unique_ptr`
- Fixes a memory leak, as the cheat file wasn't cleaned up in the destructor
* Split `AREngine` and `ARCodeFile` apart
- Suitable for frontends that have their own way of storing cheats
- Store the cheats in `AREngine` in a `std::vector`
- Apparently cheats are _supposed_ to be executed each frame; I didn't understand this until recently
too slow, not accurate enough.
we need to do a *lot* more research into the specifics of how this works with all the various aspects of the cpu's timings before we can make a good implementation
updating console, fixes crash on console type switch
EmuInstance::consoleType is already updated, so the check for whether we
should recreate the NDS to switch console type never succeeds.
On Linux, our stubbed glib functions were conflicting with the ones from
real glib, which gets used by Qt when we're running on a GTK-based
desktop. Avoid a crash by not exposing them.
non-exhaustive (but exhausting) list of changes:
* base laid for multiple window support, but will likely require more work to work correctly
* encapsulation of frontend state for proper multi-instance support
* (JIT still needs a fix for the NDS::Current workaround but we can get there later)
* new, more flexible configuration system
* nothing works yet
* don't double buffer 3D framebuffers for the GL Renderer
looks like leftovers from when 3D+2D composition was done in the frontend
* oops
* it works!
* implement display capture for compute renderer
it's actually just all stolen from the regular OpenGL renderer
* fix bad indirect call
* handle cleanup properly
* add hires rendering to the compute shader renderer
* fix UB
also misc changes to use more unsigned multiplication
also fix framebuffer resize
* correct edge filling behaviour when AA is disabled
* fix full color textures
* fix edge marking (polygon id is 6-bit not 5)
also make the code a bit nicer
* take all edge cases into account for XMin/XMax calculation
* use hires coordinate again
* stop using fixed size buffers based on scale factor in shaders
this makes shader compile times tolerable on Wintel
- beginning of the shader cache
- increase size of tile idx in workdesc to 20 bits
* apparently & is not defined on bvec4
why does this even compile on Intel and Nvidia?
* put the texture cache into it's own file
* add compute shader renderer properly to the GUI
also add option to toggle using high resolution vertex coordinates
* unbind sampler object in compute shader renderer
* fix GetRangedBitMask for 64 bit aligned 64 bits
pretty embarassing
* convert NonStupidBitfield.h back to LF only new lines
* actually adapt to latest changes
* fix stupid merge
* actually make compute shader renderer work with newest changes
* show progress on shader compilation
* remove merge leftover
* fix the pu region's end point overflowing
According to gericom it cannot overflow at all
* set a minimum and a better maximum for the pu region size
* fix pu logging
* PU regions with a size of 31 always take up the entire address space
also tweak some logging a little more
* start is actually force aligned by size, oops
* small tweaks
* hopefully more clear code
* math is for nerds
Precompute all 16 5-bit RGB palette colours into 8-bit RGBA to avoid
repeated and superfluous calculation within the nested loop at the
point of index lookup.
A speedup was observed, from ~7ms, to a consistent 1ms
(i.e. now practically instantaneous) through timing with
std::chrono::high_resolution_clock.
Also improve comprehensibility, by using meaningful names, where
appropriate, for loop counter variables.
* implement channels
* rework power-down support, fixing bugs
* fix bug when W_BeaconInterval is zero
* fix potential missing IRQs when writing to W_IE
Should remove the artifacts caused by the previous nearest
resampling. May be worth replacing with something better in the
future, but this is an improvement for now.
* check if an nds save file can be opened for writing
also add the ability to open a file in append mode
* fix multi-instance saves
also move the check for file writability into a separate function (probably uneeded?)
* implement check for gba roms
* move rom load error messages into the functions
also finish gba slot (oops)
* improve error string
* check write perms before saving path settings
* fix memory leak
* check for writability of firmware/nand/sds
* add secondary checks for nand/firmware
* add check for config file being writable
* Return the file write error as a QString to avoid the invalid char*
causing a garbled error message.
Qt wants it as QString either way.
Fixed a bug causing overlapping protection regions priority not taken into account, when access permission or cachability bits were changed only on the least priority overlap
- According to GBATek, all DSiWare games have a high title ID of 0x00030004
- Some homebrew apps set the Unitcode bits to DSi mode to enable support of DSi features
* First crack at ensuring the render thread doesn't touch GPU state while it's being serialized
* Get rid of the semaphore wait
* Add some extra fields into GPU3D's serialization
* Oops, TempVertexBuffer is already serialized
* Move vertex serialization into its own method
* Lock the GPU3D state when rendering on the render thread or serializing it
* Revert "Lock the GPU3D state when rendering on the render thread or serializing it"
This reverts commit 2f49a551c1.
* Add comments that describe the synchronization within GPU3D_Soft
- I need to understand it before I can solve my actual problem
- Now I do
* Revert "Revert "Lock the GPU3D state when rendering on the render thread or serializing it""
This reverts commit 1977566a6d.
* Let's try locking the GPU3D state throughout NDS::RunFrame
- Just to see what happens
* Slim down the lock's scope
* Narrow the lock's scope some more
* Remove the lock entirely
* Try protecting the GPU3D state with just a mutex
- I'll clean this up once I know it works
* Remove a duplicate method definition
* Add a missing `noexcept` specifier
* Remove an unused function
* Cut some non-hardware state from `GPU3D`'s savestate
* Assume that the next frame after loading a savestate won't be identical
* Actually, it _is_ worth it
* Don't serialize the clip matrix
- It's recalculated anyway
* Serialize `RenderPolygonRAM` as an array of indexes
* Clean up some comments
- I liked the dialogue style, but oh well
* Try restarting the render thread instead of using the lock
- Let's see what happens
* Put the lock back
* Fix some polygon and vertex indexes being saved incorrectly
- Taking the difference between two pointers results in the number of elements, not the number of bytes
* Remove `SoftRenderer::StateBusy` since it turns out we don't need it
- The real synchronization was the friends we made along the way
* Reload the SD card for `CartSD` and all subclasses
* Make `ROMManager::LoadDLDISDCard` delegate to `GetDLDISDCardArgs`
* Add a method overload for `CartSD::SetSDCard`
* Initialize new SD card images with the correct size
* Sync the old card to the host (if applicable) when move-assigning a new one
* Only sync the old card to the host if it's not read-only
* Remove static state in `FATStorage`
- Replace `FF_ReadStorage` and `FF_WriteStorage` with lambda functions
- Keep open and use the single `File` handle throughout the `FATStorage`'s life
* Resolve some warnings
- Their frequent appearance in the build logs is driving me nuts
* Silence warnings about `offsetof`
* Don't apply `-Wno-invalid-offset` to C, only to C++
* Align some two-element `u32` arrays as `u64`s
- To pacify "unaligned read/write" warnings from UBSan
* Align some more arrays based on how they're accessed
- When I adapted `GBACart::ParseROM` to use `unique_ptr` instead of a plain pointer, I forgot to remove the code that copied the SRAM data
- That code was made unnecessary because of the move
* Simplify the SRAM's representation in `NDSCartArgs`
- I overthought this one.
- I could've just checked `args && args->SRAM`, but then some other poor bastard might make this mistake.
- Don't mix `pair`, `optional`, and `unique_ptr` all at once, kids.
* Fix a `nullptr` read
* Fix detection of native NDS ARM BIOS images
- Instead of checking for built-in BIOS images, now the altered methods check for native ones
- The CRC32 must match exactly; patched BIOS images will result in `false`
* Encapsulate `NDS::ARM9BIOS` and `ARM7BIOS`
- Also compute the checksum only when setting the BIOS
* Allow 3D renderers to be created without passing `GPU` to the constructor
* Make the initial 3D renderer configurable via `NDSArgs`
* Fix a compiler error
* Sprinkle `const` around where appropriate
- This will make it easier to use `NDS` objects in `const` contexts (e.g. `const` parameters or methods)
* Remove the `const` qualifier on `DSi_DSP::DSPRead16`
- MMIO reads can be non-pure, so this may not be `const` in the future
Fixes Castlevania: Dawn of Sorrow's checksumming which uses crc16 swi and has garbage in the upper 16 bits of r0.
The official BIOS would seem to implicitly clear these upper 16 bits.
* Get rid of `ConfigEntry::ExternalBIOSEnable`
- Now the BIOS files themselves are checked
- The frontend's `Config::ExternalBIOSEnable` is not affected
* Add `JITArgs`
* Pass the JIT status to the `ARM` constructors
* Encapsulate `NDS::EnableJIT`
* Pass `JITArgs` to `ARMJIT`'s constructor
* Remove the `JIT_*` `ConfigEntry`s in favor of members
- Allow all the JIT args to be set with `NDS::SetJITArgs`
- Encapsulate the JIT-related parameters in `ARMJIT` so they can reset the block cache if changed
- Update the active (or newly-created) console in the frontend with adjusted JIT args
* Make audio bit depth and interpolation configurable in `NDSArgs`
- Define enums for both
- Give those settings default values in `NDSArgs`
- Remove `ConfigEntry::AudioBitDepth`
- Initialize these settings in the relevant SPU constructors
* Move the last DSi-specific logic in `Reset` to its own subclass
* Remove `ConfigEntry::DSi_FullBIOSBoot`
- Add members to `DSi` instead for getting and setting this
- Update the frontend to accommodate these changes
* Oops, missed a spot
* Remove `ConfigEntry::Firm_MAC` and `Platform::GetConfigArray`
- Also move the MAC parsing code to `ROMManager`
* Remove the last `ConfigEntry` state
- Make GDB support configurable via members
* Add some `#ifdef`s that I'd almost forgotten
* Remove `FATStorage::Open` and `FATStorage::Close`
- That's what the constructor and destructor are for, respectively
* Add `FATStorage::IsReadOnly`
* Slight cleanup of `FATStorage`
- Make it move-constructible and move-assignable
- Represent the absence of a sync directory with `std::optional`, not an empty string
- Add `FATStorageArgs` for later use
* Refactor `CartHomebrew` to accept an optional `FATStorageArgs`
- `CartHomebrew` uses it to load an SD card image
- Not passing a `FATStorage` directly because we won't know if we need to load the card until we parse the ROM
- Store the `FATStorage` inside a `std::optional` instead of a pointer
- `CartHomebrew::Reset` no longer reloads the SD card; the frontend needs to set it with the `SetSDCard` method
* Close `NANDImage::CurFile` when move-assigning
- Whoops
* Add `Args.h`
- To construct a `NDS` or `DSi` with arguments
- Mostly intended for system files
* Fix incorrect `final` placement
* Refactor how `DSi`'s NAND and SD card are set
- Provide them via a `DSiArgs` argument in the constructor
- Give `DSi_MMCStorage` ownership of the `NANDImage` or `FATStorage` as needed, and expose getters/setters
- Replace `DSi_SDHost::Ports` with a `array<unique_ptr, 2>` to reduce the risk of leaks
- Store `DSi_MMCStorage`'s disk images in a `std::variant`
- The SD card and NAND image are no longer reset in `Reset()`; the frontend will need to do that itself
* Add getters/setters on `DSi` itself for its storage media
* Remove newly-unused `Platform::ConfigEntry`s
* Use `DSi::SetNAND` in the frontend
* Add `EmuThread::NeedToRecreateConsole`
* Document `NDSArgs` and give its fields default values
* Refactor how system files are loaded upon construction
- Pass `NDSArgs&&` into `NDS`'s constructor
- Use `std::array` for the emulator's BIOS images and the built-in FreeBIOS, to simplify copying and comparison
- Initialize the BIOS, firmware, and SD cards from `NDSArgs` or `DSiArgs`
- Add a new default constructor for `NDS` (not `DSi`) that initializes the DS with default system files
- Embed `FirmwareMem::Firmware` directly instead of in a `unique_ptr`
- `SPIHost` now takes a `Firmware&&` that it forwards to `FirmwareMem`
- Add `Firmware` getters/setters plus `const` variants for `NDS`, `Firmware`, and `FirmwareMem`
- Simplify installation of firmware
* Initialize the DSi BIOS in the constructor
- Change `DSi::ARM9iBIOS` and `ARM7iBIOS` to `std::array`
* Update the frontend to reflect the core's changes
* Remove `DSi_SDHost::CloseHandles`
* Pass `nullopt` instead of the empty string when folder sync is off
* Deduplicate ROM extraction logic
- `LoadGBAROM` and `LoadROM` now delegate to `LoadROMData`
- Also use `unique_ptr` instead of `new[]`
* Oops, missed some `get()`'s
* Move `NDS::IsLoadedARM9BIOSBuiltIn` to the header
- So it's likelier to be inlined
- Same for the ARM7 version
* Remove `NDS::SetConsoleType`
* Add `NDS::SetFirmware`
* Move `GBACart::SetupSave` to be `protected`
- It was only ever used inside the class
* Rename `GBACart::LoadSave` to `SetSaveMemory`
- Same for the cart slot
* Declare `GBACartSlot` as a friend of `GBACart::CartCommon`
* Revise `GBACartSlot`'s getters and setters
- Rename `InsertROM` and `LoadROM` to `SetCart`
- Add a `GetCart` method
* Clean up getters and setters for NDS and GBA carts
* Clean up how carts are inserted into the slots
- Remove setters that operate directly on pointers, to simplify error-handling (use ParseROM instead)
- Add overloads for all carts that accept a `const u8*` (to copy the ROM data) and a `unique_ptr<u8[]>` (to move the ROM data)
- Store all ROM and RAM data in `unique_ptr`
- Default-initialize all fields
- Simplify constructors and destructors, inheriting where applicable
* Refactor GBA save data insertion
- Make `SetupSave` private and non-virtual and move its logic to be in `SetSaveMemory`
- Add overloads for setting save data in the constructor
- Update the SRAM completely in `SetSaveMemory`
* Clean up `NDSCart::CartCommon::SetSaveMemory`
- Move its declaration next to the other `SaveMemory` methods
- Move its (empty) implementation to the header
* Add some comments
* Add Utils.cpp and Utils.h
* Rename some functions in Utils for clarity
* Add `GBACart::ParseROM` and `NDSCart::ParseROM` overloads that accept `unique_ptr<u8[]>`
- The `u8*` overloads delegate to these new overloads
- Also move `SetupSave` for both kinds of carts to be private non-virtual methods
* Finalize the `NDSCart` refactor
- Add `NDSCartArgs` to pass to `ParseROM`
- Add SRAM arguments for all retail carts
- Initialize SRAM inside the constructor
- Delegate to other constructors where possible
* Replace `ROMManager::NDSSave` and `GBASave` with `unique_ptr`
* Make both cart slots return the previously-inserted cart in `EjectCart`
- Primarily intended for reusing carts when resetting the console
* Make `NDS::EjectCart` return the old cart
* Initialize both cart slots with the provided ROM (if any)
* Make `NDS::EjectGBACart` return the ejected cart
* Clean up some comments in Args.h
* Rename `ROMManager::LoadBIOS` to `BootToMenu`
- Clarifies the intent
* Add `ROMManager::LoadDLDISDCard`
* Add a doc comment
* Refactor how the `NDS` is created or updated
- Rewrite `CreateConsole` to read from `Config` and load system files, but accept carts as arguments
- Fail without creating an `NDS` if any required system file doesn't load
- Add `UpdateConsole`, which delegates to `CreateConsole` if switching modes or starting the app
- Use `std::variant` to indicate whether a cart should be removed, inserted, or reused
- Load all system files (plus SD cards) in `UpdateConsole`
- Eject the cart and reinsert it into the new console if applicable
* Respect some more `Config` settings in the `Load*` functions
* Remove `InstallNAND` in favor of `LoadNAND`
* Fix some potential bugs in `LoadROMData`
* Oops, forgot to delete the definition of `InstallNAND`
* Add functions to get `FATStorageArgs`
- Not the cards themselves, but to get the arguments you _would_ use to load the cards
* Refactor `ROMManager::LoadROM`
- Load the ROM and save data before trying to initialize the console
* Clean up `ROMManager::Reset` and `BootToMenu`
- Let `EmuThread::UpdateConsole` do the heavy lifting
* Clean up `LoadGBAROM`
* Remove some unused functions
* Set the default DSi BIOS to be broken in `DSiArgs`
* Respect `Config::DSiFullBIOSBoot` when loading DSi BIOS files
* Remove some more unused functions
* Remove redundant `virtual` specifiers
* Refactor `NDSCart::CartCommon::Type()` to return a member instead of a constant
- One less virtual dispatch
- The cart type is read in `NDSCartSlot::DoSavestate`, which is a path downstream (due to rewinding)
* Remove a hash that I computed for debugging purposes
* Make `ByteSwap` `constexpr`
* Remove an unused `#include`
* Remove unnecessary functions from the NDS carts
- Mostly overrides that added nothing
* Default-initialize all NDSCart fields
* Make `GBACart::Type()` not rely on virtual dispatch
- `GBACartSlot::DoSavestate` calls it, and savestates can be a hot path downstream
* Don't forget to reset the base class in `CartGameSolarSensor::Reset()`
* Remove redundant `virtual` specifiers
* Default-initialize some fields in `GBACart`
* Fix ROMs not loading from archives in the frontend
- Whoops
* Change how the `Firmware` member is declared
* Forgot an include in Utils.cpp
* Rename `FirmwareMem::Firmware` to `FirmwareData` to fix a build error on Linux
- One of these days I'll convince you people to let me use `camelCaseMemberNames`
* Add `override` to places in `DSi_MMCStorage` that warrant it
* Fix firmware saving on the frontend
- Remove `GetConfigString` and `ConfigEntry::WifiSettingsPath` while I'm at it
* Add a non-const `GetNAND()`
* Move `SPUChannel` and `SPUCaptureUnit` to be stored inside `array`s instead of allocated separately
* Default-initialize most of `SPU`'s fields
* Generate the interpolation tables at compile-time with `constexpr`
- Now it's faster and thread-safe
* Slight cleanup in SPU
- Iniitialize most fields in the class declaration
* Mark `SPU` as `explicit`
* Give `GPU2D::Unit` a virtual destructor
- Undefined behavior avoided!
* Add an `array2d` alias
* Move various parts of `GPU2D::SoftRenderer` to `constexpr`
- `SoftRenderer::MosaicTable` is now initialized at compile-time
- Most of the `SoftRenderer::Color*` functions are now `constexpr`
- The aforementioned functions are used with a constant value in at least one place, so they'll be at least partially computed at compile-time
* Generalize `GLRenderer::PrepareCaptureFrame`
- Declare it in the base `Renderer3D` class, but make it empty
* Remove unneeded `virtual` specifiers
* Store `Framebuffer`'s memory in `unique_ptr`s
- Reduce the risk of leaks this way
* Clean up how `GLCompositor` is initialized
- Return it as an `std::optional` instead of a `std::unique_ptr`
- Make `GLCompositor` movable
- Replace `GLCompositor`'s plain arrays with `std::array` to simplify moving
* Pass `GPU` to `GLCompositor`'s important functions instead of passing it to the constructor
* Move `GLCompositor` to be a field within `GLRenderer`
- Some methods were moved up and made `virtual`
* Fix some linker errors
* Set the renderer in the frontend
* Remove unneeded `virtual` specifiers
* Remove `RenderSettings` in favor of just exposing the relevant member variables
* Update the frontend to accommodate the core changes
* Add `constexpr` and `const` to places in the interpolator
* Qualify references to `size_t`
* Construct the `optional` directly instead of using `make_optional`
- It makes the Linux build choke
- I think it's because `GLCompositor`'s constructor is `private`
* First crack at refactoring NDS and DSi into objects
- Remove all global/`static` variables in `NDS` and related classes
- Rely more on virtual dispatch when we need to pick methods at runtime
- Pass `NDS&` or `DSi&` to its constituent components where necessary
- Introduce some headers or move some definitions to break `#include` cycles
* Refactor the frontend to accommodate the core's changes
* Move up `SchedList`'s declaration
- Move it to before the components are initialized so the `map`s inside are initialized
- Fields in C++ are initialized in the order they're declared
* Fix a crash when allocating memory
* Fix JIT-free builds
* Fix GDB-free builds
* Fix Linux builds
- Explicitly qualify some member types in NDS, since they share the same name as their classes
* Remove an unnecessary template argument
- This was causing the build to fail on macOS
* Fix ARM and Android builds
* Rename `Constants.h` to `MemConstants.h`
* Add `NDS::IsRunning()`
* Use an `#include` guard instead of `#pragma once`
* Reorganize namespaces
- Most types are now moved into the `melonDS` namespace
- Only good chance to do this for a while, since a big refactor is next
* Fix the build
* Move TinyVector to a new file
- So it's less sensitive to #include ordering
* Forgot to include assert.h
* Refactor ARMJIT_Memory into an object
* Oops, forgot a declaration
* Refactor ARMJIT to be contained in an object
* Remove an unused function declaration
* Add a missing #include
* Remove a now-unused global
* Use ARMJIT_Memory's own memory access functions
* Fix some omissions in the ARM JIT
* Move libandroid to be a member of ARMJIT_Memory instead of a global
* Default-initialize most fields in ARMJIT_Compiler.h
* Define NOOP_IF_NO_JIT
* Finish refactoring the JIT to be object-oriented
* Refactor GPU3D to be an object
- Who has two thumbs and is the sworn enemy of global state? This guy!
* Refactor GPU itself to be an object
- Wow, it's used in a lot of places
- Also introduce a new `Melon` namespace for a few classes
- I expect other classes will be moved into `Melon` over time
* Change signature of Renderer3D::SetRenderSettings
- Make it noexcept, and its argument const
* Remove some stray whitespace
* Move NDSCart-related global state into objects
- RAII will now do the heavy lifting
- Mark some methods as const or noexcept
* Move GBACart-related global state into objects (#1870)
- RAII will now do the heavy lifting
- Mark some methods as const or noexcept
- Once the `NDS` object is finalized, most of these `assert`s can go away
* Make AREngine::RunCheat public (#1872)
- I use it directly in melonDS DS to apply single cheats without using ARCodeFile
- Before the AREngine refactor I could just redeclare the function in my code
- Now I can't
- I use it directly in melonDS DS to apply single cheats without using ARCodeFile
- Before the AREngine refactor I could just redeclare the function in my code
- Now I can't
- RAII will now do the heavy lifting
- Mark some methods as const or noexcept
- Once the `NDS` object is finalized, most of these `assert`s can go away
* fix regression with facing view
Only the check for a polygon being counter-clockwise is supposed to be <=
* only use dot < 0 for 'cull front face' polygons
this is the fix.
The ExtractFileFromArchive function can sometimes return -1 on error,
however the function's return type was specified as u32, which would
mean that it would instead be represented as the maximum value.
Change the function's return type to the signed s32 instead, and
correct uses.
Free some objects that were allocated with new but not deleted, and in
one case, do not set a pointer to nullptr before deleting, as this
results in a memory leak due to memory allocated not being freed.
If a user manages to open a file as a ROM that is greater than 1 GiB,
it will cause a segmentation fault (a crash) in LoadROM due to a delete
being called on an uninitialised pointer, which is undefined behaviour.
Initialise filedata to nullptr to prevent this, as deleting a null
pointer is defined as a no-op.
In QCamera in Qt 5, the camera is required to have been loaded before
querying its settings and resolutions. Doing so without loading being
finished would result in the returned list being empty.
See https://doc.qt.io/qt-5.15/qcamera.html#supportedViewfinderSettings
Add a QEventLoop that waits for the state to change from Loading to
Loaded before supportedViewfinderSettings() is called to ensure that
valid information is returned.
(Fixes my camera being blank in preview, same issue also presumed
to occur when camera is needed in game, fix tested on a USB camera on
a Linux system and Qt 5.)
Jaklyy