[florianessl]

Support for the "Compact" version of Cherry's PowerPatch features. This version wasn't really a patch but just an external utility program named "ppcomp.exe" (sometimes renamed to ppcomp.dat or ppcomp.dll) which provided access to some new functions by accessing the memory of the running game process.

Detecting the original PowerPatch to give info about compatibility

This is not the full version of the original "Power Patch", which goes way deeper (including early Lua support) & barely found use in any games. But I also added a small heuristic check to detect the "Full" version & log a warning about limited compatibility:

• This just check for the existence of a file named "powerp.oex" in the game's root
• Another commit also checks for "warp.dll" which was bundled with PowerMode2003, another popular patch at this particular time period of RPG2003's lifecycle. (Reversing & documenting the origins/versions of that patch actually seems quite tricky - more about that in a future issue...)

Supported PPCOMP functions

• Quitting & Resetting the game ("QUIT", "RESTART")
• Calling of in-game scenes
  • "DEBUG"
  • "CALLLOADMENU"
  • "CALLSAVEMENU"
  • "CALLGAMEMENU" (+ setting the cursor position)
  • "CALLTITLESCREEN" (+setting the cursor position)
• Enabling / Disabling the "UnlockPictures" patch ("UNLOCKPICTURES")
• Load / Save operations
  • Loading from source [slot-id] ("LOAD")
  • Saving to target [slot-id] ("SAVE")
  • Checking existence of Save [slot id] ("CHECKSAVE")
  • Copying individual save files ("COPYSAVE")
  • Deleting individual save files ("DELETESAVE") (Would require an update for FileFinder -> Support File deletion through VFS  #3266 )
• GetSaveDateTime
• GetSystemDateTime
• SimulateKeyPress (Drafted, but not tested yet - allows for simulating keypresses for a targeted amount of time)
• ChangeFunctionKey
• SetMsgBoxColor
• PauseGame
• PauseTimer
• SetVar

NOT PLANNED:

• SetGlobalBrightness (Apparently this was only implemented to address a blackscreen issue caused by the original patch)
• ChangeScene (Sounds a bit too dangerous)

For the remaining features, I only found instances of "DELETESAVE" & "SIMULATEKEYPRESS" actually being used in some archived games.
-> These are still TODO

Adjustments to Ineluki KeyPatch implementation

• I refactored handling for script execution, so that the encountered script.wav files aren't solely handled in the audio mechanism inside Game_System, but will be handled inside the interpreter, when a PlaySE command comes across them there. This is, so that it is possible to get some context for the result of executed scripts & to properly set the field "_async_op" in some cases.
• Async requests: While a script file isn't yet available, the PlaySE command will result in a "YieldRepeat" async op.
• Added support for script.wav execution for the "PlayBGM" command. This is rarely seen in games, but is still possible & would have resulted in a warning by the Player when it tried to parse a script INI as Music.
• I also added support for the known program "LS.dat" which was another way many games with custom title screens used to call the Load Scene.

BetterAEP patch support

While testing this implementation, it quickly became obvious, that many games that were using ppcomp, also relied on BetterAEP, and some used this patches' special functionality to access the Load Scene or even exiting the game via command. (Even though ppcomp also had these options).

So I added support for this patch as well:
Player.exe --patch-better-aep 3350 So far I've only seen games that use the default var-id of '3350' for this.
-> Issue #1182

Some games using ppcomp

• La Leyenda Heroen (https://rmarchiv.de/games/2899)
• Mr. Hat (https://rmarchiv.de/games/183)
• Take It Cheesy (https://rmarchiv.de/games/1355)

Info

• BetterAEP: https://www.makerpendium.de/index.php/Better_AEP
• Power Patch Compact: https://www.makerpendium.de/index.php?title=Power_Patch_Compact (See PDF download for documentation)
• More info about some obscure ppcomp versions was shared by @CherryDT in this issue: Ineluki Key Patch #1181

[Ghabry]

The keystate stuff is already in maniac patch https://github.com/EasyRPG/Player/blob/master/src/maniac_patch.cpp#L674

So this could be unified in the file you proposed

[florianessl]

The keystate stuff is already in maniac patch https://github.com/EasyRPG/Player/blob/master/src/maniac_patch.cpp#L674

So this could be unified in the file you proposed

Yeah, I already moved that mapping code as part of another branch that's still WIP:
adb3f37
I included a small "FIXME" note in my last PPCOMP commit.

[florianessl]

99c559d

Needs some explanation:
While testing around with the ppcomp debug utility, I first came to the conclusion, that this command is meant to always force the Title Scene, even if something like BetterAEP (also made by Cherry) is used.
At least, I had managed this once in RPG_RT & then promptly added this new force flag to the Player, to ignore the "new-game" option set by other patches/configurations.

But then I couldn't reproduce this at all later. Using this command in a BetterAEP-patched RPG_RT seems to always skip the title screen.

Didn't mean to commit this line at first, but the flag could still be a useful feature.

[Ghabry]

hm that the Ineluki script ran async is another bug. Makes no sense to execute a script this way. Though there are almost no games using Ineluki in the web player so was never noticed 🤔

[florianessl]

hm that the Ineluki script ran async is another bug. Makes no sense to execute a script this way. Though there are almost no games using Ineluki in the web player so was never noticed 🤔

KeyPatch scripts actually do have a setting for "Wait" that is ignored by EasyRPG but I never checked the details of that one out for myself. So this could maybe lead to behavior changes in some instances :/

[giovannetti-eric]

In case a real user of BetterAEP is useful while this is in progress: Final Fantasy VII Origin (French RPG Maker 2003 fan game, RPG_RT 1.0.8.0) depends on it for two things, and both fail silently today.

It drives the patch entirely through the variable, never through Cherry's event commands — codes 5001–5005 appear nowhere in the project. Three writes to variable 3350, each followed by End Event Processing:

where	value	intended	today
title screen, parallel event	1	open the load menu	cursor does nothing
in-game custom menu	2	exit the game	black screen, quit music, never exits
title screen, "new game" branch	0	nothing — a plain teleport follows	works

The third is the useful one for testing: it is the same event and the same code path, and it behaves correctly, so a fix can be checked against a control that already passes.

Screenshots of the title screen under the current continuous build are attached, with the cursor on each of the two entries. Everything drawn works — the cursor moves, the artwork responds — so the gap is purely the action behind the entry, which is what makes it a clean test for this branch.

Happy to test a build against it — the game is not redistributable, but I can reach either behaviour in seconds and report back.

(Unrelated to this patch, I opened #3687 about the same game; mentioning it only so the two are not taken for the same report.)