Skip to content

battle_slots is seed-fragile inside the reel drive, and visual_f1 pins the luck visual_f2 just lost #240

Description

@mtklein

Two latent faults recorded while triaging #236's fallout, both left alone because they are green or out of scope today.

battle_slots fails at shifts 21 and 28 — 6 of 8 — timing out at bank1 press1 / bank1 commit inside the Slot reel drive, with zero flee battle lines in either log. Confirmed pre-existing by running the pre-change file on the shipped ROM at shift 21: identical failure. This is upstream of, and unrelated to, the win-during-flee bug that was fixed in M.fleeBattle.

visual_f1 pins H.pipWord() == 0x2173 on menu row 0. That is the same latent luck visual_f2 lost: Ot6PipStage owns exactly one HUD pip cell and moves it between rows, blanking the row it leaves, so row 0 holds a stale glyph that happens to still read right. visual_f2 now follows OT6_PIPPREV (what the flush last wrote) via H.livePipCell(); visual_f1 should do the same before a shift lands on it too.

Neither is urgent. Both are the shape this cycle kept finding — a test passing on a coincidence — so they are worth closing before something else moves the timing into them.

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    bugSomething isn't workingharnessTest harness, audits, runner, graph

    Projects

    No projects

      Milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions