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

re: the rest() fallback -- its example dissolved, the question got bigger

Browse Source 
ui-resting-pose.md built its dwell-fallback section on GP_TITLE build 7's
ptlogo_eff3.t32, listing keyframes [46, 61, 103, -] -- the STALE PARSER's output.
Corrected they are [0, 46, 61, 103], the longest gap moves from 61->103 to 0->46,
and BOTH ends of the new longest gap are a=0. The element no longer selects a
visible pose under either indexing, and build 7 renders byte-identical under the
corrected and legacy readings (0 px differ). MISSION lists this element as the one
case a Japanese capture was needed to discriminate; it is not.

But losing an example is not closing a question, so: disc-wide census. The
fallback fires on 2 305 of 13 991 elements and returns a VISIBLE pose in 1 697 of
them -- 74 %.

GP_TITLE is 5 fires, 4 visible, and all four are on the SPLASH screens:
palogo_sqex_eff and palogo_anima_eff, each [0:a0 15:a255 30:a212 45:a0], a flash
peaking at 15 and dead by 45 where the fallback returns t=30 a=212.

Independently converged on from the other side: the port, working from the JP
capture and knowing nothing of this census, found ptlogo_back2eff1's rest.t at the
peak of its own 4-unit sparkle with six staggered across the logo, so --pose=rest
fires every sparkle at once -- a frame the game never shows.

Consequence recorded as a rule: a render posed at rest is a legitimate common
reference for comparing two DECODERS and is not a frame to score against a
capture of the game.

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01Wuu56cE8vJGTBtn1ppsk8v
This commit is contained in:
sylph-decoder
2026-08-30 11:50:07 +00:00
parent 310bf8605e
commit eac80677f4
4 changed files with 204 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

77
crates/sylpheed-formats/examples/rest_fallback_census.rs Normal file
View File

@@ -0,0 +1,77 @@
//! When the resting-pose DWELL FALLBACK actually runs, does it pick a visible pose?
//!
//! `ui-resting-pose.md` argues the fallback is structurally unsound — the gap it
//! maximises is time spent *interpolating*, so neither endpoint is held. Its one
//! worked example, `GP_TITLE` build 7's `ptlogo_eff3.t32`, **no longer
//! discriminates**: under the corrected keyframe-record layout the longest gap
//! moved from `61→103` to `0→46`, and both ends of that are `a = 0`. The page's
//! listing still shows the stale parser's trailing `-`.
//!
//! Losing the example is not the same as closing the question, so: disc-wide, how
//! often does the fallback fire, and when it does, does it return something the
//! player would see? An element resting at `a = 0` is harmless whichever end the
//! rule lands on.
//!
//! cargo run -p sylpheed-formats --example rest_fallback_census
use sylpheed_formats::{pak::PakArchive, ui_layout};
use std::path::PathBuf;
fn main() {
let root = PathBuf::from(std::env::var("SYLPHEED_DISC").expect("SYLPHEED_DISC"));
let mut paks: Vec<PathBuf> = std::fs::read_dir(root.join("dat"))
.expect("dat/")
.filter_map(|e| e.ok().map(|e| e.path()))
.filter(|p| p.extension().is_some_and(|x| x == "pak"))
.collect();
paks.sort();
let (mut elements, mut plateau, mut fallback, mut fb_visible) = (0usize, 0, 0, 0);
let mut worst: Vec<(u32, String, String)> = Vec::new();
let mut per_pak: std::collections::BTreeMap<String, (usize, usize)> = Default::default();
for pak in &paks {
let Ok(ar) = PakArchive::open(pak) else { continue };
let name = pak.file_name().unwrap().to_string_lossy().to_string();
for (i, e) in ar.entries().iter().enumerate() {
let Ok(by) = ar.read(e) else { continue };
let Some(b) = ui_layout::parse_build(&by) else { continue };
for el in &b.elements {
if el.keyframes.len() < 2 { continue }
elements += 1;
// a plateau is two ADJACENT poses that are equal — the same test
// the plateau path makes before the fallback can run
let has_plateau = el.keyframes.windows(2).any(|w| {
w[0].x == w[1].x && w[0].y == w[1].y
&& w[0].scale_x == w[1].scale_x && w[0].scale_y == w[1].scale_y
&& w[0].fade == w[1].fade
});
if has_plateau { plateau += 1; continue }
fallback += 1;
per_pak.entry(name.clone()).or_default().0 += 1;
let Some(r) = el.rest() else { continue };
let a = (r.fade >> 24) & 0xff;
if a > 0 {
fb_visible += 1;
per_pak.entry(name.clone()).or_default().1 += 1;
worst.push((a, name.clone(), format!("e{i}/{}", el.name)));
}
}
}
}
println!("POPULATION: {elements} elements with >= 2 keyframes, over {} archives", paks.len());
println!("COVERAGE: {plateau} have a plateau (fallback never runs)");
println!(" {fallback} have NONE -> the dwell fallback decides");
println!(" {fb_visible} of those rest at alpha > 0 -- i.e. VISIBLE\n");
println!("PER ARCHIVE — fallback fires / of those, rests VISIBLE:");
let mut rows: Vec<_> = per_pak.into_iter().collect();
rows.sort_by(|a, b| b.1.1.cmp(&a.1.1));
for (pak, (fires, vis)) in &rows {
println!(" {pak:34} {fires:5} fires {vis:5} visible");
}
println!();
worst.sort_by(|a, b| b.0.cmp(&a.0));
for (a, pak, el) in worst.iter().take(6) {
println!(" a={a:3} {pak} {el}");
}
println!("\n--- END OF CENSUS (if this line is missing, the run did not finish) ---");
}

27
crates/sylpheed-formats/examples/rest_fallback_title.rs Normal file
View File

@@ -0,0 +1,27 @@
//! Which `GP_TITLE` elements does the resting-pose dwell fallback decide, and does
//! it hand back a visible pose? The disc-wide census says 5 fires / 4 visible here.
//! cargo run -p sylpheed-formats --example rest_fallback_title
use sylpheed_formats::{pak::PakArchive, ui_layout};
use std::path::PathBuf;
fn main() {
let root = PathBuf::from(std::env::var("SYLPHEED_DISC").expect("SYLPHEED_DISC"));
let ar = PakArchive::open(root.join("dat/GP_TITLE.pak")).expect("GP_TITLE");
for (i, e) in ar.entries().iter().enumerate() {
let Ok(by) = ar.read(e) else { continue };
let Some(b) = ui_layout::parse_build(&by) else { continue };
for el in &b.elements {
if el.keyframes.len() < 2 { continue }
if el.keyframes.windows(2).any(|w| w[0].x == w[1].x && w[0].y == w[1].y
&& w[0].scale_x == w[1].scale_x && w[0].scale_y == w[1].scale_y
&& w[0].fade == w[1].fade) { continue }
let Some(r) = el.rest() else { continue };
let a = (r.fade >> 24) & 0xff;
let ks: Vec<String> = el.keyframes.iter()
.map(|k| format!("{}:a{}", k.time.map(|v| v.to_string()).unwrap_or("-".into()), (k.fade >> 24) & 0xff))
.collect();
println!("entry {i:2} {:24} rest a={a:3} t={:?} [{}]{}",
el.name, r.time, ks.join(" "),
if a > 0 { " <== VISIBLE" } else { "" });
}
}
}

47
docs/re/data/rest-fallback-census.txt Normal file
View File

@@ -0,0 +1,47 @@
# The resting-pose DWELL FALLBACK, disc-wide -- and the element that 'exposed'
# it no longer does. 2026-08-30.
#
# examples/rest_fallback_census.rs + rest_fallback_title.rs
#
POPULATION: 13991 elements with >= 2 keyframes, over 33 archives
COVERAGE: 11686 have a plateau (fallback never runs)
2305 have NONE -> the dwell fallback decides
1697 of those rest at alpha > 0 -- i.e. VISIBLE
#
# GP_TITLE, every element the fallback decides:
# entry 7 ptlogo_eff3.t32 rest a= 0 t=Some(0) [0:a0 46:a0 61:a255 103:a0]
# entry 10 palogo_sqex_eff.t32 rest a=212 t=Some(30) [0:a0 15:a255 30:a212 45:a0] <== VISIBLE
# entry 11 palogo_anima_eff.t32 rest a=212 t=Some(30) [0:a0 15:a255 30:a212 45:a0] <== VISIBLE
# entry 13 palogo_sqex_eff.t32 rest a=212 t=Some(30) [0:a0 15:a255 30:a212 45:a0] <== VISIBLE
# entry 14 palogo_anima_eff.t32 rest a=212 t=Some(30) [0:a0 15:a255 30:a212 45:a0] <== VISIBLE
#
# 1. THE MISSION-FLAGGED DISCRIMINATOR IS GONE. ui-resting-pose.md built its
# fallback section on GP_TITLE build 7's ptlogo_eff3.t32, listing its
# keyframes as [46, 61, 103, -] -- the STALE PARSER's output, times shifted
# by one with an untimed final pose. Corrected: [0, 46, 61, 103].
# stale gaps 15, 42 -> longest 61->103, one end is a=255 at 200%
# fresh gaps 46, 15, 42 -> longest 0->46, BOTH ends a=0
# So the element no longer selects a visible pose under either indexing,
# and build 7 renders BYTE-IDENTICAL under the corrected and legacy
# readings (0 pixels differ, max delta 0). MISSION listed this element as
# the one case a Japanese capture was needed to discriminate. It is not.
#
# 2. BUT THE QUESTION IS LIVE AND LARGER. Losing an example is not closing a
# question. Disc-wide the fallback returns a VISIBLE pose 1 697 times out
# of the 2 305 it fires -- 74 %.
#
# 3. GP_TITLE's four are ALL ON THE SPLASH SCREENS: palogo_sqex_eff.t32 and
# palogo_anima_eff.t32 on entries 10/11/13/14, each [0:a0 15:a255 30:a212
# 45:a0] -- a flash peaking at t=15 and dead by t=45, where the fallback
# returns t=30, a=212. Near the peak of a transient.
#
# 4. INDEPENDENT CONVERGENCE. The port agent, working from the Japanese title
# capture and with no knowledge of this census, found ptlogo_back2eff1's
# rest.t sitting at the peak of its own 4-unit sparkle, and six of them
# staggered across the logo -- so --pose=rest fires every sparkle at once,
# a frame the game never shows. Same phenomenon, opposite direction.
#
# CONSEQUENCE: 'rest' is not a settled pose for these families. A render posed
# at rest is a legitimate common reference for comparing two DECODERS, and is
# NOT a frame to score against a capture of the game.

55
docs/re/structures/ui-resting-pose.md
View File

@@ -268,15 +268,40 @@ no pose is held, and the rule is choosing an endpoint of a movement.**
### The element that exposed it
`GP_TITLE` build 7, `ptlogo_eff3.t32` — a transient bloom:
🔴 **This listing is the STALE PARSER's, and the example it supports is dead
(2026-08-30).** The times below are shifted by one with an untimed final pose —
the pre-record-layout-fix reading. See
[CONTAINER-NOTES](../../agents/CONTAINER-NOTES.md) for the trap.
`GP_TITLE` build 7, `ptlogo_eff3.t32` — a transient bloom, **as it was printed**:
```
46: (98,42) 100%,100% a=0
61: (108,72) 0%,0% a=0
103: (108,72) 200%,200% a=255 r=80
-: (108,72) 0%,0% a=0 r=150
-: (108,72) 0%,0% a=0 r=150 ← stale: shifted, final pose untimed
```
**and as it actually reads:**
```
0: (98,42) 100%,100% a=0
46: (108,72) 0%,0% a=0
61: (108,72) 200%,200% a=255 r=80
103: (108,72) 0%,0% a=0 r=150
```
| | gaps | longest | its two ends |
|---|---|---|---|
| stale | 15, 42 | 61→103 | one is **a=255 at 200 %** — the screen-filling bloom |
| **fresh** | **46, 15, 42** | **0→46** | **both a=0** |
✅ **So this element no longer discriminates.** `rest()` returns `(98,42) a=0` —
invisible — and build 7 renders **byte-identical** under the corrected and legacy
readings (0 pixels differ, max Δ 0). ⚠️ `MISSION.md` lists this element as the one
case a **Japanese-locale capture** was needed to settle. It is not; that capture was
still worth taking, for the port's `title_jp` question, but not for this.
No two adjacent poses are equal, so there is no plateau. The longest gap is
`61 → 103` (42 units), during which the sprite grows from nothing to **200 %** at
full alpha while rotating 80°, then collapses again. The rule returns whichever
@@ -294,6 +319,32 @@ units brighter.
**The element has no resting pose.** It is a flash; after it plays there is
nothing. Neither answer is *derived* — one of them is merely harmless.
### 🔴 Losing the example did not close the question — it is larger than one element
Disc-wide ([`../data/rest-fallback-census.txt`](../data/rest-fallback-census.txt)):
| | |
|---|---|
| elements with ≥ 2 keyframes | 13 991 |
| have a plateau — the fallback never runs | 11 686 |
| **have none — the fallback decides** | **2 305** |
| **of those, it returns a VISIBLE pose** | **1 697 (74 %)** |
**`GP_TITLE`: 5 fires, 4 visible** — and all four are on the **splash screens**,
`palogo_sqex_eff.t32` / `palogo_anima_eff.t32` on entries 10/11/13/14. Each reads
`[0:a0 15:a255 30:a212 45:a0]`: a flash peaking at t=15, dead by t=45, and the
fallback returns **t=30, a=212** — near the peak of a transient.
✅ **Independently converged on from the other side.** The port agent, working from
the Japanese title capture and knowing nothing of this census, found
`ptlogo_back2eff1`'s `rest.t` sitting at the peak of its own **4-unit** sparkle,
with six staggered across the logo — so `--pose=rest` fires every sparkle at once,
a frame the game never shows.
⚠️ **The consequence, and it is a rule about how a rest render may be used:** a
render posed at `rest` is a legitimate **common reference for comparing two
decoders**, and is **not** a frame to score against a capture of the game.
### 🔴 What this retracts
Last iteration I reported the build 7 render difference as evidence **against**