Hello, the speed switch code should get the value in the address that the PC register pointed instead of the value of in register.

https://github.com/mohanson/gameboy/blob/ff9c35b3957b4c215bd817c4d2ca18de0fb7f25e/src/motherboard.rs#L21

Reference: https://gbdev.gg8.se/wiki/articles/CGB_Registers#FF4D_-KEY1-CGB_Mode_Only-_Prepare_Speed_Switch

The actual speed switch is performed by executing a STOP command after Bit 0 has been set. After that Bit 0 will be cleared automatically, and the gameboy will operate at the 'other' speed.

Cargo.lock is a generated file by cargo, but build fail if your generated Cargo.lock is present. Suggestion: delete Cargo.lock and add it to .gitignore

• adds terminal rendering w/ blockish (-t flag)
• handles input w/ crossterm_input

video demos:

zelda: https://www.youtube.com/watch?v=ou_4WSuYRHk pokemon yellow: https://youtu.be/XcrkZnIS1lM

As I've been playing with this emulator I've noticed that the clock rate seems to jump around. Sometimes the game will slow the a crawl, and other times it runs so fast that it becomes hard to control. I think this may be because my machine isn't that powerful, and it may be missing frame timing (it was taking ~30ms a cpu cycle rather than the target 16 when I threw some prints in there).

I think the following diff might fix things:

diff --git a/src/cpu.rs b/src/cpu.rs
index bb07a5e..80f3556 100644
--- a/src/cpu.rs
+++ b/src/cpu.rs
@@ -1693,7 +1693,7 @@ impl Cpu {
 pub struct Rtc {
     pub cpu: Cpu,
     step_cycles: u32,
-    step_zero: time::SystemTime,
+    step_zero: time::Instant,
     step_flip: bool,
 }
 
@@ -1703,7 +1703,7 @@ impl Rtc {
         Self {
             cpu,
             step_cycles: 0,
-            step_zero: time::SystemTime::now(),
+            step_zero: time::Instant::now(),
             step_flip: false,
         }
     }
@@ -1713,14 +1713,20 @@ impl Rtc {
         if self.step_cycles > STEP_CYCLES {
             self.step_flip = true;
             self.step_cycles -= STEP_CYCLES;
-            let d = time::SystemTime::now().duration_since(self.step_zero).unwrap();
+            let now = time::Instant::now();
+            let d = now.duration_since(self.step_zero);
             let s = u64::from(STEP_TIME.saturating_sub(d.as_millis() as u32));
             rog::debugln!("CPU: sleep {} millis", s);
             thread::sleep(time::Duration::from_millis(s));
-            self.step_zero = self
-                .step_zero
+            self.step_zero = self.step_zero
                 .checked_add(time::Duration::from_millis(u64::from(STEP_TIME)))
                 .unwrap();
+
+            // If now is after the just updated target frame time, reset to
+            // avoid drift.
+            if now.checked_duration_since(self.step_zero).is_some() {
+                self.step_zero = now;
+            }
         }
         let cycles = self.cpu.next();
         self.step_cycles += cycles;

I don't think switching from SystemTime to monotonic time is actually that important, but I think it is probably the right thing to do here since we are just using this stuff for frame timing and it doesn't really have to relate to the real wallclock time.

But the bug can take a while to manifest, so I can't really confirm yet. I'm going to run an emulator with this patch for a while and update this thread.

I'm currently running bspwm, and everything is set to tiling.

I believe the issue is that gameboy is trying to launch with specific geometry size, but since it can't due to tiling, it segfaults.

If I'm right, I can fix this by setting bspwm to float that window.

The problem though, is that I can't obtain the wm_class since the window instantly closes.

So could you tell me what the wm_class is?

Hello, there is a small typo in Hdma. 0xff43 should be 0xff53.

Reference: https://gbdev.gg8.se/wiki/articles/Video_Display#FF53_-HDMA3-CGB_Mode_Only-_New_DMA_Destination.2C_High

First, this is an awesome project! I looked into a couple of other gameboy emulators with many more commits and contributors before trying out this one, and this was the first project that worked.

I'm fortunate enough to be able to figure out the command line flags by reading sources, but I suspect that not everyone who is interested in using this project will have that option open to them so I thought I would send you a patch to try to make the command line flags a little more discoverable.

Adding the long-form flags gives argparse something to use as the argument name in its help message. The new usage message is:

Usage:
  target/debug/gameboy [OPTIONS] [ROM]

Gameboy emulator

Positional arguments:
  rom                   Rom name

Optional arguments:
  -h,--help             Show this help message and exit
  -a,--enable-audio     Enable audio
  -x,--scale-factor SCALE_FACTOR
                        Scale the video by a factor of 1, 2, 4, or 8

I also fixed a compiler warning.

Correct me if I'm wrong, but I think a mask of 0x0FFF is consistent with the 8-bit alu_add function:

self.reg.set_flag(C, u16::from(a) + u16::from(n) > 0xff);
self.reg.set_flag(H, (a & 0x0f) + (n & 0x0f) > 0x0f);

and other emulators I've checked with: SameBoy, binjgb, CoffeeGB

While I review your code in src/memory.rs, one thing makes me confused. There are two functions named as get_halfword and set_halfword. Literally, they are used to get half of a WORD length data from a specific memory address, or setting half of a WORD length data. In fact, they are designed as 16-bit version getter and setter for memory.

ARM64 and x86 family of CPU defined a WORD as 16 bits(2 bytes). Back to time of Intel 8080 and Zilog 80, a WORD stands for 8 bits. I have not found any LR35902 specifications, indicating that a WORD is 32 bits.

So I wonder if a correctly expressed version of those function names should be get_word(for widely used x86 and ARM standard) or get_dword(for 8080 and Z80 standard), the same for setter.

Thanks for your hard work.

What is the problem with Chinese README's

Firstly, we congratulate you for getting so much star by sharing this repository with humanity.

But it is very disappointing for non-Chinese speakers when one couldn't understand what a trending repository is about.

When we see such a repo on trending, our minds are blurring like Gollum's.

There is a way you can help to solve this disappointment which I believe is experienced by many people who want to know more about your valuable work and appreciate it.

What we want:

• Please add English translation of your README so you are sharing your work and knowledge with more people.

How this will help you:

• More feedback to fix and improve your project.
• New ideas about your project.
• Greater fame.
• 

“Sharing knowledge is the most fundamental act of friendship. Because it is a way you can give something without loosing something.”

— Richard Stallman

Thank you!

This issue created by us/english-please script. Please report on any error. Thank you!

I dont know if this is the right repo for this, please tell me if it is not.

Adds a dockerfile to host the emulator in tty mode as an SSH container. On ssh connection:

1. prompts for a zip HTTP url from which to download the rom
2. downloads the rom and unzip it
3. starts the emulator with the rom

https://github.com/mohanson/gameboy/blob/87865c1b7317e631ef375476311cc5c5ae9e95c8/src/apu.rs#L998 n should not be incremented by 1. Otherwise, by example, if n = 1 you will get a divisor value of 32 instead of the 16 expected.

You can also use this operation instead:

n => u32::from(n) << 4

According to the GBCPUman：C - Set if no borrow. In my opinion, the code should be：

self.reg.set_flag(Flag::C, a >= n);

when the value of register A is greater then the number n, a-n is not borrowed. But in the project, the code is:

self.reg.set_flag(C, u16::from(a) < u16::from(n));

Which I couldn't understand.

cargo run --release -- "./res/boxes.gb"

Compiling coreaudio-sys v0.2.2
error: failed to run custom build command for `coreaudio-sys v0.2.2`

Caused by:
  process didn't exit successfully: `/Users/my/Projects/gameboy/target/release/build/coreaudio-sys-8d0fefacc2d2969d/build-script-build` (exit code: 101)
  --- stdout
  cargo:rustc-link-lib=framework=AudioUnit
  cargo:rustc-link-lib=framework=CoreAudio

  --- stderr
  error: header '/System/Library/Frameworks/CoreAudio.framework/Headers/CoreAudio.h' does not exist.
  thread 'main' panicked at 'unable to generate bindings: ()', /Users/my/.cargo/registry/src/github.com-1ecc6299db9ec823/coreaudio-sys-0.2.2/build.rs:144:10
  note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace