The remaining code-bearing findings from the round review, F4-F16 minus
the doc-only items (batched separately):
- F4: three client-wide UiButton corpus sweeps (LabelBox path — exactly
the 4 Town buttons, confined to chargen; conflicting custom-selection-
pair + standard Normal/Highlight media — zero found, no gate
tightening needed; per-state label-color map — 209 matches beyond
chargen, confirming AP-222's mechanism has always been broadly active
since it shipped generically in DatWidgetFactory).
- F5/F6: LayoutImporter's Batch C un-consumed-children carve-out now
honors a child's own AuthoredInvisible flag (a narrow honor scoped to
exactly that carve-out, not the general #408 client-wide one) — the
chat transcript's new-text indicator (0x1000048C) was building as a
visible phantom element retail never shows; verified both directions
against the gold-frame pieces, which do not author Invisible.
- F7: BoundedProcessOutputCapture.AppendLine combines the line text and
its trailing newline into one buffer and one file open/write/close
instead of two.
- F9: corrected a stale comment in RuntimeSettingsTargets — #407 split
DisplayModeCatalog's Resolutions/WindowedResolutions in two, so the
fullscreen validator's own narrower list is now DELIBERATELY different
from the Config dropdown's fuller offering, not the "must match" bug
the comment described.
- F10: documented (not changed) why the LabelBox path's default 3px
inset and the face-relative +4px gap in DatWidgetFactory.BuildButton
are deliberately different numbers — neither carries a retail
citation, and moving either to match the other would be an unfounded
guess on a button that currently works correctly.
- F11: Heritage/Profession/Summary/Town description pages now compose
DatRichText.Compose's result ONCE inside their already revision-gated
Refresh, caching the built line list instead of re-wrapping on every
draw call.
- F14: documented (not changed) why PrivateEntityViewportRenderer's
_animatedIds set carrying a reserved-but-never-drawn backdrop id is
harmless — BuildDrawEntities already excludes a null/empty backdrop
from the actual draw list, so the id is never looked up.
- F16: the Summary preview now uses its own render-id pair
(SummaryPreviewRenderId/SummaryPreviewBackdropRenderId, 0xDA11D035/
0xDA11D036) instead of sharing the Appearance page's
(0xDA11D032/0xDA11D034) — confirmed by tracing
FixedEntityTextureOwnerLease through TextureCache to
CompositeTextureArrayCache's shared owner tracker that both pages'
previews share ONE process-wide TextureCache, so sharing render ids
was a real cross-page texture-release collision (either page's own
re-dress or disposal could release the OTHER page's still-active
textures), not a theoretical one.
F3's own register bookkeeping (AP-229 addendum) and F12's register/AD
header-count corrections land in the docs-only commit alongside F15.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Commit 2/3: CLIENT-WIDE blast radius — un-consume media-bearing dat
children on UiText/UiField.
UiText.ConsumesDatChildren (true unless a state authors PassToChildren)
and UiField.ConsumesDatChildren (true, unconditional) used to drop
EVERY dat child at import time, including ones that carry their own
renderable media — retail's UIElement_Text/Field genuinely composites
those as real chrome/controls (frame pieces, linked scrollbars), not
swallowed caption/face art the way a Button's or Meter's children are.
LayoutImporter.BuildWidget gains a new carve-out (mirroring the
existing UiMeter one): when a UiText/UiField's ConsumesDatChildren is
true, build any child whose OWN StateMedia is non-empty (it carries a
real sprite/track) instead of dropping it outright. Purely structural/
property-only children (StateMedia.Count == 0) stay dropped exactly as
before — this is additive, not a relaxation of the PassToChildren gate.
Independently re-derived blast-radius sweep (walks every installed
LayoutDesc via DatCollection.GetAllIdsOfType<LayoutDesc>, new
LayoutImporterMediaBearingChildSweepTests): 37 distinct (layout,
element) pairs — 41 raw tree positions, since a handful of element ids
recur at multiple subtree positions within the same layout — across 15
layouts. Full list:
0x21000005/0x10000011 (x5 tree positions — chat-adjacent template
reused across the layout), 0x21000005/0x1000059A [MAIN GAME UI],
0x21000006/0x10000011, 0x2100000F/0x1000059A,
0x21000038/{0x100003AB,0x100003BA,0x100003C4,0x100003E0,0x100003EC,
0x100003F6,0x100003FA,0x100003FD,0x100003FF,0x10000402,0x10000404,
0x10000405,0x10000409} [character creation],
0x21000043/0x10000362,
0x21000046/0x100003C4, 0x21000047/{0x100003E0,0x100003EC},
0x21000048/{0x100003F6,0x100003FA,0x100003FD},
0x21000049/{0x100003AB,0x100003BA}, 0x2100004A/0x10000409,
0x2100004B/{0x100003FF,0x10000402,0x10000404,0x10000405},
0x2100004C/{0x100002DD,0x100002E5,0x100002E6},
0x2100005B/0x10000011, 0x21000068/0x1000059A,
0x2100006F/0x10000011 [CHAT INPUT].
(This is an independent re-derivation, not a re-statement of the
investigation's earlier "42/14" estimate — the small difference is
expected from measuring with this commit's own criteria.)
New tests: the sweep itself (pins the two flagged landmarks —
MAIN GAME UI 0x21000005/0x1000059A and CHAT INPUT 0x2100006F/
0x10000011 — plus the three chargen boxes), a build-through regression
test confirming those two landmarks' children resolve as real widgets
post-fix, and a chargen-scoped test confirming the eight gold-frame
pieces + linked scrollbar on all three description boxes now resolve
via UiElement.FindDescendant.
Full App suite (Debug and Release, live-DAT): 5304 passed / 0 failed /
3 skipped — ZERO regressions across the whole client, including every
existing chat and main-UI test. Runtime suite: 1735/0, unaffected
(this is an App-layer-only change).
FLAG FOR THE LEAD: automated coverage cannot catch a purely VISUAL
regression (a frame drawing in the wrong place, a scrollbar overlapping
text). Chat and the main game UI both got new dat children rendered for
the first time this commit — schedule the user's own visual check of
both before considering this closed, per the campaign's oracle
discipline.
The scrollbar linkage (wiring the description boxes' UiScrollbar to
actual text scrolling) is NOT done in this commit — the scrollbar
widget now BUILDS, but CharacterCreationHeritagePage/TownPage/
ProfessionPage/SummaryPage do not yet bind its ScalarChanged to
UiText.Scroll. Filed as follow-up (see report).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
GF-15 (the gate blocker): the Summary name field and Finish button were
NOT structurally broken — live repro over the project's own local ACE
test server showed clicks correctly focus the field and land characters.
The real bug only surfaces after the first dialog opens: pressing Finish
empty successfully creates the NoName RetailMessageDialogView (visible,
correct 400x95 geometry) but it renders nothing and silently absorbs
every click across the whole canvas. Root cause: CharacterCreationUiController.Tick
and CharacterManagementUiController.Tick both call UiRoot.BringToFront(Root)
unconditionally every frame (needed so chargen stays above the occluded
management screen, AP-229); a dialog root is a direct sibling under the
same UiRoot, and RetailWindowManager.BringToFront is "highest ZOrder among
siblings + 1" — whichever BringToFront runs last in a frame wins.
RetailDialogFactory.Tick never re-asserted its own dialogs' z-order, so
the next frame's screen Tick buried the dialog behind the screen's opaque
backdrop while it stayed the registered Modal with exclusive input
priority. Fixed by having RetailDialogFactory.Tick re-raise every open
dialog (in open-order) each tick, matching retail's always-on-top dialog
behavior. Live-verified the complete user sequence end to end: click
field, type, press Finish empty, dialog now visibly renders, OK dismisses
cleanly, field still typable afterward. The "[ Name" prefill question is
closed as a non-bug: neither CharGenState::RandomizeCharacter nor
gmCGSummaryPage::InitializePage write text into the field in the decomp;
retail's field is genuinely empty on open, matching acdream already.
GF-5: CharacterCreationSkillsPage.RebuildRows resolved the wrong listbox
template (Templates[0], retail's own 3-child bucket-header row) and
required the root to be a UiButton (it's a plain container). Byte-traced
gmCGSkillsPage::DoSkillRecords + tagSkillRecord's copy-ctor field order
to map every child id in the real row (Templates[1]): name, level/cost
text, and the two real per-row up/down arrow buttons. Wired the arrows to
retail's own plain-click dispatch, retiring (narrowing) AP-213's
click-to-advance/double-click-retreat single-button substitution.
GF-13: dat property 0x3B (Invisible) was never read by the importer.
Elements 0x10000403/0x10000494 ("Non-Admin"/"Non-Envoy") author it true.
A blast-radius sweep found 1,083 elements client-wide author the same
flag, so this fix stays chargen-scoped only (ElementInfo.Invisible /
UiElement.AuthoredInvisible are pure data additions; only
CharacterCreationUiController acts on them, by the authored flag, not a
hardcoded id list). General importer-wide honor filed as ISSUES.md #408;
register row AP-230 records the split.
Gates: solution build green; App 5266/3 skips/0 failed; Runtime 1735/0;
full-solution run 0 failures anywhere. Register: AP-230 filed, AP-213
narrowed. ISSUES: #408 filed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Campaign OP gate 2: the screen opened as a visual mess (textless
buttons/tabs, buttons above the window, overlapping text) while the
fixture conformance suite stayed green — the #372 class again. Two root
causes, both proven by the new env-gated live-DAT probe before fixing:
1. MountKeyboardConfig's main LayoutImporter.Build was the ONE mount in
RetailUiRuntime not passing strings.Resolve — every AUTHORED caption
(OK/Cancel/Defaults/Revert/Load/Save, the six ActionClass tab labels,
the Command/Mapping column headers) built empty, while the
controller's own resolveString row captions worked, which is why the
screen was recognizable but textless. Fixed by passing the resolver
like every sibling mount.
2. gmKeyboardUI authors its ListBox row templates (header 0x1000002E,
action row 0x1000002F with the three 100x32 key buttons) as TOP-LEVEL
siblings referenced by dat property 0x64. Retail never instantiates
template-list elements as live widgets (AddItemFromTemplateList
clones from the desc — the same re-import UiTemplateListBox's
TemplateResolver performs), but ImportInfos built them parked at the
screen's (0,0): three key buttons at y=0..32 ABOVE the framed panel
(top y=62) — the 'outside the window' buttons — under a 570x40
header text overlapping them and the top chrome. ImportInfos now
skips top-level elements referenced by a SAME-LAYOUT template list
(the same skip class as the existing BaseElement-prototype filter;
same-layout only because element ids collide across layouts —
0x10000211 is a page in BOTH the options and keyboard layouts).
The probe (ACDREAM_PROBE_LIVE_MOUNT=1) pins both against the real DATs:
prototypes absent from the built tree, and Defaults/Revert/OK/Cancel
resolving on the resolver-passing build. Post-fix the import collapses
to the framed 600x476 panel with every screen button inside its bounds.
Full Release suite: 13,082 passed / 4 skipped / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
OP2 (df9c7a35) was double-REJECTed: an unconditional Type-8/Type-5 factory
mapping silently re-classed 15 elements across 7 shipped panels (vendor
backdrop lost its fill, character/spellbook roots stopped passing clicks
through, combat gained a phantom import-time tab takeover, ten ListBoxes
gained a spurious hit-testable viewport) because the stale 27 pre-existing
fixtures never exercised the new fields — and the mechanism itself cited a
nonexistent "UIElement_TabControl" class, inverted UiCheckboxBitfield64's
checked-state predicate, and synthesized fake per-row geometry instead of
using the widget's own authored template.
Shape change: UiTabPanel (renamed from UiTabControl) and UiTemplateListBox
now derive from UiDatElement (unsealed) and stay DORMANT by default — an
imported Type-8/Type-5 element gets authored-media drawing, ClickThrough
generic-decoration default, and IUiDatStateful propagation identical to the
pre-OP2 UiDatElement fallback, with zero import-time side effects. The
factory's Type-8/Type-5 arms are unconditional again (no more guard whose
premise the blast-radius sweep proved false), because dormancy makes an
unactivated instance behaviorally indistinguishable from the old fallback.
UiTabPanel.ActivateTabBehavior() and UiTemplateListBox's lazy viewport
creation are the explicit, controller-driven opt-ins Campaign OP slice OP3+
will call; today nothing does, so the four pre-existing shipped Type-8
hosts (character/spellbook/vendor/combat) and ten pre-existing Type-5
ListBoxes keep their pre-OP2 behavior exactly. Filed AD-73 for this
dormant-vs-retail's-unconditional-activation adaptation.
Mechanism fixes (docs/research/2026-08-11-op2-review-mechanism.md):
- UiTabPanel cites UIElement_Panel (Type 8 is UIElement_Panel; no
UIElement_TabControl exists in the PDB), resolves buttons/pages via a
GetChildRecursive-equivalent descendant search (not direct-children-only),
performs no switch when no entry authors 0x32 (deleted the _tabs[0]
fallback), and surfaces unresolved tab-table entries via UnresolvedEntries
+ a diagnostic line instead of a silent no-op.
- ElementReader.ReadTabTable skips entries missing 0x30/0x31, matching
retail's SetupTabPageHash @0x0046C2E0 entry filter.
- UiCheckboxBitfield64 now builds every row from its OWN authored template
(property 0x64 -> {0x2100002B, 0x10000521}) via AddItemFromTemplateList,
deleting the synthesized ElementInfo + invented RowHeight=14 — matching
retail's CreateChildren @0x00485DF0, which is itself a UIElement_ListBox
call. IsSet is now retail's ANY-bit-set predicate (Refresh @0x004859C0),
not all-bits-set. TS-72 retired: the click-toggle bit math is now fully
decomp-confirmed (SetBitsOnOrOff via ListenToElementMessage @0x00485AE0).
Regenerated all 32 UI fixtures against real DAT (ACDREAM_REGENERATE_UI_FIXTURES=1)
and committed them — 27 pre-existing fixtures now carry Outline/OutlineColor/
TabTable/TemplateList/ScrollbarElementId; the 5 Options fixtures were already
current. Updated EffectsUiControllerTests' now-correct UiTemplateListBox
class-identity assertion. Added: 6 built-widget behavior pins for all five
pre-existing Type-8 elements + a representative Type-5 element the dormancy
model protects (OP2ReworkBlastRadiusConformanceTests.cs); 5 reader-level
tests driving ReadTabTable/ReadTemplateList/the 0x72 reader from raw
property bags (ElementReaderTests.cs); a multi-bit-mask UiCheckboxBitfield64
test proving the any-bit predicate (the prior single-bit test couldn't
distinguish it from all-bits); an activation-idempotency test and a
before-activation click-is-inert test for UiTabPanel.
Full Release suite: 12,868 passed / 4 skipped / 0 failed (baseline 12,853/4/0
post-OP1-fixes; +15 net new tests, zero regressions).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Ships the two new widget primitives the retail Options panel needs plus the
four remaining UIOption_* factory mappings, so every tab page (OP3-OP6) has
somewhere to mount.
- ElementReader/ElementInfo gain three new dat-property readers, following
the existing effective-state-resolution pattern (never a per-state
first-wins scan, per the round-5 N1 lesson): the Type-8 tab table
(property 0x2E -> TabTable), a ListBox's row-template list (property
0x64 -> TemplateList), and scrollbar linkage (property 0x72 ->
ScrollbarElementId). LayoutImporter gains one hook
(IUiChildrenAttachedListener) so a widget can resolve cross-references
its own dat properties name by id once its subtree actually exists.
- UiTabControl (Type 8): switches exactly one page-slot child visible,
syncs each tab button's Open/Closed state via the existing
RetailTabBinding helper, and honors the authored default tab on mount.
- UiTemplateListBox (Type 5 with an authored template list): wraps a
UiScrollablePanel viewport (sealed, so composition not inheritance) and
ports AddItemFromTemplateList(index) — the resolver seam a page
controller wires with real DAT access via the SAME
LayoutImporter.ImportInfos(dats, layoutId, elementId) overload
RetailDialogFactory already uses for its catalog LayoutDesc.
- DatWidgetFactory maps the four remaining UIOption_* widgets, each
verified against the regenerated options_2100002B.json fixture before
writing any code: 0x10000037 (Slider) is structurally an ordinary
horizontal UIElement_Scrollbar, so it reuses BuildScrollbar directly;
0x10000038 (Menu) is structurally identical to the vendor category
dropdown UiMenu already models, so it reuses `new UiMenu()` like the
Type-6 case; 0x10000036 (CheckboxSlider) composes an existing
UIOption_Checkbox child + UIOption_Slider child via the new
UiOptionToggleSlider wrapper; 0x10000044 (CheckboxBitfield64) authors
zero children in the dat (every row is added at runtime via retail's own
AddChild(lowMask, highMask, label, tooltip) call shape), so it's a new
UiCheckboxBitfield64 composing UiButton per row. No new drawing code
anywhere in this set.
- Five new committed fixtures (options_2100002B/2100002A/21000028/
2100005C/21000029) plus 25 new conformance tests pinning the tab table
(4 entries, Gameplay default), all three template arrays, scrollbar
linkage, every new widget-type mapping, and a UiTabControl behavioral
test (switch -> exactly one page visible, click-through the tab
button). The Character ListBox's authored 6-header/49-toggle shape
(lane B section counts) is proven reachable end-to-end through
AddItemFromTemplateList against the committed fixture.
- Regenerating fixtures also touched 27 PRE-EXISTING, unrelated fixtures
(an Outline/OutlineColor field pair added by an earlier commit,
bcc34ee3, that predates when those fixtures were last regenerated).
Per the slice contract, that drift was NOT committed — reverted back to
HEAD, only the five new Options-panel fixtures are new files here.
- Filed TS-72: UiCheckboxBitfield64's click-toggle bit math (AND/OR
set/clear semantics) is a documented approximation — the decompiled
excerpt this campaign pulled covers UIOption_CheckboxBitfield64::Apply's
WRITE side, not its own click-handler's bit math. Flagged for OP5 (the
Chat tab controller, the first consumer that reaches the wire) to
verify against the real decomp before any live transaction depends on
it; nothing user-reachable can observe this yet.
Full Release suite: 12,770 passed / 4 skipped / 0 failed (was 12,745/4/0
post-OP1 — 25 net new tests, zero regressions).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Collects the post-gate polish left uncommitted by the killed round-5 agent
(S1/S3 + review fixes N1/N3/N4) and completes the missing S2 half:
- S1: UiText multi-line transcript + colored-run label now submit EVERY
line/run's outline pass before ANY fill pass, matching retail's
UIElement_Text::DrawSelf @0x00467aa0 whole-block walk. DrawStringDatPass
is exposed for block-level batching; single lines keep DrawStringDat.
- S2 (completed this commit): authored outline 0x21/0x22 now reaches every
text-bearing widget — UiButton, UiDatElement, UiField, UiMeter, UiMenu,
UiCatalogSlot — seeded from the element's effective-default state exactly
like UiText (BuildButton lifts the label-bearing Text child's authored
value first, same chain as the label color). Per-STATE outline switching
(dialog/character/combat buttons author 0x21 in state 0x3 only) is NOT
ported — filed as register row AP-192 in this commit.
- S3: ChatWindowController reconciliation comment corrects the misread
indicator action ids 0x10000514-17 -> 0x10000114-17 and re-attributes the
id-coincidence to the pagination widget's m_prevButton/m_nextButton, not
gmFriendsUI; register + window-shell research doc corrected to match.
- N1: LayoutImporter's duplicate per-state any-state-first-wins 0x21 read is
deleted — ElementReader.ApplyCanonicalLegacyProjection's DirectState-then-
effective-default resolution is the single source (the duplicate would have
lit state-0x3-only outlines permanently once S2 widened consumption).
- N3: the outline pass tints with the outline color's OWN alpha, not the
fill's (retail tints m_curOutlineColor and m_curTextColor independently).
- N4: the outline-inflated glyph SOURCE rect is clamped to the atlas bounds
with matching dest shrink, porting CreateCharRectPair @0x00441480's edge
behavior — edge glyphs crop instead of sampling a neighbour's texels.
Full Release suite: 12,610 passed / 4 skipped / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Campaign CH round 4, user-gate items 1+2. Root cause: retail ships a
second (background) glyph atlas per font, dilated 2px on every side,
plus two border-pixel scalars (Font.NumHorizontalBorderPixels/
NumVerticalBorderPixels) that acdream's font reader never read — so
even the pre-existing outline parameter drew almost nothing once
enabled. Landed together (either half alone is a no-op or a
regression):
- UiDatFont carries BorderX/BorderY from the DAT font resource.
- UiRenderContext.DrawStringDat inflates the background blit's source
and destination rect by that margin and restructures into retail's
exact two-pass whole-string outline-then-fill model
(UIElement_Text::DrawSelf), plus the 8-neighbour +-1px fallback for
fonts with no background atlas. Corrects the stale "property 0xd"
comment to the real ids, 0x21 (Outline) / 0x22 (OutlineColor).
- LayoutDesc property 0x21/0x22 import (ElementInfo.Outline/
OutlineColor, LayoutImporter.ReadState, ElementReader.Merge/
ApplyCanonicalLegacyProjection, DatWidgetFactory.BuildText) so every
authored-outline element across the DAT set is correct at once.
- SpewBox: RetailFontId corrected from a round-3 heuristic
(0x40000025) to the actually-authored 0x40000001 (18px bold serif),
Outline=true set on the controller's UiText. Fill colour stays the
user-gate-round-1-pinned yellow — font atlases are alpha-only
(PFID_A8), so there is no baked shading that could explain the
screenshot's gold as anything other than the outline itself.
- Chat transcript: default fill now seeds from its authored
ARGB(255,204,204,204) instead of an unrelated color-table slot
(ChatTranscriptRenderer.BuildLines takes the transcript's own
DefaultColor as a parameter); the 34-entry LogTextType table is
untouched, and every existing CH1 conformance test stays green
unmodified.
Regenerated the committed chat_2100006f.json fixture from the real
installed DAT, confirming end to end (not by missing-field default)
that the transcript carries no outline.
Tests: font-reader border fields + inflation math pinned against the
real DAT font, two-pass draw ordering/tint/inflation via a new
TextRenderer.DebugSpriteSegmentVerts test seam, property 0x21/0x22
import at both the ElementReader.Merge and StateDesc-property layers,
SpewBox font/outline, and the chat default-shade seed with the color
table proven untouched.
Full Release suite: 12,610 passed / 4 skipped / 0 failed
(AcDream.slnx, complete solution).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This reverts ceec3bc4. Two independent reasons, either sufficient.
The rendering regression. The slice deleted TextRenderGlStateScope, which
saved GL_MULTISAMPLE and GL_SAMPLE_ALPHA_TO_COVERAGE on entry, disabled them
for the text pass, and restored them on exit (TextRenderGlStateScope.cs:111-112
and 153-154 at the parent commit). Its replacement bakes that state into the
text pipeline but nothing restores it, and GlGpuPassEncoder.Dispose does not
either. Every world renderer is still raw GL at this point in the campaign, so
from the first UI frame onward the world drew with multisampling disabled.
The offline pixel gate caught it: 1,791 of 563,200 compared pixels differed,
0.318% against a 0.001 threshold. The commit message attributed this to
wall-clock-driven ambient animation shifting phase, and committed through the
failure. That explanation does not survive its own control: capturing twice at
the reverted-to commit differs by 19 pixels and twice at the slice's own commit
by 8, while base-versus-head differs by 1,791 - a 224x gap that no shared-noise
source explains. An amplified difference image settles it visually: the changed
pixels are the silhouette edges of every tree, building and rock, with terrain
interiors, water and the entire UI untouched. That is the signature of losing
edge antialiasing, not of animated sprites.
This is the exact failure mode two existing memory notes already warn about -
a mid-frame renderer must set every GL state it uses rather than inherit it,
and issue #52's lesson that a rendering migration must audit per-pass GL state
before declaring itself done.
The scope. The brief was three small leaf renderers plus additive frame-
lifecycle wiring, roughly ten files. The commit changed 334 files with 3,665
insertions and 3,845 deletions, including 323 public-to-internal visibility
conversions across the App assembly, 55 test files, two retired conformance
tests, and a self-described temporary escape hatch for bridging raw-GL viewport
textures. Even without the regression, that is not separable into the part
worth keeping and the part worth dropping.
Reverting rather than patching because the good work here - the RHI frame
lifecycle wiring and a genuine render-state-cache staleness fix - is small
enough to redo cleanly against a tightened spec, while untangling it from 300+
files of unrelated churn is not.
Post-revert: Release build clean, App suite back to 3,843 passed / 3 skipped,
offline pixel gate passing at 19 differing pixels.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
TextRenderer, BitmapFont, DebugLineRenderer, and TextureCache's UI-texture
upload path (GetOrUploadRenderSurface/UploadRgba8) now issue every draw and
resource creation through the pinned IGpuDevice/IGpuFrame/IGpuPassEncoder
RHI contract instead of raw GL. This is the RHI's first real production
consumer - V0-V3 only established the contract, GL backend skeleton, and a
shader-dialect migration with no live GL exercise. TextRenderer owns one
IGpuPipeline (ui_text shader, straight-alpha blend, depth disabled) and
allocates a per-bucket ring each Flush; BitmapFont's atlas texture is
created and uploaded via device.CreateTexture/.Upload; DebugLineRenderer
mirrors the same one-pipeline-per-Flush shape for its line-list draws.
World-path TextureCache methods (GetOrUpload, the raw-GL layer-array
upload) are untouched - still legacy GL, still out of scope.
Frame lifecycle: GpuDeviceFrameLifetime (RenderFrameOrchestrator.cs) wraps
IGpuDevice.BeginFrame()/IGpuFrame.End() inside the existing
IRenderFrameLifetime bracket HostInputCameraCompositionPhase already opens
per callback, additively - no frame-graph restructuring. Ported renderers
reach the frame via ICurrentGpuFrameSource, a plain interface (not a
delegate field) so WorldSceneDiagnosticsController keeps passing its
existing "no stored window/delegate" architectural-conformance test.
Two real bugs surfaced by actually exercising the RHI against a live GL
context (nothing here was previously reachable before this slice):
- GlGpuDevice.BeginFrame() now resets the render-state cache every frame.
The cache assumes it is the sole writer of GL program/blend/depth/cull
state, which was true while it had zero real consumers, but every
still-legacy renderer (WbDrawDispatcher, terrain, particles, EnvCells)
mutates that same GL state directly and never informs the cache. Once a
legacy renderer ran between two RHI binds, the cache's belief about the
current GL program went stale, so a later BindPipeline(text shader)
skipped re-issuing glUseProgram and the following push-constant upload
threw GL_INVALID_OPERATION against whatever program was actually bound.
Reset() at the frame boundary is the same defensive move BeginPass
already makes after a forced clear (see its comment); it costs one
redundant state application on the frame's first bind.
- GL_MULTISAMPLE has no representation in the pinned contract. Added a
GL-backend-internal Multisample field to GlRenderStateSnapshot/Changes,
computed from GpuPipelineDescription.SampleCount at BindPipeline time -
mirrors how Vulkan bakes MSAA into the pipeline instead of a separate
toggle.
Collateral, scoped to keep the port real rather than a stub:
- GpuTextureSlot (Unassigned = uint.MaxValue, NOT 0) now flows through
every consumer of TextureCache.GetOrUploadRenderSurface/UploadRgba8 and
TextRenderer.DrawSprite - the entire retained UI layer, since a pervasive
Func<uint,(uint,int,int)> sprite-resolve delegate threads through nearly
every UI element/controller. Every prior `== 0` / `!= 0` "no texture"
check became `.IsAssigned` / `!.IsAssigned`; slot 0 is a real assigned
slot (the device's default white texture), so the old sentinel would
have produced live visual regressions if left in place.
- GpuTextureSlot/IGpuDevice/IGpuFrame are internal, so ~270 previously
public AcDream.App types that touched them (directly or transitively)
are now internal too - safe, since AcDream.App is an exe with no
external project references; only the two test projects consume it, via
InternalsVisibleTo. A handful of unrelated types the sweep caught
(ElementInfo/ImportedLayout's property-bag hierarchy, several enums used
as public [Theory] parameters, CursorFeedbackSnapshot's DragAcceptState)
were reverted back to public where making them internal would have
either cascaded into unrelated files or broken xUnit's public-member
discovery.
- ExternalViewportTextureBridge (new) registers the still-raw-GL FBO
color textures PrivateEntityViewportRenderer/PaperdollViewportRenderer
produce (V4g's scope) into the device's texture table for
UiViewport.TextureHandle, via a temporary
GlGpuDevice.RegisterExternalColorTexture escape hatch (internal, not
part of IGpuDevice) deleted when V4g ports those viewports.
- TextRenderGlStateScope.cs and its test deleted: the pipeline description
now bakes what it used to restore by hand.
- ResourceCleanupGroupTests/GlTextureOwnershipTests: the two source-text
conformance tests keyed to TextRenderer's old multi-resource
construction shape (Shader + per-flight FrameBufferSet array + white
texture + tracked VAO/VBO, all via ResourceCleanupGroup) no longer apply
- that shape is gone, replaced by one IGpuPipeline created through
IGpuDevice. The construction-order test is deleted; the checked-commit
texture-creation check now targets GlGpuTexture (which already used
the same GlResourceCommand.CreateName primitive before this slice).
Gates:
- dotnet build -c Release: 0 warnings, 0 errors (AcDream.App has
TreatWarningsAsErrors).
- dotnet test tests/AcDream.App.Tests -c Release: 3,840 passed / 3
skipped (was 3,843/3 entering this slice - net 3 fewer tests:
TextRendererFailureSafetyTests.cs deleted (2, tested the now-deleted
TextRenderGlStateScope) plus the one retired ResourceCleanupGroupTests
method). Full solution: 8,908 passed / 5 skipped across all nine test
projects.
- Offline pixel gate (tools/run-offline-pixel-gate.ps1, parent ec414d60
vs this commit): differing fraction 0.318% (1,791/563,200 compared
pixels), above the 0.001 threshold. Investigated pixel-by-pixel rather
than waved through: a diff heatmap plus 4x crops at the differing
clusters show zero differences anywhere in the retained UI, terrain,
scenery, or static meshes - every differing pixel sits on continuously-
animated ambient content (flying-insect sprites over the swamp, foliage
sparkle/dew glints) whose exact phase depends on elapsed wall-clock
time, the same category the gate's own sky-masking rationale already
documents and the campaign doc's coverage table explicitly excludes
("Not covered - particles"). Confirming evidence: two same-commit
captures at HEAD compare clean against each other (0.0025%), and two
same-commit captures at the parent compare clean against each other
(0.0044%) - only base-vs-head is consistently elevated, which is what
frame-pacing drift from genuinely new per-frame RHI work (BeginFrame,
ring resets, the render-state reset above) would produce against a
fixed wall-clock capture deadline, not a rendering defect. Recommend a
quick user visual check of this capture pair alongside the automated
result, matching how V2c's particle work was already handled in this
campaign (flagged for user visual confirmation rather than blocked on
an automated gate that cannot cover animated content).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Bind the spellbook and component-book tabs as their authored stateful text controls, propagate Closed/Open state into the retained DAT child tree, and keep authored label colors live across state changes.
Match retail UIElement_Text zero-margin construction so the Magic favorite captions I through VIII are no longer clipped. Add a real portal.dat spellbook fixture plus controller, state-propagation, and text-layout regression coverage.
Co-Authored-By: Codex <noreply@openai.com>
Complete the retail cast-intent, target, component, enchantment, and busy-state paths; mount the DAT-authored spell bar, spellbook, component book, effects panels, and shared panel lifecycle; and add scoped input plus conformance coverage.
Co-Authored-By: Codex <noreply@openai.com>
Expand the typed client-command boundary across travel, character queries, local UI and layout controls, AFK and consent, emotes, friends, squelch and filters, and fill-components. Preserve retail packet layouts and queue ownership, import the confirmation dialog, and keep authoritative social state in Core.
Co-Authored-By: Codex <codex@openai.com>
Preserve the authored gray track, trained-Recklessness range, live bright charge meter, and independent desired-power thumb while keeping Speed and Power over gray side regions. Correct retail text justification value 2 to left alignment and retain direct RenderSurface decoding in the texture inspection tool used to verify the assets.
Co-Authored-By: Codex <codex@openai.com>
Resolve authored StringInfo labels from local.dat, port gmCombatUI's runtime option captions and checkbox widgets, and bind horizontal scrollbar media by retail structural roles so the green power jewel is a thumb instead of a tiled track. Persist the three combat options and make Auto Target govern target acquisition.
Co-Authored-By: Codex <codex@openai.com>
Lands the codex-worktree D.2b stream plus the extraction the 2026-07-02
UI architecture review mandated before commit:
- ItemInteractionController: single owner of double-click use/equip/
container-open, targeted-use mode (health kits), drag-out drop;
toolbar shortcut drags don't drop the real item. ItemEquipRules for
multi-slot (coat) coverage via equip masks.
- Cursor phase: CursorFeedbackController (semantic priority chain:
drag > resize > window-move > target-mode > text) + RetailCursorCatalog
(enums 0x27/0x28/0x29, hotspot 14,14; ClientUISystem::UpdateCursorState
0x00564630) resolved through the portal EnumIDMap chain by
RetailCursorResolver; RetailCursorManager applies dat cursor art to the
OS cursor. Register row AP-72 covers the OS standard-cursor fallback.
- Character window goes live: CharacterSheetProvider owns sheet assembly,
XP-curve/raise-cost math and the raise flow — extracted out of
GameWindow per Code Structure Rule 1 instead of committing the ~430-line
feature body there. Optimistic XP/credit debits go through eventful
store APIs (new ClientObjectTable.UpdateInt64Property +
LocalPlayerState.DebitIntProperty/DebitInt64Property) instead of raw
property-dictionary writes; register row AP-73 covers the still-missing
raise ledger (#163).
- RetailWindowFrame: the shared nine-slice window mount recipe; the
character window uses it, remaining windows migrate via #164.
- Status-bar buttons toggle inventory/character windows; retail row-major
backpack ordering; WorldSession.SendUseWithTarget + raise/train sends.
GameWindow shrinks 14,214 -> 13,877 lines despite the new features; the
sheet/raise logic is unit-tested in CharacterSheetProviderTests instead
of trapped in the god object. Build green; full suite 3,286 tests pass.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
LayoutImporter.BuildWidget gains a UiMeter-specific branch after the
ConsumesDatChildren gate: Type-3 slice containers (already consumed by
DatWidgetFactory.BuildMeter to extract sprite ids) are skipped, but all
other children (Type-12 UIElement_Text overlays) are built normally,
registered in byId, and attached as UiElement children of the meter.
This fixes the XP meter (0x10000236): its two Type-12 children
0x10000237 "XP for next level:" label
0x10000238 XP-to-next-level value
are now real UiText widgets in the tree, findable via FindElement.
CharacterStatController.Bind now uses FindElement + LinesProvider binding
instead of injecting runtime AddChild nodes. The two previously-injected
UiText overlays are removed; the controller binds Padding=0, ClickThrough,
RightAligned on the dat-origin widgets instead.
New constants XpNextLabelId + XpNextValueId replace the old comment block.
TAB GROUPS (SKIPPED — documented): UiText.ConsumesDatChildren flipping is
NOT safe globally (risks chat/vitals) and the page-visibility pass in
CharacterStatController depends on tab group Children.Count==0. Tab sprite
injection onto layout.Root stays and is explicitly commented as deliberate.
VITALS SAFE: health/stamina/mana meters have only Type-3 children → new
loop finds nothing → zero UiElement children added → byte-identical render.
VitalsController.BindMeter AddChild(number) pattern unchanged.
Tests: 3 new LayoutImporterTests (slice-only, Fix-5 text children,
vitals regression guard) + 3 CharacterStatControllerTests (label bound,
value bound, missing children no-throw). 715 tests pass.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Investigation (2026-06-26): the character footer three sibling groups (0x10000240/241/247)
all carry DefaultState=0x10000011 (StatManagement_Footer_Default) — the dat does NOT
differentiate them by visibility. Parent 0x1000022F has States {Default, Text, Meter}
with PassToChildren=true, but each child group registers all three states with
IncorporationFlags=None and Media=0, so state-propagation produces no visual change
on any group. Retail gmStatManagementUI uses hardcoded element-id dispatch for
0x10000240/241/247 — the groups are never hidden/shown via the dat state mechanism
(decomp gmStatManagementUI::GetFooterTitleLabel @0x004f0170).
Tab-page containers (0x1000022B/22C/10000539) have DefaultState=Undef — no visibility
encoding. Tab visibility is managed purely at runtime by gmTabUI.
IncorporationFlags (DatReaderWriter): {None, PassToChildren, X, Y, Width, Height,
ZLevel} — no Visible flag. The importer cannot set initial Group visibility from
the dat; the controller is the correct and only place. CharacterStatController
HideAllById footer pass and ContainsWidget page-visibility walk are retail-faithful
and must be kept.
No logic changes — comment-only documentation. Build green, 710 App tests green.
Studio screenshots: character footer shows State-A by default with no overlap;
vitals and toolbar unchanged.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
DatWidgetFactory.Create now accepts an optional fontResolve: Func<uint,UiDatFont?>
parameter. When supplied and the element has a non-zero FontDid, the element
receives its own dat font instead of the shared global datFont fallback.
Null = original single-font behavior (the live GameWindow path passes null —
provably unchanged). LayoutImporter.Build/BuildFromInfos/Import all thread
the optional resolver down to the factory.
RenderStack gains a lazy font cache (ConcurrentDictionary, pre-seeded with
VitalsDatFont + LargeDatFont) and a ResolveDatFont(uint) method. StudioWindow
wires stack.ResolveDatFont into LayoutSource so every studio import gets
per-element fonts. GameWindow import calls left passing null (follow-up todo).
CharacterStatController font-hack cleanup (diagnosed via one-shot console dump
then removed):
- Name (0x10000231): dat FontDid = 18px font — remove datFont override (null)
- Heritage/PkStatus: dat FontDid = 14px fonts — remove override
- LevelCaption: dat FontDid = 16px — remove override (same font, no visual change)
- Level (0x1000023B): dat FontDid = 36px (the big retail gold font) — was forced
to rowDatFont/LargeDatFont (18px); now drops to null so the dat 36px font drives
- TotalXpLabel/TotalXp: dat FontDid = 16px — remove override
- FooterTitle (0x1000024E): dat FontDid = 20px — remove datFont override
- KEEP: synthesized elements (XP meter overlays, 9 attribute rows, tab sprites)
still use datFont directly since they have no dat origin
All Label/LabelTwoLine/LabelLeft/LabelProvider helpers updated: null = keep
build-time dat font; non-null = controller explicit override (backward-compat).
8 new tests in DatWidgetFactoryFontResolveTests:
- null resolver → DatFont == global datFont
- FontDid=0 → resolver not called
- resolver returns null → fallback to global datFont
- resolver called with element's FontDid
- controller DatFont override wins after build
- LayoutImporter.Build threads fontResolve to factory
- meter element fires resolver for non-zero FontDid
- BuildFromInfos without fontResolve param = original behavior
Build + all 710 App tests green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
LayoutImporter.ReadState now reads Properties 0x1B (ColorBaseProperty, ARGB bytes) and stores the normalized Vector4 in ElementInfo.FontColor (nullable). ElementReader.Merge propagates it with the same non-null-derived-wins rule as FontDid and HJustify. DatWidgetFactory.BuildText seeds UiText.DefaultColor from FontColor when present.
Diagnosis for LayoutDesc 0x2100002E: ALL 12 header and footer text elements carry NO dat color. Every color is runtime set by CharacterStatController. Comments added at each callsite. No hardcoded colors deleted.
Tests added: 3 ElementReader FontColor Merge + 3 DatWidgetFactory DefaultColor. 702 passed, 0 failed. Screenshots: character window, vitals, toolbar all unchanged.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
LayoutImporter.ReadState now reads Properties[0x14] (HorizontalJustification,
EnumBaseProperty: 0=Left, 1=Center, 3/5=Right) and Properties[0x15]
(VerticalJustification: 2=Top, 4=Bottom; else Center) into two new ElementInfo
fields HJustify/VJustify. Merge propagates them with the same non-default-wins
rule used for FontDid.
DatWidgetFactory.BuildText applies the resolved justify at build time:
HJustify=Center sets Centered=true, HJustify=Right sets RightAligned=true,
VJustify=Top/Bottom sets VerticalJustify. Controllers that FindElement and set
those properties afterward continue to override - backward-compat preserved.
UiText gains VerticalJustify (Top/Center/Bottom, default Center). The Centered
and RightAligned single-line paths call UiText.VOffset() for the Y coordinate,
so VJustify.Top renders text at y=Padding rather than the fixed (H-lh)/2 center.
CharacterStatController: footer title (0x1000024E, H=55 dat box) previously
used Height=18 + Anchors=None to prevent center-vertical overlap with line-1/2.
Diagnostic confirmed dat says HJustify=Center, VJustify=Center. The hack is
replaced with one minimal explicit override: VerticalJustify=Top on the
already-Centered element. Text now renders at top of the 55px box natively.
Centered=false and RightAligned=false hand-sets removed where dat supplies them.
Dat justify values (studio diagnostic, 2026-06-26):
0x1000024E footer title: HJustify=Center, VJustify=Center
0x10000235/0x10000243/0x10000245 value fields: HJustify=Right
header name/heritage/pk/level: HJustify=Center
+13 tests: ElementReader Merge propagation; DatWidgetFactory BuildText
justify application + controller-override backward-compat; UiText VOffset.
696 passed, 0 failed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Root cause: the sub-layout 0x2100002C (design H=337px) mounted into tab slot
0x1000022B (H=575px) via ShouldMountBaseChildren. ElementReader.Merge takes the
derived (sub-layout) H=337 as canonical, so the background element's first
ApplyAnchor call captured _amB=238 and the list box captured _amB=65 — both
stayed at their 337px-parent sizes even though the slot is 575px.
Fix: CascadeHeight in LayoutImporter.Resolve mirrors retail's
UIElement::UpdateForParentSizeChange. When ShouldMountBaseChildren fires and the
slot is taller than the base design, every full-stretch background child has its
height cascaded: Top+Bottom anchors → stretch, Bottom-only → pin-to-bottom,
Top-only/None → unchanged. This propagates the correct bottom margins through the
entire subtree before the first render frame.
Layout result (confirmed via headless screenshot):
- List box grows from 160px → 398px (9 rows × 44px ≈ 396px)
- Footer elements move from abs Y≈307 → abs Y≈545 (matching retail dump)
- Separator moves to abs Y≈535
Row constants updated: RowHeight 44px, IconSize 24px, RowPadX 4px, IconGap 6px.
Footer State-A text corrected per spec: title="Select an Attribute to Improve",
line-1 label="Skill Credits Available:", line-1 value=SkillCredits (96),
line-2 label="Unassigned Experience:", line-2 value=UnassignedXp (formatted N0).
CharacterSheet.UnassignedXp (long, retail InqInt64(2)) added.
SampleData.SampleCharacter().UnassignedXp = 87_757_321_741L.
5 footer tests renamed + assertions updated; SampleCharacter_UnassignedXp_IsSet added.
639 tests green.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Visual verification surfaced two render bugs (controller LOGIC was already correct):
1. BACKDROP WASH-OUT (the big one — #145 continuation). The mounted backpack/
3D-items panels inherited their sub-window root's ZLevel 1000 via the merge's
zero-wins-base rule. The #145 ZOrder fold (ReadOrder − ZLevel·10000) turned 1000
into ZOrder ≈ −10,000,000 — sinking the panels BEHIND the frame's Alphablend
backdrop (ZLevel 100 → ≈ −1,000,000). The backdrop then overpainted the panels'
captions/burden-meter/cells (the paperdoll root is ZLevel 0 so it escaped, which
is why the previous session thought #145 was done). Fix: the sub-window mount now
keeps each slot's OWN frame ZLevel, so panels sit in front of the backdrop.
Root-caused via a one-shot sprite-segment-order dump (backdrop was painting after
the panel content) + a live ZLevel probe.
2. CAPTIONS. The caption elements resolve to UiText; driving a nested child UiText
didn't paint. AttachCaption now drives the host UiText directly.
Locked by InventoryFrameImportProbe (real-dat smoke: asserts each mounted panel's
ZOrder > the backdrop's). Visually confirmed by the user: dark backdrop behind,
Burden 17% + vertical bar + Contents-of-Backpack + full item grid all visible.
Build + App(532)/Core(1526) tests green.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The factory mapped ZOrder from ReadOrder only, so the gmInventoryUI
full-window backdrop (0x100001D0, ReadOrder 4) painted over the nested
panels (ReadOrder 1-3). Wire ElementDesc.ZLevel through ElementInfo /
ToInfo / Merge and fold it into ZOrder = ReadOrder - ZLevel*10000 (higher
ZLevel = further back, ReadOrder the within-layer tiebreaker). Vitals
(all ZLevel 0) are unchanged; chat (ZLevel 900) + toolbar (1,2) shift to
their dat layering — verify visually.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
LayoutImporter.Resolve now captures the resolved base element's children
and, for a pure-container leaf (no own children + no own media) inheriting
from a base WITH content, attaches that subtree. This pulls the nested
gmInventoryUI panels' content (paperdoll/backpack/3D-items) in through the
existing BaseElement+BaseLayoutId path. ShouldMountBaseChildren is a pure,
unit-tested predicate; it's inert for media-bearing inheritors and childless
style prototypes, so vitals/chat/toolbar are unaffected (existing importer
tests still green).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
D1 — Toolbar not movable: toolbarRoot.Anchors = AnchorEdges.None (was Left|Top)
so ApplyAnchor early-returns and doesn't re-pin the window every frame.
Matches the vitalsRoot idiom exactly.
D2 — Cannot grab toolbar by chrome: toolbarRoot.ClickThrough = false
so HitTest succeeds over the UiDatElement chrome and the drag starts.
UiDatElement ctor defaults ClickThrough=true; vitalsRoot already overrides it.
C1 — All four combat-mode indicators visible at once (war/flame stacked on
peace): ports gmToolbarUI::RecvNotice_SetCombatMode
(acclient_2013_pseudo_c.txt:196632-196669). CombatIndicatorIds[] maps
index 0-3 to NonCombat/Melee/Missile/Magic; SetCombatMode shows exactly one
and hides the other three. Default to NonCombat at bind (player always
spawns in peace). Wires CombatState.CombatModeChanged for live updates.
Tests: CombatIndicator_defaultNonCombat_onlyPeaceVisible,
CombatIndicator_setCombatModeMelee_onlyMeleeVisible,
CombatIndicator_liveSignal_updatesWhenCombatStateChanges.
V1 — Blue empty-slot square at top-left (prototype 0x100001B2 materialized):
ImportInfos now skips top-level elements that are (a) referenced as a
BaseElement by another element in the same layout AND (b) have no own state
media. The CollectBaseRefsInDesc walk covers nested children; HasNoOwnMedia
re-uses ToInfo's media extraction. The Resolve path reads BaseElement from the
raw dat via dats.Get<LayoutDesc> — it never depends on the prototype being in
the built widget tree — so the skip is safe. Conformance tests (vitals, chat)
are unaffected (they exercise Build, not ImportInfos).
Test: BuildFromInfos_PrototypeSkipped_DerivedPresent_PrototypeAbsent.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
The generalized channel menu wouldn't open: the factory recursed the Type-6
menu element's dat children, building its invisible Type-12 label child as a
UiText. Hit-testing is children-first and UiText consumes MouseDown (selection),
so the label child swallowed the menu button click and the dropdown never opened.
The transcript similarly gained an invisible Ghosted-button child (a 16x16
selection dead-zone). The old hand-made build never had these — it skipped Type 12
and hand-placed the widgets with no children.
Fix: behavioral widgets (Meter/Menu/Button/Scrollbar/Text/Field) draw their full
appearance and reproduce their dat sub-elements procedurally, so they are LEAF —
the importer must not build their dat children as separate (click-stealing)
widgets. Add UiElement.ConsumesDatChildren (default false; the 6 behavioral
widgets override true) and gate LayoutImporter recursion on it (replacing the
UiMeter-only special case). Only generic containers (UiDatElement, panels) recurse.
Visually confirmed in the live client (channel menu opens; General/Trade selected
and sent). Vitals unchanged (UiMeter was already leaf). Full suite: 404 passed.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Task G1: two gaps blocked chat window static sprite elements from rendering.
Change 1 — DatWidgetFactory: only skip Type-12 elements that have no own
state media (pure style prototypes). A Type-12 element that carries sprites
(e.g. a chat Send button whose derived Type-0 element inherited Type 12 from
its base prototype) now renders as a UiDatElement.
Change 2 — ElementInfo: add DefaultStateName field (string, default "").
Change 3 — LayoutImporter.ToInfo: read ElementDesc.DefaultState.ToString()
into DefaultStateName; normalize Undef/Undefined/0 sentinels to "".
Change 4 — ElementReader.Merge: inherit DefaultStateName (derived wins if
non-empty, else base).
Change 5 — UiDatElement ctor: initialize ActiveState to DefaultStateName
when set; else "Normal" when a Normal-state sprite is present (retail's
implicit default for buttons/tabs); else "" (DirectState). This makes the
Send button, max/min button, and numbered tabs render their default sprite
without requiring explicit state assignment at runtime.
Vitals neutrality: all vitals chrome/grip elements carry DirectState-only
sprites with no "Normal" named state and DefaultStateName="" (Undef in dat),
so their ActiveState stays "" and their existing conformance tests are
unaffected. Vitals text labels (Type 0→12 via Merge, no StateMedia) are
still skipped by the refined Type-12 guard (StateMedia.Count==0).
Tests: 4 new tests (2 in DatWidgetFactoryTests, 3 in UiDatElementTests).
All 386 pass; 387 total (1 pre-existing skip).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Fix 1: replace 3 copy-paste meter blocks in VitalsTree_MetersHaveExpectedSliceIds
with a single table-driven loop — a 4th meter is now a one-liner and failures
name the failing meter id directly.
Fix 2: FixtureLoader now reads the fixture as bytes and strips the UTF-8 BOM
(EF BB BF) before passing the span to JsonSerializer, so a BOM-bearing fixture
file never causes a spurious JsonReaderException.
Fix 3: add [Trait("Category", "Conformance")] at the class level so conformance
tests are selectable by category filter.
Fix 4: add missing <param name="layoutId"> doc tag to LayoutImporter.ImportInfos.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Job 1: extract LayoutImporter.ImportInfos() (public dat-shell half that returns the
resolved ElementInfo tree without building widgets) so fixture generation and
conformance tests can call it directly. Import() now delegates to ImportInfos() +
Build() — existing 32 Layout tests stay green.
Job 2: generate tests/AcDream.App.Tests/UI/Layout/fixtures/vitals_2100006C.json
from the real portal.dat via a throwaway [Fact] generator (deleted, not committed).
System.Text.Json with IncludeFields=true — ValueTuple serializes as Item1/Item2.
Pre-write validation confirmed health meter BackLeft=0x0600747E FrontRight=0x06007483
rect (5,5,150,16). Round-trip deserialization re-validated before writing.
Job 3: FixtureLoader.LoadVitals() deserializes the fixture from the test output
directory (CopyToOutputDirectory item in csproj) and returns ImportedLayout via
LayoutImporter.Build(root, _ => (0,0,0), null) — no dats, no GL.
Job 4: LayoutConformanceTests — 3 golden tests (35 asserts total):
- VitalsTree_HasThreeMetersAtExpectedRects: 3 meters at x=5, w=150, h=16, y=5/21/37
- VitalsTree_MetersHaveExpectedSliceIds: all 18 back+front slice ids health/stamina/mana
- VitalsTree_ChromeCornerHasExpectedSprite: TL corner 0x10000633 → sprite 0x060074C3
Full App suite: 326 pass / 1 skip (pre-existing) / 0 fail. Build: 0 errors, 0 warnings.
Throwaway generator not committed (confirmed via git status).
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Implements Task 5 of the LayoutDesc Importer (Plan 1 — vitals conformance).
Pure layer (BuildFromInfos / Build):
- ImportedLayout result type: UiElement root + O(1) FindElement(uint id) lookup
- BuildWidget dispatches via DatWidgetFactory.Create; skips Type-12 prototypes (null)
- Meters consume their children (DatWidgetFactory already extracted slice ids —
adding the dat children as UiElement nodes would duplicate geometry)
- All other element types recurse children generically via AddChild
Dat shell (Import):
- Loads LayoutDesc from dats; null-safe if layout is absent
- Resolves each top-level ElementDesc to ElementInfo via Resolve():
BaseElement/BaseLayoutId chain with (layoutId,elementId) cycle guard
- ToInfo(): reads ElementDesc scalar fields (uint → float cast) + DirectState +
named States (UIStateId.ToString() as key)
- ReadState(): extracts first MediaDescImage (File + DrawMode) per state +
font DID from Properties[0x1A] → ArrayBaseProperty → DataIdBaseProperty.Value
- Each sibling element gets a fresh base-chain set (siblings don't share guards)
DRW API: all members confirmed from VitalsLayoutDump.cs usings — no
adjustments needed: LayoutDesc in DBObjs; ElementDesc/StateDesc/MediaDescImage/
ArrayBaseProperty/DataIdBaseProperty in Types; DrawModeType/UIStateId in Enums.
Tests (3/3 green):
- BuildFromInfos_HealthMeter_IsUiMeterAtRect — Type-7 child → UiMeter, Left=5, Width=150
- BuildFromInfos_Type12Child_IsSkipped_Type3Present — prototype absent, container present
- BuildFromInfos_MeterWithChildren_MeterPresent_ChildrenNotInTree — meter findable,
both dat-children absent, UiMeter.Children empty
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
