the disc
Testing my own caveat instead of leaving it hanging. I had worried the
plate-free title capture at t=4.0s was too early, since elements carry
keyframes out to t=600 and I had judged "settled" from mean luminance,
which cannot see a thin sprite still moving.
The plate sits at y 550-600 and the swoosh band at y 112-225, disjoint,
so a LATE capture works even with the plate present. The band correlates
0.7342 at t=4.0s and 0.7353 at t=21.5s -- identical to a thousandth over
17.5 seconds, with the band mean flat at 127. So the band is settled by
4 seconds, the capture I handed the port agent is sound, and my caveat is
withdrawn. It also corroborates the earlier finding that groups hold:
nothing crosses that band in 22 seconds.
That was the seventh candidate for the swoosh and the last one that was
not a format field. Pivot is inert at scale 100 and there is no
displacement anyway; fade is white-with-alpha; tint is white; the texture
is blue-leaning; additive via +0x04 bit 0x02 makes everything worse; the
capture is settled. The residual is stable and modest -- band mean +1.83,
edge-correlation 0.70 against 0.92 frame-wide.
So I have classified it rather than generating an eighth candidate:
undecodable from the disc, with the reach written out. And I have named
where a next attempt should start, which is not another field -- a
per-draw GPU capture of the running guest reads the actual blend state,
vertex colours and draw order for that band instead of inferring them
from the file. That is where static RE stops here.
residual is smaller than I said
Testing the candidate I raised last iteration rather than carrying it.
Blending bit-0x02 sprites additively moves every measure the wrong way --
whole-frame mean diff +0.55 to +1.04, swoosh-band mean +1.83 to +3.98,
band edge-correlation 0.6971 down to 0.5578. So the bit is real and
independent but does not select an additive blend. I reverted the
experiment and kept the word as T8adImage::flags, documented and not
acted on; the render is byte-identical to before.
Second refutation: the swoosh is not displaced. Shifting the band over
plus or minus 80 by 8 pixels peaks sharply at zero, 0.7342, falling to
0.22 at 24 px. So the pivot story is dead twice over -- inert at scale
100, and no displacement to explain anyway.
And I have restated the residual, because earlier sections overstated it.
The +16 to +34 band tiles I quoted were measured WITHOUT --primitives.
With the dim drawn the band's average is nearly right at +1.83; what is
wrong is its structure, tiles running -38.6 then +33.8 and cancelling.
Six candidates eliminated now and none confirmed.
One caveat I owe the port agent about the capture I gave them: it is at
t=4.0s, roughly 174 keyframe units into a screen whose elements have
keyframes out to t=600. I judged "settled" from mean luminance, which
cannot see a thin sprite still moving. It is settled for the bulk of the
screen and not proven settled for every element -- which is a live
alternative explanation for a structural difference in exactly the band
the sweeps cross.
METHOD: cargo build passing does not mean cargo test compiles. Adding the
field built the library in 1.48s and broke two test-only struct literals;
cargo test failed with exit 101.
Continuing the swoosh. Last iteration ended with "the next step is
finding where a blend mode would be encoded, and I do not know the format
carries one". It does carry a candidate.
The kind field is not it -- the swoosh sprites are kind 0x0, the same as
ordinary ones. But the T8aD header word at +0x04 splits the title's
sprites exactly along effect versus normal: pteff01, pteff03a,
ptlogo_back2eff1..5 and both ptlogoall_eff are 0x8832, while ptlogo1,
ptlogo2, ptlogo_tm, ptbase2, ptlogo_back2 and ptcopyright are 0x8830.
One bit, 0x02.
Disc-wide it behaves like a real flag rather than an artefact: 19216
sprites, 18 distinct values, bit 0x02 set in 27.1 percent, and it toggles
against otherwise identical words -- 0x8830 against 0x8832, 0x0830
against 0x0832, 0x0810 against 0x0812, 0x0030 against 0x0032.
Marked as correlation and not decode, because nothing here shows the bit
MEANS additive. The one thing that makes it more than a guess is
ptlogo_back2eff, which carries 0x8830 despite having eff in its name --
so the split is the field's and not my pattern-matching on names. The
test is to blend bit-0x02 sprites additively and re-correlate the title
against the plate-free capture.
METHOD gets the trap that cost the first attempt: searching for a
sprite's name and taking the next T8aD returns the SAME header for every
sprite, because the names all live together in the declaration table. It
failed its own control at once -- different sprites reporting identical
dimensions -- and the fix was to match on width and height instead of on
proximity.
Their question was which screens the other four of "six" elements are on. The
six span the whole 12-screen export: ptframe1/ptframe2 on main_menu and
main_menu_jp (trailing run alpha 0xff, VISIBLE -- their rule accepts these, and
they are the four it revealed disc-wide), and pteff02 on title and title_jp
(trailing run alpha 0x00, TRANSPARENT -- their rule excludes it).
That exclusion is correct, and their own measurement is what proves it: pteff02
is the 25 % dim quad, and they measured the title render going from +13.14 to
+0.55 against the plate-free capture once it is drawn. So `rest` must stay at
0x40 and must not move to the transparent trailing run -- which is what their
alpha rule does. Two investigations converging from opposite directions, and no
third discriminator is needed.
Also recorded: the reference renderer was stale for three diff runs and reported
success throughout, and the shape worth naming is that a build system reporting
success is not evidence the artifact you are about to trust is the code you
pinned. The RE agent hit the same class of trap this session from the other side.
BLOCKED.md: the loop/hold question is answered (groups hold -- ptloop01/ptloop02
park off-screen at x=1521 and x=-839), and the rest_plateau entry is closed as
fixed, noting the adopted rule is theirs and not the looser one this port
proposed.
The reference renderer was a revision behind for three consecutive diff runs and
nothing said so.
After the pin bump, `build-reference-cli` reported success at rev b436e5a and
`verify-screen` showed main_menu jumping 3/255 -> 72/255. The natural reading was
that the port had regressed. It had not: the port was right and the REFERENCE was
stale. The shared CARGO_TARGET_DIR still held a `sylpheed-cli` built from
47f423f, cargo said "Finished in 0.13s" and left it there. A clean target dir
built a binary resolving ptframe1 to (440,108) t=62; the shared one still said
(620,108) t=16.
The old check -- does `screen list` run -- cannot catch this. A stale binary runs
perfectly.
So: build into `$CARGO_TARGET_DIR/reference-cli/$rev`, keyed by the pin, where a
new revision has no artifacts to reuse. Then verify the binary AGAINST export/:
both come from the same pin, so if the CLI resolves ptframe1's rest differently
from what the exporter wrote, the two halves of the verification are not the same
revision and it fails loudly. It compares the two rather than asserting a
literal, so it stays true when the pin moves again.
docker/bin is baked into the image, so this needs an image rebuild to reach
PATH; until then invoke the repo copy by path.
P2 shipped the wrong rule and the title is the counter-example. I had reasoned
that the exit is the final untimed keyframe, so anything timed was still the
entry. It is not: `pteff02` on the title holds at t=46 with the 25 % dim quad at
alpha 0x40 and then ramps to 0x00 by t=236. The exit can be a long run of TIMED
keyframes, and running to the end drops the dim and leaves the whole screen
~13/255 too bright -- exactly the luminance excess P2 recorded (oracle 64.1,
rest 62.8, timeline 80.0) and filed as an unexplained "glow slab".
A group is pre-roll -> ramp in -> hold -> ramp out -> post-roll, and a screen
that has arrived sits on the hold. `settle_units()` is now `rest.t`, the
decoders' own identification of that hold.
The check is that a disagreement vanishes: on ALL TWELVE screens the settled
timeline is now byte-identical to the `--pose=rest` render, where before this
four differed by up to 247/255. The timeline's endpoint should BE `rest` -- the
animation is what it adds, not a different destination -- so this is the
property to want, and it holds with no special case.
Credit where due: this came out of the RE agent measuring the title's dim quad
against a plate-free capture, in pursuit of a different question.
What I wanted from it: `ui_layout::rest()`. At 47f423f a trailing run of
identical keyframes was always treated as the exit, so an element with no exit
animation rested at its invisible pre-roll -- `ptframe1`/`ptframe2`, the main
menu's circuit bracket, which a capture of the running game plainly shows.
Pinned at b436e5a rather than 5e6cf0d where the fix was written, because
b436e5a is where it carries its disc-wide check: 30 of 13 991 elements move, 4
become visible, 0 become invisible. The rule shipped is the RE agent's, not the
condition this port proposed -- mine was too loose and would have erased the
word PAUSE on `pgptitle.rat`, whose trailing run is two identical TRANSPARENT
frames. A trailing run is the hold exactly when it is VISIBLE.
The re-export shows the change was contained: two files differ, and within them
exactly four `rest` blocks -- ptframe1/ptframe2 on both main menus, t=16 alpha
0x00 at the pre-roll position becoming t=62 alpha 0xff at the arrived one.
Every diff line pairs. The other ten screens are byte-identical, no sprite
changed, and `pteff02` correctly did not move.
candidate with the check I named
Last iteration I offered the pivot mismatch as a quantified candidate for
the swoosh rendering too thick, and said explicitly that whether it bites
depends on which pivot the compositor uses and that I had not checked. I
checked.
blit takes the drawn size from the TEXTURE and uses the pivot only as the
scale anchor, kf.x minus pivot times scale-minus-100 over 100. At scale
100 that term is zero. And every one of the seven swoosh elements --
ptlogo_back2, its glow and the five eff segments -- reports a single
scale of (100,100) across all its keyframes. So the mismatch, real as it
is in the data, cannot move or resize the swoosh in our render. Refuted,
under my own name.
It is not harmless everywhere, and I have said where it does bite:
ptlogo1 and ptlogo2 run 100, 101, 103, 112, 150 during the build-in, so
there the wrong pivot really does displace them -- during the animation
rather than at rest.
What that leaves is the blend, by elimination rather than by evidence.
Position and size are the texture's own and right, fade is
white-with-alpha, tint is white, the texture is blue-leaning rather than
pink. Seven overlapping sprites at 5 to 23 percent opacity stacked with
plain alpha-over saturate toward opaque white, which is what we draw and
would read as thicker beside the game's thin coloured stroke. Recorded as
the remaining candidate and explicitly not as a diagnosis -- no blend
mode has been identified in the data.
The scripted HANDOFF edit asserted on text an earlier edit had rewrapped,
so the pivot and colour findings landed in docs/re and not on the page
the port agent reads. Third time this session; METHOD now says the rule
was insufficient and names the mechanical fix.
not in any field I can find
Chasing the swoosh defect. The colour half is a clean negative and the
geometry half now has a number.
Ruled out for the colour: every keyframe of every swoosh element carries
fade 0x??ffffff -- white RGB with only the alpha varying -- and no tint
is anything but white. The texture is not pink either; ptlogo_back2
decodes blue-leaning at (175,174,198) and its glow warm at (255,253,234).
So the magenta edge the game draws comes out of blending, and I have said
that rather than leaving "wrong colour" as an undifferentiated mystery.
The geometry half: the declared pivot does not match the element's own
texture. Build 4's ptlogo_back2 pivot is (500,117), which is exactly half
the JAPANESE sprite at 1000x234, not half its own English one at
1118x262 -- off by 59 px, the right order to make the swoosh too thick
and extend too far right. The layout record is authored once and shared
while the sprites are swapped per language, which ui-rat-layout.md warns
about in general; this is that warning measured on the screen where our
render disagrees with the capture. It cuts both ways -- the Japanese
build's ptlogo1 is off by 40.
Marked as a quantified candidate rather than a diagnosis, because whether
it bites depends on which pivot our compositor uses and I did not check.
One METHOD line, because the first version of that check was misleading:
a pivot-equality test conflates rounding with real breakage. Equality
flagged 74 percent of title elements. Split by magnitude it is 17 exact,
54 off by a pixel from odd dimensions, and 24 genuinely off -- and those
24 are the finding.
lands on the known tie-break group
Chasing the residual I left open last iteration, and it turns out I had
attributed the port agent's slab to the wrong thing.
First the refutation. My hypothesis was that our dim is applied over the
whole frame rather than beneath the UI where its layer key puts it. If
that were true the logo would render too dark. It reads +2.36 against a
background of -0.74, so the compositor honours the paint order and the
hypothesis is dead.
What the residual actually is: with the dim in place the error is not
uniform but concentrated in one band, running -38.6 in one tile and +33.8
three tiles to its right. Cropping that band from capture and render
shows it immediately -- the game draws the logo's Z swoosh thin with a
pink edge, and we draw it thick and solid white. That, not the missing
dim, is the washed-out slab over the title logo the port agent reported.
The dim explains a uniform +13; this explains the slab.
The elements are ptlogo_back2, its glow, and the five ptlogo_back2eff1..5
segments -- which are also the group carrying this corpus's known
unsolved paint-order tie-break at key 0x8083, on the same screen. I have
recorded that as a lead and explicitly not as a diagnosis, because a
blend-order swap is a poor explanation for white instead of pink and I
would expect a second cause.
MISSING dim quad
Both halves of what the port agent could not settle, answered against the
plate-free capture rather than against another renderer.
Rendering build 4 and edge-correlating to the capture of the real screen
gives 0.9163 at zero shift, 0.92/0.78/0.93 per band. The geometry of rest
is the title's arrived pose, so the timeline is not needed there.
The slab is the more useful half, and it is inverted from how it looked.
rest alone renders uniformly +13.14 too bright, R +12.35, G +13.58,
B +13.48. Drawing the .prm primitives collapses that to +0.55. The
element is pteff02.prm, the 25 percent dim at rest t=46 with fade 0x40,
and --primitives is OFF BY DEFAULT. So the washed-out cyan slab is not
something being drawn that should not be; it is a dim that should be
drawn and is not, and because the title art is blue-dominant the
shortfall reads as a cyan wash. Any consumer of screen render that omits
--primitives on this screen gets it.
Two residuals recorded rather than tidied. With the dim drawn the mean is
essentially exact but per-pixel agreement gets slightly WORSE --
edge-correlation 0.9163 down to 0.9066 and pixels off by more than 20
rising from 108051 to 162636 -- so the dim's average contribution is
right while its application is not exactly the game's. And separately the
per-tile map shows a +25 to +38 hot spot on the upper-right Z swoosh,
which a capture from the first iteration of this work already showed pink
in the game and white in ours. Same region, still undiagnosed.
port agent
Closing the two things I flagged as unconfirmed when I committed the fix.
The unit gate is green: cargo test -p sylpheed-formats with SYLPHEED_DISC
set gives 131 passed, 0 failed across six binaries including the
disc-gated ones. The earlier background run that produced an empty log
had died with SIGTERM, which is why it looked like nothing happened.
And the disc-wide check the classification requires. Reimplementing both
rules over every RATC bundle on the disc -- 2859 bundles, 13991 elements
with at least two keyframes -- the fix moves rest for 30 elements, which
is 0.21 percent. Four go invisible to visible. Zero go visible to
invisible, which is the safety property I wanted and did not have when I
committed. The four are ptframe1 and ptframe2 in GP_TITLE entries 5 and
8, the same pair once per language.
That surfaces a discrepancy worth handing back rather than smoothing
over. On the English main menu exactly TWO elements satisfy the port
agent's pose-equality condition, not six, so their six must span the
whole twelve-screen export. That fits their own observation that the
timeline and rest differ in exactly one region, the bounding box of
ptframe1 and ptframe2 and nothing else. But if any of their other four
have a TRANSPARENT trailing run, my alpha rule leaves them alone
deliberately -- that exclusion is what protects PAUSE -- so I have asked
which screens they are on and whether a capture shows any of them drawn.
If one is, the alpha rule is incomplete and needs a third discriminator.
VISIBLE, not when its pose repeats
The port agent's report was right about the defect and about which
elements it hits, and its proposed condition does not survive the case
the exclusion was written for.
Their test -- the final untimed keyframe has the same pose as the last
timed one -- is also true of pgptitle.rat, whose trailing run is two
identical transparent frames. Adopting it would erase the word PAUSE
again, which a committed capture disproves. What separates the two is
alpha: an exit fades the element out so its last keyframe is transparent,
while an element with no exit ends on the pose you can see. So
rest_plateau now accepts a trailing run exactly when it is visible.
Verified against a capture rather than against another renderer, which is
the point their own report made: ptframe1's rest moves from (620,108)
t=16 to (440,108) t=62; the changed pixels are 10082 in a bounding box of
x 440-839, y 108-577, which is exactly the 400x470 at (440,108) they
predicted; and correlation against main-menu-oracle.png over that region
improves from 0.9596 to 0.9748. The PAUSE wordmark is unchanged across
all three pause builds.
This also closes a question ui-paint-order-key.md has carried for a
while, that ptframe1 and ptframe2 rest at alpha 0 while the capture shows
the menu frame plainly.
One trap cost most of this iteration and goes in METHOD: CARGO_TARGET_DIR
is redirected in this container, so ./target/debug/sylpheed-cli is hours
stale and every render I made against it was old code. Byte-identical
before and after reads as "no effect" when it actually means "you ran the
old binary".
The full disc-gated test suite was still executing when this was
committed; the verification above is artifact-based.
than loop
The port agent ranked a plate-free capture of build 4 above any further
static RE, so that came first.
B from the main menu returns to the title and the plate fades in a beat
later, which opens a clean window. Recorded at 20 fps from the press: the
art appears at 1.10 s, builds in to 3.70 s, sits settled and unobstructed
until 5.00 s, and the plate arrives at 5.10 s -- the band jumps from 282
to 3755 bright pixels. Committed the frame at 4.0 s as the reference for
the cyan glow slab they report drawing and the game not having.
Their other sub-question -- whether a keyframe group loops or holds --
falls out of the decoded sweeps plus a measurement I already had, and the
two agree. ptloop01's final keyframe parks pteff03.t32 at x=1521 and
ptloop02's parks pteff03a.t32 at x=-839, both off-screen on a 1280-wide
design; and over 18 s of settled title the centre tiles sit at sd <= 0.01
when a looping group would recross the screen every 7.5 s. So groups HOLD
at the last keyframe. The loop*.rat name is misleading.
Also recorded, in the corpus rather than only in their report: the
rest_plateau bug, with their exact identifying condition -- the final
untimed keyframe has the same pose as the last timed one -- the six
elements it misses on main_menu, and the bracket it drops. That closes an
open question ui-paint-order-key.md has carried for a while about
ptframe1 and ptframe2 resting at alpha 0 while the capture shows the
frame plainly. Same two elements, same cause. Not fixed yet; the change
is in ui_layout's rest().
And a METHOD line I would not have written myself: two renderers agreeing
is not evidence the field is right. Their composite and screen render
matched to 3/255 on main_menu and both omitted two elements the game
draws, because both read one field through one decoder.
assumed
Continuing the pulse's cycle length, which I had called a structural
limit of the format. It is, and now I can show it rather than assert it.
Dumping past what I thought was the final keyframe found an EIGHTH block:
fade 0x00ffffff at the same position, so the glow returns to fully
transparent and the pulse is a closed cycle rather than a one-shot ramp.
That is worth having on its own -- the port knows the animation ends
where it starts.
And the eighth block's time slot contains the four bytes "end ", the
record's ASCII terminator. So the corpus's rule that a group's last block
has no time of its own holds here in a second form: not the next group's
element index, but the chunk terminator. The value does not exist. That
goes into ui-rat-layout.md, because it generalises beyond this record.
The measured 2.3 s would need a final step of about 33 units. I have
written that number down explicitly labelled as fitted to the measurement
rather than read from the file, so nobody later re-derives it as a
decode.
One thing I had half-assumed and can now rule out: the word at +0x004 is
not a keyframe count. It reads 60 here with 8 keyframes and 30 in the
loop records with 3. Unknown, and marked so.
Closing the residual I left explicitly unidentified last iteration, and
correcting my own claim twice over in the process.
The cheap test first: is the oscillation global? No -- the bottom-right
corner is flat at sd 0.003 and uncorrelated with the whole frame. So a
per-tile amplitude map over an 8x6 grid, which localises it hard: sd 7.65
in the band x 318..954, y 560..672 at lag 2.3 s, against 0.06 on the
wordmark. That band is the PRESS (A) BUTTON plate's rest position.
Decoding ptbtn00f.rat, the plate's highlight variant, gives the pulse
itself: alpha 0x00 -> 0x06 -> 0x4a -> 0x50, held, then back down through
0x4a to 0x06 across t=6..105. A glow that fades in and out, which is
exactly what a press-start prompt does.
So "the title screen loops at 2.2 s" was wrong in both halves. The title
ART is near-static apart from the two decoded sweeps at 7.5 s and 9.5 s;
what pulses is the PLATE, and the plate is build 2, not build 4. REFUTED
carries it under my own name.
Still amber, and said so rather than rounded off: the cycle LENGTH is not
readable. The keyframe group's last block has no time -- that slot
belongs to the next group -- so the declared span is at least 105 units,
1.75 s, against a measured 2.3 s. Consistent with a final block
extending the tail. Not confirmed.
Second iteration running where a scripted HANDOFF edit failed its
assertion on rewrapped text, leaving the port's page stale while docs/re
had the finding. Patched by reading the file first, and METHOD now says
to check the commit's file count for a silently dropped edit.
Two of my own claims from last iteration die here, and the decode is the
thing that kills them.
Searching for the SHAPE instead of an assumed count found it immediately:
fade words matching 0x??ffffff at alternating strides of 24 and 16, which
is a 40-byte block carrying fade at +0 and tint at +24 -- exactly the
build layout, starting at +0x68. So "a leaf record's keyframes are not in
the build's 40-byte layout" is withdrawn. They are. My scan failed
because I demanded 29 strictly-increasing times, having read the word at
+0x004 as a keyframe count; the records hold THREE.
Decoded, both are horizontal light sweeps. ptloop01 runs pteff03.t32
from x=-639 to x=1521 over t=150..600, which is 450 units or 7.5 s;
ptloop02 runs pteff03a.t32 the other way over 570 units or 9.5 s.
Which refutes the second thing. Last iteration I attributed the measured
2.2 s oscillation to these elements because build 4 declares them --
association, not evidence. A 7.5 s cycle would give about three peaks in
a 22 s capture and I counted eight. So the loops are not what I measured,
the 2.2 s stands as a measurement, and its source is now explicitly
unidentified rather than wrongly assigned.
METHOD gets the general form, because it is what cost the previous
iteration: do not hard-code an expected element count into a structure
scan. Scan for the shape and let the count fall out.
The previous commit's handoff edit failed its own assertion -- the text
had been rewrapped since I wrote the pattern -- so the finding landed in
docs/re and REFUTED but not on the page the port agent actually reads.
An answer not reachable from HANDOFF.md has not been delivered.
Continuing the one gap from last iteration -- turning the title's
measured 2.2 s loop into a decoded cycle. It did not turn.
The record itself is found and partly confirmed. ptloop01.rat is an
opt-linked leaf RATC at 0xbb5966 in build 4, placing pteff03.t32, with
0x1e -- thirty -- in the high half of the word at +0x004, which is the
~30 keyframes ui-rat-layout.md predicts for a loop record. Its pivot
fields read 200 and 90, matching exactly what screen info prints for
ptloop01, so this is the right blob and the header offsets hold.
The times are not there. A build's keyframes are 40-byte blocks with the
time at +36; scanning every 4-byte-aligned start across a 0xC0 window for
29 strictly increasing values at that stride finds nothing, in either
loop record. So a leaf record's keyframe layout is NOT the build
placement layout, which is now a line in REFUTED because it is the
obvious first assumption and it is wrong.
The 2.2 s stays measured and the port hardcodes it. What the next attempt
inherits is the record's address, a confirmed pivot, and one layout ruled
out -- which is the useful part of a negative.
did not work
I set out to do for the title what worked for the splash: measure the
build-in and check it against the declared keyframes. That part failed,
and I am recording the failure rather than the number.
It failed for two reasons, both mine. The title was reached by skipping
the movie with A, which cuts to black and brings the title up on a path
that may not be the normal one. And the visible rise, about 2.7 s, is a
luminance curve -- which screen-transitions.md already establishes is not
the fade quad's ramp. So 2.7 s neither confirms nor contradicts build 4's
declared 16 -> 261, four point oh eight seconds. They are not measuring
the same thing, and I have said so instead of reporting a discrepancy.
What the recording did establish is worth having anyway: the title never
settles. Over 22 s the mean luminance oscillates continuously, peaks at
5.6, 7.8, 10.0, 12.7, 15.2, 17.1, 19.2 and 21.3 s, mean interval 2.24 s.
Marked as about 2.2 s plus or minus 0.4 rather than a precise period,
because peak-picking a low-amplitude signal is coarse.
The mechanism was already decoded and I did not re-derive it: build 4
declares ptloop01.rat and ptloop02.rat, and loop*.rat is a looping sprite
animation rather than a composition. What is new is the period, and that
it runs forever. For the port that matters -- drawing the resting pose
and stopping will look wrong.
Not confirmed, and flagged: that 2.2 s is the loop's OWN declared cycle.
screen info prints the element's placement keyframes but not the frame
count inside the .rat, and I did not open it.
Closing the caveat I left last iteration -- the splash timing was +/-0.5s
because I sampled at 2 fps, and at that rate I could not see ramps at
all, only plateaus.
Re-recorded at 10 fps: the splash FADES, both in and out, rather than
cutting. And the bundles declare it. palogo_sqex.t32 carries keyframes
[15 30 235 239 251 255] and palogo_gamearts.t32 [15 30 190 194 206 210],
each with an _eff glow child on [15 30 45]. Under Q1's 1 unit = 1/60 s
that is a 0.25 s ramp in, a 3.42 s or 2.67 s hold, and a 0.33 s fade out
-- against measured holds of about 3.5 s and 2.4 s and fade-outs of about
0.3 s. The constant 0.35 s offset between declared and measured start is
just that my recording's t=0 is when the WINDOW appears, not when the
guest starts drawing.
So the first screen's animation moves from measured to decoded: the port
reads it off the disc instead of trusting my stopwatch.
It also explains something the capture showed and I had no account for.
The brightness overshoots on the way in -- peaks at 0.8 s, settles by
1.1 s -- which reads as a bloom. It is the _eff glow child, whose
keyframes run 15 -> 30 -> 45, ramping in after the logo and back down
while the logo holds. Mechanical, not a rendering artifact.
Closing the amber I raised last iteration: the splash renders had no
framebuffer capture to diff against. Recording the boot from the moment
the window appears, at 2 fps, catches it before the movie.
Both halves match their own render at zero shift, each with the OTHER
half as a negative control: publisher frame 0.9146 against entry 10 and
0.027 against entry 11; developer frame 0.9792 against entry 11 and
-0.033 against entry 10. The EN/JP twins score 0.876 and 0.966, so the
test tells a screen from a different screen but cannot tell a pair apart,
and I have said so rather than implying it discriminates.
The recording also times the front of the boot, which Q6 and Q7 only had
from the title onward: SQUARE ENIX from about 0.5 to 4.0 s, black,
developer logos from about 5.0 to 7.5 s, black, then ADV.wmv from about
9.5 s.
One trap avoided and written into METHOD. The frames around 9.5 to 12.5 s
show SQUARE ENIX again in cyan, and read from the capture alone that is a
third logo screen. It is not -- it is the intro movie's own opening,
which the milestone-2 notes describe as white SQUARE ENIX plus cyan glow
plus red diamonds. A logo appearing twice may be two different things.
Two pages of this corpus disagreed. MISSION says the splash "is the RATC
screen, which already renders"; ui-paint-order-key.md says it "cannot be
rendered by screen render at all". Running the tool both ways settles it.
screen list --all shows all 16 GP_TITLE entries instead of 12, and the
four the default listing drops are the splash, each half shipped twice:
entries 10 and 13 are the white SQUARE ENIX publisher logo, entries 11
and 14 the GAME ARTS / SETA / studio anima developer logos. Entry 11's
seven elements are the three logos, their three _eff glows and the
palogo_eff0.prm backdrop -- exactly the composition ui-paint-order-key.md
had measured for the splash without being able to draw it.
So the "cannot be rendered" line is wrong and is corrected in place. What
is true is narrower and worth keeping: the splash is invisible to the
DEFAULT listing because is_build wants a .rat child, so anyone who does
not pass --all concludes it is missing. That goes in METHOD -- a default
filter can hide a whole screen and the corpus will record it as absent.
The payoff is for the port rather than for the RE: the first of the five
screens now has a reference composite, which it did not have. Marked
amber on one point -- there is no framebuffer capture of the splash in
this repo, so the match to the running game is by description against the
milestone-1 notes, not by pixels.
because the test probed the wrong state
Nine stores to the phase field this+132, control-gated on the 136 query
returning at least the 18 known ones. Attributed to the handler each
lives in, they give the phase graph: entry -> 2, then 2 -> 0 (splash),
2 -> 3, 3 -> 4, and 4 -> 2 on event 0.
That last edge is the problem with what I did last iteration. The event-0
block sets TWO fields one instruction apart -- stw r28,136(r30) for state
0 and stw r11,132(r30) with r11 = 2 for phase 2. So B from the menu lands
in PHASE 2, the same phase as the boot title, and the test I ran
believing it probed phase 4 state 0 probed nothing of the kind.
So the refutation is withdrawn. Worse for me and better for the idea: A
working on the B-returned title is exactly what the hypothesis predicts,
since phase 2 is the phase that references BUTTON. The hypothesis is back
to untested, now consistent with two observations rather than one, and
what it still needs is the attract-returned title's phase -- which no
test so far has read.
The measurement from that iteration stands, because it does not depend on
the phase: the B-returned title accepts A, so only the ATTRACT-returned
title is inert. I have said so explicitly in the page rather than letting
the withdrawal take the good half down with it.
METHOD gets the lesson: before testing "state X behaves like this", check
your route actually reaches state X. When a transition writes more than
one field, read the whole block and not just the store you were looking
for.
this corpus thought
Last iteration I proposed that the boot title is phase 2 and the
attract-returned title is phase 4 state 0, and that only phase 2 wires up
A -- a candidate answer to the oldest open question in
canary-scripted-input-traps.md. I marked it untested and named a test
that needed reading this+132 at runtime.
There was a cheaper one. The decoded graph says B from the main menu is
the edge 4 -> 0, so the B-returned title IS phase 4 state 0. If that
state were the inert one, A would not work there.
A works. Control first -- A on the boot title opened the menu -- then A
on the B-returned title opened it too. So phase 4 state 0 handles A and
the hypothesis is refuted, one iteration after I raised it. Recorded in
REFUTED under my own name, and the conditional stated: the edge is
decoded, but which title it produces is my inference.
The test earned something better than it cost, though. That page frames
the problem as "the boot title accepts A, a later title does not", and
that is too broad -- the B-returned title accepts A perfectly well, a
case nobody had tried. Only the ATTRACT-returned title is inert. Whatever
causes it is tied to the attract cycle specifically rather than to the
title having been shown before, which is a much smaller haystack.
Also reproduced in passing: 4 s after B the plate is absent and fades in
after, the same near-miss this corpus has already been caught by once.
to this corpus's oldest input puzzle
Continuing Q6 on the phases left unread. Strings each phase handler
references, plus whether it carries its own jump table:
phase 0 sub_821C5690 LOGO no switch
phase 2 sub_821C5818 BASE_INFO, BUTTON, TITLE_SCREEN no switch
phase 3 sub_821C5EC0 (none) one switch
phase 4 sub_821C6458 BASE_INFO, LOADING, TITLE_MENU, TITLE_SCREEN
Phase 0 referencing LOGO is a second independent confirmation that it is
the developer splash -- the iterate3E notes reached the same function
from the guest side and named the splash's LOGO items. Phase 2 draws the
title WITH the PRESS A plate, which the archive side had already
established is a build of its own.
Which produces something worth more than either: the title is installed
from TWO places, phase 2 and phase 4 state 0. Same screen, different
code. canary-scripted-input-traps.md has recorded for months, and never
explained, that the boot title accepts A while the attract-returned title
accepts nothing, with the giveaway that a draw capture in each is
identical. Two code paths installing one screen is exactly that shape. I
have written it into that page as a candidate with the cheap test named
-- read this+132 on each title -- and marked it untested, because it is.
One query in this iteration failed its own control and I threw its half
away: counting stw rX,136(r30) per phase returned 0 for phase 4, which
has 18, because the operand text has a space the pattern did not allow.
The bctr half passes its control and is reported. METHOD gets the
underlying trap: instructions.function is unpopulated for most rows, so a
query scoped on it silently returns nothing.
confirmed three ways
Chasing where the event code comes from, sub_821C5580 turned out to do
two things worth having.
For Q6 it arms the outer gate: li r11,3 ; stw r11,16(r28) is exactly the
value sub_821C7850 tests before dispatching, which answers the "what does
this+16 == 3 gate on" question I left open last iteration. It then passes
r5 = [r27+4], an event read out of a structure, so the vocabulary is
still not enumerable. Of the four callers of sub_821C7850, three pass
constants -- 0, 0 and 5 -- and this one does not.
The accident is a few instructions earlier: li r5, 1103 into a sound-play
call. 1103 is a BGM cue id, BGM_103. That closes a residual I had written
off as undecodable: which bank the menu plays. The cue table genuinely
cannot say, since its BGM entries are numeric -- but the code can, and it
checks out three independent ways. The census says BGM_103.slb is two
waves of 3876864 and 3930112 bytes; the XMA probe at the main menu saw
two stereo streams of 3876864 and 3930112 bytes. Byte for byte on both.
That also corrects me. bgm-two-stems.md said those observed sizes matched
no bank's declared waves and inferred the decoder gets a window rather
than a whole wave. Wrong on both counts -- I had checked only the BGM_0xx
rows of the census because that was the block on screen. METHOD gets it:
check a measurement against the whole reference set, not the part you
happened to be reading.
Chasing where the event code comes from turned up the level above and
corrected my own framing three sections running.
sub_821C6458 has exactly one caller, sub_821C7850, and that caller is
itself a dispatcher -- on a SECOND, outer state field at this+132, five
phases, gated on this+16 == 3. Phase 4 is the title/menu machine I have
been decoding. Phase 0 is sub_821C5690, which the iterate3E notes had
already fingered as the splash mechanics from a completely different
direction. That agreement is the useful check here.
So GamePart_Title has two nested state fields: a phase at this+132
choosing which sub-machine runs, and the state at this+136 inside phase
4. Everything I decoded over the last three iterations -- ten states,
eighteen edges, the event dispatch -- is phase 4 ONLY. Phases 1, 2 and 3
are untouched. I have called that out in REFUTED against my own earlier
wording rather than quietly restating it.
And the event code is forwarded, not created: all five phase handlers are
called with the same (this, r29, r28), where r28 is sub_821C7850's own
third argument passed through untouched. So the event vocabulary is
defined at least one level further up, and finding what 3, 5, 8, 10 and
25 mean means going up again. Not done.
Continuing Q6's last item: the condition on each edge.
sub_821C6458's third argument is an EVENT CODE, and state 4 -- reached
straight after the main menu is installed, so the input-waiting state --
dispatches on it through a second jump table at 0x821c6b7c, 26 cases.
Six are handled and twenty fall through changing nothing:
event 0 -> state 0 (TITLE_SCREEN, back to the title)
events 3, 5, 8, 25 -> state 8 (LOADING)
event 10 -> state 5
So the edges I listed last iteration as "4 -> 0, 5, 8x4" are each now
attributed to a specific event, and the input state's graph is complete.
One thing I deliberately did NOT promote. The main menu has five items, B
returns to the title, and this table has one event to the title, four to
LOADING and one elsewhere. That is a tempting fit -- B plus four items
that load an external archive plus EXTRAS staying inside GP_TITLE -- and
it matches everything measured. It is also a count-match and nothing
more: the events are unnamed, nothing shows event 3 is a particular menu
row, and state 5 installs no named screen so the EXTRAS half has no
support at all. Written down as an observation with that said plainly,
because the same shape of reasoning already cost me DIFFICULTY and
EXTRA_MENU two iterations ago.
Still open: what the event numbers mean, the conditions on edges out of
the other states, and what states 1/3/5/6/7/9 do.
Q6's last open piece was which state leads to which, and it turned out to
be a plain switch.
state = this+136, compared against 9, dispatched through a jump table at
0x821C6498. The ten words at that address disassemble as lwz r16,N(r28)
instructions and are nothing of the kind -- they are the table. That is
the same "self-reference" I flagged two iterations ago as looking like a
jump table; it is one, and the disassembler was decoding its data.
Three states install a named screen: 0 is TITLE_SCREEN, 2 is TITLE_MENU,
8 is LOADING. Eighteen transitions, every one a literal li/stw pair into
the state field, giving the full graph.
It lines up with the behaviour measured weeks of iterations ago without
either side knowing about the other: boot reaches state 0 and A opens
state 2; B at the main menu returns to the title, and 4 -> 0 is the ONLY
edge back to state 0, reached from 2 -> 4; entering a submenu goes
through LOADING and comes back, which is 4 -> 8 at four separate sites
then 8 -> 2. I have marked that as corroboration and said plainly it is
me matching a graph to observations -- the conditions on the edges are
not decoded, so nothing here proves which input picks which branch.
Still open and written down as such: the condition on each edge, what
states 1/3/5/6/7/9 do, and the fact that state 3 is never a destination
in this function, so something outside sets it.
own corroboration from last iteration
Last iteration I listed a candidate name vocabulary from "upper-case
strings the callers reference", marked it amber because those are not
proven arguments, and named the check: confirm per call site which string
actually lands in r5. Ran the check. It fired.
The first extractor FAILED ITS OWN CONTROL -- a backward scan for
addi r5 recovered 7 of 48 sites and missed both title sites I had already
read by eye, because there the name is built in r27 and reaches r5 via a
later mr. Rewritten as a forward register simulation it reproduces all
three known sites and recovers 46 of 48, giving 28 distinct names.
Two of my own claims die. DIFFICULTY and EXTRA_MENU are NOT arguments at
any site -- they are strings that merely live in the same functions, so
the "independently corroborated by measurement" line I wrote was an
artifact of the loose filter. Only TUTORIAL_MENU survives. And
sub_821CC860 is not a screen factory at all: its real arguments include
BG, BLACK, FADE, FILE, KEY, PAD, SOUND and GAMMA_RGB. It is a generic
name-keyed lookup, mostly config, which the title part happens to call
with its three state names.
The title finding itself is untouched -- sub_821C6458 really does
transition by calling this with TITLE_SCREEN, TITLE_MENU and LOADING.
What is gone is the inference that the 28 names are a screen vocabulary.
Two METHOD lines, both earned here: strings a function references are not
arguments it passes, and a backward scan for a destination register
misses arguments staged through a scratch register.
agrees with the timeline
P2's gate is the buttons sliding in, and `tools/screen-strip` renders the strip
that shows it. But the useful result came out of checking where the animation
settles.
On 8 of 12 screens the settled timeline is BYTE-IDENTICAL to the declared `rest`
pose -- the port walks the keyframes with an authored time unit and arrives, to
the pixel, where the pinned decoders independently say the screen rests. On
`main_menu` the two differ in exactly one region, 400x470 at (440,108): the
bounding box of `ptframe1` and `ptframe2` and nothing else.
`rest` puts both at their first keyframe, off-position and transparent. The
capture of the running game shows them -- the bright circuit bracket around the
menu. Cropping the same region from the capture and from both renders puts the
ring and its elbow trace in the timeline render pixel-aligned with the game's,
and absent from the rest render. Geometry, so it does not depend on the
capture's gamma or on its having been taken with NEW GAME focused.
`ui_layout::rest_plateau` excludes a trailing run of identical keyframes because
it is normally the exit. On an element with NO exit animation the trailing run IS
the hold. The condition that identifies these exactly, with no false positives
here, is "the final untimed keyframe has the same pose as the last timed one" --
six elements, and `rest()` misses all six. Filed in BLOCKED.md for the RE agent:
the decoders are pinned and are not this port's to fix, and `sylpheed-cli screen
render` is missing the bracket too.
Worth saying plainly what this does to P1: the port and the reference agreed on
`main_menu` to 3/255 and BOTH were missing two elements the game draws. Two
renderers reading one field through one decoder agreeing is not evidence the
field is right. BLOCKED.md had already said that about the pivot; here it bit.
The title is NOT settled and P2 does not claim it. `rest` and the timeline
disagree there by 142-247/255, the only live title capture composites the PRESS A
plate over build 4 so it cannot be diffed against the title alone, and both of
the port's modes draw a cyan glow slab the game does not have -- a third problem,
P3's. Recorded as an open question rather than resolved by tuning.
Also reconciled against the RE agent's new work: Q8 is answered -- the SE waves
are located in `Static.slb` (move/confirm/back), which unblocks P6's audio; and
the title's transitions are a lookup by NAME, giving P3/P5 the game's own screen
vocabulary as candidate `goto` targets, marked as the name match it is.
P2. A keyframe is the start of a linear ramp toward the next; `ScreenView` walks
them at `time_units` and `boot.gd` advances that in real time, or freezes it with
`--time=<seconds>`.
`authored/timing.json` holds the ONE constant this needs. HANDOFF Q1 is answered
-- linear, 2 units per rendered frame, 1 unit = 1/60 s -- but that conversion was
MEASURED off the running game, not read from a file, so it is authored rather
than exported and it says so at length. Expressed as units-per-second, because
60 is exact and 0.01666... is a decimal a reader has to recognise.
The timeline stops at the last TIMED keyframe and never plays the exit. Every
group's final keyframe carries no `t` -- across this export it is a fade-out for
116 of 134 elements, a scale-and-slide exit for 12, and identical for 6 -- so
playing into it would mean inventing how long the ramp takes. That duration is
the screen transition, it is measured at ~0.4 s, and it is P3's to author with
its own evidence. `exit_ramp_seconds` is therefore null on purpose, not missing.
`--pose=rest` keeps the P1 behaviour available: since the port's default is now
the timeline and the two DISAGREE, renderer-vs-renderer diffing has to be able to
ask for the same assumption the reference renderer makes.
The interpolation is checked by where it lands: on 8 of the 12 screens the
settled timeline is byte-identical to the rest render.
Continuing Q6's last residual -- the control flow I left unread.
All three state-name sites in sub_821C6458 compile to the same shape:
load the name, then bl 0x821CC860 with it in r5 and 0 in r6, then hand
the result to sub_82187B78. So a transition is lookup-by-string then
install: sub_821CC860(this+88, this+24, "TITLE_SCREEN", 0). Not a numeric
id and not a table index. That also explains something that had been
sitting unexplained -- GP_ADVERTISE_DEMO having zero xrefs -- because at
this level the screen graph is keyed by name rather than by GamePart id.
sub_821CC860 has 28 callers, and the upper-case identifiers in that
neighbourhood split cleanly into screen names and config keys. Three of
the screen names are corroborated by measurements I took before ever
opening this function: DIFFICULTY is what NEW GAME opens, EXTRA_MENU is
the EXTRAS submenu, TUTORIAL_MENU the lesson list. That is the static
side agreeing with the dynamic side on names neither knew about the
other.
Held at amber deliberately. The 35 strings are what those callers
REFERENCE, not proven arguments, and the list plainly mixes screen names
with things like TEXT_FONT and GAMMA_RGB. Confirming it means checking
per call site which string actually lands in r5, and I did not do that.
Still unread: which state leads to which. The three lookups sit in
different branches and TITLE_MENU's is guarded by a cmplwi/bne, but I did
not trace the branch structure, so the ORDER still comes from measurement
rather than from the code.
TITLE_MENU, LOADING
Taking the one residual I had ranked as worth doing: sub_821C6458, which
the previous note left explicitly unread.
It is the title part's screen-state function -- 4460 bytes, has EH,
called from exactly one place, sub_821C7850, which sits beside the
creator the registration site points at. Its string references are
TITLE_SCREEN, TITLE_MENU, LOADING and BASE_INFO. Those are precisely the
states measured off the running game, in the game's own words, and
BASE_INFO is this corpus's own marker for a screen-config lookup rather
than a table read.
That sharpens Q6's negative rather than overturning it. The title part
DOES ask for configuration keyed by TITLE_SCREEN and TITLE_MENU, and
config.ini -- the disc's only config -- has no such sections, only an
empty [SYSTEM] and [LANGUAGE]. So the game asks the question, the shipped
disc does not answer it, and the defaults live in code. The port can
transcribe the state names even though it must author their contents.
Marked amber, not green: this is characterisation from string references
and call counts, not a read of the control flow. Two self-references
inside the function look like jump tables, which is what a state machine
compiles to, but I did not confirm that and nothing here shows which
state leads to which.
One refutation recorded because it looked like a find for several
minutes: the {func, func, ptr} triples at 0x820a3b48 read convincingly as
a state/handler table and are nothing of the kind -- the bytes before
them are the tail of a RegisterToFactory diagnostic string and the data
column is zero-filled descriptors. Static-init records, not dispatch.
located, and two of them reproduce
Continuing Q8 rather than opening anything new. The gap that mattered was
the confirm cue: a menu needs a sound on A, and I had only cursor and
cancel.
The fix was where I was counting from. The first run started counting
streams at the main menu, so the confirm cue had already fired during
boot and showed up as "played before any input". Counting from the TITLE
instead attributes it cleanly: the A that advances title -> main menu
fires the 12288-byte wave at 0x5d6c0, together with the two stereo BGM
stems, which is the menu's music starting.
So move (0x1ec0, 0.533 s), confirm (0x5d6c0, 1.016 s) and back (0x0ec0,
0.344 s) are all located and decodable. Move and back came back with
IDENTICAL head bytes and sizes on a second independent boot, so the
dedup-keyed method is stable and those two are now n=2 rather than n=1.
Two honest limits recorded rather than smoothed over. The A press both
confirms and opens a screen, so its wave could be the cue the vocabulary
calls DECIDE or the one it calls SUB_WIN_OPN -- the port gets the sound
the game plays, not a name. And left/right fired no new stream, which
excludes a DISTINCT invalid cue but cannot exclude them quietly replaying
one of the three already heard, because the probe dedups on head bytes.
Finishing the step I left open last iteration rather than starting
something new. Offsets and packet counts were in hand; what was missing
was proof they are actually waves.
slb_extract_wave.py wraps a (bank, offset, packets, channels, rate) slice
in a synthesized XMA1 RIFF, following the layout the Rust decoder already
uses. The three located cues decode to 0.533 s, 0.344 s and 1.016 s of
mono 48 kHz audio, audible from sample 0, each with the percussive
attack-and-decay envelope of a UI blip. Bitrates come out at 12-15 kB/s,
about half the stereo BGM rate, which is what mono should be.
The control matters more than the results. The SAME wrapper applied to
BGM_001's first wave decodes to 173.808875 s -- identical to the duration
that bank's own on-disc RIFF header produced back when Q10 was answered.
So the header I synthesized is not approximately right, it reproduces a
known-good decode exactly, and the cue durations are trustworthy for the
same reason.
I did not commit the decoded audio. Three commands regenerate it from the
disc, and the corpus's job is measurements and tooling rather than
extracted game assets. The offsets, the packet counts and the tool are
the deliverable.
`main_menu` -- the milestone's gate -- agrees with the reference renderer to
within 3/255 on every channel of every pixel, RMSE 0.38 %, no pixel above 4 %.
Nine of the twelve screens are at or under that; two are byte-identical.
Three exceed it, and each gets a cause rather than a wider tolerance:
* `title` (max 6): every disagreement is INSIDE A TIE -- the derived order and
the CLI's measured-off-the-game order differ only among elements with
identical layer keys. That is exactly the residual HANDOFF Q3 documents and
the export already declares in `unresolved: paint_order_ties`. Cost: 904 px at
4-6/255 in one glow band. The port keeps the stable sort; fitting it to one
screen's capture would be tuning.
* `title_jp` (max 154): `ptlogo_eff2` at 125 % is the only drawn element in the
whole export at a scale that is not a whole multiple of 100 %, and `title_jp`
is the only screen over 6/255. Same fact twice. `ui_layout::blit` samples the
source at the destination pixel's top-left corner; a GPU samples at its
centre, and at 125 % those disagree on one column in five.
I think the CLI is the one that is wrong -- corner-sampled nearest is a
half-pixel bias toward the top-left that no rasteriser produces. But that is a
reading, not a measurement: it needs a framebuffer capture of the Japanese
title screen, so it is filed in BLOCKED.md as a question. The port is NOT
changing to match, because matching would mean reproducing a half-pixel offset
on purpose to make a number smaller.
* `extras` (max 4): two pixels.
Also reconciled: the pivot question predicted a P1 diff could not distinguish
the declared pivot from half the texture, because both renderers use the
declared one. That held. Recorded so the agreement is not later mistaken for
evidence -- and P2 will not settle it either.
BLOCKED.md's `/reborn` HEAD updated to 690683d, with a note that the mount is
read-only by design and `git -C /reborn pull` fails rather than being skipped.
The P1 gate is a pixel comparison, so it needs to be a command rather than a
paragraph. `tools/verify-screen` renders each screen in the manifest both ways
and reports the largest per-channel difference anywhere in the frame.
Both renderers are held to the same inputs on purpose:
* the reference CLI is the one `build-reference-cli` builds from the revision
the exporter is PINNED to -- not `/reborn/target/`, which is a live mount of
the other agent's checkout and moves mid-iteration. A pixel disagreement
against a moving decoder has a free variable in it and proves nothing.
* `--black`, because the screen carries its own background and the CLI's default
dim slate stands in for a 3D scene that is not on this screen.
* `--primitives --animated`, because the port draws every element at rest and
those two flags are what make the CLI draw the same set. NOT `--focus`:
nothing is focused at rest.
The threshold is 3/255 -- what integer-truncating compositing in the CLI and
float rounding on a GPU differ by. Above that is a placement, order or colour
disagreement that needs a reason, and the script says DIFFERS rather than
pretending a wider tolerance is a result.
P1. The project reads only `export/` -- the manifest, a screen's JSON and its
PNGs -- and draws every element at `rest`, in the export's own `paint_order`.
Three choices worth the words:
* `ExportTree` addresses screens by manifest NAME, never by path, and checks
`format` on the manifest and on each screen before drawing. Textures are
decoded from bytes at runtime rather than Godot-imported: `export/` is
gitignored and regenerated wholesale, and a `.import` per sprite would be
derived state next to derived state, invalidated on every re-export.
* One CanvasItem draws the whole screen. `paint_order` is already back-to-front,
so honouring it is a loop; spreading it across sixteen nodes' z-indices would
hide the one unresolved thing about that order -- the ties -- behind Godot's
sibling rules.
* The screen renders into a SubViewport sized to the export's `design` rect.
Capturing the window instead gave 1280x720 of screen minus a window manager's
title bar: 1235x695. A gate that rescales that to compare against a 1280x720
composite is measuring the compositor.
Nearest-neighbour filtering, because the export is a 1:1 copy of the disc's
texels, elements draw at up to 500 %, and it is what `ui_layout::blit` does --
so a filter difference cannot masquerade as a placement difference in the diff.
No keyframe interpolation and no focus state: both depend on constants that are
MEASURED rather than decoded (HANDOFF Q1, Q5), and a pixel-diff gate must not
have one of those inside it. P2 and P5.
extracted", and the tool was in the build all along
Two iterations ago I closed Q8 by declaring the SE audio undecodable:
Static.slb has no wave boundaries, there is no XACT container anywhere,
and I said the index "exists only at runtime" as though that put it out
of reach. The disc half of that stands. The conclusion did not.
This build of Canary carries a cvar called xma_param_probe, added by this
project, whose own comment says it logs each XMA stream's parameters and
head bytes so raw sound.pak entries can be matched to real decode params.
It has been sitting in the startup CONFIG DUMP of every log I have read
this session.
Run with it, driving the main menu: a d-pad move spawns a new mono 48 kHz
stream of 4 packets / 8192 bytes, and B spawns a different one of 2
packets / 4096 bytes. Searching their logged head bytes in Static.slb
finds each at exactly one offset -- 0x1ec0 and 0x0ec0 -- and the two are
contiguous, 0x0ec0 + 4096 = 0x1ec0. So the bank is a packed run of whole
2048-byte packets with no delimiters, which is precisely why the seek
scan found nothing: there is nothing to find. A wave is (offset, packet
count) and nothing else.
That splits Q8's binding cleanly. Event -> WAVE is now measured: the port
can have the audio. Event -> cue NAME is still a name match on the
authors' identifiers, and the page says so.
The same run settled something for Q10 too. Sitting on the main menu, TWO
stereo 48 kHz streams were decoding simultaneously. bgm-two-stems.md said
that observation was what it needed and that this container could not
make it; it can, and a music bank's two waves are now measured as
simultaneous rather than only inferred.
METHOD gets the general lesson, because it cost two iterations: check
what instrumentation the local build already has before declaring a
question blocked on tooling.
ambiguous, the sampling was
Continuing last iteration's amber rather than starting something new. The
fix was already named there: stop using 5-second screenshots and record
the display.
Recorded with ffmpeg x11grab at 10 fps and matched every 0.5 s. Across
the 25 consecutive samples from capture 5.0 s to 17.5 s the S00A playhead
is strictly non-decreasing -- 1.0 through 11.0 s, advancing at
essentially real time -- with scores at or above 0.96 and four of them at
0.999 or 1.000 against a runner-up in the 0.78-0.83 range. S00A is the
top match on 23 of the 27 frames carrying signal. So MS00A -> S00A.wmv is
decoded from the manifest AND measured off the game, and the intro begins
about 4.5 s after A on the save slot.
The previous attempt's failure is worth its own METHOD line, because it
did not look like a sampling problem: it looked like weak evidence for
the wrong film. Sparse sampling produced contrast-23 frames, a playhead
that would not join up, and one frame preferring ADV. Sampling does not
weaken a signal gracefully; it turns it into noise shaped like a
different answer.
One aside recorded and not chased: in the S00A 5.5-10 s window, ADV also
scores 0.97-0.99 at its OWN monotone playhead of 33.5-37.5 s. Two films
matching strongly with two consistent playheads is not noise -- it reads
as the boot movie being a trailer cut from the story cutscenes, which
also explains why the sparse run kept flipping between them.
this container kills long runs
Three results, and the middle one is deliberately not upgraded.
First: NEW GAME -> DIFFICULTY -> NORMAL -> SELECT DATA -> pick a slot ->
a movie plays, with no crash. The previous run's throw at 0x82307128 is
therefore not inherent to the menu path -- the same six presses got
through it -- which fits the incomplete-cache trigger the corpus already
names. n=1 either way, and the note says so. Worth recording because one
observation had already hardened once into "A on NEW GAME hangs", and it
would have hardened again into "the new-game path crashes".
Second: with the path open, the manifest's MS00A -> S00A.wmv finally got
a runtime test, and it did NOT reproduce the quality of the attract-movie
identification. The first four samples give a clean monotonic S00A
playhead of 2, 6, 9, 13 s and the strongest frame matches at 0.916, but
the run as a whole is not one advancing timeline, most frames sit at
contrast 23-37 where this page's own rule says discrimination is poor,
and one high-contrast frame prefers ADV. So it stays amber: consistent
with S00A, corroborating the manifest decode, not confirming it. A clean
answer needs the 30fps x11grab route rather than 5-second screenshots.
Third, and it bounds the rest: the emulator vanished at ~145 s with no
crash line in its own log, last entries routine MEM-WATCH at 1.15 GB RSS.
Second time this session -- the other was ~50 s into a boot. That is an
external kill on a box with ~1 GB free and swap exhausted, not a guest
fault. METHOD and the handoff's residue table both now say dynamic
experiments here have to fit in about two minutes of guest time, which is
why the DELTASABER plates went unseen again: they need a mission load and
the run never got there.
It set `git config --local credential.helper "store --file=$HOME/..."`, which
PERSISTS in the repository. The repo is a bind mount the host also uses, so the
host's git inherited `store --file=/sylph-home/port/.git-credentials` -- a path
that exists only inside the container -- and every host push then failed with
`unable to get credential storage lock: No such file or directory`.
Same root as the EBUSY credential error, from the other side: a tool that
configures a shared repository to suit itself breaks every other user of that
repository. Now applied with `-c` to the single push invocation.
Cleared the leaked setting from both checkouts by hand. Syplheed-Reborn had it
too, from the RE container's copy of this script.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
1/60 s, and the 60 Hz reading is excluded
ui-keyframe-time-unit.md named its own re-test and nobody had run it: time
300 submitted frames on the IDLE title, where nothing is streaming from
the ISO. Reaching the title is cheap now that one A skips the intro, so
this cost one boot.
Two clean trials: 300 frames in 10.40 s and 10.60 s, i.e. 28.8 and 28.3
fps. That is the same rate as the 27.6 fps measured during the loading
splash, which is exactly the page's own criterion for "the game is 30 Hz
and the unit is 1/60 s". The 60 Hz alternative is now excluded rather
than disfavoured: it needs the emulator at 47% of real time while sitting
idle on a screen the capture says costs 1526 draws over 300 frames, about
five per frame. Nothing there halves an emulator.
So the durations stand as written -- title 4.2 s, main menu 1.1 s, EXTRAS
0.87 s -- and Q1 leaves the handoff's open-residue table. It stays
MEASURED, not decoded: no field on the disc says sixtieths of a second.
One trap, and it nearly cost the conclusion. My first attempt reported
39.5 fps, which sits between the two hypotheses and would have left the
question open. It was a harness bug: I polled the log for a "done" line
rather than for the COUNT of them increasing, so it matched a capture
that had already completed and timed 0.1 s of nothing. METHOD gets the
general form -- a log-polling probe must compare a count, not test for
presence.
Three fixes from the port agent's first infrastructure report.
**A. The credential file is written, so it cannot be a read-only mount.**
`credential.helper=store` rewrites its file after a successful auth: temp file,
then rename over the target. Renaming onto a bind-mount point gives EBUSY, which
surfaces as `fatal: unable to write credential store: Device or resource busy`.
The push succeeds anyway, and that is the real hazard -- a `fatal:` line that is
routinely wrong teaches the reader to ignore the one that is real. It also fired
intermittently, so it read as flakiness rather than as a mount.
Fixed by mirroring the pattern already used for .claude.json: mount it as
`.git-credentials.host:ro` and have the entrypoint copy it to a writable
~/.git-credentials at 600. Mounting rw would also silence it, but then the
container can clobber the host's real credential file; copying cannot.
**B. `git -C /reborn pull` can never work, and should not.** /reborn is a live
read-only mount of the RE agent's working tree -- it updates itself, and pulling
would move another agent's checkout. The prompt now says so, and adds the
consequence the agent found the hard way: because the mount is live, HANDOFF can
move mid-iteration, so anything copied out of it (BLOCKED.md especially) may
already be stale and must be re-checked rather than trusted.
**C.** CARGO_HOME moves to a named volume; it was on the container overlay, so
the pinned decoder source was re-fetched from the network on every fresh start.
Also adds SYLPH_PORT_REPO, so this launcher can be run from a worktree without
repointing the agent's checkout -- which is how these edits were made, the agent
being mid-iteration on auto/p0-exporter in the shared tree.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The port agent's first published arc. Merged from auto/p0-exporter so the image
build below carries its docker/bin fixes -- build-export follows the exporter's
new subcommand shape, and build-reference-cli makes the pinned reference CLI
reproducible rather than a hand-built artifact in a volume.
residue in one place
Every MISSION row is answered, so this iteration checked the deliverable
itself rather than opening new scope: an answer that is not reachable
from HANDOFF.md has not been delivered, and I have been editing that page
by surgical replacement for a dozen iterations.
Four things were wrong or stale:
- the Q5 row still said initial focus was 2x TUTORIAL, 1x NEW GAME while
its own bullet said four boots -- I updated the body last time and not
the row;
- the Q2 bullet still described the other four main-menu destinations as
"likely by name -- an inference, not a measurement", which Q4 has since
MEASURED, and it listed GP_MISSION_SELECT as a main-menu destination
when MISSION SELECT lives under EXTRAS;
- the Q1 bullet claimed that if the game presents at 60 Hz "nothing else
on this page changes". That is now false: Q7's declared fade-in lengths
are keyframe-derived and would halve. The same bullet also missed that
the transition capture is a SECOND independent measurement supporting
1 unit = 1/60 s -- the quad is declared black for 12 units, 0.20 s
under the conversion, and the capture measured the plateau at
0.17-0.23 s. Recorded as corroboration, not proof, because the plateau
spans two screens' fades and cannot separate them;
- bgm-two-stems.md linked ../port/MISSION.md from inside structures/,
which resolves nowhere.
Also named the two videos actually in scope, since "3.3 GB of video, two
files are in scope" is not much use without the filenames.
And added a What Is Still Open section. With every row green the residue
was scattered across eight pages; it is now one table with why each item
is stuck. Two entries share a single unblocker -- an emulator whose audio
path can be observed -- and they are the only ones that would change a
port decision.
