Little update on LJX

[hansschmucker]

You can draw forms just fine in ARM, the only problem is that when you need access to a struct it has to be realigned and swapped, which is a bit problematic when you have a function that accepts a pointer to a struct that you're supposed to allocate, because structs are not necessarily the same size in 68k and in ARM, but aside from that it's quite forware.

Question: Do you also want to use standard menues or do you want to do it through buttons again?

[_Em]

32X and SCD would be quite an undertaking... currently the only code I know for these is written in IAX86 ASM.

[tgwaste]

hey tinnus,
does code warrior cost money? can you please tell me the exact version you use? are there any other tools required on the PC to compile LJP? Did I ask enough questions?

i wanted to play around with it a little just for grins.

thanks much!

[metaview]

CW costs money and even worse: it isn't available anymore
You need the latest 9.0 ? with all updates: 9.3

[tgwaste]

maybe one of you will send me yours?
how much is the key?

[vilmos]

CodeWarrior for PalmOS is still available here

[tgwaste]

$399.00!! wow.. LJX will be dev'd in a 'free' environment right?

[Tinnus]

Sorry, wrong. I'm just too used with the stuff we use to do the 68k/ARM interfacing. That's ONLY FOR PALM though! The common framework and the emulators are written in 100% portable code (heck, it works in PalmOS) so if you just want to fix/improve something that's not strictly Palm-related you'll probably be able to program and test it in the Windows/Linux/GP2X/whatever version. Moreover all other versions *will* be programmed and compiled in a free (gcc) environment.

Although... add some more quotes in 'free' and you might be able to work on it

BTW, I really am working on the Palm launcher side of things right now, ask Vilmos It's cool to do the launcher because I can test and debug it in the Simulator (PC).

[vilmos]

Anyone who is adamant that LJX be coded in a free environment would be welcome to convert it to gcc. It is completely possible but such a pain in the arse that I can't see it happening.

[Tinnus]

That's the pretty part of LJX. Anyone wanting to do a Palm-gcc backend would be able to do so without breaking compatibility with the main tree

In my current setup I can cross-compile (that means: I tried, but it's certianly more) for Windows, PalmOS and PSP. I'm pretty sure others like Linux, MacOS and GP2X would work out-of-the-box as well.

[_Em]

Just remember that there are some optimization incompatabilities between gcc and the Metrowerks compiler. I used to have a list of them sitting around. However, if it compiles on gcc for other targets, there shouldn't be an issue. Also, anything that compiles on gcc seems to compile OK on MWC; just not the other way around.

[Tinnus]

I'll get a Pocket PC version done if there's some free environment with emulator to develop for it. If not, someone with a Pocket PC would be able to

FYI, I'm currently working on the PalmOS launcher and the SMS/GG module. I'll get the launcher here for you folks to test the UI as soon as it's finished Actually, once the launcher is done, it's a short way from that to actually launching the emulator (once I get controls & sound working OK in the Palm-emulator framework).

The greatest part of all of this, is that the more you have, the more something you do will give you. For example, if I have the PalmOS and PC frameworks, and I develop SMS, that will give us SMS for Palm and PC. If I have several emulators implemented and I implement a new platform, that new platform will have all of the old emulators working right away (well, mostly).

[semi]

Tinnus, just curious, with the re-write and organized codes for LJX, how much percent do you think it's faster than the current LJP?

[Tinnus]

Just at start, something small since the framework is a lot more clean. In just a little more time, I will optimize the renderers/blitters since some systems draw the screen like 3 times when could just do it in 2 or even one pass. After that I will start to hand-optimize the code of the emulators themselves since I know a lot more about emulators now

That's for the second release or so, though. Remember that the primary goals of LJX are stability and extensibility.

[_Em]

...and portability.

[Tinnus]

That's under "extensibility"

[galaxygirl]

Which game systems will be emulated by LJX? I want to know:P!!

I kindly request that Videopac (Odyssey 2 & 3), Spectrum, Intellivision, and Vectrex be emulated on LJX. I also would kindly request that black&white support be emulated on any Atari 2600 emulators.

Lastly, has anybody else had problems with "World Heroes 2", "Xevious", and "Mach Rider" - NES - "Killer Instinct" and "Mr. Do" on Original Gameboy when using LJP??

all apologies if this post has offended anybody. I love LJP, LJX, and all the people who post!!!

[tgwaste]

galaxygirl, reading this may answer some of your questions:

yoyofr.proboards44.com/index.cgi?board=general&action=display&thread=1139617631

[galaxygirl]

OK! i would like to know how to code! please give me advice on how to create emulators for palm os. i will create my own emulators!! OOH LA LA!!

why not?? this is not a joke. give me advice on how to code so i can make emulators. i love all forms of videogames. your advice will bring honour to my heart. this is not a joke.

sweet kisses to all!!

[_Em]

Good to have new enthusiasm on-board!

If you want to write emulators for PalmOS, I have a few good places to start.

First, I'd suggest looking through the MAME source code for the platform of your choice -- a lot of stuff in there is well documented, and it will show you how all the different parts of a system can be emulated. Note: MAME code is a bit of a mess for actually WRITING an emulator, but it's a great way to DOCUMENT a machine architecture. You might want to check out the MESS sources as well.

After doing that, try out something simple on the platform of your choice, such as a gameboy emulator for Windows. Start from the available documentation, look at some open source gameby emulator code for inspiration if you need to (eg, gnuboy), and see if you can get something running.

Next, it's time to delve into the world of PalmOS coding... for this, you'll either need a copy of MetroWerks' CodeWarrior for PalmOS 5, or you'll need to figure out how to use one of the other development environments/compilers such as the gcc hack for PalmOS.

Write a simple "hello world" app for the Palm, and begin to figure out its various issues. Note that when you're writing an emulator for the PalmOS, you generally have to do an end-run around the OS-provided API and GUI. Write yourself an interface framework, and a little app to run inside it.

Next, you'll have to go into learning mode in order to get something playable: you'll need to become proficient at ARM assembly, as well as gain intimate knowledge of the hardware and assembly for the architecture you eventually want to emulate. Learn how to virtualize all the registers etc. in ARM code, and figure out a toolset of swapping techniques etc. you'll need to use over and over again to make your system run on the limited hardware found in the PalmOS.

Right, you're almost there Hopefully you're still raring to go, and haven't been sidetracked by some other project you found interesting while getting to this point.

From here, you'll want to start implementing all the documented routines from the emulated system. Do them one by one, and test that they work correctly. ROM-based consoles are good for this, as you can use samples of production ROM code to test out your various routines.

Once you've got them working, you'll need to put them all together, and see how it works. Due to glitches in the original hardware/firmware and undocumented "features", you'll probably find that things won't look exactly right at first, and you'll need to adjust some timings, register handling, etc. until everything works.

At this point, you should have something that can emulate at least one piece of software (the one you were testing with) for the emulated hardware. Now, try to make a few more software titles work -- best to choose ones that are likely to do things in a different way (choose something from a different software developer, published in a different year, that does something noticeably different on the system). Get this working too, and then try a few more titles. Get 5 or so working reasonably.

Now for the fun part You'll probably have noticed that while you now have 5 or so titles that at least sort of work (are usable), they're REALLY slow. Now that assembly knowledge you picked up above comes into play: you have to start optimizing your code, finding shortcuts where you can, skipping parts of the emulated process that it seems nobody actually implemented in their software, etc. By the time you're done, you should have a very usable, fairly accurate and stable emulator that works on the PalmOS!

If you decide to do this, please start a weblog documenting your adventure; it will not only be beneficial for the community, but it'll be useful for you as well -- you'll be able to go back and review what you have done, and use that knowledge to help do things better as you progress.

[tgwaste]

i cant believe you answered that Em..

[_Em]

I figure most people just don't know what it takes to make an emulator; education is a good thing.

Plus, I'd REALLY like to see a new batch of emulation coders; the pool of active emulation coders for all platforms has been dwindling in recent years, and anything we can do to reverse that is a good thing.

[tgwaste]

ya, gotta a agree with you there. i think most people just coded something and then said 'good enough' and stopped. tinnus at least is trying to break that trend. imho something can always be coded better.

[Tinnus]

Too bad I haven't had time to work on LJX the last month.

I *am* still alive and reading the board though ;D

[tgwaste]

but we just had a 3 day weekend.. youre telling me while we were all getting hammered you werent at home working on the project? SHAME!

[metaview]

Tsss, all these your people: drinking comes first

[varnil]

Right well, regardless of your work schedule we all still appreciate your efforts on the emulator, i do however have 2 questions. First, with the development of homebrew on the slim and light psp do you plan on converting your eboots into m33's compatible format? If not, it has been known that the 1.00 version of your SNES TYL v .3 works near flawlessly (apart from that version's problems) so i was wondering is there any explicit requirenment for the 1.5 kernel in the .4 non me version? and would that be possible to remedy? and given the inability to run the 1.5 kernel do you believe your new batch of emulators will still function properly?

[Tinnus]

Uh... this one board is for "Little John PalmOS". As for SnesPSP TYL, apart from being in another board (in any case, you probably are aware that this is not it), it's developed by yoyofr and Laxer3a, not me (assuming you directed the post at me, that is )

Anyway, about LJX for PSP, I'll make sure it works in my fat PSP, in the firmware I have at the time it's programmed. I'll try to leave it compatible with everything else but can't promise compatibility with what I cannot test with.

[jhoff80]

Stupid question, but are you planning on including the bluetooth gamepad support in LJX as well? If I knew it'd be supported in that, I'd be more motivated to hunt one down I think.

[metaview]

I'm sure LJX will include everything from LJP and more. If Tinnus should forget the BT gamepad, I will add it