etxmato / emma_02 Goto Github PK

I found a bit confusing that instead of the clock setting being the clock speed of the processor. It's the clock speed before it's dived by 2.
Most of the other systems in the emulator are the clock speed of the processor. Which really makes more sense as no one is thinking about a pixel clock in an emulator.

When I click the "VideoSetup" button in the "RCA" tab, Emma 02 crashes with the following error:

This is Emma 02 V1.32 running on Windows 10. All the other VideoSetup dialogs work as expected.

on Linux: COMX-35 emulation misses a lot of my keystrokes if I type even at moderate speed (reported by Mark)

• This doesn't happen on all occasions
• Also noticed on windows (on work laptop) that it hangs for about 30 seconds
• Possibly caused by the update check; maybe internet access is blocked by 'something', firewall or other?
• Maybe this should be 'default' off

As soon as the 1802 trace is started in the debugger Emma 02 crashes with a segmentation fault

At least not the default values, when re-selecting it works ok.
Only on Linux

• Not happening all the time
• Could this be caused by some leaks?
• After debugging on windows I know there is at least one leak maybe more, possibly linked to Main::setConfigurationMenu, Main::buildConfigMenu, Main::setChangeNoteBookEvent or Main::onElfChoiceBook

i just started cosmac vip with the default config and then this hapenned. this is my emma02 window:

(emma_02:15558): Pango-CRITICAL **: pango_layout_get_cursor_pos: assertion 'index >= 0 && index <= layout->length' failed

No clue what this is as yet, reported by Mark

Instead it should be in ~/emma_02 on Linux

On windows maybe it should stay as is - to be looked into
On OS X either the same as on Linux or maybe emma_02.ini should be in the application/emma_02 folder

First thanks for the emulator! It is nice that there is one that takes care of so many exotic systems whit this nice CPU. :-)

While testing VP590 support in my own emulator and comparing behavior of CHIP-8X I found that the Bxyn opcode is not working correctly. Debugging the issue showed that the interpreter file in data/VIP/chip8x.ram has a bug, I guess it is based on OCR of a hex-dump and somehow the byte at offset 0x228 in the file is 0x9c but should be 0x8c, as clearly readable in the VP580, VP585, VP590, VP595 Instruction Manual page II-3.
With that byte changed Bxyn works as described.

I mean there ain't no 1802 emulators on Android yet.

On linux only

I've tried a few tools to convert the DEB and RPM files to run on Manjaro, but with no luck. Would you be able to package Emma_02 up in a compatible format, or release it as Flatpak or AppImage? I've even tried the Windows version on my Steam Deck but that doesn't run either.

Does anyone else have the knowledge to do this?

When using Memory Access in the comx-35, it should change the EOP to point at the end of the loaded space when SpecifyLocation is check.

EOP is commend in basic and returns the end of program. It's also used by PSAVE to tell it when to end.

The EOP value is located in memory at 0x4283 high byte and 0x4284 low byte.

Just so I don't forget:

• Check why default window size doesn't open large enough, possible workaround to save size of window and re-open same size
• Set size of wxChoiceBooks to size window
• Update help to use new GUI layout description for: COMX Plot / STD button, Elf Disk vs. File button for IDE.

Seen by Mark, I haven't seen it for a while on Fedora, needs to be debugged

I'm experimenting a bit with how the random numbers are generated on the superchip vm, and...

Seem the numbers generated there do following some patterns and it's pretty short and pretty deterministic. Doesn't look good...

Expected results are where the Hex mixt and Framed MK1 fills the whole screen:

I'm proposing to make use of LSFR [linear shift feedback register] which I think it's good for being simplistic and having more randomness to it. I consider 16-bit LSFR is good, and should be shifted 8 times when fetching one byte from it. (still having period of 65535)

pseudocode

seed = 12345 # 16-bit initial feedback value, could be any number

def shift(seed):
 return ((((seed>>1)^seed)<<15)|(seed>>1))&0xFFFF

def shiftByte(seed):
 return shift(shift(shift(shift(shift(shift(shift(shift(seed)))))))) # shift 8 times

# print random bytes 10 times
for n in range(10):
 seed = shiftByte(seed)
 print(seed&255)

One may shift it to only few to save cycles, or by making it perform shifts for every chip8 instr executed. It depends on implementation.