Repository navigation
battle: decompile the hit and particle effects - #268
Conversation
| extern u16 D_80163B80; | ||
| extern void (*D_80163B84[60])(void); | ||
| extern DR_MODE* D_80163C74; // TODO might be a generic u_long*, not DR_MODE* | ||
| extern void* D_80163C74; |
There was a problem hiding this comment.
This to me looks like the Gpu* type from main.h which encompass the various polygon types.
There was a problem hiding this comment.
This type is returned by several functions (func_800D29D4, func_800D4D90, BattleEffectSpriteAdd) and each seems to return a different pointer (DR_MODE, POLY_FT4, 0x24-byte model packets...), so I think we can already tell it's not DR_MODE. I also tried Gpu but it didn't match (battle.exe and choco0.exe break). So I thought leaving it void* for now, at least until we know more.
There was a problem hiding this comment.
Fair. Let's not have this PR blocked then, but we need not to forget this detail.
| otz = RotAverage4(&D_800F1914, &D_800F191C, &D_800F1924, &D_800F192C, (long*)&quad->x0, (long*)&quad->x1, | ||
| (long*)&quad->x2, (long*)&quad->x3, &p, &flag); | ||
| if (otz > 0) { | ||
| quad->tag = 0x09000000; |
There was a problem hiding this comment.
tricky low-level optimization.
| quad->tag = 0x09000000; | |
| #ifdef PLATFORM_PSYZ | |
| setlen(quad, 9); | |
| #else | |
| quad->tag = 0x09000000; | |
| #endif |
this is required, because PsyZ splits a tag into addr and len.
| *(u32*)&quad->r0 = 0x2E202020; | ||
| } | ||
| quad->clut = 0x78C7; | ||
| quad->tpage = 0x3A; |
There was a problem hiding this comment.
Can you please decode this as a getTPage?
There was a problem hiding this comment.
It didn't match with the current getTPage, because it calls GetGraphType(). I tried the version from before #147 and that one matches, but it breaks field and world, which need the current one. So I'm not sure what to do here.
There was a problem hiding this comment.
Thanks for looking into it. I forgot about GetGraphType, but it still doesn't sound right. I'll look into the macro shortly.
| frame = slot->frame + 1; | ||
| slot->frame = frame; | ||
| if ((s16)frame == 16) { |
There was a problem hiding this comment.
| frame = slot->frame + 1; | |
| slot->frame = frame; | |
| if ((s16)frame == 16) { | |
| if (++slot->frame == 16) { |
doesn't this match? it would also allow you to remove the local var frame
| u8 anim; | ||
|
|
||
| p = (BattleKeyframeParticleSlot*)&D_801621F0[D_801590D4]; | ||
| desc = (SpriteRenderDesc*)0x1F800000; |
There was a problem hiding this comment.
| desc = (SpriteRenderDesc*)0x1F800000; | |
| desc = (SpriteRenderDesc*)getScratchAddr(0); |
| m = &D_800F10B8; | ||
| desc->frames = D_800F0B14[anim]; | ||
| } else { | ||
| m = (MATRIX*)0x1F80000C; |
There was a problem hiding this comment.
| m = (MATRIX*)0x1F80000C; | |
| m = (MATRIX*)getScratchAddr(3); |
this will likely crash on 64-bit builds, but let's keep it this way for now.
Use getScratchAddr for the scratchpad pointers, setlen under PLATFORM_PSYZ since psyz splits the tag into addr and len, and drop the temporary in BattleHitFlashGrowTick. Move the getScratchAddr fallback for psyz into game.h so battle and world share it.
2 participants
Continuing from #261, this decompiles the battle functions that spawn and animate small visual effects, like hit flashes, sparkles and trails.
It also names the matrix helpers they share and the world view matrix behind them. Two existing names didn't match what the functions do, so I renamed them. That's why most of the magic overlays are touched.