Fixes the CT7 gate finding: on the Titles tab, the authored divider
0x10000530 escapes the Character window above its top edge at the
CT6-correct 372px mounted default (computed Y ~ -178, matching the
owner's screenshot). Retail clips child rendering to the intersected
ancestor clip-rect chain -- UIRegion::DrawHere @0x0069FA30 takes the
element's screen Box2D plus a SmartArray<Box2D> of inherited clip
rects, intersects them (the min/max clamp loop @0x0069FAA7..0x0069FB82),
and draws EraseSelf/DrawChildren/DrawSelf with the intersected rect
only when non-empty (the var_24 gate @0x0069FB8E). Our UiElement draw
walk rendered children unclipped by default, so any authored element
relying on clipping -- this divider, and the chat input row at small
window sizes (the owner's earlier "text input sticks out on resize"
report) -- became a visible artifact.
Mechanism (element-level, reusing the existing clip-rect-stack
infrastructure in UiRenderContext.PushClip/PopClip):
- UiElement.ClipsChildren now defaults to TRUE for every element
(was an opt-in used only by UiScrollablePanel/UiItemList). Each
element's children draw AND hit-test clipped to the intersection
of its own rect with the inherited ancestor clip; an element
positioned outside its parent's box silently disappears, matching
retail's non-empty-intersection gate. HitTest's existing early
bounds check already implemented this shape for ClipsChildren=true
elements -- flipping the default aligns hit-testing with the new
draw-clip default in one property, per the plan's own point 4.
- UiElement.ExpandsClipForPopup (default false) is the one opt-out:
retail spawns a menu popup as a SEPARATE top-level region
(UIElement_Menu::MakePopup), clipped only by the screen; acdream
draws UiMenu's popup inline from the owning button in a second
traversal (OnDrawOverlay, pre-existing -- its own doc comment
already says "regardless of this element's position in the tree").
DrawOverlays now resets the accumulated clip to unbounded
(UiRenderContext.PushClipUnbounded, sharing the existing clip
stack) for exactly the OnDrawOverlay call of an opted-in element.
UiMenu overrides ExpandsClipForPopup=>true, paired with
ClipsChildren=>false so its own out-of-bounds OnHitTest union (the
popup occupies ly<0 or ly>=Height depending on open direction)
stays reachable through the same early-bounds gate that now
defaults on for every other element.
Opt-out audit (grep for OnDrawOverlay overrides + negative/overflow
OnDraw coordinates across src/AcDream.App/UI): UiMenu's popup is the
ONLY OnDrawOverlay override client-wide, so it is the only element
needing ExpandsClipForPopup. RetailTooltipPresenter's popup and
UiRoot's drag ghost both already escape structurally -- the tooltip
mounts as an ordinary UiRoot CHILD (sibling of every window, clipped
only by the canvas), and the drag ghost is drawn directly by UiRoot
outside the tree entirely -- neither needed a code change, both are
covered by new tests proving the invariant. UiResizeGrip and
UiNineSlicePanel's frame/bevel draw entirely within their own
[0,Width]x[0,Height] (grip flush at the window's own edges; the
window's own Width/Height already represents the OUTER frame
including its 5px bevel, so its ClipsChildren push already covers
the frame's own content children correctly -- no negative insets
found). UiScrollbar draws entirely within its own bounds (confirmed
by reading OnDraw).
Hit-testing: aligned with the new default via the single
ClipsChildren flip (see above); UiMenu's own opt-out override keeps
its popup hit-test union working, verified by the full UiMenuTests
suite staying green.
Divergence register: AD-113 filed for the ExpandsClipForPopup
adaptation (inline popup drawing vs retail's separate top-level
region).
Fixed two pre-existing test-harness gaps the new default surfaced
(both real bugs in the harnesses, not workarounds around the fix):
- ChatLayoutConformanceTests' bottom-right-grip grow test read a
STALE (pre-shrink) grip screen position because it drove two resize
gestures back-to-back with no intervening Draw pass -- the only
place UiElement.ApplyAnchor/LayoutPolicy.Apply run. A real frame
draws every tick, so production never hits this; the test now
inserts a real DrawSelfAndChildren pass between the two gestures,
matching a real frame boundary.
- VendorUiControllerTests' hand-built Items/Buying/Selling page
containers were left at their bare 0x0 UiElement default (the
harness never runs a real DAT-driven layout pass) -- harmless
before ancestor clipping existed, but now hides every child of an
unsized page. Sized them to the window's own content root, matching
production's shape (a tab page fills the window body).
Tests (all confirmed as genuine regression pins by temporarily
reverting the relevant default/override and observing the exact
predicted failure, then reverting back):
- CharacterTitlesControllerTests.TitlesPage_Divider_ClipsAwayAtThe
CT6Default_AndAppearsWhenTheWindowGrowsTaller: the literal gate
repro against the real character_2100002E.json fixture through
RetailWindowFrame.Mount at the CT6 372px default -- the divider
renders nothing (computed Y ~ -173, matching the owner's ~-178);
growing the window to 600px renders it at its authored spot.
- ChatLayoutConformanceTests.ResizingTheWindowSmall_NoInputRowQuad
RendersOutsideTheWindowRect: no input-row quad escapes the chat
window rect at three small sizes (300x100 sanity control,
120x40/80x30 genuine pre-fix overflow -- verified failing without
the fix at Y=38/55 past the window edge).
- UiAncestorClipTests (new file): the core mechanism against plain
synthetic elements (culled-outside / clipped-at-the-edge / hit-test
parity), UiMenu's popup escaping a tiny owning window (and staying
clipped while closed), and the tooltip's structural immunity
(mounts as a UiRoot sibling, unaffected by a tiny ancestor window).
Verification: full solution build green; hermetic suite green
(--filter "Lane!=InstalledDat&Lane!=PreparedPackage&Lane!=Live&
Lane!=Manual&Lane!=Timing&Lane!=Windows&Lane!=Linux&
Lane!=SystemFont&Purpose!=Diagnostic&Status!=KnownFailure",
14,000+ tests across every project); InstalledDat lane green
(ACDREAM_RUN_INSTALLED_DAT_TESTS=1, Status!=KnownFailure,
205+34+3+172 tests). CharacterTitlesControllerTests' existing suite
and the full UiMenuTests/UiScrollbarTests suites are unaffected.
src/AcDream.App/UI/UiRoot.cs carries an unrelated, pre-existing
uncommitted owner probe (ACDREAM_PROBE_UI_HOVER) -- untouched by
this change and deliberately left out of this commit.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Opus dual-lens review of CT6 (ec50455a) found 1 blocker, 4 should-fix, 5
notes. All applied:
BLOCKER B1 — the shared gmPanelUI host (0x100005FE) IS retail's own
outer window frame, not a content element: its authored 310/372/310/1000
already include the 5px bevel on every side. RetailWindowFrame.Mount was
adding the NineSlice wrapper's OWN 10px chrome inset on top of that
already-chrome-inclusive source, clamping MinWidth to 320 while the
window's actual mounted outer width stayed 310 — silently below its own
minimum until RetailWindowManager.ResizeTo forcibly widened it despite
ResizeX=false. Fixed with a new
RetailWindowFrame.Options.DatConstraintSourceIsOuterFrame opt-out
(chrome inset = 0 for constraint resolution only, value stays
DAT-sourced); MountCharacter sets it true. Mounted clamp is now exactly
the host's four raw values: width fixed 310, height 372..1000. Added a
mount-time invariant (throws if the mounted outer extent falls outside
its own just-computed clamp) that would have caught this at the first
test run.
S4 (campaign-lead ruling) — the window must MOUNT at retail's authored
default, outer 372 (content 362, matching the host's own content parent
0x10000180), not 0x2100002E's own 300x600 content-authoring canvas
(which produced a stale 610px default pre-fix: 600 + 10 chrome inset).
372 is exactly the host's own authored MinHeight — retail opens at its
resize floor and can only be dragged taller. MountCharacter now sets
ContentHeight=362f explicitly. At this default the 9 attribute/vital
rows (180px) overflow the 160px list immediately — retail-correct, not
a regression.
S2 — 0x1000023E and 0x10000533 both author property 0x79
(HideWhenDisabled) TRUE (fixture-verified: BoolValue=true on both). A
fitting list HIDES the scrollbar entirely; it does not draw a full-track
"disabled" thumb. The code was already correct; four wrong descriptions
(plan ledger, CharacterStatController comment, CT7 script, test comment)
are corrected, plus a new IsPresentationVisible assertion pair in the
resize test.
S3 — CharacterTitlesController's `if (listBox.LayoutPolicy is null)`
Anchors fallback was unreachable on both the real DAT and the fixture
(0x10000532/0x10000539 both author HasOriginalParentSize=true, so
LayoutPolicy is always assigned). Deleted; added an InstalledDatFact pin
guarding the deletion against DAT drift.
N4 — renamed NineSlice_ChatShapedConstraints_... to
NineSlice_ContentShapedConstraints_InsetArithmeticClampsProgrammaticResize
(it tested inset arithmetic on a content-shaped source, not chat's real
contract) and added a true chat-contract pin mounting Chrome=Imported
with chat's real 300/100/2000/2000 constraints, asserting no inset
applies.
N5 — corrected the "nothing inferred, no register row" sentences in the
ground-truth doc and plan ledger: they were false pre-fix (the mounted
clamp WAS an inferred double-counted composition); true now that B1
removes the composition.
CT7 script §4 rewritten with exact clamps (no "≈"), the corrected
default-overflow scrollbar behavior, and an absolute starting-height
statement.
Verified: full hermetic solution suite green (15,441 tests, Release,
Lane exclusions per the release gate), InstalledDat lane green across
the whole solution (414 tests, ACDREAM_RUN_INSTALLED_DAT_TESTS=1,
Status!=KnownFailure) including two new pins
(TitlesListAndPage_AuthorHasOriginalParentSize,
Imported_ChatContract_ClampsAtAuthoredBoundsWithNoChromeInset).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
CT6 (Campaign CT slice 6): the resize clamp source is the SHARED
gmPanelUI host (0x100005FE in LayoutDesc 0x2100006E), not 0x2100002E's
own root and not the Character/Skills slot 0x1000018E either — live
probe confirmed the host authors MinWidth=MaxWidth=310 (fixed — no
horizontal Resizebar authored), MinHeight=372, MaxHeight=1000, and that
the bottom Resizebar (0x10000660) and top Dragbar (0x1000065C) are
direct children of the host, not the content parent. Decomp chain:
UIElement_Resizebar::StartMouseResizing @0x0046B7E0 calls
UIElement::StartResizing(this->GetParent(), ...), stashing drag state on
that parent; UIElement::MouseResizeElement @0x00461130 then reads
GetAttribute_Int(this, 0x3C..0x3F) off that same element every
mouse-move.
RetailUiRuntime.MountCharacter now imports the host element and passes
it as RetailWindowFrame.Options.DatConstraintSource, matching the
existing MountSideVitals pattern.
CharacterStatController.RebuildActiveList now wraps BOTH the Attributes
and Skills tabs' rows in the same UiScrollablePanel viewport (previously
only Skills got one; Attributes rows had no clipping/scrolling and the
shared scrollbar was force-hidden — owner report item 2). The shared
scrollbar is now always bound + visible; UiScrollbar's own
IsPresentationVisible/IsModelDisabled already draw the correct
full-track "disabled" thumb when content fits. This surfaced and fixed
a real #372/#412-class anchor-baseline bug: the viewport's
Left|Top|Bottom anchor was capturing its baseline margins lazily on its
own first ApplyAnchor call, which happens AFTER the ListBox has already
grown from its raw DAT height to its mounted height, permanently
capping the viewport short on every later resize. Fixed with an eager
CaptureCurrentAnchorBaseline() call, mirroring UiTemplateListBox
.Viewport's own lazy getter.
CharacterTitlesController.Bind gained the same defensive
Anchors = Left|Top|Bottom fallback for the Titles ListBox that
CharacterStatController already had (a no-op on the real DAT — both the
Titles page and its ListBox already carry a real authored LayoutPolicy
that stretches correctly).
Standardization audit: UiElement.MinWidth/MinHeight/MaxWidth/MaxHeight,
set once at RetailWindowFrame.Mount, are the ONLY clamp fields — read
identically by interactive drag, RetailWindowManager.ResizeTo, and
RetailWindowLayoutPersistence's restore clamp. No gaps found; no
register row (every number is a live-probed authored DAT value or a
structural correctness fix, nothing inferred).
Tests: CharacterStatControllerTests
.CharacterWindow_ResizesYWithinAuthoredHostClamp_AndReflowsListAndScrollbar,
CharacterTitlesControllerTests
.TitlesList_ReflowsWithWindowResize_AndScrollbarOverflowFlips,
RetailWindowFrameTests
.NineSlice_ChatShapedConstraints_ClampProgrammaticResizeAtAuthoredBounds
(shared-mechanism regression pin), CharacterPanelLiveDatTests
.PanelHost_AuthorsFixedWidthAndBottomOnlyResizeContract (InstalledDat
pin). Existing attribute-row tests updated from list.Children to
Descendants(list) for the new nested-viewport shape (the pattern skill
rows already needed).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Opus review of CT5 (f532f28c) found 0 blockers, 4 SHOULD-FIX, 6 NOTEs; all
applied here.
SHOULD-FIX 1 (visible retail gap): unselected attribute/skill rows now draw
the row template's Normal-state media (0x06004CC2 — pinned by
CharacterPanelLiveDatTests.AttributeRowTemplate_...) instead of drawing
nothing. Independently decoded against the installed DAT: PFID_A8R8G8B8,
48x48, uniform (0,0,0,175) — a ~69%-opaque black tile the native-size
copy-or-tile blit repeats across the row. Wired at all three sites
(HandleRowClick, ApplySkillSelectionVisuals, AddRow). Selected rows keep
0x06000F93 (RowHighlightSprite) unchanged.
SHOULD-FIX 2: added Bind_AttributeRow_/Bind_VitalRow_/Bind_SkillRow_
geometry tests asserting the authored template pixels against BUILT rows
(not DAT pins) — width, height, icon/name/value column positions. The
skill-row case reproduces the real production scrollbar (X=281, per
CharacterPanelLiveDatTests.StatListBox_AuthorsFiveRowTemplatesInSharedLayout)
to prove the documented 281px clamp (scrollbar.Left - list.Left), one pixel
short of the attribute/vital rows' 282px ceiling.
SHOULD-FIX 3: AddRow's name-column Padding corrected from 1f to 0f — the
authored template carries no margin on 0x1000012A; Padding=1f re-created
the X=26 glyph-start bug this slice existed to fix.
SHOULD-FIX 4: reworded both UiPanel.BackgroundSprite doc comments — the
draw is a native-size copy-or-tile blit (UV-repeat), never a stretch.
Decoded 0x06000F93 as exactly 282x20 (matches the row natively, draws as a
plain copy) vs 0x06004CC2's 48x48 tile. Retail's UIRegion::SetImageByDID
(@0x0069F960) decompiles to a pure BlitMode selector switch — param_2==2 ->
Blit_3Alpha, ==3 -> Blit_4Alpha, else Blit_Normal — with no width/height
touched anywhere in the function, answering CT1's open "draw mode 3"
question: it's an alpha-blend selector, not a resize flag.
NOTEs:
a. AddRow's nameEl now sets OneLine=true so the authored VJustify=Center
takes the same single-line vertical-centering path the value column
already uses.
b. Tempered the "row width is 282" wording in the SkillContentWidth /
RowContentWidth doc comments — that's a ceiling attribute/vital rows
land on, not a fact true for skill rows (281, via SkillViewportWidth's
scrollbar-gutter measurement).
c. Reworded the section-header (RowPadX) comment — CT1 verified only the
four header SPRITES; the caption label's own authored margins were
never checked. Recorded as an open residual, not a cleared divergence.
d. AttrRows/VitalRows are now internal (InternalsVisibleTo("AcDream.App.Tests")
already covers AcDream.App.Tests); CharacterPanelLiveDatTests iterates
them directly instead of a re-typed duplicate array, and now also
asserts the vitals 2/4/6 current-enum aliasing claim the doc comment
made but never enforced.
e. Deleted the stale pre-CT5 0x06001397 narrative in
CharacterPanelLiveDatTests; the pin's comment now describes the
post-CT5 state (a regression guard, not an open bug).
f. Unified ApplySkillSelectionVisuals' selected-branch SpriteResolve
wrapper closure with HandleRowClick's direct assignment.
Build green; full hermetic solution suite green (Release,
Lane!=InstalledDat&...&Status!=KnownFailure filter); InstalledDat lane
green (ACDREAM_RUN_INSTALLED_DAT_TESTS=1, Lane=InstalledDat&Status!=KnownFailure).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Aligns the hand-built attribute/skill rows in CharacterStatController with
the authored shared row template 0x10000248 (LayoutDesc 0x21000045,
InfoRegion::InfoRegion @0x004F1450 template index 0 — the same template
gmAttributeUI and gmSkillUI both instantiate):
- Row geometry replaced with AUTHORED PIXEL VALUES instead of derived
fractions: icon flush left 20x20 (was 16x16 at X=4, vertically
centered), name column X=25 W=150 fixed (was RowPadX+IconSize+IconGap
offset with a width*0.60 fraction), value column X=175 W=100
right-justified (its right edge sits 7px short of the row's 282px
right edge — the authored gutter the owner reported). Row width itself
now clamps to the authored 282px template width (RowContentWidth)
rather than the ListBox's raw 300px container width. Attribute-row
height fixed at 20px (was 22px, no dat basis); SkillRowHeight folded
into the same RowHeight constant since both row kinds share H=20.
- RowHighlightSprite corrected from 0x06001397 to 0x06000F93 — CT1's
ground-truth research sealed the verdict that gmAttributeUI::
UpdateSelection @0x0049DEE0 (SetState(6) -> InfoRegion::SetState
@0x004F0EE0) swaps the row's Highlight-state media (0x06000F93), a
full-row background swap. 0x06001397 belongs to a different mechanism
entirely (the spellbook row's UIElement_UIItem::SetSelectedState
overlay child) and SpellbookRowStyle.cs is untouched.
- UiClickablePanel.UseSelectionBars/SelectionBarHeight retired outright
(UiPanel.cs): they existed only to emulate 0x06001397's dark-bars art;
the correct retail rendering is the full-panel sprite stretch the base
UiPanel.OnDraw already performs, so the override is dead code once the
correct sprite is used. No consumer existed outside
CharacterStatController.
- Per-attribute/per-vital icon DIDs now resolve through the live
DBObj::GetDIDByEnum chain (RetailDataIdResolver.Resolve, AP-235's
unification seam) when a resolver is supplied — RetailUiRuntime.
MountCharacter wires one under the shared DatLock — falling back to
the hardcoded AttrRows/VitalRows column otherwise (tests, no dat).
gmAttributeUI::PostInit @0x0049DB70 read verbatim: attributes resolve
via category 0x10000002 (statId order 1,2,4,3,5,6, matching AttrRows'
authored display order exactly); vitals via category 0x10000003.
Live-DAT-verified: every hardcoded fallback value already matched the
resolved DID byte-exact (new InstalledDat pin
AttributeAndVitalIconDids_MatchTheRetailEnumMapperChain).
- RetailAppraisalNameResolver.ResolveHeritage's independent
re-implementation of the 2/5/13 heritage overrides deleted; it now
delegates straight to CharacterIdentityText.HeritageGroupDisplayName
(which already bakes in the same overrides) — one owner, byte-identical
behavior. AP-235's register row updated to reflect the single-owner fix
(the underlying hardcoded-vs-live-DAT mechanism divergence itself
stays open — out of CT5's scope).
Hand-built-vs-template ruling: rows stay HAND-BUILT rather than
converting to UiTemplateListBox instantiation. The hand-built path hits
every authored number byte-exact (proven by the CT1 InstalledDat pin
AttributeRowTemplate_IconIsFlushLeftTwentyPixels_NameAndValueAreFixedColumns),
while conversion would touch ~15 call sites (raise-button affordability,
footer State A/B, per-row tooltip, section bucketing, live-refresh,
selection-highlight) for a geometry-only slice — smaller-risk path per
the task's own judgment-call guidance.
Tests: CharacterStatControllerTests' sprite/UseSelectionBars assertions
corrected to the authored geometry; new InstalledDat pin for the icon-DID
chain. Full hermetic solution suite green (App/Core/Runtime/Headless/
Launcher/Content/etc., 0 failures) and the full InstalledDat lane green
(203 App.Tests pins, TowerAscentReplayTests' known Status=KnownFailure
case excluded per the acceptance filter).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Opus dual-lens review of ed652ed8 found 2 blockers + 5 should-fix. All applied.
BLOCKERS:
- Bind the luminance pair (0x100005C5/0x100005C6): caption "Luminance:"
(UTF-16, PE-byte-decoded from the gmStatManagementUI vftable-adjacent
data at @0x007c3dd4) and value "<available> / <maximum>" (narrow
"%s / %s" @0x007c3dcc) — both literals independently re-derived from the
raw acclient.exe bytes and confirmed byte-exact against the review's
claim. Numbers format through a new shared FormatXp helper
(.ToString("N0", InvariantCulture) — retail's ExperienceSystem::XPToString
equivalent), also now used by Total XP / XP-to-next-level (previously an
un-invariant bare "N0"). Hide path switched from Visible=false to
retail's own UIElement_Text::ClearAllText mechanism
(@0x004f0e31/@0x004f0e3c — empty LinesProvider, leave layout); each
LinesProvider re-reads data() on every draw, so no separate refresh call
is needed.
- CharacterIdentityText.StripLeadingArticle deleted: retail AppendText's
the resolved title VERBATIM (@0x004f0990); 26 real ACE CharacterTitle
entries begin with "The" and were being mangled. The dead
CharacterSheet.Race fallback is deleted alongside it — retail's
InqGenderHeritageDisplay creature-type argument is a hardcoded literal 0
(@0x004f08db), no producer exists.
SHOULD-FIX:
- PK line re-sourced: classifies off the live ClientObject.PublicWeenieBitfield
PWD bits (0x20 IsPK / 0x02000000 IsPKLite — ACCWeenieObject::IsPK/IsPKLite
@0x0058c8b0/@0x0058c8a0) instead of a bitwise test against raw
PropertyInt 134, which carries ACE's own PlayerKillerStatus enum bit
layout, not the PWD layout. PropertyInt 134 already drives the correct
bits via the existing PlayerKillerStatusBitfield.Apply; this is a
re-source, not new wiring. Deleted the 0x4|0x8 combined-flag test case,
which asserted a non-retail answer.
- Register AP-109 row: restores CT3's Titles-page narrowing paragraph
(CT4's edit had compressed it to a bare pointer phrase), corrects the
rank-prefix source to PropertyInt 0x1E (AllegianceRank) read live off
the qualities bundle — not RuntimeAllegianceState, which is a different
UI's (SocialAllegiancePageController) own documented substitute —
corrects the title-table size from an estimated 22 functions/~200
strings to the actual 17 functions/~170 strings (AllegianceSystem::GetTitle's
dispatch switch read directly), and downgrades the evidence claim.
Filed AP-235 for the gender/heritage hardcoded-table-vs-live-EnumMapper
mechanism divergence, pointing at the ALREADY-EXISTING
RetailDataIdResolver.Resolve helper as CT5's unification seam.
- CharacterPanelLiveDatTests.HeaderElements_AuthorExpectedFontsAndColors
extended with the luminance pair's own occurrence-count + font/color
pins, matching every other header id's pattern.
Also landed: an InstalledDat pin
(GenderHeritageDisplayNameTables_MatchTheRetailEnumMapperChain) proving
CharacterIdentityText.GenderDisplayName/HeritageGroupDisplayName match the
live retail EnumMapper chain (master map category 1 ->
ClientEnumToID[0x10000001]/[0x10000002] -> EnumMapper DIDs
0x2200000A/0x2200000B) byte-exact, including the two entries the review
flagged as unverified guesses (10 "Penumbraen", 12 "Olthoi" — both
correct). CharacterSheetProvider.BuildSheet's level read switched from a
GetInt+ContainsKey double lookup to one TryGetValue. Plan ledger's
test-provenance sentence corrected (Bind_HeaderElements_... predates CT4,
extended to cover PkStatusId).
Tests: CharacterStatControllerTests (verbatim title incl. "The Noob",
luminance content/gate, luminance text binding, extended
Bind_HeaderElements_... covering PkStatusId), CharacterSheetProviderTests
(PK status driven through ClientObjectTable.UpdateIntProperty instead of
a raw property write), CharacterPanelLiveDatTests (luminance pin, gender/
heritage EnumMapper pin). Full hermetic solution suite green under Release
(0 failures, 15 projects); InstalledDat pins green (197/197, excluding one
confirmed pre-existing unrelated failure — TowerAscentReplayTests, verified
to fail identically with these changes stashed out).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Retires the rest of AP-109's UI half: the character panel's Name/Heritage/
PkStatus/Level header identity block is now live and DAT-faithful on both
Attributes and Skills pages (verified: CharacterStatController.Bind already
scopes Label/LabelAuthoredColor to the ONE physically-visible page container,
so both tabs share the same bound widgets).
- Name/Heritage/PkStatus/Level switch from hand-picked Body/Gold runtime
colors to the widget's own authored DefaultColor (LabelAuthoredColor) —
CT1's live-DAT pin (HeaderElements_AuthorExpectedFontsAndColors) confirmed
all four already carry the correct FontColor (white/white/white/pale-gold
with Outline); the former "runtime color, dat carries none" comment was
false.
- PkStatus resolves through StringTable 0x23000001 by key
(ID_StatManagement_Header_PKStatus_PK/_PKL/_NPK) with a bitwise
IsPK/IsPKLite test (gmStatManagementUI::UpdatePKStatus @0x004F00A0) instead
of the prior exact-equality switch, which silently dropped combined-flag
PlayerKillerStatus values. Live-DAT-verified strings: "Player Killer" /
"Player Killer Lite" / "Non-Player Killer" (new InstalledDat pin
PkStatusKeys_ResolveExpectedAuthoredStrings).
- Level shows "%d"-formatted InqInt(0x19) or the PE-recovered literal "???"
when absent (CharacterSheet.Level is now int?).
- Heritage line appends CT2/CT3's resolved RuntimeCharacterTitleState
display title through CharacterTitleResolver, refreshing live on both
TableReplaced (0x0029) and DisplayTitleChanged (0x002B) —
CharacterSheetProvider's ChangeBinding now subscribes to both.
- Name-line ruling: ships the PLAIN-NAME case only. Retail's allegiance
rank-title prefix (AllegianceData::GetFullName @0x005B6950 ->
AllegianceSystem::GetTitle @0x005B8DD0) needs a ~200-string, 22-function
heritage x gender table (verbatim decomp literals, e.g.
GetAluvianMaleTitle @0x005B7BC0's Yeoman/Baronet/.../High King) judged out
of reasonable size for this slice. RuntimeAllegianceState already carries
the local player's own rank; only the string table is missing. Registered,
not silently omitted.
- Luminance pair (0x100005C5/0x100005C6): CharacterSheet.AvailableLuminance/
MaximumLuminance (PropertyInt64 6/7) already flow generically through both
the PlayerDescription snapshot and the live 0x02CF private-update parsers
(no wiring gap). The retail show/hide gate (Level >= 200 &&
MaximumLuminance != 0, UpdateExperience @0x004F0A70) is wired and toggles
Visible on both elements every sheet refresh; the exact caption/value text
could not be recovered this slice (retail's SetText source resolves
through a Binary-Ninja-mislabeled data pointer, not a StringTable key — a
DAT string-table sweep found no match), so content stays unbound rather
than guessed.
- AP-109 narrowed accordingly (register row amended in the same commit).
Tests: CharacterStatControllerTests (heritage composition + live title
update, name stays plain, level int/"???" with authored — not constant —
color across 3 cases, PK line shows resolved text in authored color across
3 statuses, luminance visibility across 5 level/luminance combinations) and
CharacterSheetProviderTests (PK key-by-status resolution including a
combined-flag case, no-resolver leaves PkStatus null, Level null-vs-present,
title resolution + live refresh on both title events + unsubscribe-on-
dispose, luminance Int64 read-through). Full hermetic solution suite green
under Release (0 failures across all 14 test projects); InstalledDat pins
green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Opus dual-lens review of CT3 (03e073b7) found 1 BLOCKER + 2 SHOULD-FIX + notes.
BLOCKER: CharacterTitlesController never ported Refresh @0x0049abc0's own
SetSelectedItem(nullptr, 1) (@0x0049ac5a) — retail clears the current
title selection UNCONDITIONALLY on every Refresh() call, regardless of
whether the previously-selected id is still earned. OnTableReplaced
(0x0029) and OnDisplayTitleChanged (the display half of 0x002B) are
retail's two Refresh() call sites, so both now clear _selectedTitleId
before rebuilding/re-highlighting. OnTitleAdded (0x002B's add half) is a
DIFFERENT retail method — RecvNotice_AddCharacterTitle @0x0049a990 splices
one row without ever touching m_pSelectedItem — so it deliberately still
preserves selection. Net effect: after the user sets a display title and
ACE echoes 0x002B, the previously-highlighted row now goes dark and the
Set-as-Display button re-ghosts, matching retail; earning a new title
while a row is selected still leaves that selection alone.
SHOULD-FIX: ported AddTitleToList @0x0049A840's early-outs
(@0x0049a873/@0x0049a914) — an id of 0, or an id CharacterTitleResolver
fails to resolve, now produces NO row at all. The "Unknown" fallback
literal belongs only to the display-title text (Refresh @0x0049abc0's
other half), never a row — this was previously ported backwards.
SHOULD-FIX: rows and the display text now use their UiText's own authored
DefaultColor instead of a hardcoded Vector4.One, and each LinesProvider
now returns a cached UiText.Line[] built once per text change instead of
allocating a fresh array literal every draw call (pattern:
CharacterCreationSkillsPage.cs:829).
Notes (all ruled in): corrected two CharacterStatController comments that
falsely claimed the Titles page authors its own copies of the raise
buttons (verified against the fixture — it does not; the hide loop that
comment guarded is a defensive no-op given Visible's draw/click cascade,
kept only for the contentPage-not-found fallback); switched the row sort
from List.Sort to a stable OrderBy/ThenBy (ties broken by title id) so
equal-text rows keep retail's insert-after-equals order; wrapped the
title-resolver delegate in RetailUiRuntime.MountCharacter with the same
DatLock the row-template resolver already takes (DatCollection is
documented not thread-safe); set the list box's authored 24px row height
so wheel/line scroll lands row-aligned; kept the bind-time display-text
refresh with a comment explaining why the pre-notice "Unknown" frame is
unreachable in live play (ACE always sends 0x0029 before this panel can
open).
Tests: inverted TableReplaced_SelectedTitleStillEarned_KeepsSelectionHighlighted
into TableReplaced_ClearsSelection_EvenWhenTheSelectedIdIsStillEarned (cites
@0x0049ac5a), added its DisplayTitleChanged twin, and added
TitleAdded_PreservesSelection (the case most at risk from the blocker fix).
Inverted Rows_UnresolvedTitle_ShowsRetailUnknownLiteral into
Rows_UnresolvedTitle_ProducesNoRow (cites @0x0049a873/@0x0049a914) and added
Rows_TitleIdZero_ProducesNoRow for the other early-out. Extended
ClickingSetDisplay_..._AndMutatesNothingLocally to assert the row set and
selection are untouched by the click. Added
Fixture_PageCaptions_ResolveToNonEmptyText, which rebuilds the committed
character_2100002E.json fixture with a stub string resolver to pin this
class's own claim that the two page captions (0x1000052E/0x10000531) carry
a resolvable authored StringInfo.
Verified pre-existing/unrelated: the full hermetic suite run surfaced 2
failures in AcDream.App.Tests (LiveEntityNetworkBranchRoutingTests IL-shape
assertion, GameWindowRenderLeafCompositionTests IL-shape assertion) that
also fail with these five files stashed back to their pre-fix-round state —
confirmed unrelated to this change.
Build green. CharacterTitlesControllerTests: 24/24 (was 21, +3 net after
one invert-and-split and two new facts). Full hermetic solution suite
(Lane!=InstalledDat/PreparedPackage/Live/Manual/Timing/Windows/Linux/
SystemFont, Purpose!=Diagnostic, Status!=KnownFailure): only the two
pre-existing IL-shape failures above; every other project green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Titles tab (AP-109's known-inert gap) now switches to a real page and
CharacterTitlesController binds it entirely through UiTemplateListBox/
UiScrollbar/UiButton — zero bespoke widgets, matching every other
social/options row-list page in this codebase.
Retail anchors: gmCharacterTitleUI::PostInit @0x0049A610; AddTitleToList
@0x0049A840 + FindSortedInsertPosition @0x0049A760 (rows sorted by
resolved display text — this port rebuilds the full sorted set on every
change rather than a positional splice, since UiTemplateListBox has no
insert-at-index primitive and no other consumer needs one either);
InfoRegion::SetState(selected?6:1) (row Highlight/DirectState swap, the
same mechanism CT1's SEALED VERDICT confirmed for the stat rows);
UpdateButtons @0x0049A500 CORRECTED direction (Ghosted unless a row is
selected whose id differs from the current display title — no selection
IS the Ghosted case); Refresh @0x0049abc0 (display-title text, including
the hardcoded "Unknown" fallback, refreshed on both TableReplaced and
DisplayTitleChanged per CT2's review anchor 1); Event_SetDisplayCharacterTitle
@0x006a5720 (wire-only TitleSet 0x002C send, no local mutation).
CharacterStatController.Bind now three-way switches Attributes/Skills/
Titles — Titles is a genuinely separate, non-duplicated page container
(CT1 ground truth §3), unlike Attributes/Skills which share one mounted
page and only rebind content.
The two page captions (0x1000052E/0x10000531) are left untouched:
LayoutImporter.BuildText already resolves every element's authored
StringInfo caption at import time, so no controller-side string lookup
was added.
New IGameRuntimeCommands.SetTitle seam on DeferredGameRuntimeStateCommands
(InteractionUiRuntimeSources.cs) mirrors the existing Advance() shape.
CharacterRuntimeBindings gains Titles/TitleResolver/SendSetTitle;
CharacterTitleResolver (CT2) is constructed once at composition time and
its .Resolve method group is passed to the controller as a delegate
(not the concrete DAT-backed type) so the controller stays hermetically
testable without a live IDatReaderWriter.
Tests (tests/AcDream.App.Tests/UI/Layout/CharacterTitlesControllerTests.cs):
binding-seam coverage against the REAL committed character_2100002E.json
fixture (verified this session to already carry the Titles page subtree,
including the ListBox's own authored TemplateList=[(0x2100005E,
0x10000536)] entry — RowTemplateResolver_ReceivesTheFixturesOwnAuthoredTemplateIds
proves the controller reads that authored pair, not a hardcoded one); a
hand-authored ElementInfo standing in only for the row template itself
(a separate LayoutDesc with no committed fixture yet — CT1 was a live-DAT
probe only); sorted-row order, Unknown fallback, row selection/highlight,
the ghost truth table (no selection / selected==display / selected!=
display), click-sends-exactly-one-SetTitle-and-mutates-nothing,
click-while-ghosted-sends-nothing, TableReplaced rebuild (including
selection survival when the id is still earned), TitleAdded single-row
growth, DisplayTitleChanged text+ghost refresh, and Dispose
unsubscription. CharacterStatControllerTests updated for the Titles tab
no longer being ClickThrough, plus a new tab-switch visibility test.
Register: amends AP-109 (docs/architecture/retail-divergence-register.md)
to record the Titles-page half as LIVE; the header identity block and
luminance fields remain open for CT4.
Suites: full solution 15,405 tests / 0 skips (App 6,130) green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Opus dual-lens review of CT2 (bcfddc97) found 4 SHOULD-FIX + notes; this
applies the campaign lead's rulings.
F1 (the important one): retail's client-side table add is DEDUPED —
gmCharacterTitleUI::RecvNotice_AddCharacterTitle @0x0049a990 walks
mTitleList and returns without effect when the id is already present,
only inserting on a miss. The server-side SendNotice_AddCharacterTitle
broadcast is unconditional, but RuntimeCharacterTitleState.ApplyUpdateTitle
models the CLIENT receive side, so TitleAdded now fires only on a genuine
new membership. Inverted the pin:
ApplyUpdateTitle_AlreadyEarnedId_DoesNotFireTitleAddedOrBumpRevision.
F3: removed the send-side titleId==0 rejection from both command
adapters. Retail's own send path (Event_SetDisplayCharacterTitle
@0x006a5720) packs whatever id it is handed, and ACE accepts id 0
(CharacterTitle.Invalid is a defined enum value) — retail's real
protection is the UI ghost-when-current gate (CT3's job), not a
send-side rejection. No register row: this makes acdream MORE
retail-exact.
A2: ResetSession now publishes TableReplaced unconditionally and
DisplayTitleChanged when the display id was non-zero before the clear,
matching the LocalPlayerState.Clear() precedent (publish every category
even when Clear is repeated, so a failed reset can converge on retry).
A3: RuntimeCharacterState.CaptureOwnership reads the new non-allocating
Titles.Count instead of EarnedTitleIds.Count; EarnedTitleIds now carries
an XML warning that every read allocates.
A4/A5: ReplaceTable/ApplyUpdateTitle now mutate under one _gate hold with
change flags computed inside the lock and events raised after release;
every revision bump is gated on an actual state change (a no-op wire
resend produces zero revision edges), matching the change-gated
RuntimeMovementSkillState precedent. TableReplaced itself still fires
unconditionally per retail's own Refresh() dispatch on 0x0029.
A1/A6/A7/A8: CharacterTitleResolverLiveDatTests honors ACDREAM_DAT_DIR
first (CT1 fix-round pattern); documented the EmitResult
primaryObjectId-as-title-id precedent inline; corrected the "third
consumer" comment (CT1 §5 already records gmAttributeUI::PostInit's
icon-DID lookup — CT5 factors the shared GetDIDByEnum helper); added a
titleId -> resolved-string memo to CharacterTitleResolver, the DAT-static
equivalent of retail's lazy-hash cache on the string buffer.
Appended a "CT3 anchors from the CT2 review" list to the plan doc's CT2
ledger entry for CT3 to consume.
Build green. Runtime (102), Core.Net (12), and App (27 + 3 InstalledDat
pins under ACDREAM_RUN_INSTALLED_DAT_TESTS=1) title-scoped tests pass.
Full hermetic solution suite (Lane exclusions per the release gate) is
green: 0 failures across all 15 test projects.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Campaign CT slice CT2: the client now learns the character's earned
titles and current display title from the server, owns that state in
Runtime, and can send a display-title change. No UI (CT3/CT4).
Wire (Core.Net):
- GameEvents.ParseCharacterTitleTable (0x0029 CharacterTitle): retail
CharacterTitleTable::UnPack @0x005c6e90 skips a leading u32 into no
field — its own Pack @0x005c6e40 always writes the literal 1 there,
matching ACE's unconditional Writer.Write(1u) — then reads
displayTitleId, then a count-prefixed PList<uint> of earned ids.
- GameEvents.ParseUpdateTitle (0x002B UpdateTitle): titleId +
setAsDisplay, per CM_Social::DispatchUI_AddOrSetCharacterTitle
@0x006a54c0 -> Handle_Social__AddOrSetCharacterTitle @0x00564260,
which ALWAYS adds (SendNotice_AddCharacterTitle, unconditional) and
additionally sets display only when setAsDisplay != 0
(SendNotice_SetDisplayCharacterTitle, gated).
- SocialActions.BuildTitleSet / WorldSession.SendSetTitle: outbound
TitleSet (0x002C), u32 titleId, matching ACE's GameActionSetTitle.
- GameEventWiring gains onCharacterTitleTable/onUpdateTitle delegate
holes (Core.Net cannot reference AcDream.Runtime directly).
Runtime:
- New RuntimeCharacterTitleState (RuntimeCharacterState.Titles): earned
title id set + display title id, TableReplaced/TitleAdded/
DisplayTitleChanged events matching retail's unconditional-add /
gated-display-set contract, clears at generation reset.
RuntimeCharacterOwnershipSnapshot/CaptureOwnership/IsConverged and
RuntimeCharacterSnapshot extended (trailing optional fields, no
existing call site broken).
- IRuntimeCharacterCommands.SetTitle: generation-gated, sends
TitleSet only — NO optimistic local mutation. Verified against
retail's own CM_Social::Event_SetDisplayCharacterTitle @0x006a5720,
which sends the wire message and touches no local field; the display
title updates only from the server's own echo (the CA-campaign
lesson: never re-add an optimistic write). Implemented on both hosts
(DirectGameRuntimeCommandAdapter direct-send;
CurrentGameRuntimeCommandAdapter via LiveCommandBus /
LiveSessionCommandRouter's new SetTitleRuntimeCmd).
- LiveSessionEventRouter wires the two inbound events unconditionally
(RuntimeCharacterState.Titles is a required child, not an optional
sibling like Fellowship/Allegiance).
App (non-UI plumbing + resolver):
- CharacterTitleResolver (src/AcDream.App/UI/Layout/): ports
CharacterTitleTable::GetCharacterTitleFromID @0x005c6ed0 — titleId ->
EnumMapper(0x22000041) canonical key -> compute_str_hash ->
StringTable(0x2300000E) localized text. Runtime stays id-only; CT3/
CT4 consume this for display. DIDs hardcoded per the RetailKeyNames
precedent (CT1 verified them end-to-end).
Register: no new row. Retail's send path is non-optimistic and so is
ours — no deviation to record for this slice.
Tests: wire conformance (byte-exact + truncation) in
CharacterTitleEventsTests.cs + SocialActionsTests.cs; Runtime owner
unit tests in RuntimeCharacterTitleStateTests.cs plus integration in
RuntimeCharacterStateTests.cs; a no-local-mutation command test in
DirectGameRuntimeCommandAdapterTests.cs; an InstalledDat pin
(CharacterTitleResolverLiveDatTests.cs, ids 0/1/2/3/5/13/14, run green
with ACDREAM_RUN_INSTALLED_DAT_TESTS=1). Full solution build green;
hermetic filtered suite green (15,380 passed / 0 failed).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Applies the Opus dual-lens review corrections to CT1's DAT ground-truth
research (docs/research/2026-08-24-campaign-ct-dat-ground-truth.md):
- Window constraints (BLOCKER): replaced the "likely a hardcoded
ResizeTo/SetMinSize" guess with the verified mechanism —
UIElement::ResizeTo clamps only via element attributes 0x3C-0x3F,
nothing writes them at runtime, and retail resizes the SHARED
gmPanelUI host (LayoutDesc 0x2100006E, slot 0x1000018E) rather than
0x2100002E's own content root. Flags the unresolved 300x600-vs-300x362
size tension for CT3/CT6 and marks the host elements NOT PROBED by
CT1.
- RowHighlightSprite upgraded from a flagged hedge to a SEALED VERDICT:
the stat row's selected-state media is 0x06000F93
(gmAttributeUI::UpdateSelection -> InfoRegion::SetState on template
0x10000248), not 0x06001397 (which is legitimately the spellbook
row's separate selected-overlay mechanism). Falsifies the matching
comment in CharacterStatController.cs and dated-corrects the older
2026-06-26 doc at the spot that originated the wrong sprite id.
- Replaced the "18px gutter + 7px = 25px" derived story with the bare
authored rectangles (the numbers don't compose cleanly: 300-281=19,
and the 282px row overlaps the 281px scrollbar band by 1px) — CT5
must implement the authored numbers directly, never a derived
listWidth-18 formula.
- Plan doc: corrected the UpdateButtons ghost rule (no selection ->
Ghosted, not "ghosts when selected == current") and added the
AddTitleToList row-write contract for CT3.
- Pins: CharacterPanelLiveDatTests now honors ACDREAM_DAT_DIR first
(matching InstalledDatFactAttribute and its sibling live-DAT test
classes), hoists five vacuous bare-foreach assertions to counted
.ToList() pins, and adds the stat ListBox + scrollbar rect pins that
CT5/CT6 depend on.
- Doc hygiene: marked several probe-session observations (header
geometry "identical" claim, 0x06004CC2 characterization, the
master-map/category-map dump) as unpinned inference vs. committed
fact, corrected the 0x1000052D "throwaway container" mislabel, and
stated the header table's parent-relative coordinate frame.
- Recorded the CT5 gold this round found: InfoRegion::InfoRegion's
icon-DID lookup (a third GetDIDByEnum consumer, category
0x10000002) and gmSkillUI::RebuildSkillList's section-header order
confirmation, plus the RowHeight=22-vs-authored-20 divergence for
attribute rows.
Verified: ACDREAM_RUN_INSTALLED_DAT_TESTS=1 CharacterPanelLiveDatTests
filter 9/9 green; hermetic App suite filter (CI's Lane exclusion list)
6111/6111 green. No production code changed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Campaign CT slice CT1 (docs/plans/2026-08-24-character-panel-parity-campaign.md):
establishes the authored ground truth for LayoutDesc 0x2100002E that CT2-CT6
build against, so those slices port against verified DAT facts instead of
guessing. No production code changed.
Header findings: the PK line is authored pure white (the CT4 bug is in
CharacterStatController's runtime color choice, not a DAT gap); the level
color is a pale-gold (1, 0.949, 0.498) WITH an authored outline, diverging
from the current hardcoded Gold constant. The stat ListBox's row-template
list (LayoutDesc 0x21000045) is unreachable via the whole-layout
ImportInfos overload (the #375 same-layout template-list skip filter) --
the targeted ImportInfos(dats, layoutId, elementId) overload is required,
same as UiTemplateListBox's TemplateResolver already uses. The shared
attribute/skill row template (0x10000248) authors a 20x20 icon flush at
X=0 (current code: 16x16 at X=4), fixed 150px/100px name/value columns at
X=25/X=175 (current code: a width-fraction split), and a 7px gap between
the value's right edge and the row's own edge -- the scrollbar-gutter
margin the owner reported missing. The Titles page roster, row template
(LayoutDesc 0x2100005E), and window constraints are also pinned; the
character window's root authors NO min/max size properties at all (unlike
chat's self-contained window layout), and RetailUiRuntime.MountCharacter
never wires DatConstraintSource -- correcting the plan's "already in-tree"
claim for CT6.
Also derives and pins the full CharacterTitleTable::GetCharacterTitleFromID
chain (title id -> EnumMapper(0x22000041) canonical name -> compute_str_hash
-> StringTable(0x2300000E) localized text), resolved via the two-level
DBObj::GetDIDByEnum master-map indirection (0x25000000 -> category map ->
target DID) and verified end to end against ACE's CharacterTitle.WarMage=13
-> "War Mage". This independently cross-validates RetailKeyNames' existing
0x2300000A/0x2300000B/0x23000007 constants, which turn out to be the same
category-4 map's enum 4/5/3 entries.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Owner screenshot pair (2026-08-24): retail centers 'Chat' in the whole
46x17 face; we drew it at the synthetic 20px left indent. The caption
child (0x10000015) spans the full button with H=Center/V=Center and no
margins (live-DAT probed) — the controller now derives
ButtonTextCentered from the authored HJustify instead of leaving the
indent default.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Owner report (2026-08-24, post gate-pass): the chat channel button was
bigger than retail and both its caption and the Send caption were warm
gold instead of retail's near-white.
Size: retail never resizes the talk button — HandleSelection
@0x004cd540 only swaps the caption string; the authored 46x17 element
stands, and the authored SHORT captions ('Gen', 'Fell', ...) fit it —
that is why retail abbreviates. Our content-widening reflow (grow the
button to its label, shift the input) was a compensation for the
now-retired invented long captions, measured with the wrong font on
top. Deleted; the authored row layout stands.
Color + font: the button caption child (0x10000015) and the Send
button (0x10000019) both author pure white text with their OWN FontDid
0x40000002 (live-DAT probed) — different from the transcript font,
which is the other half of why 'Chat' fits 46px. UiMenu gains a
ButtonDatFont for the caption (popup rows keep the menu font);
the controller reads both elements' authored FontColor/FontDid instead
of the invented (1,.92,.72) constants.
The old widening pin is rewritten to the retail contract; a new
conformance test pins authored width, white captions, and the authored
font DID being requested for both buttons.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Four of the five owner-reported chat deltas (2026-08-24), each traced to
its retail mechanism:
1. Missing gold separator left of the input: the chat input authors two
1px Type-3 rail CHILDREN (0x10000017 at X=0, 0x10000018
right-anchored) whose only media is Normal_focussed (0x06004D67,
live-DAT probed). UiField consumes its DAT children, so the rails
were swallowed and never drawn. The factory now folds them into the
field, which draws both while focused.
2. Button says "General", retail says "Gen": the talk button's short
caption comes from per-target ID_Chat_ChatTargetMenu* strings
(HandleSelection @0x004cd540, StringTable 0x23000001 via
compute_str_hash — recovered from the raw binary after BN elided the
ids into name-hash globals). Authored set: Chat/Tell/Fell/Pat/Mon/
Vas/Alg/Gen/Trade/LFG/RP/Soc/Olt. Menu rows + squelch/tell specials
resolve from the same table (ID_Chat_TellTo*); production resolves
through DatStringResolver, fallbacks ARE the authored EoR English.
ChatStringsLiveDatTests pins the whole set against the installed DAT.
3. Channel button stayed green while the popup was open: retail's
pressed face is the momentary physical press ("flicks"); the OPEN
state drives only the arrow-cap child's StateDesc swap
(UIElement_Menu::UpdateState @0x0046cad0 writes attribute 0xe).
UiMenu now keys the face on the press, not on IsOpen.
4. Window-title/button text "vibrates" while dragging windows:
DrawStringDatPass snapped glyphs with MathF.Round — banker's
rounding. A centered label with a constant .5 fraction alternates
round-up/round-down across successive integers, double-stepping then
sticking while the background glides. Half-up Floor(v+0.5) snaps
every tie one way: uniform 1px steps in lock-step with sprites.
The fifth report (input row sticking out on window resize) did not
reproduce: a controller-bound fixture resize at 220/300/600px keeps the
whole input row inside the window (test added) — awaiting the owner's
exact gesture.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Owner report + live [ui-hover] probe (2026-08-24): hovering a
content-fits scrollbar did nothing because IsModelDisabled made the
whole bar hit-TRANSPARENT — every hover over it reported
widget=<none>. Retail's arrows and thumb are real child elements whose
Normal_rollover hot-tracking keeps running while the scrollbar is
disabled (UpdateLayout @0x004710d0 only hides the page-click regions,
children 4-7, and — with attribute 0x79 — the whole bar); scrolling
stays inert through geometry, not an input gate: a full-track thumb has
zero travel and the line/page steps clamp against nothing.
OnHitTest and the input path now gate on presentation visibility only.
A visible disabled bar hover-highlights and consumes clicks without
scrolling; a HideWhenDisabled bar stays inert. New root-level hover
tests drive real UiRoot hit-test dispatch (bare widget + the mounted
production character fixture) so this class of "state machine green,
pointer never arrives" bug fails loudly.
User-verified live 2026-08-24 ("bar works now").
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Owner report (2026-08-24): our scrollbar arrows pointed the wrong way,
the thumb vanished when there was nothing to scroll, and neither the
thumb nor the arrow buttons reacted to hover/press.
All three are one retail mechanism we had not ported:
1. Seating: UIElement_Scrollbar::UpdateScrollingArea @0x00470AA0 moves
the INCREMENT designee (attribute 0x77) to the top/left corner and
the DECREMENT designee (0x78) to the bottom/right, ignoring authored
positions. The vertical base skin (0x10000455 in layout 0x2100003E)
authors the DOWN-arrow decrement at Y=0 and the UP-arrow increment
at Y=32 (live-DAT probed; sprite art visually verified from decoded
PNGs), so our authored-Y ordering drew both arrows upside down.
DatWidgetFactory now seats by designation; the hand-wired sites
(CharacterStatController, ExternalContainerController, the
Config/Vendor menu chrome) share the new RetailScrollbarChrome
catalog instead of local constants.
2. Full-track thumb: UpdateLayout @0x004710d0 sizes the thumb from
proportion attribute 0x88, which DEFAULTS to 1.0 — a content-fits
bar shows a thumb filling the whole track; disabled only removes
input and the page regions. Our draw skipped the thumb entirely on
!HasOverflow.
3. States: every arrow button and thumb slice authors Normal (red gem /
dark navy), Normal_rollover (amber gem / bright blue) and
Normal_pressed (gold highlight / dark) media. The widget now tracks
thumb hover and selects rollover media on hover and pressed media
while dragging; the factory extracts the thumb-state media for both
the 3-slice and single-sprite thumb shapes.
ScrollbarSkinLiveDatTests pins the designations and state media against
the installed DAT so a revision or importer regression fails loudly.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Third owner-screenshot round: acdream fit one more word per line than
retail and drew glyphs flush against the popup's right border. Retail's
InqSizewMargins @0x00469660 wraps the glyph list at
(bound - m_margL - m_margR) and adds the margins back into the measured
width; the popup skins' shared text child 0x10000396 authors margins
L=2/R=2 (U=2/D=2 on three of the four skins — live-DAT probed). The
presenter now subtracts the horizontal margins from both wrap passes,
re-adds them into the measured width used for root sizing, and counts
the vertical margins in the measured/re-wrapped heights; the widget's
own draw already insets by all four margins (UiText ContentOffsetX +
the top/bottom inset), so the right-side spacing returns for free.
TooltipSkinLiveDatTests pins the authored margins per skin alongside
the P0x3D=256 wrap bound; a new presenter test proves margins shrink
the wrap bound and survive onto the widget.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Owner screenshots vs retail at the CA5 re-check caught both:
Wrap width: the popup skins' shared TEXT CHILD (0x10000396) authors
P0x3D=256 on all four skins — live-DAT probed, now pinned by an
installed-DAT test. Retail's InqSizewMargins UITS_MAX_WIDTH reads the
text element's 0x3D BEFORE the display-width fallback, so retail wraps
tooltip text at 256px; our measure pass used the display width because
TS-85's 'zero elements author P0x3D' sweep had only covered hover
TARGETS, never the popup skins. ApplyTooltipText now measures and
re-wraps at the text child's authored bound, falling back to the display
width only when none is authored.
Alignment: tooltip text rendered centered where retail hugs the left
edge. The skin authors no justification; retail's unauthored default is
Left, our importer's ElementInfo default is Center — the same
wrong-default class as #410's VJustify finding, now recorded there as the
horizontal sibling. Point-fixed in the presenter exactly as the chat
transcript already does; the client-wide default flip stays #410's scope.
The two-pass sizing test now models the real skin (max width on the text
child) and asserts left alignment. Full hermetic suite 15,332 passed / 0
failed; the new live-DAT pin passes against the installed DATs.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two corrections from the owner's retail-render oracle at the CA5 re-check,
both against readings the TS-85 register row had recorded as settled:
Compose (skill tooltips): retail is formula + newline + description —
GetTooltip @0x004f1fe0's operator+ has the InqSkillFormula output as the
LEFT operand; the old '"\n" + formula, no separator' reading had the
operand order backwards and produced a leading blank line with the formula
and description glued on one line. A formula-less skill (Salvaging) shows
the bare description, matching the failed-InqSkillFormula branch.
Sizing (ALL tooltips, per the owner's direction): retail sizes a tooltip
in TWO passes (StartTooltip @0x0045DE90) — measure-wrap at the max width,
resize the root through the authored ResizeTo clamps, then
RecalculateGlyphList RE-WRAPS the text at its final clamped width and a
second resize grows the root's HEIGHT for the extra lines. The branch the
register called 'a structural no-op' IS that second pass; without it a
description longer than the clamped popup stayed one clipped line, where
retail shows three. ApplyTooltipText now ports the full chain, so every
tooltip surface (items, options rows, character panel, world hover, map)
wraps and grows exactly as retail.
Pinned by BuildTooltip_FormulaFirstThenNewlineThenDescription,
BuildTooltip_FormulaLessSkillShowsBareDescription, and
LongTooltip_RewrapsAtTheClampedPopupWidth_AndGrowsHeightForTheExtraLines.
TS-85 carries both dated corrections. Owner visual re-check owed: skill
tooltip shows formula on line one, description below, long descriptions
wrapping to three-plus lines inside the parchment. Full hermetic suite
15,332 passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two defects the owner found at the CA5 drive, one shared theme: the data
was right and the presentation seam was dead.
#430 (tooltips): the TS-85 Batch-B port set runtime TooltipText on the
runtime-built attribute/vital/skill rows but never gave them a popup
locator, and RetailTooltipPresenter.OnTooltipShow refuses any widget with
AuthoredTooltipRootElementId == 0 — the tooltip could never mount, on any
row, ever. (The register's 'live-verified on the Character tab' was the
OPTIONS panel's Character tab — authored elements with authored locators;
a different surface.) Rows now carry the shared popup skin
0x10000395/0x21000041 — live-DAT probed as the ONLY locator pair the
character layout references, and the same inference UiItemSlot already
ships for runtime-built widgets. TS-85's row carries the dated correction.
#440 (train row stuck): training a skill debited credits on screen but
left the row in the untrained section until the NEXT click — because the
sheet-changed subscription only refreshed the captured sheet, and row
STRUCTURE rebuilt exclusively in click handlers (the raise 'completed'
callback runs after SEND, before the server answers; the owner's second
click was simply the first rebuild after the record landed, and ACE's
rejection of that second train — 'Failed to train', no credit change —
matches the owner's report exactly). The same gap kept CA4's
awaiting-ghost from visually releasing. CharacterStatController.Bind now
returns the data-changed refresh and MountCharacter invokes it on every
authoritative sheet change, mirroring retail's quality-change broadcast
(InfoRegion::OnQualityChanged @ 0x004F0EB0).
Pinned by DataChangedRefresh_MovesATrainedSkillToItsSection_WithoutAClick
and Rows_CarryTheSharedTooltipPopupLocatorAndDescriptionText. Owner
visual re-check owed next session (hover-dwell a row; train a skill and
watch it move immediately). Full hermetic suite 15,329 passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Retail sends a raise and WAITS: one request in flight, the raise
controls ghost, and displayed state changes only when the authoritative
quality-change record lands (gmStatManagementUI @ 0x004F03F0 family,
pinned in docs/research/2026-07-10-retail-panel-behavior-pseudocode.md
§5, whose own conclusion names ApplyLocalRaise as the thing to remove).
The optimistic layer predates the inbound parsers — it existed so the
panel showed anything at all — and with CA2 delivering server truth it
became strictly harmful: against ACE, a wrong TrainSkill cost fails
SILENTLY, so the optimistic promote-and-debit could show a trained
skill the server refused with nothing to ever correct it.
Deleted: CharacterSheetProvider.ApplyLocalRaise + both spend helpers,
and LocalPlayerState's six optimistic mutators (ApplyAttributeRaise,
ApplyVitalRaise, ApplySkillRaise, ApplySkillTraining, DebitIntProperty,
DebitInt64Property) with their tests. Added: the one-in-flight latch in
HandleRaiseRequest, CharacterSheet.AwaitingRaise ghosting all raise
controls, and gate release on every authoritative quality signal
(attribute/character/player-property events unconditionally; vital
events only release-and-refresh while a raise is in flight, so regen
ticks stay out of the sheet-rebuild path). Panel unmount resets the
gate — retail's awaiting flag lives on the panel instance.
AP-73 NARROWS rather than retires: retail's release on a rejection that
produces NO quality change is statically unverifiable, and ACE sends
chat-only (Raise*) or nothing (RaiseSkill/TrainSkill) on failure; until
the CA5 live check, a silently-rejected request leaves the controls
ghosted until panel reopen — recorded with its observable symptom.
Also verified for CA4: the train button sends the DAT-exact TrainedCost
(ACE's silent exact-match rule), and there is correctly NO panel
specialize send — retail/ACE specialize only via the SkillAlterationDevice
item-use + confirmation round-trip, whose client seams
(SendConfirmationResponse 0x0275, the 0x028B WeenieErrorWithString chat
routing) already exist. Provider tests now pin the retail contract:
send-without-mutation, one-in-flight, release-on-record, release-on-
unmount, and the regen-tick rebuild guard. Full hermetic suite 15,327
passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The r6 soak hard-fails without 'move-truth OUT' lines, yet the only
automated guard was a text assertion that the SCRIPT sets the env var —
it stayed green while #435 part 2 deleted the emitter, and the breakage
would have surfaced as a misleading connected-gate failure. The new
contract test asserts all four links of the chain in one place: the
script greps the pattern, MovementTruthDiagnosticController still emits
it, RuntimeOptions still parses the flag, and GameWindow still wires it
through. Deleting any link fails here, at build time, with the reason.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Attacking with no valid target has told the player nothing since Campaign V
slice V11 orphaned the DebugVM toast the message was wired to (#434 found
the drop; this closes it retail-faithfully).
Ground truth from the Ghidra decompile of
ClientCombatSystem::ExecuteAttack (0x0056bb70): retail writes
"You must select a valid combat target before attacking" via
ClientSystem::AddTextToScroll(..., 0x1A, true, 0) — the ClientLocal
SpewBox channel this codebase already routes every other client-local
refusal through. And retail has ONE message, not the two we carried:
attacking outside melee/missile modes is silent (ExecuteAttack is
unreachable there), so the invented "Enter melee or missile combat first"
text is deleted rather than rerouted, and the invented "No monster
target" is replaced by the retail string, which joins ClientTextRefusals
with its decomp citation.
Wiring: CombatFeedbackSlot gains the sibling BindOwned session-lifetime
shape, and SessionPlayerComposition.CompleteSessionPlayer binds it to
RuntimeCommunicationState.AddText(ClientLocal) with session-owned
teardown — a torn-down session's slot returns to its silent unbound
state. A binding-seam test
(CompleteSessionPlayerBindsCombatFeedbackToTheClientLocalSpewBoxRoute)
inspects the compiled composition for the BindOwned call and its
AddText-routing lambda, so the slot can never again pass its unit tests
while production leaves it unbound — the exact failure mode that hid
this defect. The two tests that pinned the invented strings now pin the
retail contract (exact string; silence for the unsupported-mode case).
Full hermetic suite 15,325 passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Part 1 deleted probes whose owning issues were closed. These 14 named no
issue at all, so each was traced to its introducing commit
(git log -S) instead of guessed at. Attribution split them three ways:
DELETED (7, investigations closed): ACDREAM_A8_DUMP_PV and
ACDREAM_DUMP_LIVE_SPAWNS (Phase A8), ACDREAM_DUMP_CLOTHING (#37),
ACDREAM_DUMP_EDGE_SLIDE (#32), ACDREAM_DUMP_STEPUP (L.2.3d-f),
ACDREAM_DUMP_VENDOR (the vendor campaign, 25 call sites across 8 files),
ACDREAM_DUMP_VITALS (#5, four independent read sites). VendorDiagnostics.cs
went entirely.
RECLASSIFIED (8, tools misfiled as probes): the DUMP_CELLS/DUMP_GFXOBJS
fixture-extraction family (replay-harness tooling with a roundtrip test),
PROBE_CELL (standing cell-transit tracer, pair of the permanent
PROBE_RESOLVE), DUMP_SKY and HIDE_PART (generic isolation tools), and
DUMP_STEEP_ROOF — which looked like an L.4 relic but observes LIVE
divergence-register row AD-56; deleting it would have removed the only
runtime lens on an active divergence. All moved to Permanent diagnostics
with their attribution recorded.
RESTORED (1): ACDREAM_DUMP_MOVE_TRUTH was deleted and un-deleted the same
day. It is not a probe — the canonical nine-stop soak
(run-connected-r6-soak.ps1) hard-fails every destination without its
'move-truth OUT' records, with a message that would misdirect the next
operator. Under the no-workarounds rule the gate's mechanism is restored,
not left broken with an IOU (#437, closed). Process lesson recorded on
both issues: a closed owning issue is NOT sufficient to delete a probe —
grep tools/ and the contract tests for consumers first.
Also lands the owner-requested default-off invariant: every diagnostic in
the codebase is inert until its env var is explicitly set. Exactly four
flags default ON and none is a diagnostic — RETAIL_CHASE, CAMERA_COLLIDE,
CAMERA_ALIGN_SLOPE, RETAIL_CLOSE_DEGRADES are retail behaviors wearing an
A/B off-switch. That set is now FROZEN by
LaunchOptionsDocumentationTests.OnlyTheFourRetailBehaviorFlagsDefaultOn;
docs/launch-options.md's Conventions and CLAUDE.md state the rule, and
CLAUDE.md now binds future probes to a documented row in the same commit.
The client reads 137 environment variables (161 at audit start); 40
temporary probes remain, every one attributed. Full hermetic suite 15,322
passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Each of these was temporary apparatus added to chase one bug, and each was
supposed to be deleted in the commit that fixed it. Fourteen closed issues
later they were still here: #337's support/wire-mesh trio, #171's sticky
timeline, #119's viewer and entity dumps, #113's phantom probe, and a dozen
more. 3,493 lines removed; the client now reads 144 environment variables
instead of 161, and 47 temporary probes remain instead of 64.
This is not only tidying. Every probe leaves a branch on its hot path when
unset, several re-read the environment per call rather than caching, and
the volume buries the diagnostics that are actually load-bearing. It is
also a headless correctness matter: HeadlessStaticStateAudit reflects over
PhysicsDiagnostics' flags to refuse a multi-session host when any is set,
and cannot see probes that live outside that owner.
Four files went entirely — WalkMissDiagnostic.cs, CollisionMeshWireframe.cs
and two test files whose only subject was a deleted probe.
TransitionTypes.SetContactPlane also sheds its CallerMemberName /
CallerLineNumber parameters, which existed solely for #337's cpSrc=
attribution and carried the instruction to strip them with the probe
family; no call site passed them, so no behavior changes. F2's collision
overlay survives and reverts to its proxy-cylinder form, which is what
removing the ACDREAM_WIRE_MESH upgrade means.
LaunchOptionsDocumentationTests earned its keep here: it refused the
deletion until docs/launch-options.md moved the 17 rows into Retired and
the frozen direct-read counts came down (PhysicsEngine.cs to zero,
TransitionTypes.cs 3 to 2). The documentation could not drift during a
cleanup this wide.
The 14 probes that name no owning issue are deliberately NOT deleted.
Nothing records when they became safe to remove, and guessing is how a
future investigation loses apparatus it needed; #435 stays open for their
attribution.
Full hermetic suite 15,321 passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
DebugPanel and DebugVM have not been constructible since Campaign V slice
V11 removed the ImGui frontend that hosted them: nothing in src/ ever
called their constructors, only a test did. Two consequences, both fixed
here — 35 environment reads inside them were unreachable, and roughly forty
XML doc comments across the diagnostics owners promised a runtime checkbox
that no longer exists. A flag documented as runtime-toggleable when it is
startup-only sends the next investigation down a path that cannot work.
Deleted DebugPanel.cs (340 lines), DebugVM.cs (548) and DebugVMTests.cs
(327). Corrected the surviving claims in PhysicsDiagnostics,
RenderingDiagnostics, CameraDiagnostics, PhysicsEngine and GameWindow to say
what is actually true: these flags are set from the environment at startup
or by direct assignment.
The one real dependant was CombatFeedbackSlot, whose binding target was
DebugVM. It now takes a plain Action<string>, which removes the dependency
without changing behavior — and makes visible that there is no behavior:
nothing binds the slot, so the combat refusals it carries ("No monster
target", "Enter melee or missile combat first") have been discarded all
along. Filed as #436 and pinned by a test, rather than papered over with an
invented chat message; the retail text and channel need the oracle first.
Deliberately untouched: F1's AcdreamToggleDebugPanel binding, which
GameplayInputCommandController consumes as a documented no-op so the key
does not fall through to a lower input scope; and the
DebugVmRenderFactsPublisher / DevToolsRuntimeSources chain, which is still
wired into production composition and deserves its own dead-code pass
instead of being pulled into this one.
Full hermetic suite 15,333 passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The client reads 161 ACDREAM_* environment variables across 79 files. Only
about 25 were written down, and the audit found the documentation drifting
in both directions: CLAUDE.md still advertised ACDREAM_RUN_SKILL /
ACDREAM_JUMP_SKILL (deleted; skills are server-authoritative now, and the
jump fallback is 300, not the documented 200), while flags with real
side effects had no description at all.
docs/launch-options.md documents every one by lifecycle — production,
command line, measurement, automation, permanent diagnostics, temporary
probes, deprecated, retired — with a mandatory side-effects column. That
column is the point: #432 cost three days of taxed measurements because
ACDREAM_AUTOMATION_ARTIFACT_DIR reads like an output path and also builds
a per-frame diagnostics referee, and ACDREAM_STREAM_RADIUS silently
measures a streaming window production never uses. Rows now say so. Other
surprises the audit surfaced and recorded: ACDREAM_DUMP_SCENERY_Z swaps in
a duplicate scenery-placement path rather than only logging,
ACDREAM_PROBE_VIS silently also enables ACDREAM_PROBE_ENVCELL, and
ACDREAM_DUMP_ENTITY's id list doubles as an unrelated probe's watchlist.
LaunchOptionsDocumentationTests enforces it, because a hand-maintained list
of 161 flags is stale within a week: an undocumented flag fails, and so
does a documented row whose read site was deleted. It scans string literals
rather than GetEnvironmentVariable call shapes — the startup path reads
through an injected delegate, so a call-shaped pattern silently missed
ACDREAM_LIVE, ACDREAM_PAK_PATH and every other production flag. A third
test freezes per-file direct-read debt by exact count (20 files outside the
owner classes), so structure rules 4 and 5 can be paid down but not
regressed.
CLAUDE.md's 94-line env-var section becomes a 16-line pointer, and its
stale test-character paragraph is corrected.
Also fixed, all doc-vs-code mismatches the audit proved:
- RenderingDiagnostics.FrameProfEnabled described a GPU-query self-disable
that Campaign V slice V11 deleted.
- Two comments named ACDREAM_RENDER_BACKEND as a live co-requisite; it died
with the OpenGL backend.
- EnvCellRenderer.CollectCellAuditLines and its ACDREAM_A8_AUDIT doc: the
method had no caller anywhere and its documented caller never existed.
Filed rather than fixed, to keep this a documentation change: #434 (the
DebugPanel/DebugVM surface is never constructed, so ~40 "runtime-toggleable"
comments are false and 35 env reads are unreachable) and #435 (17 temporary
probes outlived their closed investigations; 14 more name no owner).
Full hermetic suite 12,202 passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Second instance of the #429 (ad695589) boxing-comparer defect class, this
time in the measurement harness rather than production:
ACDREAM_AUTOMATION_ARTIFACT_DIR (with retained-UI screenshots) constructs
CurrentRenderSceneOracle, whose presence as partition observer re-enables
the G5-retired legacy InteriorEntityPartition every frame with per-entity
fingerprinting. Complete() then sorts one fingerprint per RESIDENT entity
(~60k across the streaming window), and the comparer's first key bound
x.ProjectionClass.CompareTo(y.ProjectionClass) to Enum.CompareTo(object),
boxing both operands. The 3-value enum almost always ties, so the boxing
ran on essentially every comparison: a measured ~6.2 MB and ~14 ms per
frame, everywhere — not town-specific and not view-triggered, which is
also why it appeared to "latch" (the resident set drives it, not the view).
Comparing the underlying integral value keeps the identical order. Hermetic
gate: one warmed observed partition of 20,000 entities allocated 15,876,088
bytes before, and passes a <64 KiB bound after
(OracleObservedPartitionAllocationTests).
Ordinary play never constructs the oracle, so no player-visible behavior
changes; what changes is that captures taken with the automation artifact
directory set are no longer taxed. The #429 acceptance data is unaffected
(owner drives and the deciding A/B arms ran with the artifact dir null).
The temporary [pview-alloc] attribution probe that localized this is
retired in the same commit; the gate test now guards the defect.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The directional-shadow topology rebuilt on every streaming-churn frame
and was the measured body of the run-hitch stalls (701 of 708 baseline
stalls alloc-correlated):
- The draw sort comparer's enum-vs-enum CompareTo bound to
Enum.CompareTo(object) and boxed BOTH operands on every comparison —
a constant ~38.9 MB of garbage per topology rebuild (~4M boxes),
handing the GC a forced gen0 collection mid-frame. The full ~100k-draw
sort is replaced outright: draws hash-group by exact DrawKey in one
O(n) pass over retained chained-index arrays, and only the
few-thousand DISTINCT group keys sort (order-preserving packed
material|cull|firstIndex|baseVertex + count|slot|layer|foliage keys,
first-appearance tie-break) — bit-identical emission order to the old
stable sort, near-zero allocation, and no per-draw comparisons at all.
- The caster frame sorts 4-byte indices keyed on SortKey.Value instead
of shuffling multi-hundred-byte records through a boxing comparer.
- Owner-approved pipelining: on a frame whose shadow inputs just changed
(the same frame already paying frame-view/landscape rebuilds), the
caster-frame and prepared-draws topology rebuilds defer to the next
quieter frame, capped at two consecutive deferrals — inside the GPU
fence depth, so retained draws never reference a released arena range.
First build, generation change, caster BuildSequence change, and
journal overflow force the immediate path; deferred refreshes skip
identity-mismatched journal rows.
Owner-accepted in both presentation modes: stall frames 5.8/s -> ~0.45/s
uncapped (0.49/s capped), median stall 20.3 -> 13.7 ms, >25 ms frames
near zero, 275 fps uncapped baseline restored. Allocation gate: a warmed
topology rebuild must allocate <2 KiB (DirectionalShadowPreparedDrawTests).
docs/ISSUES.md carries the full evidence trail; the residual
content-proportional rebuild milliseconds are filed as the
incremental-topology successor, and the pre-existing town-view scaling
latch is filed as #432.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
UploadGfxObjMeshData built every completed mesh's index data three-plus
times over in LINQ transients (per-batch Indices.ToArray copies plus an
unsized SelectMany growth) on the render thread, up to the per-frame
upload budget. The conversion now fills one exact-size retained
CPUIndices array (the same one the B.4b pick path keeps) and hands the
shared arena (offset, count) segments of it; CPUPositions fills by a
direct pre-sized loop; the Sum/Any/FirstOrDefault transients are gone.
GlobalMeshBuffer.UploadMesh takes the segment form — the staged bytes
per batch are unchanged. Gate: a warmed completion must allocate near
its retained-copy size (MeshPipelineDeviceSeamTests).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The owner measured 63% GPU at night vs 28% by day: the starfield
lattice was the whole difference. Three lossless cuts:
- one packed-multiply hash round per cell instead of ~9 chained PCG
rounds (position/presence bit-sliced from one result, brightness/tint
from a second);
- the 3x3x3 neighbourhood sweep becomes the 2x2x2 block around the
sample — a star's visible footprint (a few px) is far smaller than a
lattice cell on screen, so any star outside that block is over a full
cell (>=15 px) away and contributes nothing;
- background mottle drops its invisible third octave.
Same densities, sizes, colours, and look; star positions reshuffle
(procedural layout, not authored). sky.frag.spv re-pinned.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Rotation (user-directed): the procedural starfield rotates once per
Dereth day (~2 real hours - constellations visibly wheel through a
night) about a celestial pole ~41 deg above the northern horizon, plus
dayOfYear/360 of seasonal drift so the 360-day year changes the night
sky. One SkyParams float (272-byte block, layout test re-pinned)
carries dayFraction + dayOfYear/360 from the world clock; sky.frag
applies a Rodrigues rotation to the sample direction so stars and
mottle turn together. Impossible with retail's static stretched layer.
Fade rework (the 2026-08-23 two-screenshot gate finding): the
per-vertex vTint signal carried the sun-facing product and blanked
stars across the entire twilight half of the sky. The fade now reads
the UNIFORM ambient term - identical star visibility in every compass
direction, same dusk-to-dawn schedule - with one deliberate exception:
a thin suppression band hugging the low sky toward the sun's azimuth
while the sun term is strong, so stars still wash out inside the
actual twilight glow.
Guards updated (rotation anchor, uniform-fade anchor, 272-byte layout);
both sky SPIR-V hashes re-pinned.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The enhanced sky multiplied the full 4-tier star lattice by a night
factor of 0.0 all day long. Early-out when the star layer's lighting
product says daytime, so the pack's night sky costs nothing outside
dusk-to-dawn. sky.frag.spv re-pinned.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The first night-sky gate (2026-08-23 screenshot) showed three defects:
glowing beams along the cube-face boundaries (fwidth blowup where
adjacent pixels land on different faces lights every neighbourhood star
solid), diffraction-spiked standouts the user rejects ('that is in a
photo only, not in real sky'), and ellipse-stretched stars from scalar
length(fwidth) sizing at oblique view angles.
One rewrite removes all three: stars now live on a seamless 3D lattice
over the unit sphere (no faces, so no seams by construction), each star
resolved through an exact tangent-plane -> screen-pixel 2x2 solve of
the direction derivatives (perfectly round, true pixel sizing at every
view angle, sharper cores), spikes deleted in favour of a soft round
halo on the bright tiers. Guard test updated to pin the new anchors and
forbid both fwidth-face grids and spikes; sky.frag.spv re-pinned.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
User-directed enhancement ('I want the night sky to look very good'):
retail's star layer is one small texture stretched over a 10-poly dome
cap, so stars smear regardless of source-image quality. With the
Atmospheric render pack active, sky.frag now renders the star layer
(GfxObj 0x010015EF, identical in all 20 Dereth day groups) as a fully
procedural sky computed from the view direction: hash-derived stars on
a cube-face grid in three density tiers plus sparse diffraction-spiked
standouts, sized in SCREEN pixels via derivatives so they stay crisp at
any resolution and FOV, over the user-approved 0.4-1.3% cool mottle
(gen_starfield2.py seed 11, approved 2026-08-23). The draw is forced
additive; the day/night fade rides the star layer's existing retail
lighting product so the schedule matches the authored keyframes. Pack
inactive = retail look byte-untouched.
EnhancedNightSkyRuleTests pins the uParamA gate, the exact star-layer
id, the forced-additive draw, and the pack-runtime wiring; sky shader
SPIR-V recompiled and re-pinned. Hermetic App suite green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The aurora was never missing data — it was a missing mechanism plus a
misread. New decompile evidence closes the April-2026 contradiction:
retail plays the sky carriers' PES through the Setup's own DefaultScript
(GameSky::MakeObject @0x00506EE0 -> CPhysicsObj::makeObject @0x00513970
sets state|=0x80000; animate_static_object @0x00513DF0 ticks
ScriptManager + ParticleManager). The pes_id column stays dead — that
half of the April finding stands; the ids are byte-equal mirrors.
- SkyPesFrameController is now the production owner (ACDREAM_ENABLE_SKY_PES
deleted): script ids resolve from the Setup DefaultScript
(SkyObjectData.DefaultScriptId; the pes_id column is a one-time-logged
cross-check), slots persist by (index, gfx id, properties) per
CreateDeletePhysicsObjects @0x005073C0 — a day-group swap keeping the
carrier no longer restarts its emitters — and stale slots stop before
replacements claim the slot-derived owner id.
- RetailParticleFacing ports calc_draw_frame @0x0050DFA0: degrade mode 2
faces the viewer roll-free (set_vector_heading) instead of the camera
plane; modes 3/4/5 spin the authored frame around one local axis
(rotate_around_axis_to_vector) — Dereth authors 54 mode-5 emitters that
previously got no facing at all; 1,583 mode-2 emitters get the exact
law; authored/mode-1 paths are unchanged.
- The 2026-08-23 'whole-sky tint' was the Rainy-group lightning/thunder
PES playing at the debug anchor inside their 0.03-0.19 window, not the
aurora: the aurora is nine faint viewer-facing glows pulsing on
6.7/15/55-minute rebirth cycles, in every day group, all day.
Research: docs/research/2026-08-23-sky-default-script-port.md.
Register: AD-112 filed (camera-anchored synthetic owners vs sky-cell
physics objects). ISSUES #2 corrected (the playback ban is lifted by the
new evidence); #28 fix landed pending the connected night gate.
Tests: RetailParticleFacingTests (16), SkyPesFrameControllerTests (6);
hermetic suites App 6,076/0, Core 4,905/0.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Two April stand-ins, neither registered, met at the horizon from altitude as
a hard line between the dome's rim and the fog-coloured clear:
1. sky.frag fogged every non-additive sky layer with a 0.2 floor. Retail's
GameSky::Draw @0x00506FF0 disables fixed-function fog around the whole
sky draw unless an AdminEnvirons fog override is active (SetFFFogEnable(
LScape::m_override_enabled ? 1 : 0)); additive layers stay unfogged via
SetFFFogAlphaDisabled(1) at D3DPolyRender::SetSurface 0x59c882. The sky
pass now sets ApplyFog only for (override active && !additive), with no
floor.
2. WorldRenderFrameBuilder overwrote the authored fog range with one derived
from the streaming window (538..2189 m always). Retail sets FOGSTART/
FOGEND straight from the keyframe's MinWorldFog/MaxWorldFog
(SkyDesc::GetWorldFog @0x00500CE0 -> SetFFFogProperties @0x005A2F70) with
no draw-distance scaling; zfar is a constant 4000 m. The builder now
leaves SceneLightingUbo.Build's values alone; ACDREAM_FOG_START_MULT /
_END_MULT are deleted from RuntimeOptions.
Guards: SkyFogRuleTests (source-level, the sky renderer has no hermetic
harness); sky.frag.spv re-pinned in VulkanShaderManifestTests with the
reason. Research note 2026-04-23-sky-fog.md carries a correction banner.
App hermetic 6,070/0, Core 4,707/0 (Release). Owner look gate owed: night
and rain fog are now retail's shorter authored ranges.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Live Holtburg at 2560x1440: the Low preset needed 67,368,164 resident bytes
(screen-sized HDR/depth/ray targets are 44 MB of that) against an absolute
64 MiB ceiling that had only been validated at 1080p, so Options -> Apply
fell back to the default path; every later Apply was then refused by the
controller's failure memo, which treated the user's deliberate choice like
automatic re-activation.
RenderPackResidentBudget.Effective scales the declared 1080p figure by the
viewport's pixel-count ratio (never below 1), still capped by the hardware
MaxPackResidentBytes; both pack graphs use it and the performance-matrix
tool judges its resident column by the same rule (contract test updated).
RenderPackController.Request gains explicitUserChoice, which clears the
memo for that selection; RenderPackSelectionBinding passes it on every
display edge (Apply, including resolution changes) and keeps the memo for
the startup request.
Tests: RenderPackResidentBudgetTests (1080p/720p keep the declared
ceiling, 1440p = 16/9x, 4K = 4x, hardware cap wins, zero extent rejected);
controller explicit-retry; the binding test now proves the user's next
Apply activates once the cause is gone. App hermetic lane 6,068/0 (Release).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
GLFW auto-iconifies an exclusive-fullscreen window on focus loss; for one
frame the window size reads 0x0 while the swapchain is still created, so
GameWindow.OnRender's PrepareFrame guard let a zero-area RenderFrameInput
through. The retail path tolerated it silently; the render-pack controller
correctly rejects a zero activation extent, which surfaced the latent frame
as an unhandled ArgumentOutOfRangeException during the owner's VM6/VM3 gate.
RenderFrameOrchestrator.Render now returns RenderFrameOutcome.ZeroArea
before BeginFrame when either dimension is <= 0 (no GPU frame, phase,
measurement, diagnostics or recovery runs) and GameWindow skips
NoteFrameClosed for it. Test: ZeroAreaViewport_SkipsTheFrameBeforeAnyGpuWork.
Verify: App hermetic lane 6,059/0 (Release).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Round-5 narrow re-review of 754d59d9: APPROVE (verified to the SPIR-V
disassembly; shadows-on path proven opcode-identical). Four
non-blocking nits.
N1 (test robustness): T3 (ReceiverVertexShadersFallBackToThePlainLightDirectionWhenTheShadowFlagIsClear)
was satisfied by comment prose -- round 5's own explanatory comments
quote the exact plain-pipeline direction substring
("-uLights[i].dirAndRange.xyz, ... matched bit-for-bit"), so the
substring-only assertion passed even with the code mutated. Fixed:
comment lines are now stripped (new StripLineComments helper) before
any assertion, and a new ordered regex per shader asserts the exact
branch SHAPE -- shadowGatedOff ? -uLights[...] : normalize(uShadow...)
-- not just substring presence. Mutation-tested locally against the
rewritten test: (1) swapping the ternary's true/false operands --
FAILED (previously passed); (2) deleting the fallback entirely,
collapsing to the pre-round-5 buggy expression -- FAILED (previously
passed). Original file restored and reverified passing after each
mutation.
N2 (doc accuracy): "numerically the plain pipeline" overstated the
round-5 fix in three places (plan doc, mesh_atmospheric.vert,
terrain_atmospheric.vert). The direction expression is bit-for-bit;
the SUM is not, because the atmospheric shaders' split ambient+point
vs directional accumulation (and terrain's two varyings vs the plain
pipeline's one) reassociates float summation order by ~1 ulp -- which
is exactly the measured mean |Delta| 0.007 the coordinator's own
pixel-proof evidence already recorded (well under the 65 px noise
floor). All three rewritten to say the receiver "matches the plain
pipeline to within float summation-order rounding (measured mean
|Delta| 0.007 on the offline scene)."
N3 (coverage): RenderPrepared's own cascadeCount == 0 exit (the F2
fix) had no direct test even though RenderPrepared already has 8
direct call sites in this file. Added one: ResidentMaximumReachMeters
at/below CameraNearMeters, passed straight to RenderPrepared with an
otherwise-fully-valid environment (so the fitter, not the environment
gate, is what returns zero cascades), asserting IsBindableFor true /
IsValidFor false.
N4 (latent): RenderPrepared's OWN "if (!environment.ShouldRender)"
exit is a fourth bufferless-disabled path -- unreachable via Render
(whose own gate already validates ShouldRender first) but the same
shape, and RenderPrepared is called directly by tests and any future
caller. Took the preferred fix: publishes the disabled binding there
too, via the same helper, so every exit on a frame that draws the
world publishes when the pack's AtmosphericFrame is bound. Also made
EvaluateGateAndPublishDisabledBinding self-contained: it now resets
_currentFrameBinding to Disabled on its own entry instead of relying
on Render having done so first (idempotent with Render's own reset).
Regenerated SPIR-V: mesh_atmospheric.vert and terrain_atmospheric.vert
are comment-only changes (N2), so only the manifest's source hashes
changed -- compiled .spv bytes are unchanged, consistent with round 3's
precedent for comment-only shader edits.
Verify: Release build 0 warnings/0 errors. App hermetic-lane filter
6,055/0 failed. Full hermetic-filtered solution: 15,283/0 failed
across 15 projects.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Round 4 (eec95535) fixed wind but introduced a new lighting bug: the
receiver VERTEX shaders source the sun direction from the shadow block,
not only the shadow visibility term. mesh_atmospheric.vert's
accumulateLights read uShadowLightDirectionAndSource unconditionally for
every directional light; terrain_atmospheric.vert did the same for its
single sun term. Round 4's PublishDisabledReceiverBinding writes
direction (0,0,1) into that block on every shadow-gated-off frame (user
sun-shadow-strength 0 in daylight, indoor/portal cover, night), so every
such frame was lighting outdoor terrain and objects from straight
overhead instead of the authored sun. Publishing the environment's real
direction would not have restored parity either -- the celestial shadow
source direction (sun/moon disc) is not the authored light direction.
F1 (BLOCKER): fixed in the shaders themselves, exact parity with the
plain pipeline. Both receiver verts now branch on the same flag bit
acdreamDirectionalShadowVisibility already reads
((uShadowTextureAndFlags.w & 1u) == 0u) and, when clear, use the EXACT
plain-pipeline expression instead of the shadow block's direction:
-uLights[i].dirAndRange.xyz in mesh_atmospheric.vert (matching
mesh_modern.vert, hoisted out of the light loop as a uniform branch);
-uLights[0].dirAndRange.xyz in terrain_atmospheric.vert (matching
terrain_modern.vert's sunDir/-sunDir form). The (0,0,1) word in the
disabled block stays as the documented normalize()-cannot-NaN guard; its
comment now says so explicitly since it is no longer read as a light
direction when the flag is clear.
F2: RenderPrepared's cascadeCount == 0 return is a third bufferless-
disabled path reachable from a frame that already passed Render's own
two gates (the cascade fitter can still find zero usable cascades) --
publishes the same disabled binding now, via the same
PublishDisabledReceiverBinding helper (re-signatured to take a bare
AtmosphericFrameBufferBinding so all three call sites -- Render's two
early-outs plus this one -- share it).
F3: removed a stray duplicated " -- Closeout and merge" fragment under
the plan's VM7 heading.
F4: corrected the false "the flag bit makes it numerically the plain
lighting sum" claim in the plan's round-4 paragraph and in
WbDrawDispatcher.DirectionalShadowReceivers.cs -- the flag bit alone
only fixed the shadow VISIBILITY term (already correct before round 4);
it took both that AND round 5's light-DIRECTION fallback to actually
match the plain pipeline.
T1: extracted Render's gate prologue (environment evaluate -> two
early-outs -> PublishDisabledReceiverBinding) into internal
EvaluateGateAndPublishDisabledBinding(frame, in input, out environment,
out environmentGateTicks), behaviour-preserving, called by Render before
it touches world/terrain -- the ArgumentNullException.ThrowIfNull(world)/
ThrowIfNull(terrain) calls keep their exact position relative to the
gate. No test in this suite constructs a real WbDrawDispatcher +
TerrainModernRenderer pair (still true), so this extraction is what
makes the gate itself testable; two new tests drive it directly with
PlayerInsideCell: true and with ResidentMaximumReachMeters <=
CameraNearMeters, asserting TryGetCurrentFrameBinding true / IsValidFor
false for both.
T2: proves the actual composition WbDrawDispatcher.PipelinesFor and
TerrainModernRenderer both use -- TryGetCurrentFrameBinding feeding
ShouldSelectReceiverPipeline -- selects the receiver pipeline for the
atmospheric world pass once a disabled binding is published, and still
refuses a non-atmospheric pass name.
T3: shader-source guard (same style as AtmosphericPostProcessGraphTests'
existing shader-text tests) pinning that both receiver verts contain the
flag-gated fallback and reference the same uLights expression the plain
verts use, so a future edit that drops the fallback fails this test
instead of only showing up in a pixel capture.
T4: the (0,0,1) test's doc comment and an inline assertion comment now
say the value is a NaN guard, not a light direction.
Regenerated SPIR-V: mesh_atmospheric.vert and terrain_atmospheric.vert
recompiled to different bytes this time (a real code change, not a
comment); manifest updated to match.
Verify: Release build 0 warnings/0 errors. App hermetic-lane filter
6,054/0 failed. Core.Tests 4,695/0 failed. RenderPackValidator 30/30.
Full hermetic-filtered solution: 15,282/0 failed across 15 projects.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The reviewer's offline pixel apparatus found a real design defect, not a
test artefact: foliage wind was welded to "directional shadows rendered
this frame." Evidence: offline High preset, sun-shadow-strength=0,
wind-strength 2 + lean/branch 1 m — wind-on vs wind-off at the same
pinned clock differed by only 49-65 px, inside the apparatus's own 22 px
run-to-run noise floor (no measurable motion). A CPU probe independently
confirmed ResolveFoliageWind was correct (first advance snaps to Clear
0.25/0.15, gate 1, one graph) — the correct uniform never reached the
world pass.
Root cause: DirectionalSunShadowRenderer.Render's two early-out paths
(!environment.ShouldRender, ResidentWindowUnavailable) left
_currentFrameBinding at its pure Disabled (no-buffer) default.
WbDrawDispatcher.PipelinesFor and TerrainModernRenderer's matching
selection logic only choose the atmospheric receiver pipeline
(mesh_atmospheric, the only pipeline that #includes foliage_wind.glsl)
when TryGetCurrentFrameBinding returns true; with no buffer it always
returned false, so the world pass silently fell back to the plain
mesh_modern pipeline, which has no wind code at all. Because the shadow
gate is ActiveDayGroupMultiplier = dayGroupPolicy x elevationResponse x
strength, this killed wind every night (elevation response -> 0), at
user sun-shadow-strength 0, and under the portal/login cover.
Fix (decouple, not patch): DirectionalShadowFrameBinding gained
IsBindableFor ("a real current-frame allocation exists") separate from
IsValidFor ("...and it is Enabled with real shadow content" -- kept
exactly as VolumetricShaftRenderer's own gate needs it).
TryGetCurrentFrameBinding now returns IsBindableFor. When the built-in
pack supplies an AtmosphericFrame binding (declared packs never do, so
their receiver shaders -- which never declare set 3 binding 5 -- are
unaffected), Render's two early-out paths call a new
PublishDisabledReceiverBinding: it allocates one real ring slice and
writes a DISABLED DirectionalShadowUniforms block -- every matrix
Identity, every control/bias term zero, TextureAndFlags all zero (bit 0
clear is exactly what directional_shadow_receiver.glsl's
acdreamDirectionalShadowVisibility already reads as "no shadow, full
visibility" via its existing early return 1.0), and a unit light
direction (0,0,1) so a fragment shader's normalize() can never produce
NaN. BindDirectionalShadowReceiver and TerrainModernRenderer's
shadow-buffer bind now check Buffer is not null instead of Enabled, so
the disabled block actually gets bound once it is selected.
PublishDisabledReceiverBinding is internal (not private) specifically so
it is testable without standing up a real WbDrawDispatcher/
TerrainModernRenderer pair -- no test in this suite constructs either.
New tests: (a)/(b) PublishDisabledReceiverBinding is bindable-not-valid
with a bound AtmosphericFrame and a genuine no-op with an unbound one;
(c) BindDirectionalShadowReceiver emits both UniformDirectionalShadow and
UniformAtmosphericFrame binds for a disabled binding; (d)
VolumetricShaftRenderer's gate still reports NoCurrentDirectionalShadow
for a disabled binding. ShouldSelectReceiverPipeline itself is untouched
and its existing tests (parametrized directly on bindingValid) remain
valid; no existing test asserted the old "disabled shadows -> plain
pipeline / no binding" behaviour in a way this fix invalidates -- every
existing caller either bypasses Render (calls RenderPrepared directly)
or uses a stale-serial binding IsBindableFor still correctly rejects.
Verify: Release build 0 warnings/0 errors. App hermetic-lane filter
6,050/0 failed. Core.Tests 4,695/0 failed. Full hermetic-filtered
solution: 15,278/0 failed across 15 projects.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Narrow re-review of a82959f1: APPROVE, with follow-ups. All items landed.
N2 (structural): (a) extracted the ONE shared InstanceGroup-from-key
construction seam, WbDrawDispatcher.CreateGroupFromKey(key, registration,
frame) — before this there were two near-identical `new InstanceGroup
{ ... }` initializers (GetOrCreateInstanceGroup and GetOrCreatePackedGroup)
that had already drifted once (the round-2 F1 bug). Both routes call it now;
CreateGroupFromKey's own `new()` is the only production InstanceGroup
construction site repo-wide, same precedent as AppendPackedInstance. (b)
GroupKey.FoliageFlags lost its `= 0u` default and moved before CullMode in
the declaration (CullMode keeps its default, C# requires optional params to
trail required ones), so a `new GroupKey(...)` that omits it is a compile
error. Fixed every real construction site the reorder/requirement touched:
the 2 production sites, ToKey (a reconstruction from InstanceGroup the
review didn't count but the reorder broke), and 5 test sites (one more than
the review's "4" — InstanceGroupClearTests had a second, implicit
target-typed `MakeKey` factory the original count missed). Verified by a
full solution build.
N1: added CreateGroupFromKey_CopiesFoliageFlagsFromTheKey
(InstanceGroupClearTests) — a key carrying FoliageFlags 0x2 in, the created
group's FoliageFlags 0x2 out. That test plus N2b's required field are what
actually guard the round-2 F1 blocker; reworded PackedDispatcherOracleTests'
existing test comment to say what IT proves (the classification-to-
BuildIndirectArrays-to-BatchData.flags path), not that it guards the
classifier.
N3: corrected the plan's round-2 paragraph — folding FoliageFlags into the
G2/G3 digest is correct and symmetric, but CompareClassifiedOutput only
runs from RenderScenePViewFrameProductController.BuildAndCompare, which has
no production caller anywhere in src/AcDream.App/, and both of
RenderScenePViewFrameProductTests's own callers construct the controller
without the optional dispatcher argument — so the fold catches nothing
until that oracle is wired to an actual caller.
N4: the plan's F6 note now names both classification caches — the classic
route's EntityClassificationCache.EntityCacheEntry (self-heals per entity
on its own next eviction) and the packed route's
PackedProjectionClassificationEntry/PackedClassifiedBatch.Key
(PackedProjectionClassificationCache.BeginFrame clears its entire cache in
one shot on a RenderSceneGeneration change) — and notes neither mechanism
is keyed to a pack switch specifically.
N5: deleted the now-unused single-generic ComputeEntityHasCutoutSubset<T>
overload; its 4 test call sites now use the two-generic, zero-alloc
overload with an unused int context and a static (_, value) => value
lambda, so there is exactly one ComputeEntityHasCutoutSubset to keep
correct.
A6 (reviewer-filed): ResolveFoliageWind's _windMean/_windGust started at 0
and always eased toward the weather target by clock delta, with no
distinction for a graph's first-ever advance. A pinned clock
(ACDREAM_SKY_PHASE_SECONDS, the offline pixel gate's determinism pin) never
advances between calls, so the wind reached only whatever fraction the
first (1-second-clamped) step produced and sat there forever; live, the
first 10 s after a graph is constructed (pack selection / login) spun up
from dead calm even though the weather already IS what it is. Fixed at the
root: the first advance (_windFrameSerial == -1, the constructor sentinel)
now snaps _windMean/_windGust straight to the target; every later advance
eases over WeatherSystem.TransitionSeconds exactly as before. Added a
SetWindClockSecondsOverrideForTesting seam (_windClockSecondsOverride is no
longer readonly) so a hermetic test can advance the pinned clock by an
exact amount between two resolves without a real-time Thread.Sleep; two new
tests prove the first-advance snap is exact and a second advance still
eases at the normal rate. The three existing indoor/wind-disabled/amplitude
gate tests pass unchanged.
Verify: Release build 0 warnings/0 errors. App hermetic-lane filter
6,046/0 failed. Core.Tests 4,695/0 failed. Full hermetic-filtered solution:
15,274/0 failed across 15 projects.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Narrow re-review of 43e3abed found every A1-A5/A7/A8 item resolved but
one new blocker in the production packed classifier.
F1 (BLOCKER): RetailPViewPassExecutor.DrawPackedProductionRoute — the
route production world geometry actually draws from — never computed
FoliageFlags at all. WbDrawDispatcher.PackedOracle.cs's
ClassifyPackedBatches built its GroupKey with the field defaulting to
0u, and GetOrCreatePackedGroup never copied it onto the created
InstanceGroup, so production BatchData.flags bits 1/2 were always zero
for every scenery entity: the world geometry never swayed even though
the independently-classified shadow caster did, so shadows visibly
swayed under rigid trees. Both classifier call sites now compute
FoliageFlags via the identical FoliageWindClassification.Classify call
and entity-scoped HasCutoutSubset OR the classic (non-packed) path
uses, and GetOrCreatePackedGroup copies it exactly like
GetOrCreateInstanceGroup always has. The G2/G3 classified-output
digest (AddOpaqueSubmissionGroup/BuildTransparentSubmissionDigest) now
also folds GroupKey.FoliageFlags into its hash — present in the key
since round 1 but never actually read by either digest function, so a
content-level (not just group-count-level) classic-vs-packed
divergence is now caught.
F2 (medium): the delayed-alpha replay path (PrepareDeferredAlphaDraws)
hardcoded Flags = 1, dropping bits 1/2 for any group replayed through
it — a trunk instance promoted into the alpha-blend group mid-fade
(the #188 translucency-promotion case) would stop swaying for the
duration of its fade. Now 1u | key.FoliageFlags.
F3 (nit): ComputeEntityHasCutoutSubset's three call sites (classic,
caster, and the newly-fixed packed classifier) each allocated a
closure over _meshAdapter per Setup entity per frame. A new
context-taking overload passes the mesh adapter as an explicit
argument to a static lambda instead, letting the compiler cache one
delegate for the method's lifetime rather than allocating fresh ones.
A3 test gap: WbDrawDispatcher.BindDirectionalShadowReceiver is now
internal so DirectionalShadowGpuTests can drive it directly with a
bare RecordingGpuDevice pass encoder, proving it emits
UniformAtmosphericFrame with the exact buffer/offset/size a
DirectionalShadowFrameBinding carries — paired with the existing test
proving that binding carries the caster's real bind forward untouched.
F1's missing test: PackedDispatcherOracleTests chains
FoliageWindClassification.Classify (called with the packed
classifier's exact argument shape) for a real 0x8... scenery entity id
through BuildIndirectArrays — the same shared, already-tested
production step both classic and packed group lists feed into BatchData —
proving the resulting flags word carries bit 0x2. Driving
ClassifyPackedBatches/GetOrCreatePackedGroup directly was not a "cheap
test": both are private instance methods reachable only through the
full RetailPViewPassExecutor route, which needs a real IGpuDevice,
world-pass scope, mesh manager, and compiled pipelines to construct —
no test anywhere in the App test project stands one up.
Nits: F4 corrects foliage_wind.glsl's header comment from "bit 31" to
the top-nibble test; F5 documents at the receiver bind site that the
caster's own AtmosphericFrameBufferBinding has its seven ABI v1
members zero/Identity by construction (only the two v2 wind members
are valid) — safe today because mesh_atmospheric.vert reads that
binding solely for wind displacement, flagged as a footgun for a
future v1-reading addition to that shader; F6 notes in the plan
(rather than fixes) that EntityCacheEntry does not proactively
invalidate when FoliageWindExclusions changes on a pack switch —
harmless with the pack off, self-heals on the entry's next natural
eviction.
foliage_wind.glsl's F4 comment-only change updated the SPIR-V
manifest's source hashes for the five includers (mesh_atmospheric.vert
+ four directional_shadow_world_* casters); the compiled .spv bytes
are byte-identical since comments do not affect bytecode.
Verify: Release build 0 warnings/0 errors. App hermetic-lane filter
6,043/0 failed. Core.Tests 4,695/0 failed. RenderPackValidator 30/30.
Full hermetic-filtered solution: 15,271/0 failed across 15 projects.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Opus dual-lens review of the three VM6 commits (0930c35d, 39e8408c,
6cc5e183) found two blockers and two should-fix issues; all landed here
along with the review's nits and documentation corrections.
Blockers:
- A1: the procedural-scenery classifier tested bit 31 alone instead of
the full top nibble (0xF000_0000 == 0x8000_0000), so it also matched
LandblockStaticEntityIdAllocator's 0xC... namespace (fences/gates/
building shells with a cutout subset), the 0xDA11_D0xx paperdoll id,
and the 0xFFFF_FF01 portal-tunnel id as procedural scenery — all
three would have swayed. ProceduralSceneryIdAllocator.IsInNamespace
now does the exact top-nibble test; FoliageWindClassification
delegates to it.
- A2: GroupKey (the receiver's instance-batching key) did not carry
FoliageFlags while the caster's dedup key already did, so a scenery
instance and a non-scenery instance sharing a mesh subset coalesced
into one receiver InstanceGroup whose flags were last-writer-wins —
disagreeing with the correctly-keyed caster. GroupKey now carries
FoliageFlags, computed before key construction and set exactly once
at group creation; the imperative re-stamp is gone, and CachedBatch's
now-redundant FoliageFlags field is removed.
Should-fix:
- A3: the world receiver pass bound UniformAtmosphericFrame only by
accident (leftover from the caster pass, which runs first each
frame, since Vulkan binding state isn't reset between passes).
DirectionalShadowFrameBinding now carries the caster's exact
AtmosphericFrameBufferBinding and BindDirectionalShadowReceiver binds
it explicitly.
- A4: a Setup-composed tree's opaque trunk part never got the trunk
flag because HasCutoutSubset is cached per GfxObj part, not per
entity. FoliageWindClassification.ComputeEntityHasCutoutSubset now
ORs HasCutoutSubset across an entity's resolved sibling parts once
per entity, threaded into ClassifyBatches/AddDirectionalShadowBatches
via a new optional override parameter.
Nits: A5 hashes the per-vertex flutter seed relative to the instance
origin instead of absolute world XY (fp32 sin() precision loss at far
landblock corners), mirrored in both foliage_wind.glsl and
FoliageWindModel; A7 documents the max(maxHeight, 0.5) divide-guard as
a deliberate pseudocode divergence; A8 switches FoliageWindExclusions'
construction to ToFrozenSet() and softens the "never stale" doc
comment to "no slower than one frame behind."
Tests added: top-nibble classification (0xFFFFFFFFu now correctly
false), GroupKey inequality across entity-driven scenery/landblock-
static classification, a caster-batch test proving the same pairing
never coalesces, ComputeEntityHasCutoutSubset unit + end-to-end
two-part-Setup tests, the caster→receiver AtmosphericFrame binding
carry-through, flutter-hash translation invariance relative to
instance origin, and a Storm-wind mid-height displacement floor
guarding against a "no motion" regression.
Docs: plan VM6 body corrected to the five-row WeatherKind table, "bits
1 and 2", "all four" caster shaders, and top-nibble wording throughout;
the owner gate checklist's Rain/Storm step; the stale v1-only shader-
interface compatibility entry; semantic-bindings-v1.md's v2 members
folded into the main 192-byte block; the IA-25 register row's top-
nibble wording; AtmosphericFrameInputs.cs's ABI size reference.
foliage_wind.glsl's A5 change recompiled exactly the five shaders that
include it (mesh_atmospheric.vert, the four directional_shadow_world_*
casters) plus the manifest; no other .spv changed.
Verify: Release build 0 warnings/0 errors. App hermetic-lane filter
6,041/0 failed (no environment-specific failures this run).
RenderPackValidator 30/30. Full hermetic-filtered solution: 15,269/0
failed across 15 projects.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
FoliageWindByDayGroup / FoliageWindDayGroupPoint(int ActiveDayGroup, ...)
becomes FoliageWindByWeather / FoliageWindWeatherPoint(string WeatherKind,
...) in AtmospherePolicyDeclaration (Plugin.Abstractions is BCL-only, so
the key is the exact member name of AcDream.Core.World.WeatherKind rather
than the enum itself). The raw activeDayGroup index carries no weather
meaning by itself; WeatherState.cs already classifies each day group's
authored DAT name into one of five real weather kinds, and that fact was
already threaded through AtmosphericFrameInputs.Weather / uAtmosphereWeather.x
— this reuses it instead of guessing an index-to-category mapping.
Built-in table (BuiltInAtmosphericRenderPack.AtmospherePolicy()): Clear
0.25/0.15, Overcast 0.60/0.35, Rain 0.85/0.60, Snow 0.35/0.20, Storm
1.00/0.75 — all five WeatherKind members declared, the invented "Cloudy"
row dropped. RenderPackAtmospherePolicyEvaluation.FoliageWind now takes a
WeatherKind and matches by weather.ToString() (ordinal) against each
declared point's name; a kind absent from the table falls back to the
declared Clear row, then to (0,0) if Clear itself is undeclared. The
delta-seconds EMA interpolation (EaseTowardTarget) is unchanged.
AtmosphericPostProcessGraph.ResolveFoliageWind and its two callers
(RenderPostProcess via inputs.Weather; RenderDirectionalShadows via
foundation.Atmosphere.Kind) now pass WeatherKind instead of the day-group
int.
RenderPackValidation.ValidateAtmosphere (runs for every pack declaring an
AtmospherePolicy, not gated to Tier2/shadow packs) now rejects an unknown
or non-exact-case weather-kind name and a repeated kind, mirroring the
existing ActiveDayGroupMultiplier duplicate-key check.
Tests: RenderPackAtmospherePolicyEvaluationTests rewritten for the
kind-keyed API (all five kinds resolve to their declared row, an unlisted
kind falls back to Clear, ordinal exact-case matching, null-table
handling); RenderPackSpirvValidatorTests gains four descriptor-validation
cases (unknown name, wrong case, duplicate kind, the five-kind table
accepted); AtmosphericPostProcessGraphTests' three foliage-wind cases now
select WeatherKind.Storm via `with` instead of an assumed day-group index.
Spot-check (per the coordinator's ask, not changed here): yes —
ActiveDayGroupMultiplier / EvaluateDayGroupPolicy (pre-existing, Campaign
AR/VM3-era — BuiltInAtmosphericRenderPack.AtmospherePolicy()'s three rows
`new ActiveDayGroupMultiplier(0, 1.0), (1, 0.35), (2, 0.20)`) key the
sun-ray/shadow/volumetric day-group strength multiplier by the same raw
activeDayGroup index with an undocumented assumed meaning (0=brightest ...
2=dimmest), the identical class of issue this commit fixes for foliage
wind. Left unchanged per instruction; flagging for the coordinator to file.
Full solution Debug and Release builds green. App hermetic filter
6024/6026 — the same 2 pre-existing failures as VM6a/VM6b. Both were
re-run in isolation per the verification ask: both still fail alone (not
a load-flake in this environment) — confirmed via git stash earlier this
session that both already fail on the unmodified pre-VM6 baseline, so
they are pre-existing and unrelated to this change. Core.Tests hermetic
4697/4697. RenderPackValidator.Tests 30/30. No shader/spv changes in this
commit (pure C#/docs fix).
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Procedural-scenery foliage (trees/bushes — entity ids in the
ProceduralSceneryIdAllocator's 0x8XXYYIII namespace) sways with weather in
mesh_atmospheric.vert and all four directional_shadow_world_* caster vertex
shaders, both calling the identical new foliage_wind.glsl include so the
shadow moves with the leaf by construction.
Classification (FoliageWindClassification, AcDream.App.Rendering.Wb): two
new BatchData.flags bits, computed once per (entity, subset) from four
inputs — entity id (bit 31 for procedural scenery), the pack's declared
FoliageExclusions membership, the subset's TranslucencyKind, and
ObjectRenderData.HasCutoutSubset (computed once per mesh at build time, not
per frame). Bit 1 marks an alpha-cutout leaf subset; bit 2 marks an opaque
trunk subset (only when its own mesh also owns a cutout subset, so rocks
stay still). WbDrawDispatcher.ClassifyBatches (world receiver) and
AddDirectionalShadowBatches (caster) call this with the same four inputs, so
casters and receivers classify identically without needing to share state.
Retail's mesh_modern/terrain_modern/mesh_detail pipelines never read these
bits, so pack-off output is unaffected.
Motion model (foliage_wind.glsl, mirrored bit-for-bit in the new
FoliageWindModel for hermetic CPU tests): height-squared-scaled slow lean
for every foliage subset, plus branch swing and per-vertex-hash-decorrelated
flutter for cutout subsets only. AtmosphericPostProcessGraph.ResolveFoliageWind
resolves the wind block once per frame.Serial — advanced by whichever of
RenderDirectionalShadows (which runs first) or RenderPostProcess is called
first that frame, with the second reading the already-advanced state, which
is what keeps the caster and receiver reading byte-identical clock/strength
values. The per-day-group mean/gust target (AtmospherePolicyDeclaration.
FoliageWindByDayGroup, keyed by the same day-group index convention
ActiveDayGroupMultipliers already established: Clear/Cloudy/Overcast/Rainy)
eases toward its target over WeatherSystem.TransitionSeconds (10s) so a
weather change never snaps; wind-enabled off or indoor instead gates the
OUTPUT to an exact zero (not an asymptotic approach) so a settings toggle or
cell transition is immediate. The wind clock is a Stopwatch started at graph
construction (monotonic, session-relative magnitude for GPU sin() accuracy),
overridable by the same ACDREAM_SKY_PHASE_SECONDS pin SkyRenderer already
uses, for deterministic offline gates.
New settings: wind-enabled, wind-strength, wind-direction-degrees (225°
default — no authored retail wind direction exists to read),
wind-lean-metres, wind-branch-metres, wind-flutter-metres (0 on Low),
wind-canopy-height-metres.
Register row IA-25 files this as an intentional, strictly opt-in divergence:
retail applies no per-vertex wind displacement to any geometry. Known,
accepted limitation: classification is per mesh-subset (one BatchData.flags
word per indirect-draw batch), not per entity instance, so the rare case of
one mesh subset being reachable from both a procedural-scenery and a
non-scenery placement would classify all of that subset's instances alike.
Tests: FoliageWindClassificationTests (the full classification matrix),
FoliageWindModelTests (identity on non-foliage/calm-wind/base-vertex,
canopy-top displacement bound, z-never-increases, trunk has no flutter
term), RenderPackAtmospherePolicyEvaluationTests (exact day-group lookup,
no interpolation across day-group ids, easing convergence without overshoot
or discontinuity), AtmosphericShaderAbiTests (each of the five shaders calls
acdreamFoliageDisplace exactly once; mesh_modern/terrain/mesh_detail call it
never), and four AtmosphericPostProcessGraphTests additions (indoor/disabled
exact-zero gating, settings-to-UBO wiring, same-frame-Serial idempotency —
the last proxies the caster/receiver agreement invariant without needing
this hermetic harness's WbDrawDispatcher/TerrainModernRenderer dependency
chain to exercise RenderDirectionalShadows directly).
App hermetic filter: 6015/6017 (the same 2 pre-existing failures as VM6a,
confirmed unrelated). Core.Tests hermetic: 4697/4697. RenderPackValidator.Tests:
30/30. Full solution Debug and Release builds green. Shader recompile
touched exactly the 5 edited files' .spv (plus manifest); the retail oracle
set and every other pack shader are byte-identical.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Erik