dirkwhoffmann / virtualc64 Goto Github PK

Some games manually trigger a lightpen interrupt. Implementation is missing yet

Some UI elements do not appear in the expected order.
When hitting TAB, the focus should switch to the next visible UI element 

Encountered an issue with text fields that are controlled by a formatter. The 
text field keeps the 
focus until a *valid* string is entered. Better: Invalid strings should be 
discarded. 

To do: Check all illegal opcodes against their addressing modes (read, 
read-modify-write or write)

build V64 from command line. (coding with a external editor vim, emacs)

Handling of user settings needs to be cleaned up.

Custom colors are not displayed correctly or unsafed
PAL/NTSC value is inconsistent during startup

Code that synchronizes the C64 thread is brittle

Idee: New API functions in VirtualComponent

setSpeedinMhz(float mhz); // negative value for "run as fast as possible"
float getSpeedinMhz;
synchronizeSpeed();

When overtaking another robot, the digital structure always looks the same. 

Image is captured in a wrong way (worked in Tiger, changes in Leopard?)

Warpmode is somehow intransparent, because both the user and the virtual 
machine can activate it.

Suggestion: 
Two internal variables: warpLoad, userWarp

warpMode = 1 iff (loading && warpLoad) || userWarp

Preference panel only affects warpLoad variable
Speed button only affects userWarp variable

Freezes while peeking CIA register dd00

When resizing the window, it can gets much smaller (something is wrong in the 
code that tries to 
maintain the aspect ration when resizing)

Step into, step over, etc. should only be pushable when the simulator is halted

screen geometry is broken

Problem: The current VC1541 simulation cannot simulate real disk changes.

In a real C64, the disk drive recognizes an ejected disk by continously polling
the write protection LED. To simulate this behaviour, we need to toggle the 
protection
LED for a certain amount of time (e.g., half a second).

Game freezes with a black screen

border is drawn too late.
24/25 row mode is probably checked in the wrong cycle

What steps will reproduce the problem?
1.
2.
3.

What is the expected output? What do you see instead?


Please use labels and text to provide additional information.

Fast resets should be more robust:

Idea:

fastReset()

- Check, if reset image is available
- If not, perform a hardReset and create reset image automatically.
- Otherwise quickload reset image

Listener callbacks should be moved to a seperate Objective-C file

Some games (e.g., Fort Apocalypse or QBert) exhibit bitmap errors in the upper 
screen area 

The PC should be user editable when the emulator is paused

There is a warpLoad variable in both the C64 class and the CPU class
One needs to be eliminated

Goal: Implement the GUI completely in Objective-C

Directory list should fill the entire window

The progress indicator should only be animated when disk data is transfered.

Fullscreen mode needs to be implemented

Joystick icon contains white non-transparant pixels. 

On a real C64, a disk can be formatted via

OPEN 1,8,15,"N:TEST,00"

To reproduce:

Enable warp mode -> Diasable warp mode

Symptom:

It takes more than a second until the C64 runs at its native speed again.

Tasks:

1. Create toolbar in interface builder
2. Cleanup toolbar code

GUI does not update values when changing the CIA chip in pause mode.

Credits.rtf Links to http://www.dirkwhoffmann.de/VirtualC64 which doesn't exist.

Change Link to http://www.dirkwhoffmann.de/software.html or configure webserver 
accordingly?

only a white screen is displayed. game seems to run (sound is playing etc.)

Static disturbing noise occures cyclical in sound output.

What steps will reproduce the problem?
-> Play any game (e.g. PACMAN) and listen attentive.

Tested with iMac (Alu, Mid 2007) and MacBook Air under Leopard. The older 
version from early 
2007 compiled for Tiger didn't have this issue.

The cyclic appearance of the noise may be an indication for some kind of 
overflow in the circular 
buffer storing the audio samples. Compiling for Tiger an testing again could 
exclude a (unlikely) 
Leopard-Issue

Loading a tape archive into the emulator allows the user to mount the archive 
(as opposed to 
flashing it). Unfortunately, mounting a tape archive most certainly causes a 
failed assertion along 
the way. The attached patch does four things:

* It prevents the emulator from crashing if the user attempts to mount a tape 
archive
* It removes the user's ability to mount a tape archive in the first place.
* Alert the user if the archive could not be mounted.
* Fix a minor typo in a comment.

Some callback functions cause an "autoreleased with no pool in place" error.
The functions need to create an own autorelease pool due to multithreading.

The runStop restore combination does not work yet.

NMI is already initiated, but it sill does not work.
I guess, the key combination needs to be held down for some microseconds (use 
NSTimer?)

Problem:

reset() changes some internal values without notfifying the listener

Solution:

- Register the listener inside the C64 constructor
- Send messages during reset()

marble madness requires exact cia timing to run.

Broken.

Probably a timing issue or a bug in the VIC implementation

On a real C64, the BA line can is used by the VIC to freeze the CPU.
Will replace the current hack with "dead cycles"

A new background image is created for each virtual machine.
Make it a static variable and create it only once.

Debug messages are currently displayed in a tableview with a somehow 
complicated update 
mechanism. Needs cleanup and should be handled by a standard textview 

code seems buggy. needs cleanup

Class SID uses API functions that are deprecated in Leopard.

Right now,  joystick actions are polled whenever the correspondig CIA memory 
location is peeked. 
This causes a slowdown in some games.
Polling needs to be replaced by a message based implementation.

Need to delete the info string when launching the emulator

In International karate, each fighter is assembled via three sprites.
There is a blank line between the sprites, most probably caused by a VIC.
Might be a simple bug in the Y-stretching of sprites.

The sanity check inT64Archive::getSizeOfItem() is always true. What's worse is 
that the assignment 
affects the method's return value, making it wrong in almost all cases. See 
attached patch for a fix.

This is probably just a typo but I don't feel comfortable enough to commit this.