Now:PdHud
draws the converted BtMenu sprites at the exe’s
one-view layout (own panel, health segments,
aura ring, aura recharge bars (decoded
(aurapips): frac(aura) × 10 + 1 bars
below the level, none at it, tint 0xff5a1e02),
deck box, clock, lock-on marker (see the rows
below), result logo); F3 /
--debug-hud swaps to the old debug
HUD (SlotHud)
Now:PdMenu._fit() (and the title, its
child) in the same frame: H / 480 per unit, the
640 area at frame x + 106.625 units
Where:scripts/menu/pd_menu.gd_fit()
Lead: the rectangle (24,
12, 616, 468) handed to the device (L: safe
area)
3-D view aspect
on 16:10
Status: decoded,
implemented (uifix build)
Now: The player camera
keeps the game’s projection over the frame (75°
across its width,
PdFrame.apply_camera every tick);
black bars on a top CanvasLayer
(main.gd)
Where:scripts/pd_frame.gd,
player.gd
Lead: the stage intro
(--intro) and char view cameras do
not use the frame yet
HUD text font
Status: MISSING (docs
11-10)
Now: Godot’s default font
with an outline, digits centred in the decoded
16-px cells; the name plate text colour and size
and the level number size are guesses
Where:pd_hud.gd_text(),
_digits(), NAME_TINT
Lead: decoded (docs
11-10): the width table 0x1402dd2c0, the
code-point → cell map 0x1402dd3c0, the glyph
quad and scale rule (8 / 16 = 1:1), corner
colours, the font2 blend, HUD
digits = atlas row 116 (FORMATS “Bitmap font and
text drawing”). Left: implementation (converter
font.json + atlas PNG through a BC
library, PdFont in Godot)
HUD
text positions (level number x, slot cost, plate
text)
Status: INFERRED
(hud)
Now: Placed by eye.
Decoded: health (+0xf, 4), aura (+0x11, 0xc),
deck (+7, +1) as 16-px cells 12 apart, level
number y +0x21, clock (+0x12), and the colours
of health (tier), aura, level and deck
numbers
Where:pd_hud.gdlayout(),
_slot()
Lead: level x in
0x14015249b; the slot text calls in 0x140152620.
Text metrics are decoded (docs 11-10),
so plate text width = the game’s width function
0x140247550
HUD slot icon
positions
Status: decoded (uifix
11-14, M), implemented (uifix
build)
Now: Icon top-lefts from
the ring base: top (19, −18), left (−18, 18),
right (55, 18), bottom (19, 57)
Where:pd_hud.gd SOCKETS
Lead: frame-by-frame check
against footage
HUD slot cost
number
Status: decoded (uifix
11-14, M), implemented (uifix
build)
Now: Black, 8 × 16 cells,
centred on icon x + 9 at icon y + 2 (glyphs:
Godot’s font, row “HUD text font”)
Where:pd_hud.gd_slot()
HUD name
plate side and layout
Status: text anchors
decoded and used (uifix build, M);
plate pieces INFERRED
Now: Text at the decoded
anchors (left slot right-aligned, 7 px per
character); plate caps just outside the text,
middle across it, 2 above the decoded plate
y
Where:pd_hud.gd_slot(),
NAME_AT, PLATE_Y
Lead: the plate piece
arithmetic in 0x140152b09 (caps at −0x1b / +5 /
−0x21)
HUD
lock-on marker: ring by range, aim point
Status: decoded
(lockring), built (lockhud
11-15)
Now:PdHud._lock() draws the ring of the
lock grade (15 orange ≤ 80 u, 20 + dot 23 yellow
≤ 300 u, 25 green beyond) from
PdLockMarker (the per-view record),
at the target’s torso root node
(DataSkill.target_point, no added
height), hidden when the lock tracks nothing or
the target is off screen; locked ally: sign 75.
Split-view sprites (120..134) not used (no split
screen)
Status: decoded, built
(lockhud 11-15); glyphs and name
PLACEHOLDER
Now: Track 14, needle 77
turned p × 1.8125 − 116° (clockwise positive,
settled on footage), underline 76 and number
int((d − 10) × 100) at the 65-point table,
advances 9 / 7 from the width table: done.
PLACEHOLDER: the number and
name glyphs are Godot’s font
(LOCK_NUM_FONT 13,
LOCK_NAME_FONT 14), and the name is
the body’s node name
Where:pd_hud.gd_lock(),
_lock_number()
Lead: glyph sheet of the
2-D context (cells 16 × 16, FORMATS “Lock-on
marker” draw step 4) and the rich string font;
player name table 0x1404bca30
HUD
lock-on marker: closing-in and ring swap
Status: decoded, built
(lockhud 11-15)
Now: 10-tick closing pieces
(e = 2·k², grade 2 in 0xffffd503), common cue 4
when closed (the same counter,
PdBattleCues.lock_tick →
PdLockMarker.tick), replay on a
switch / back on screen, 5-tick ring swap with
the scale-and-recentre offsets
Where:pd_lock_marker.gd,
pd_hud.gd
Lead: done; closing-in and
swap are code-only (no footage frame)
Skill
range test (targeted auto kind)
Status: decoded, built
(lockhud 11-15)
Now:PdLockOn.range_test (0x140273c80:
set 0 class 0 in range < 300 u, other classes
/ sets ≥ 300 u, set 0 class 2 always) applied by
AutoAttack.start for set 0,
behaviour 4, 0x7C = 0, from the start point to
the target’s torso point; out of range ends it
at cast (no playable record has 0x7C = 0, so
only the test drives it)
Now: Not drawn: the slot
plates show without the coloured card behind
them
Where:pd_hud.gd_slot()
Lead: 0x140151cb0 draws the
card-back quad (name string 0x1402f1538) in the
type colour of 0x1402f12f0 (FORMATS “Slot
colours”); its quad and shader are not
traced
HUD
aura point sound: who hears it
Status: INFERRED
(aurapips)
Now: Decoded: common bank
cue 5, no position, when the regen tick changes
the aura’s integer part (0x14024dee7 →
0x14024df1d). It is gated by flag
+0x30 of the player’s battle menu
object ([player+0x61a0]), read as
“this machine’s player”: we play it for the
HUD’s own body only
Now: Ours (the exe has no
netplay): a client regenerates locally and takes
the host’s aura from snapshots, so a point is
counted only when int(aura) rises above the
highest point already counted. That mark falls
to int(aura) on a local spend and otherwise
follows int(aura + margin) down each tick. The
margin is AURA_MARK_TICKS = 3 ticks of the
current regen (snapshots up to 100 ms late, the
INTERP_TICKS window; at most ≈0.16 at the
fastest decoded factor). The net autotest
measures the client-ahead gap per snapshot:
about 1 tick alone (0.04), up to about 3 ticks
in a loaded suite run, with 0–3 of ~1500 samples
just over the margin (each a possible double
sound). A spent point arriving by snapshot
lowers the mark whatever the drop size. A
snapshot that jumps past a point plays it (the
host’s regen crossed it). A tick or snapshot
crossing several points plays one per point (the
game’s own regen can’t cross two in a tick). A
skill gain (gain_aura: aura_boost,
the aura_blocked add-on, status effect 2) raises
the mark with the aura, so it plays nothing, as
in the game. Limits: (1) a skill gain on the
host reaches a client only by snapshot, so the
client plays one sound per point it crosses (the
snapshot doesn’t carry the mark). (2) A
snapshot-only spend that lands within the margin
below a point keeps the mark, so that point gets
no sound. (3) A snapshot more than the margin
behind the local regen (over 3 ticks late)
lowers the mark below a point that already
sounded, so the point sounds again
Lead: a mark (or a point
event) in the snapshot, so host-side skill gains
and late snapshots stay silent; footage of a
double or a late sound
Skill
card when standing at a capsule
Status: decoded
(capsulecard)
Now:PdHud._card(): the first own
capsule (by slot) within 1.5 m of the feet after
the 0.9 m drop, none with another player that
close, never for enemy capsules; the WusMenu
back, type band, school icon, cost / STR / dash
labels, uses icon and range label at the decoded
offsets, fields from skills.json
Where:pd_hud.gd,
<data>/wusmenu/
Lead: done (FORMATS.md
“Skill card”)
Skill card
text and numbers
Status: PLACEHOLDER
(capsulecard)
Now: Name, description and
glyphs in Godot’s font (sizes 14 / 12 / 16, no
outline) at the decoded spots; the description
wraps by Godot’s metrics at 0xe5
Where:pd_hud.gd_card()
Lead:System/dbfont.ddx (row “HUD text
font”); decoded (docs 11-10): 8×16 is
the atlas at 1:1, 7×14 is 7/8 scale, the 14 px
numbers are the wide set / digit strip of atlas
rows 116–119, word wrap rule 0x14028d630
(FORMATS “Bitmap font and text drawing”)
Skill card
DEF label sprite
Status: INFERRED
(capsulecard)
Now: Sprite 2 for defenses
(by its art; the defense branch’s register is
set outside the lines read)
Where:pd_hud.gd CARD_DEF
Lead: 0x14016902c and the
code before it
Skill
card uses glyphs, cost 20+, special values
Status: INFERRED / MISSING
(capsulecard)
Now: Uses as digits from x
0x8d only when the record has uses (table
0x1402f2890 not read); costs of 20 and up placed
like 10..19; non-plain damage (0xffff) and the
defense value −2 show no number; range classes
4/5 of the second classifier draw no label
Now:PdHud.aura_record(): skill id 0 =
record 0 (text index 301, cost 0) from the
re-converted skills.json; a stand-in only when
the data has no record 0
Where:pd_hud.gd,
converter/skilldata.py
Lead: — (data must be
reinstalled)
Aura
card name without record 0 (uifix r2
11-14)
Status: PLACEHOLDER
(fallback only)
Now: The aura card’s name
(HUD name plate, debug slot list, aura capsule
label) and card are record 0’s
(PdHud.aura_name() /
aura_record()). Only when the
loaded skills.json has no record 0 (old or
missing data) do they fall back to a stand-in
record named “Aura”, type aura, cost 0
Lead: none needed: install
the re-converted skills.json (record 0); drop
the fallback once every data root has it
Records
without a text entry
Status: decoded,
implemented (uifix build)
Now:skilldata.py text_entry(): a record
whose text index has no entry gets the table’s
first entry (two more usable records come out of
the 1.3 file this way)
Where:converter/skilldata.py
Skill
card number strip and frame (menus)
Status: strip decoded and
drawn (uifix build); frame PLACEHOLDER;
rarity marks MISSING
Now:PdHud.draw_card(style 1, frame):
the strip box and the three-digit number
(Godot’s font); a thin box for the window frame;
an empty place shows a “-” name (PLACEHOLDER for
the game’s one character)
Where:pd_hud.gddraw_card()
Lead: the window frame
routine 0x14016cf30 and its sprites; the rarity
mark sprite handles at object + 0x2a8 / +
0x2ac
Skill
card for a held slot button
Status: decoded
(capsulecard2)
Now:PdHud._card() /
held_slot(): the first of the four
slots whose button is held this tick (no hold
time), whatever it holds, is the one looked at:
with a skill it shows that card ahead of the
capsule card, also with another player near;
empty, it falls through to the capsule card (a
later held slot with a skill is not used);
casting is unchanged
Where:pd_hud.gd
Lead: done (FORMATS.md
“Skill card”)
Skill card gates
Status: INFERRED / MISSING
(capsulecard2)
Now: Modelled: the battle
menu flag bit 0x10 (menu_flags, on
at init; which scripts switch it is not traced,
so it stays on). Not modelled: the per-player
choose-from-the-deck menu (+0x3fc;
no skill of ours opens it), a held slot’s
override index +0x384 ≥ 0 (no card;
not identified, we never set it), the per-player
nibble +0x118 (hides the capsule
card only; not identified); the battle-loaded
state (== 3) always holds while our HUD
runs
Where:pd_hud.gd_card()
Lead: 0x1401582f0 (menu
opener, called at 0x140271221), script command
13 of the native table at 0x1403f1d50
(0x1402169a0, which scripts call it), writers of
the slot field +0x384 and of menu +0x118
Skill card
state 3/4 cases
Status: MISSING
(capsulecard)
Now: States 3 (list
0x14058bd30) and 4 (set at 0x14024e692) are not
identified
Where:pd_hud.gd
Lead: 0x140155160,
0x140251550
Debug
capsule spawn distance
Status: PLACEHOLDER
Now: 1 m ahead on the floor
(debug only; inside the card’s reach)
Where:player.gdspawn_capsule()
Lead: none (debug
tool)
HUD result
logo position
Status: decoded
(menus) in menu sessions; PLACEHOLDER
(centred) otherwise
Now: Menu sessions: the
battleend script draws it through the decoded
banner objects (win / loss: banner A + the
sprite at (127, 150), banner B at (513, 150);
draw: banner B at (320, 150)) and the prompt
line at (320, 412) fading in;
pd_hud.gd_result()
steps aside. Debug / chooser starts: our centred
logo
Where:pd_hud.gd_result(),
scripts/menu/pd_menu.gd group
8
Lead: the non-menu
starts
HUD
animations, status icons, other players’ panels,
slot flashes, range line
Status: MISSING
(hud)
Now: Not drawn (the
aura-state colour blend of +0x38c,
a status flash with a 16-tick timer
+0x390 (0x140150acd); its lerp is
0x140153640 and the health drain of
+0x2bc not applied; health colours
are the decoded steady state)
Where:pd_hud.gd
Lead: FORMATS.md “Battle
HUD”, Not decoded yet. Status icons and the
colour blend helper are decoded (docs
11-10) (row “Status cues, character effects
and HUD status icons”); the rest is still
open
Loading
screen (battlestart)
Status: INFERRED
(menus)
Now: Menu sessions: the
user’s battlestart script runs on
PdMenu over the stage load (tips and pictures
from the user’s files at run time) and closes
when the battle is ready (0x41CC); debug /
chooser / net starts: none
Where:scripts/menu/menu_flow.gd,
main.gd
Lead: the
wait-for-all-players part (one machine
here)
Results
list (places, knockouts, deaths, money)
Status: PLACEHOLDER
(menus)
Now: Decoded
(menus): the split-screen versus never
shows it (its scenes close on the confirm); only
online / system link do. Its data built-ins
answer from the battle (places, the knockout
tally from who went down in the 1-on-1, empty
slots 2 and 3); money and funds read 0 and names
are empty (no save profiles). Not reached by our
build (no online menus)
Lead: the money code;
profile names; the item types’ frame drawing
(table 0x1402dcaf0)
Rich
text escapes, inline button icons, wide capital
font
Status: MISSING (docs
11-10)
Now: Not supported by our
text drawing
Where:pd_hud.gd
Lead: decoded: 0x07 escapes
(columns, centre, right-align, mode), glyph
modes 0–4, atlas rows 116–122; the inline icon
set (0x140533c08, table 0x1403fa230) is only
located
Countdown,
KO and time-up art
Status: MISSING (docs
11-10); the countdown part done
(menus), row “Countdown cues and
sprites”
Now: HUD notes only (KO /
time-up)
Where:pd_hud.gd,
match.gd
Lead: not in the
battlestart /
battleend scripts; leads:
battle-object functions 0x140226420 and
0x140220a50 (HUD sprite handles 83 / 84), start
/ result logo sprites 81, 82, 84, 86