Entry 234

The symbol's light reaches 1000

2026-09-27render

Filescrates/piney-world/src/field_world.rscrates/piney-battle/src/gimmick.rs

Reported from play: on Mutation's first field (Θ Chosen Hopeless Nothingness, area 1 field 27), the lighting flickered and looked wrong.

Finding it

• The flicker. Frames of the field with Kite standing still showed his body's brightness jumping every frame, between 37.7 and 42.4. The field's own light is fixed: frame 0 of the background's animation.
• The source. A debug print of the group the characters are lit by (FieldWorld::cast_lights) showed one extra light: the field's symbol (ccGimSymbol, row 17), about 2,500 from Kite. symMain sets its intensity each frame to 1 + ccRandF(0.2). That is the game's own fire flicker.
• The mistake. The port gave the light a reach of "full to 1000, nothing past 1000000", white. So the whole field was lit by a white light that flickered.

The game's light

ccGimSymbol's constructor (INF gcmn 0x0045a0d0, the same values on Mutation at 0x0046fc0c) writes the omni light's fields:

ccOmniLight::CheckRange (0x00139830) compares the squared distance with +0xcc. It then fades the light linearly from farDownStart to farDownEnd. So the port had read the squared limit as the end, and the end as the start. The colour 0x7fff is red 255, green 127, blue 0 (ccSetColor puts the low byte in red). That is an orange, the symbol's flame.

The fix

symbol_light in field_world.rs builds the light as the game does:

• orange (1, 127/255, 0);
• full at the flame;
• nothing 1000 away.

The springs' lights (ccGimEtc) and the effects' (prim::OmniLight) already had the three fields right.

In the same run after the fix, Kite's brightness stands at 25.3 (±0.3). Characters on the field are darker, lit by the field's own light alone until they come near the symbol.

The new test a_symbol_lights_only_near_it puts the light 500 above a model and checks the colour: half its orange, doubled as VU1 does. At 1000 and at 2500 it checks for none.