Oracle: timing-dependent bytes $0630-$0632 (field scanline probe) #15

New issue

Open

opened 2026-10-29 10:17:14 +00:00 by Gatchan · 0 comments

Gatchan commented

2026-10-29 10:17:14 +00:00

Owner

Open question from the RAM map slice 1 (refs #7).

scanline_h, scanline_v, scanline_v_max ($0630-$0632) are written by FieldLoop from the PPU H/V counters, so their values depend on CPU timing. Once field code runs as C (on), the per-frame WRAM comparison will differ there unless the timing service reproduces them. Decide: exclude them from the oracle (explicit, justified exclusion) or have the timing service record/reproduce them. See port_note of scanline_h in re/ram_map.toml.

Open question from the RAM map slice 1 (refs #7). `scanline_h`, `scanline_v`, `scanline_v_max` (`$0630-$0632`) are written by FieldLoop from the PPU H/V counters, so their values depend on CPU timing. Once field code runs as C (`on`), the per-frame WRAM comparison will differ there unless the timing service reproduces them. Decide: exclude them from the oracle (explicit, justified exclusion) or have the timing service record/reproduce them. See `port_note` of `scanline_h` in `re/ram_map.toml`.

Gatchan added the

validation

task

labels

2026-10-29 10:17:14 +00:00

Gatchan referenced this issue from a commit

2026-10-29 10:20:20 +00:00

docs: RAM map slice 1 in re/ram_map.toml, format and decisions in docs/ram_map.md (refs #7)

Gatchan referenced this issue

2026-10-29 10:37:17 +00:00

RAM map slice 1, generated state accessors, peek/poke by name #18

Gatchan referenced this issue