Add Maniac Patch Event commands

[Ghabry]

SetMousePositon surprisingly doesn't add a new chunk to the savegame, which means the offset is not stored.

Not fully sure what "EndLoading" is supposed to do, doesn't match with the documentation. waiting for a response on twitter.

[carstene1ns]

I think this should involve a generator change that applies some "Maniac Patch" comment here.

[Ghabry]

jadehpy wrote me that End Loading reduces the blackscreen after load from 6 frames to 1 frame but the blackscreen here is around 0.5s long, so I can't really notice this 5 frame difference o.O.

[fmatthew5876]

I agree with carstene1ns, these need to clearly be marked as non-standard in some way.

Also #272

[Ghabry]

I agree with your objections.

Note that this patch fulfills the "doesn't conflict with normal games" policy, otherwise I wouldn't even consider adding it.

[fmatthew5876]

Maybe even a separate fields_maniac.csv or something like that?

[Ghabry]

I find these commands so useful that I would even consider them for the EasyRPG Editor.
Which spawns the discussion where to put (potentially) new Event commands invented by us.

In this case enums_maniac.csv (no new fields) and *_easyrpg.csv feels more future proof.

[fmatthew5876]

*_easyrpg.csv makes sense to me for custom stuff.

In the long term, we need to decide on some uber-format to support all patches, but there is no reason we can't use exactly the maniac convention to start with. If some other patch implements the same thing we'll just have to translate it at the liblcf level to how maniac does it.

[fmatthew5876]

For EventCommand::Save, what is the value of SaveSystem::screen in some of those maniac saved games?

Normally in RPG_RT, all saving goes through the save menu, and therefore in save games screen == 5 (filemenu) always.

If this command lets you save anywhere, I'd bet screen == 0 (map). Can you save in battles?

[Ghabry]

When saving in Battles the screen is "3" and loading them puts you on the map. You can even save in BattleTests (but you can't load it).
Loading these saves corrupts something, RPG_RT crashes when closing.

Edit: Moving it to *_maniac.csv is surprisingly difficult because the generator needs many updates o.O

[Ghabry]

new solution. opinion?

[fmatthew5876]

My vote is for enums_easyrpg.csv instead of having separate files for different patches.

In other words, we have base RPG_RT compatibility layer, and then our easyrpg customization layer which also incorporates support for patches (with data transforms if needed).

[Ghabry]

I'm okay with this solution 👍

[fmatthew5876]

Apologize for being so nitpicky, but if they are easyrpg_enums.csv then I think the enum tags themself should also be easyrpg and not Maniac_.

How about an EZ_ prefix? Like EZ_Save? Or EasyRPG_Save if you want to be longwinded.

[Ghabry]

Well the problem is that these event commands are specific to Maniac Patch and we decided to only have one file.

I would be fine with "EZ_", doesn't really matter for me.
Waiting for @fdelapena and @carstene1ns favourite, then I will choose the winner. xD

[fmatthew5876]

To give more context of where I am coming from, I'm thinking to make a PR add these custom equipment settings. This is why I made equipment_setting an enum and not a bool.

EZ_EquipmentSetting_either
EZ_EquipmentSetting_both

(where EZ can be whatever convention we decide)

That would be a pure EasyRPG extension.

Yours are specific to Maniac but I see them as EasyRPG extensions first, which also happen to be compatible with maniac patch.

[fdelapena]

I agree once EasyRPG Editor adopts this feature as own, keeping the patch name prefix may not be worth, adding a comment or documentation somewhere about the origin and design decision of those numeric IDs is still useful, though.

[carstene1ns]

approved now, can merge after rebase :P

[Ghabry]

soooooo, what is the result? Do you want Maniac, EZ or Easy-prefix now?

[carstene1ns]

Since all of these additions are good ideas to support later, should be okay to use our prefix, but may be a good idea to clarify the origin with the current maniac prefix.

[fmatthew5876]

Why don't we have 2 conceptual name spaces?

EasyRPG and Maniac.

Each custom command that's compatible with a patch shall have an alias in that patch's unique namespace. It also needs a id in the EasyRPG namespace in order to be compatible with our tools.

The right place to store information about patch lineage is in the code itself. This will also be the foundation of parsing logic needed later to convert incompatible map patch customization to EasyRPG.

[fmatthew5876]

Regarding this save command. How does it work? Does it save immediately or does it wait till the end of the frame and save at the same time opensavemenu would be called?

If it's the former then I don't support adding this command. It's useful for debugging and research for sure, but there are many local variables mid frame which don't have LSD chunks.

One that comes to mind is the loop counter in the interpreter that goes up to 10000.

The result will be heisenbugs with broken saves done at the wrong time. They would so rare and hard to track that fixing them would be next to impossible. I would never use such a feature in a game as practically it would just result in game instability and corrupted save games.

I'm not convinced the maniac author has thought this through carefully.

[Ghabry]

That's a good point, unfortunately there is no event tracer for 2k3 Steam due to lack of DynRPG to figure this out easily.
I would expect an interpreter yield, otherwise this would be really a bad design choice, will try to figure this out.

[Ghabry]

While testing this I directly found a bug with the normal scene calling:

Parallel Process 1:

Loop
 V1 += 1
 if V1 > 1000
  Save
 endif
Repeat

PP2:

Message: V[1]

A blank non-transparent message box appears right before it fades to the save scene and after closing the scene it draws the text. The next boxes are all correct. Must be a problem with initialisation and fade, is fixed by a Wait 0.0.

@fmatthew5876 You were right.
I simply use the "execution limit" of Parallel Processes to test this: I save and look at the message box after loading:
The output is 3000 (Maniac) vs 5000 (Scene) which means it doesn't use the same way because the output is different.

I still like this command but we can't support it as fully intended (but was likely just not properly thought about a said)

[fmatthew5876]

Here's another,

Parallel events are allowed to continue running while a message box is up. So you could save your game while a message box is active.

I'd like to see what happens when you try to load that save..

[Ghabry]

Finally rebased it and added all commands up to version 190920