LMS

Interesting findings! I noticed that the internal animal control task had changed, resulting in the changes in control scheme you mentioned with the horse controls now working for animals. This is actually more a fix by Rockstar since they are now assigning the correct control scheme for animals when playing as such. Beforehand you had the normal on foot player control scheme which is why many more actions such as attacking or crouching were available. It also means birds lost the ability to fly. Luckily all of this can be fixed easily, but it remains interesting to see what R* has in store.

It should work also without that being set. Are you sure you are on the latest version?

The gypsy cobs definitely work, I tested it myself. Try one of these models: A_C_HORSE_GYPSYCOB_PALOMINOBLAGDON, A_C_HORSE_GYPSYCOB_PIEBALD, A_C_HORSE_GYPSYCOB_SKEWBALD, A_C_HORSE_GYPSYCOB_SPLASHEDBAY, A_C_HORSE_GYPSYCOB_SPLASHEDPIEBALD, A_C_HORSE_GYPSYCOB_WHITEBLAGDON Also make sure you are on the latest OutfitChanger version in order for them to show!

I suppose it really depends on how it is done. It could be that R* manually sets up the scene or that there are some scenarios to switch. You could try finding the associated mission script and seeing if you can figure out if everything is done manually (which should be fairly obvious).

Horses, Animals and models should work just fine. Just items and weapons will need my MP weapons fix which, as stated above, can glitch out your game.

You would need to pass the reference, so &s, but you tried that before too iirc, so probably also not going to work. And no problem.

Good catch, I had renamed the variables from Unknown to what I think they are for the sake of this post, but forgot to change the assignment name below. In the actual code it is using the fourth parameter, though. I have fixed the snippet to reflect the code executed. Why it still does not work for you is beyond me, though.

I am genuinely confused as to why it is not working for you - it all looks good to me. I am doing it in C# so the syntax will be different, but this is my code to test (normally I'd just use Game.LocalPlayer.Character.PlayAmbientSpeech obviously): [StructLayout(LayoutKind.Explicit, Pack = 1, Size = 96)] internal unsafe struct CSpeechStruct { [FieldOffset(0)] public sbyte* SpeechName; [FieldOffset(8)] public sbyte* VoiceName; [FieldOffset(16)] public uint UnknownUInt1; [FieldOffset(24)] public uint SpeechModifier; [FieldOffset(32)] public uint UnknownUInt2; [FieldOffset(40)] public bool UnknownBool1; [FieldOffset(48)] public int UnknownBool2; [FieldOffset(56)] public uint UnknownUInt3; } [ConsoleCommand] private static unsafe void Command_PlaySpeech() { CSpeechStruct s = new CSpeechStruct(); IntPtr speechNameAddress = Marshal.StringToHGlobalAnsi("GENERIC_THANKS"); s.SpeechName = (sbyte*)speechNameAddress; s.SpeechModifier = 0xE7176FB2; Game.CallNative<uint>(0x8E04FEDD28D42462, Game.LocalPlayer.Character, (IntPtr)(&s)); Marshal.FreeHGlobal(speechNameAddress); }

All the other parameters I left untouched, i.e. their default values (0 and false). The hash is some speech modifier, try it with that one.

I changed my model to U_M_M_ValSheriff_01 and only filled the first parameter (speech name) and the fourth: GENERIC_THANKS and 0xE7176FB2 respectively. The rest I left as their default values. It played just fine for me when calling the native with player ped and the params. RPH mods do not use C++, but C# so it is definitely quite some work to port it.

Yes, you need the latest version.

Unfortunately I have no idea why it is not working for you. That's all I am doing (but from RPH). I will test you code tomorrow in my game as well and see if I can narrow it down.

Not sure how boost::any works internally, but I'd just use "raw" memory for it. It seems someone also defined a type for it though, so you can try this: struct ScriptedSpeechParams { const char* speechName; const char* voiceName; alignas(8) int v3; alignas(8) Hash speechParamHash; alignas(8) Entity entity; alignas(8) BOOL v6; alignas(8) int v7; alignas(8) int v8; }; static_assert(sizeof(ScriptedSpeechParams) == 0x40, "incorrect ScriptedSpeechParams size"); Example: ScriptedSpeechParams params{"RE_PH_RHD_V3_AGGRO", "0405_U_M_M_RhdSheriff_01", 1, 0x67F3AB43, 0, true, 1, 1}; _PLAY_AMBIENT_SPEECH1(PLAYER_PED_ID(), (Any*)&params);

I believe those are pulled from data files and not set via natives. There might be a way to override them, though.

From game memory, if you are not familiar with what that means I'd recommend sticking to normal mod development first as it's a bit of its own world. And no problem 🙂

I read all currently loaded doors from memory, then get their positions and compare them to the player location. I don't think you can get the hash via natives (at least I don't know how).

You can find all the default information in update/common/data/relationships.meta including all groups and their relationships. Since you set the relationship group per ped, it is definitely worth trying to work with it. We did it for RDRFR and it seems fine for the most part.

The first is the ped, the second is the container. And there is no "type" for this, you can define your own struct or just use new memory and pass a pointer to that.

The hashes for the two doors (they are single doors) are: 55C8DDE4 (right) 5F12F078 (left) Example when setting the state to (in doors view, so reversed order):