Rendermode internal simplification

[sauraen]

Like some of my other PRs, this started as a simple fix--make the fog controls show up under Sources if fog color or fog alpha are being used in the rendermode, even if fog is not enabled in the geometry mode. This turned into three main changes:

• fast64 now sets all the individual options for the advanced rendermode (while they're hidden) based on which rendermode presets are selected. This conversion happens at .blend file load and when changing the presets. This does not affect export (the strings for the presets are still written if the presets are used).
• Code dependent on the rendermode now is significantly simplified, because it only has to check the individual advanced rendermode settings, rather than have one codepath for that and another to extract the relevant bits from the selected presets.
• A GUI warning is added if the selected rendermode presets are not valid (exactly one must set the rendermode flags). The warnings about RDP silicon bugs are moved from being shown for advanced mode only, to also be shown when the presets are in use. The combination of these warnings will cover most cases of guiding users towards setting up the blender correctly.

[Yanis002]

I still need to try this but code looks ok to me, I don't have a lot to say about it

[Dragorn421]

Good changes!

While checking the material presets to see if they're consistent with the warnings you added, I noticed

oot_shaded_multitexture_lerp_transparent_vertex_alpha
f3d_mat.rdp_settings.rendermode_preset_cycle_1 = 'G_RM_OPA_SURF'
f3d_mat.rdp_settings.rendermode_preset_cycle_2 = 'G_RM_AA_ZB_XLU_SURF2'

oot_shaded_texture_transparent_vertex_alpha
f3d_mat.rdp_settings.rendermode_preset_cycle_1 = 'G_RM_OPA_SURF'
f3d_mat.rdp_settings.rendermode_preset_cycle_2 = 'G_RM_AA_ZB_XLU_SURF2'

oot_shaded_vcol_multitexture_lerp_transparent_vertex_alpha
f3d_mat.rdp_settings.rendermode_preset_cycle_1 = 'G_RM_OPA_SURF'
f3d_mat.rdp_settings.rendermode_preset_cycle_2 = 'G_RM_AA_ZB_XLU_SURF2'

oot_shaded_vcol_texture_transparent_vertex_alpha
f3d_mat.rdp_settings.rendermode_preset_cycle_1 = 'G_RM_OPA_SURF'
f3d_mat.rdp_settings.rendermode_preset_cycle_2 = 'G_RM_AA_ZB_XLU_SURF2'

And unrelated but this is weird as well? noop makes clr_in 0 for the next cycle

oot_water_lerp_specular_fresnel
f3d_mat.rdp_settings.rendermode_preset_cycle_1 = 'G_RM_NOOP'
f3d_mat.rdp_settings.rendermode_preset_cycle_2 = 'G_RM_AA_ZB_XLU_SURF2'

oot_water_mult_specular_fresnel
f3d_mat.rdp_settings.rendermode_preset_cycle_1 = 'G_RM_NOOP'
f3d_mat.rdp_settings.rendermode_preset_cycle_2 = 'G_RM_AA_ZB_XLU_SURF2'

EDIT: actually these last two have rendermode_advanced_enabled = True so I guess the preset doesn't matter, but it still looks weird 😅

[Dragorn421]

for the merge conflicts, same suggestion as #280 (comment)

[sauraen]

I apologize for what's going to be a delay--I see you have a couple more comments which didn't show up in the code viewer page for some reason, which I haven't addressed. I'm going to be out of town for a bit soon so it'll be a while before I can get back to this.

[sauraen]

@Dragorn421

f3d_mat.rdp_settings.rendermode_preset_cycle_1 = 'G_RM_OPA_SURF'
f3d_mat.rdp_settings.rendermode_preset_cycle_2 = 'G_RM_AA_ZB_XLU_SURF2'

From gbi.h:

#define RM_OPA_SURF(clk)                                        \
    CVG_DST_CLAMP | FORCE_BL | ZMODE_OPA |                      \
    GBL_c##clk(G_BL_CLR_IN, G_BL_0, G_BL_CLR_IN, G_BL_1)

So this is doing G_RM_PASS for the first cycle, CVG_DST_CLAMP and ZMODE_OPA are both 0 so they don't affect anything when being OR'd in, the only thing it actually sets is FORCE_BL which is also set by all XLU_SURF rendermodes. So practically it is fine, but theoretically it's wrong because it's two rendermodes which set the flags, I've changed all these to G_RM_PASS instead.

noop makes clr_in 0 for the next cycle

No? It sets the four entries to "enum item 0" for each, which is G_BL_CLR_IN, G_BL_A_IN, G_BL_CLR_IN, G_BL_1MA. This is not the same as G_BL_0 which has a value of 3. RM_NOOP is functionally equivalent to RM_PASS though not actually the same configuration.

[Yanis002]

I reviewed again and tested, import works but fog preview has issues I think

Here's what Fast64 shows me with the intro setup (first cutscene header of hyrule field)

Here's how it is in-game:

I'm not sure if this was already broken or not but I noticed preview looked different so I wanted to try

Otherwise the code itself is ok to me

[sauraen]

The difference between Fast64 and in game is present in main as well, so that won't be addressed in this PR. However, there was a bug where rendermode bits were not updated from rendermode presets when creating a material from a preset in code, which led to collision meshes not having fog on this branch. This has been fixed, now the fog appearance in Blender of an imported scene looks the same between main and this branch.

[Dragorn421]

@Yanis42 ?

[Yanis002]

sorry for the delay