Entry 295

Fidchell's spells drawn, its prediction's text and voice, the skills' names

2026-09-28renderaudiobattlevolumes

Filescrates/piney-effect/src/fidchell.rscrates/piney-effect/src/boss.rscrates/piney-effect/src/lib.rscrates/piney-effect/examples/boss_probe.rscrates/piney-battle/src/boss/fidchell/eff.rscrates/piney-battle/src/boss/fidchell.rscrates/piney-world/src/combat/boss.rscrates/piney-world/src/field_world.rscrates/piney-desktop/src/anm.rscrates/piney-game/src/fx.rscrates/piney-game/src/area.rscrates/piney-game/src/session/tests/fidchell.rscrates/piney-gen/src/manifest.rstools/test_boss_effect_rs.pydocs/engine/boss-fidchell.mddocs/engine/boss.md

What 294 left named but not shown now shows and sounds: the four spells' pictures, the prediction's text and voice, a skill's name in the cinema, and a boss's effSkillStart. The reference is docs/engine/boss-fidchell.md ("The spells' pictures", "The prediction's text, its voice and the skills' names").

One simulation, two drivers

The spells' rules draw every ccRand (the Fidchell harness runs them natively and counts it), so the pictures must not draw again. The rules (fidchell::eff) now keep what each Draw drew: a meteor moved this Draw, a bolt's ten points, how many segments it drew, their transparency and each sprite's pattern draw, a tower's place and transparency before it moved. They run on a small eff::Host (ccRand, the frame, ccGetDist, the calls). In the fight the boss's Cx is that host and piney-effect's new fidchell draws each spell from the rules' state after the boss's task (BossEffects::sync_spells). For the harness, piney-effect's manager runs the same rules itself on its own generators (Make::MeteoSworm etc.), making the calls beside in the game's order. The pictures are the game's: a meteor is particle's ANM_x300 and the glow CMP_x100 turning pi/30 a Draw; a segment is CMP_x012 stretched to its length over 900 and laid along it, then an EFF_x011 sprite; a tower is CMP_x101b ("abcd"[1] in its constructor: always b). The swarm's light is an omni light at the last meteor, full to nothing at 500.

ccSetColor has an HSV half

Outbreak's disc failed Skeith's own magic square in the boss effect harness (frame 75, the light's colour 0.984 grey against the port's black). Outbreak's ccBossEffLight::Draw turns its count into an unsigned with an inline cvt.w.s where Infection and Mutation call fptoui (0 for a negative): a count fallen below 0 keeps its high bits, and ccSetColor (OUT main 0x00136f90) reads a colour with an alpha byte as HSV (hue in sixths of the low byte, saturation, value). 0xfffb0000 is value 251, saturation 0: grey. The meteors' light, 0x80ffffff, is hue sector 5 at full value: red with 250/256 of the way to magenta, blue 0.023. Both are ported (boss::set_color, to_uint by volume), and the light's modes 1 (the thunders' square: a ccRandF(0.5) tilt drawn and lost, grey and orange by turns, its light moved one Draw in three) and 2 (the meteors': a tilt of -1.5393804, the turn falling behind after 15, pulsing red), with magic squares 1 and 2 (their SEs by game.field, rows 2 and 6, 5).

The text is on the screen

m_predTxt's pos and dirc are written in step 2 and never read anywhere in boss04.cpp: ccAnm::Draw places ANM_ex41txtN by the animation's own F_Camera records, on layer 254 (sysLayer's view) with SetRenderState(CCRS_ZENABLE, 0). The world draws it with the desktop's ccAnm (zenable added) set when Pic::Text first comes. The voice is ccEvVoiceRequest(-40, n), which the audio driver already routes to the field voice groups (-40 is Fidchell's).

A skill's name

OUT gcmn 0x00472550 is a switch on game.field (+0x24) and the skill: fields 2, 4, 5, 6, 7 and 8 each have a table of CINEMASKILLNAME_T rows (field 4: 157, 158, 160, 203, 219, 227, 259, rows 8-14 of x41's TEX_fid_skl). piney-gen finds it as the first call of the function after CinemaOn__19ccBossEffCinemaFadeFi (Outbreak's CinemaOn with a skill; on Infection and Mutation CinemaOff follows, so nothing) and runs it in eemu for fields 0-31 and skills under 512: cinema_skill_rows in the combat group (data version 17). SetupSkillName is Infection's, with its x01-x04 for fields 9-12.

Checks

tools/test_boss_effect_rs.py now hooks genrand (behind ccRand and ccRandF) with a 32-bit LCG on both sides instead of a constant ccRand, and ccSeOn; it builds no particle control, so the probe has no free particle either (the tower's effSmoke first differed there: the port's particle took a pattern rand() the game's never reached). test_fidchells_spells (Outbreak): meteo 192, storm 312, tower 436, square 1 288 (fields 4 and 2), square 2 184 frames agree on every draw, sound, generator, light, slot, rand and genrand count. The Skeith cases pass on Outbreak and Infection. The Skeith, Innis, Magus, Kyvia and Fidchell battle harnesses pass. With the pictures' own rand() draws now made in the field, the story pilot still ends Outbreak's 207 (120,000 frames) and Mutation's 108 (150,000, the gomoras' effSkillStart) done.

session::tests::fidchell::fidchell_spells_show brings event 207 to block 9 (area 73's WORLD_MAN, field 4 of town 3) and forces the prediction, each spell and a skill from act 0: the text shows 420 frames, voice (-40,

1. and (-40, 0) are asked, every spell's objects are drawn (ANM_x300, CMP_x100, CMP_x012, EFF_x011, CMP_x101b, CMP_x201d, CMP_x033), effSkillStart makes its rings at Fidchell, and the skill's cinema has its name. fidchell_shots wrote the frames to /mnt/data/claude/scratch/fidchell-fx: the text mid-screen under "Oracle: Ice Storm", the meteors falling and landing round Kite, the zig-zag bolts from the sky, the towers rising round him seen from above, the ice's photons, the skill's name. The first run forced acts in the middle of others (camera 3 left on from the prediction), which put the tower camera inside the rocks; forcing only from act 0 does not.