Pocket Heaven

[thac0nfus3done]

I was looking through my roms the other day, getting the proper ones and I was deciding whether or not to keep the PD demos (Public Domain i think) Most of them are just demos created by the various homebrew organizations back in the day. The groups would challenge other groups by developing demos that would push the systems to more extreme limits.

This got me to thinking, since these roms are small easy to sort through, and use a variety of different techniques to do things on roms, could we use them to easily find slowdowns/bugs in the SnezziDS code? I'm hoping yes. If anyone would like a rar of the PD roms, I will gladly upload them.

PS. PD roms are not illegal to distribute as they have no copyright status and were never meant to cost money.

[tepples]

Not all the "PD" roms are public domain. Many are copyrighted yet still freeware, just like any DS or GBA homebrew rom.

A lot of freeware roms were tested only on emulators during development and thus don't work properly on flash cards. This was a big problem in the first few years of the NES homebrew scene, when if a .nes file ran in Nesticle, it was deemed worthy of release, even if its content was completely incompatible with the NES hardware. The point is that ideally, we want to test the behavior of freeware roms on emulators against their behavior on a Super NES, not against their behavior on other emulators. So which roms have been tested on a Super NES game copier and which haven't?
_________________
-- Where is he?
-- Who?
-- You know, the human.
-- I think he moved to Tilwick.

[thac0nfus3done]

What does PD stand for then sir Tepples .

[Dwedit]

It's a misname, the stuff isn't public domain, but the ignorant folks who name the roms think they are.

And I think the many of the SNES demos predate emulators, and were intended to run on copiers.
_________________
Here come the fortune cookies! Here come the fortune cookies!
They're wearing paper hats!

[tepples]

Background: In the 1980s, some bulletin board systems had warez and others didn't. Some BBSs allowed only "public domain" software due to copyright concerns, and this caused little problem because any work first published in the United States without a copyright notice entered the public domain under the laws in effect at that time. In 1989, the US Congress amended the copyright statutes to grant an automatic copyright for any work first published in the United States, so some software authors just licensed the copyrights in their programs to allow free redistribution ("freeware") or even free modification and redistribution ("free software"). The boards that had once held only PD software began to fill up with copyrighted freeware, and users began to misunderstand what "PD" means.

So the (PD) in GoodCodes appears to stands for nothing in the same way that the (8) stands for nothing. Perhaps "permitted distribution" would be the backronym.
_________________
-- Where is he?
-- Who?
-- You know, the human.
-- I think he moved to Tilwick.

[bubble2k]

Back to the original question, using any form of ROMs to discover slowdown or bugs in SnezziDS is possible, that's what we are already doing actually. The only difference is that we use production games.

But why freely distributable (that's what I assume you meant) ROMs in particular? Because they are small, it's likely that the instructions and operations they execute is likely to be much less exhaustive, if not used differently, than full-blown games.

Actually the best way is to write a test SNES ROM in assembly to test all running instructions in all forms of boundary cases that one can think of. Compare its output to something that is known to be fairly accurate. I used to run this test ROM in the early days of Snezziboy and compared it with the debugger output from Snes9x line-for-line, instruction-for-instruction to see that they match. If they don't match, then some CPU opcode emulation must be buggy and I correct the problem. These days I use it less and less, because there are less bugs now occurring on the CPU core, and more on timings, interrrupt handlings, graphics, etc. But if I hadn't had this help from the output of Snes9x, Snezzi would still probably extremely buggy from the emulation side of things.

A test ROM to test CPU core is not as difficult as it is tedious. But what is difficult to test with a test ROM are the finer intricacies of the hardware like timings issues, handling of DMA, and hardware bugs (yes!), that if not properly emulated can cause many games to run into problems. Byuu (author of bsnes) has written a set of test ROMs that tests some of these intricacies; they run correctly on hardware, and they are meant to test if the emulator can run them correctly. But these test ROMs may not be sufficiently exhaustive, at least to ensure all games will run properly.

So the best way, at least I feel, is to test it with full games. Unless some kind soul was willing to write an exhaustive test SNES ROM to test all fallable scenarios.

[thac0nfus3done]

Excelent, that now set aside, I'll delete most of these PD roms. That draws this topic to a close. Back to testing all the english roms I can.

[tepples]

[bubble2k]