Adding PPSSPP as ported emulation core for Sony PSP

[SergioMartin86]

Adding https://github.com/TASEmulators/ppsspp/tree/ported as new ported core in Bizhawk

• Core compiling and running on Bizhawk
• Add proper cd callbacks
• Correctly handle shutdown
• Add save/load
• Add inputs
• Add memory domains
• Load ini with compatibility flags
• Print UI messages to CoreComm.Notify
• Handle disc swap
• Handle thread create on save
• Preserve NVRAM
• Load language ini file
• Load atlas font file
• Add sound
• Find out why some games crash on load (whereas they don't for https://github.com/SergioMartin86/headlessPPSSPP)

Check if completed:

• I have run any relevant test suites
• I, the committer, have read the licensing terms for contributors (last updated 2024-06-22) and am compliant

[SergioMartin86]

Second thing to note as above is related. Some titles can crash out in BizHawk. Example title being used is Jetpack Joyride (minis) where the game will be able to boot

This should be fixed now.

[SergioMartin86]

The core is now mostly functional and I reduced the diff wrt upstream the most I could.

Still, there seems to be some games (Crash Tag Team Racing (US).iso) that fail to load. These games DO work on my headless playerhttps://github.com/SergioMartin86/headlessPPSSPP, so no idea why putting this into bizhawk could be a problem.

I haven't figured out how NVRAM is saved so that is still pending.

[vadosnaprimer]

Which branch?

[SergioMartin86]

ported2

[vadosnaprimer]

It's better to always keep hawk specific changes rebased on top of upstream branch that we rely on. And each time we pull upstream and rebase anew we would increment our branch version like this: wbx-1, wbx-2, etc, so every rebase is a new branch, because it basically changes history, which we want to preserve it on every update. This is better than merging upsream changes because that makes our changes harder to find over time.

And yeah please explicitly put the branch into this file, that way it's clear where the code is coming from. Only pointing to commit can result in detached head and that commit can belong to different branches, making future maintenance more annoying.

[SergioMartin86]

added

[vadosnaprimer]

I can see the wbx branch in the repo now but not in .gitmodules...

[SergioMartin86]

Uh, this is strange. This core isn't wbxed; it's ported. This every must be a residual from when I tried to wbx it

[vadosnaprimer]

Ah LOL indeed. Then the usual name for it is bizhawk.

[Spikestuff]

I truncated out the issues from my above post.
Just read the History for what's been fixed (good shit eien).

In terms of new issues (by the autobuilds here). Using f1a8d12.
The major one besies the wrong font is being called compared to PPSSPP for some reason.
Is that the "™" is missing from "Memory Stick™":

Compared to PPSSPP:

And according to KAGE-008 at the very least this font issue affects Ghost in the Shell, Assassin's Creed Bloodlines, Crimson Room: Reverse, Jeanne D'Arc and Dissidia.

I was able to validate this missing text issue with Crimson Room: Reverse.

Another title having issues beside Crash Tag Team Racing is Xyanide Resurrection according to KAGE-008 as well.
You're able to boot into the game, but you land into a loading loop before the gameplay starts.

And I just noticed that there's "Right Analog" controls.
PSP only had a single analog stick.

And you're currently unable to access PSP > Settings.

(At least all these are issues with autobuilds)

[YoshiRulz]

Missing glyphs aside, could the difference in font metrics cause any problems? Would a scrollbar appear if the message was slightly longer?

[Spikestuff]

I don't believe it would be a situation where for example you're missing the Japanese Locale and the entire game would run differently.
But I also haven't run into a situation where it might have happened to me.

[SergioMartin86]

Regarding unstability for some games. I suspect there is some memory corruption that causes issues on Bizhawk+Windows. When running the core on my own driver (https://github.com/SergioMartin86/headlessPPSSPP), it works perfectly for the games that crash

[InfamousKnight]

I'm curios why this is not waterboxed. I know waterbox would affect performance, but I don't like tasing while worrying about desyncs.

[SergioMartin86]

I'm curios why this is not waterboxed. I know waterbox would affect performance, but I don't like tasing while worrying about desyncs.

The interpreter crashes when waterboxed. We don't know exactly why, but for now a waterboxed implementation is out of the question. The core seems to have solid load/save state procedures so I'm confident it can still be used ported directly.