fabi/Sylpheed
1
0
Fork 0
Code Issues 9 Pull Requests Actions Packages Projects Releases Wiki Activity

Re-propose the F5/F6 findings — the docs/re f5-*/f6-* corpus — onto main #8

New Issue
Closed
opened 2026-09-04 15:48:07 +00:00 by fabi · 3 comments
fabi commented 2026-09-04 15:48:07 +00:00
Owner
Copy Link

~20 pages of answered F5/F6 work sit on auto/frame-blend-draw-path, 495 commits from main, and main has none of them:

f5-a-press-snaps-the-plate.md         f6-unit1-238-is-an-exit.md
f5-artwork-window-prereg.md           f6-unit9-sweep-period-and-onset.md
f5-code-route-no-literal-target.md    f6-what-the-human-is-watching.md
f5-snap-target-undecodable-...md      f6-plate-identity-and-clock-...-resolved.md
f5-verified-with-full-quad-reader.md  f6-out-of-sample-RESULT.md   (+ 9 more)

This is why the runbook's own list read F5 and F6 as open: written from main's view, where the answers are invisible.

Carry the corrections, not just the conclusions. f5-a-press-snaps-the-plate.md first refuted clock: "shared" off a 5-frame window that sat inside the 11–12 frame input lag, and a pre-registered wider test refuted that. The page that lands must be the corrected one, with the correction legible — a reader who sees only the conclusion cannot tell which way it was reversed.

Pass: the corpus on main as a reviewable PR, and the Port's boot.gd citations resolve against it.

Blocks the Port's re-proposal.

~20 pages of answered F5/F6 work sit on `auto/frame-blend-draw-path`, 495 commits from `main`, and **`main` has none of them**: ``` f5-a-press-snaps-the-plate.md f6-unit1-238-is-an-exit.md f5-artwork-window-prereg.md f6-unit9-sweep-period-and-onset.md f5-code-route-no-literal-target.md f6-what-the-human-is-watching.md f5-snap-target-undecodable-...md f6-plate-identity-and-clock-...-resolved.md f5-verified-with-full-quad-reader.md f6-out-of-sample-RESULT.md (+ 9 more) ``` This is why the runbook's own list read F5 and F6 as open: written from `main`'s view, where the answers are invisible. **Carry the corrections, not just the conclusions.** `f5-a-press-snaps-the-plate.md` first refuted `clock: "shared"` off a 5-frame window that sat inside the 11–12 frame input lag, and a pre-registered wider test refuted that. The page that lands must be the corrected one, with the correction legible — a reader who sees only the conclusion cannot tell which way it was reversed. **Pass:** the corpus on `main` as a reviewable PR, and the Port's `boot.gd` citations resolve against it. Blocks the Port's re-proposal.
fabi added this to the Title screen milestone 2026-09-04 15:48:07 +00:00
fabi added the agent/decoderstate/proposedkind/item labels 2026-09-04 15:48:07 +00:00
fabi added a new dependency 2026-09-04 15:48:22 +00:00
#7 Re-propose the F5/F6 port work (auto/port-p6-audio, tip 0148cb8)
fabi added state/approved and removed state/proposed labels 2026-09-04 15:59:18 +00:00
sylph-decoder added state/needs-human and removed state/approved labels 2026-09-11 19:25:57 +00:00
sylph-decoder commented 2026-09-11 19:26:01 +00:00
Collaborator
Copy Link

Opened #23. auto/frame-blend-draw-path (514 commits, 502 files, +76 780/−1 320)
onto main, mergeable (had one real conflict against main moving ahead
under me — a formatting-only clash in sylpheed-cli/src/main.rs from the
rustfmt sweep, resolved by taking main's formatting and keeping the added
byte-rate line; cargo build -p sylpheed-cli confirms it compiles).

Look at this: the PR body walks the F5/F6 corrections (not just conclusions),
confirms the Port's boot.gd/screen_view.gd citations resolve against the
paths landing here, and flags — without blocking on — the ~113 MB of
docs/re/captures/ game-screenshot evidence for a size/policy look, since
that directory name wasn't on anyone's radar before either.

Moving to state/needs-human; I can't merge or close this myself.

Opened #23. `auto/frame-blend-draw-path` (514 commits, 502 files, +76 780/−1 320) onto `main`, mergeable (had one real conflict against `main` moving ahead under me — a formatting-only clash in `sylpheed-cli/src/main.rs` from the rustfmt sweep, resolved by taking `main`'s formatting and keeping the added byte-rate line; `cargo build -p sylpheed-cli` confirms it compiles). Look at this: the PR body walks the F5/F6 corrections (not just conclusions), confirms the Port's `boot.gd`/`screen_view.gd` citations resolve against the paths landing here, and flags — without blocking on — the ~113 MB of `docs/re/captures/` game-screenshot evidence for a size/policy look, since that directory name wasn't on anyone's radar before either. Moving to `state/needs-human`; I can't merge or close this myself.
sylph-decoder commented 2026-09-11 20:26:53 +00:00
Collaborator
Copy Link

#23 was red — a sylpheed-viewer build break the merge conflict never pointed
at (ComposeOptions grew at, the Explorer's literal predated it).
cargo check --workspace --all-targets catches what cargo build -p sylpheed-cli couldn't. Fixed with at: None (7c5a4a6), which is what the
field's own doc comment says every caller wants — no Explorer behavior change.
cargo check --workspace --all-targets and cargo build --workspace are both
clean now. Still state/needs-human — I can't verify CI or merge.

#23 was red — a `sylpheed-viewer` build break the merge conflict never pointed at (`ComposeOptions` grew `at`, the Explorer's literal predated it). `cargo check --workspace --all-targets` catches what `cargo build -p sylpheed-cli` couldn't. Fixed with `at: None` (7c5a4a6), which is what the field's own doc comment says every caller wants — no Explorer behavior change. `cargo check --workspace --all-targets` and `cargo build --workspace` are both clean now. Still `state/needs-human` — I can't verify CI or merge.
fabi added state/in-progress and removed state/needs-human labels 2026-09-12 11:07:00 +00:00
fabi commented 2026-09-17 19:02:12 +00:00
Author
Owner
Copy Link

Done — closing. #23 (re: land the F5/F6 title-clock corpus) merged into main on 2026-09-13, and main carries 24 docs/re/f5-* / f6-* pages today.

The size-and-policy question #23 raised alongside it — docs/re/captures/ is 82 MB of game screenshots — is not this item and is now #49.

**Done — closing.** #23 (*re: land the F5/F6 title-clock corpus*) merged into `main` on 2026-09-13, and `main` carries **24** `docs/re/f5-*` / `f6-*` pages today. The size-and-policy question #23 raised alongside it — `docs/re/captures/` is 82 MB of game screenshots — is not this item and is now #49.
fabi closed this issue 2026-09-17 19:02:12 +00:00
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agent/decoder
Owned by the Decoder (disc to meaning; runs the emulator)
 agent/port
Owned by the Port (disc to playable; no RE)
 kind/ask
One agent asking the other for something it cannot answer in role
 kind/bundle
A bundle the human defined; agents decompose it into items
 kind/defect
Found by a play-test or a check
 kind/item
One unit of work, small enough to finish in a single session
 state/approved
Human approved the shape; an agent may start
 state/blocked
Waiting on another item; use the Depends-On field, not prose
 state/in-progress
An agent is working it now
 state/needs-human
Done as far as an agent can tell -- a person must look. The body says what to look at
 state/proposed
Agent proposed this item; awaiting the human's approval to start
No Label agent/decoder kind/item state/in-progress
Milestone
No items
No Milestone Title screen
Projects
Clear projects
No project
Assignees
Clear assignees
sylph-decoder sylph-pi sylph-port fabi
No Assignees
2 Participants
Notifications
Due Date
No due date set.
Blocks
#7 Re-propose the F5/F6 port work (auto/port-p6-audio, tip 0148cb8)
fabi/Sylpheed
Reference: fabi/Sylpheed#8
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?