Dual-lens review of CC4 (0e71d3b8) returned architectural FAIL (F1, F6)
and retail-fidelity PASS-with-reservations (F2, F3, F4), plus LOW
findings F5, F7-F12. F13 (TS-82's merge collision with campaign-cc6a) is
merge mechanics for the orchestrator, not addressed here.
F1 (HIGH, blocking): CharacterCreationUiController never released
UiRoot.FixedCanvasSize, on a FALSE premise that CharacterManagementUi-
Controller does a per-tick set (it does not — it sets once on activation
and nulls on Deactivate/Dispose). Root cause: RuntimeCharacterCreation-
State had no CompleteEnter() analogue to RuntimeCharacterSelectionState's,
so the creation view reported IsActive=true for an entire in-world
session. Added CompleteEnter(), wired at both LiveSessionController
in-world edges (StartCore, EnterHighlightedCore); made Open/Close/
Deactivate/Dispose set/null the canvas symmetrically; corrected the false
comment and ledger claim; added FixedCanvasSize test coverage.
F2 (MEDIUM-HIGH, blocking): the attribute-slider scalar mapping was not
retail's. Fixed display to value/100f (UpdateAttributeValues @
0x0048251d) and the drag inverse to truncate+clamp-low-only, no rescale
(ListenToElementMessage @ 0x004829c0, independently re-verified against
the decomp). Added tests at scalar 0.5/0.0 plus a display-direction test.
F3 (MEDIUM, blocking): ported the unported heritage-button tab-restore
arm (ListenToElementMessage @ 0x004e9450) — SHOW/HIDE id sets independently
re-derived from the decomp, including the genuine Lugian (0x100005f1)
no-restore quirk, reproduced faithfully. Wired via a new HeritagePage
click callback; added restore + quirk tests.
F4 (MEDIUM): ported SetTown's (@ 0x0047c360) separate per-town page-root
state literal (Holtburg->0x10000034 etc.), independently re-derived from
the decomp's tail-merged branches; wired via the existing
IUiDatStateful.TrySetRetailState seam; added a test.
F5 (MEDIUM): softened AD-103's unmeasured pixel-equivalence claim.
F6 (MEDIUM, blocking): DECISION — install ChargenOptions in the headless
content path (chosen over marking headless creation out-of-scope).
HeadlessSessionHost now calls InstallOptions off the shared content
lease's Dats, beside the existing InstallSpellMetadata call.
F7: AP-213 already named the label format and click/double-click
substitution explicitly on inspection — no edit needed.
F8: AP-212 now names all six DoRandom primitives with a known landing site.
F9: AD-101 retirement corrected to precede CC5's Finish un-ghosting.
F10: merged ItemAppraisalTextFormatter's duplicate <summary> block.
F11: fixed TS-82's wrong AP-211 cross-reference.
F12: cached the chargen DatStringResolver once per composition instead of
per ResolveText call.
Runtime 1713/0, App 5125/13 skips (+8 new tests), Headless 165/0, full
solution Release build green. Live-DAT probes 7/7 under
ACDREAM_PROBE_LIVE_MOUNT=1.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Mounts gmCharGenMainUI (enum 0x10000039, root 0x100003CC) via
CharacterCreationUiController/CharacterCreationUiMountCoordinator,
cloning CharacterManagementUiController's recipe. Master shell ports
SetProgressState @0x004e7a10 (Olthoi tab-hide + redirect) and
ListenToElementMessage @0x004e9450 (Back/Next/Finish/Help/Exit/Random
nav) verbatim, with free tab navigation over all six pages. Heritage,
Profession, Skills, and Town pages bind to CC3's
RuntimeCharacterCreationState commands; Appearance and Summary mount as
content-inert placeholders for CC6b/CC5.
Live-DAT probing (CharacterCreationLiveDatTests) found two widget-
mapping surprises the decomp's DynamicCast hints don't predict: the
Profession slider's value field imports as an editable UiField (wired
for direct numeric entry), and the avail/health/stamina/mana/credits
displays author as UIElement_Button hosts whose Type-12 value child is
swallowed by UiButton.ConsumesDatChildren — substituted with the
button's own Label. No new DatWidgetFactory widget types were needed.
Threads the installed DAT's real ChargenOptions into Runtime via the
new RuntimeCharacterCreationState.InstallOptions, called from
ContentEffectsAudioCompositionPhase.Compose (mirrors
InstallSpellMetadata's pattern); headless keeps ChargenOptions.Empty
unchanged. Wires CC3's F14 status-hook gap (ApplyCharacterCreated/
ApplyCreationFailed) to SessionStatusWriter for both graphical and
headless hosts, and adds the CharacterCreation view/command seam
through CurrentGameRuntimeAdapter and DeferredGameRuntimeStateCommands
alongside CharacterSelection's existing shape.
Register: AD-101/102/103, AP-212/213, TS-82 filed for the auto-gender-
select interim default, the omitted ToD-account gate, the button-Label
widget substitution, the Random-button approximation, the flat-listbox
Skills simplification, and the Appearance/Summary placeholders.
Runtime 1713/0 (was 1707), App 5117/13 skips (was 5101/6), Headless
165/0 unaffected, full solution Release build green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The CC3 narrow re-review returned CLOSED (both lenses PASS, merge
recommended) with residual R1: the post-create wire-slot assignment read
the cached wire CharacterList count, which ACE never refreshes after a
create — correct for the first create, off by one for a second create in
the same session (reachable via create Ok -> server-rejected guid enter
-> ReturnToSelection -> create again), the same wire-contract failure
class F2 fixed. Root fix now rather than carried: a
creates-since-CharacterList counter (the equivalent of retail's own
CharacterSet growing via AddIdentity per create), reset on every fresh
wire CharacterList apply and at generation reset, applied only to the
cached-wire branch since the display-roster fallback already contains
prior appends. Regression test drives the full
create->Ok->rejected-enter->create-again flow and pins wire slots
0/1/2/3.
Docs: CC3 ledger row flipped to REVIEW-CLOSED with real shas (re-review
R2); CC7 risk item 8 downgraded to LATENT with measured installed-DAT
data (user-prompted): every heritage's single cost override is Arcane
Lore at NormalCost=0/PrimaryCost=2 vs global 4/6, so ACE's over-deduction
(= NormalCost = 0) cannot fire with end-of-retail data — the earlier
"may be rejected" claim was inferred from code without measuring.
Runtime 1707/0 Release.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Opus dual-lens review of CC3's RuntimeCharacterCreationState passed on
retail fidelity but failed the controller integration: the post-create
log-straight-in indexed the CACHED wire CharacterList, which ACE never
resends after a create (it only appends server-side and replies Ok) —
with zero pre-existing characters this throws, with N it can silently
enter the WRONG character. The same stale-index problem corrupted every
pre-existing character's delete slot on roster re-sort. Fixes all four
blocking findings plus a credit-gate correctness bug (retail warns and
lets the user confirm through unspent credits; it does not force a full
spend) and eight lower-severity findings from the same review round.
F1 (blocking): WorldSession gained a guid-based EnterWorld(uint,string,
TimeSpan?) overload sharing EnterWorldCore with the index-based one;
ILiveSessionOperations gained a default EnterWorldByGuid method.
LiveSessionController factored EnterSelectedCore/the new
EnterCreatedCharacterCore through a shared EnterHighlightedCore so the
post-create enter sends by the exact guid the 0xF643 Ok reply carried,
never by a roster index.
F2 (blocking): RuntimeCharacterSelectionState gained a real
AppendCreatedCharacter primitive that preserves every existing entry's
ActiveIndex (a wire contract — SendDeleteCharacter sends it as the
CharacterSet slot) and assigns the new entry's from the pre-create wire
roster count, instead of round-tripping the post-create roster through
ApplyRoster's name-sort-and-renumber.
F3 (blocking): retail's DoFinish(this, arg2) gate is
"arg2 != 0 && remainingAtrbCredits > 0" — the ordinary click warns and
refuses, but the warning dialog's own confirm re-invokes DoFinish(this,
0), which sends anyway with credits unspent (ACE accepts this).
TryBeginFinish/Finish gained a confirmedUnspentCredits parameter; the
plan doc's "retail FORCES full spend" line is corrected in the same
commit.
F4 (blocking): a stale out-of-range template index surviving a heritage
switch to a heritage with fewer templates now clears to TemplateUnset,
matching ConstrainAllByHeritage's clamp.
F5/F9/F10: three register-row/doc citation corrections (AP-207's real
FitTemplateToCharacter call sites — a fourth one the original filing
also missed; the Slot field's real retail assignment source; AP-209's
classID branch table for Olthoi/OlthoiAcid). F6: ApplyCreationResponse
no longer publishes from inside the owner lock. F7: two new tests pin
BalanceAttributes' persistent donor cursor (successive-overspend
advance, Self-to-Strength wrap). F8: ResetSkillLevels' doc corrected to
retail's real both-costs->=0 gate. F11: the integration test fixture
captures guid-based enter calls and uses two pre-existing characters
whose wire order differs from alphabetical order, so the roster
assertion actually exercises F2 instead of coinciding with it by
accident. F12: filed register row AP-211 for the client-side RosterFull
slot-cap refusal (no retail DoFinish-layer counterpart). F13: narrowed
Finish's bare catch to InvalidOperationException/SocketException and
bound _scope to a local. F15: RandomizeStartAreaLocked leaves the start
area unchanged on an empty list instead of forcing -1, matching retail.
Runtime 1706/0 (was 1701), Core.Net unchanged at 994/0, full solution
Release build green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Ports retail's CharGenState as the one Runtime-owned character-creation
state machine, mirroring RuntimeCharacterSelectionState's exact pattern
(snapshot/delta/event-stream, borrow-only view, generation-gated
commands, one mutable owner, no App types). Every command ports a named
retail function: SetHeritageGroup, SetGender, SetTemplate/ApplyTemplate
(Custom = template 0, Olthoi force-lock), the six attribute setters plus
GetAbsRemainingCredits/BalanceAttributes (retail's literal round-robin
order and fairness cursor), SetSkillLevel plus ResetSkillLevels' free-skill
baseline (reusing CC1's ChargenSkillCreditMath two-tier cost lookup
verbatim), RandomizeStartArea, and DoFinish's complete gate sequence
(empty name / unspent attribute credits / already-Pending / client-side
roster-vs-slotCount cap).
LiveSessionController gained a sibling IRuntimeCharacterCreationCommands
implementation, a CreateCharacter wire hook, and a response handler that
reuses existing machinery rather than inventing new paths: the Ok
identity is appended to the roster via RuntimeCharacterSelectionState's
own ApplyRoster, and the "log straight in" behavior reuses the private
EnterSelectedCore. ILiveSessionLifecycleHost gained two default-no-op
hooks (ApplyCharacterCreated/ApplyCreationFailed) so AcDream.App needs
zero changes to keep compiling; wiring them to the status stream is a
CC4 follow-up.
Filed four divergence-register rows for the corners deliberately not
ported: the FPU-unrecoverable FitTemplateToCharacter auto-detect (AP-207,
ACE only reads the field for title text), the per-style color-count
approximation (AP-208, CC1's model has no per-style palette data), the
classID DAT-DID placeholder (AP-209, ACE ignores the field), and
ApplyTemplate's atomic-vs-sequential attribute apply (AP-210).
34 new tests: full state-machine coverage (every Finish gate, every
rejection-code mapping, duplicate-NameInUse tolerance, Olthoi lock,
attribute balance/lock interaction, uncostable-skill rejection) plus a
LiveSessionController integration suite proving the wire send is exactly
55 skill slots (decoded from a real WorldSession + GameMessageCapture)
and the full Ok/rejection round trip through WorldSession.ProcessDatagram.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Both narrow re-reviews returned CLOSED. This closeout takes the two cheap
re-review residuals before CC3 takes references to the shared model:
R2: every array handed into the typed chargen model is now wrapped in
Array.AsReadOnly at the projection seam — a T[] behind IReadOnlyList<T>
was still downcast-mutable, and ChargenOptions is a process-shared
singleton graph.
R3: the no-Chorizite-leak guard now also walks public fields; every
current type uses properties, but a public field would have slipped
through the property-only walk.
Ledger: CC1 fix-round sha corrected to cb4703e8 (the cell previously
cited the pre-amend 459a87f2), CC1/CC2 rows flipped to REVIEW-CLOSED
with the re-review outcomes, R1 (retail refunds +1 credit on a
both-tier cost miss; port charges 0 — unreachable via retail's own
listbox, noted for CC3) and the Olthoi-locked-to-template-0 decomp fact
recorded for CC3/CC4.
Core.Tests 4736/1 skip, Content.Tests 145/0, Release.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Implements all six Opus review findings against 04450041 (Campaign CC
CC1 chargen data layer):
- F1 (HIGH, blocking): ChargenTemplate's doc claimed "Custom" has no
ChargenTemplate entry and cited two nonexistent addresses. Verified
against the named retail decomp: gmCGProfessionPage::UpdateProfession
@ 0x004821b0 resolves BOTH the highlighted button and the description
string from CharGenState.template_ 0..6, and case 0 is button
0x100003d9 / ID_CharGen_CustomText. Custom IS template index 0 (the
"Adventurer" row CC1 already found sitting at the attribute floor).
CharGenState::SetTemplate @ 0x005C5A60 confirms every button (including
Custom) calls CharGenState::ApplyTemplate @ 0x005C5080 when committing,
so selecting Custom resets the sliders/skills to that row rather than
leaving them untouched.
- F2 (MEDIUM): retail's skill-cost lookup is two-tiered
(ACCharGenData::GetSkillTrainedCost/GetSkillSpecializedCost @
0x005C26D0/0x005C27D0 fall through to the global SkillTable,
portal.dat 0x0E000004, on a heritage-list miss — confirmed against
ACE's identical PlayerFactory.cs precedence). ChargenTableReader now
also projects the global SkillTable into
ChargenOptions.GlobalSkillCostsBySkillId, and
ChargenSkillCreditMath.ComputeSpent/RemainingCredits check the
heritage list first and the global list on a miss. Added an
installed-DAT completeness assertion recording reality: the global
table prices 38/54 advancement skill ids, every one of the 13
installed heritages ships exactly one heritage-specific override
(always also priced globally), and 16 ids are genuinely uncostable in
both tiers. Also filed a CC7 risk-item note: ACE's own heritage-
override branch over-deducts on Specialize (PlayerFactory.cs:184-211)
— a retail-legal build may be rejected by local ACE at the CC7
connected gate; that is an ACE bug, not an acdream defect.
- F3 (MEDIUM): every collection ChargenTableReader hands into the
record model is now frozen at projection (ToFrozenDictionary/ToArray,
matching MagicCatalog's house pattern), including both
ChargenOptions.Empty dictionaries.
- F4 (LOW): added a reflection guard test
(ChargenNoChoriziteLeakTests) that walks every public
AcDream.Core.CharGen member (property/indexer/constructor/method
types, recursively through generic arguments) and fails if any
resolves to the DatReaderWriter or a Chorizite* assembly.
- F5 (LOW): ChargenGenderOptions.HasAnyAppearanceOptions's doc now
states precisely what the installed-DAT gate proves (an OR across
eight lists, for at least one gender per heritage) rather than the
stronger claim it previously made, and explicitly calls out the three
omitted color lists. Added a second installed-DAT gate that records
per-list reality across every gender of every heritage — found
complete, no empty lists anywhere in the installed DAT today.
- F6 (LOW): ChargenOptions.TryGetHeritage/TryGetStarterArea now use
[MaybeNullWhen(false)] instead of null! suppression, matching the
house pattern already used elsewhere in the test suite. Fixed every
call site this surfaced (more than the five originally estimated,
since Content.Tests has TreatWarningsAsErrors).
Core.Tests: 4737 passed / 1 skip (pre-existing, unrelated).
Content.Tests: 145 passed / 0 skip.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
F1 (MEDIUM): the correlation-latch docs claimed replies are never
misattributed; in truth an overlapping send OVERWRITES the latch and the
first reply routes to the newest request's event. Narrowed all three doc
sites to the exact contract (single outstanding request; overlap refusal
is CC3's Runtime verification gate, retail's DoFinish UNDEF-state rule)
and pinned the overwrite behavior with
OverlappingSend_OverwritesTheLatch_ReplyRoutesToNewestRequest.
F2 (LOW): filed register AD-100 for the drop-unless-armed deviation —
retail's Handle_CharGenVerificationResponse@0x0055E8B0 has no armed gate
and processes whatever arrives against its persistent verification state.
F3 (LOW): doc note in CharacterCreate.cs — ACE double-sends NameInUse
(IsCharacterNameAvailable runs twice; the first callback's return exits
only the lambda), so the second reply hitting the drop path during a
connected gate is EXPECTED, not a defect.
F4 (LOW): creationFailed's enum-member key renamed name -> reason and the
ATTEMPTED character name added as name, before any consumer shipped —
one status vocabulary must not give the same key two meanings
(characterCreated.name is a character name). Contract, writer, tailer,
and shape-pinning tests updated in lockstep.
F5 (LOW): the thread-id probe-note pointer now cites
ProbeNetLogOutbound's doc comment, where the note actually lives.
Fidelity fold (reviewer's positive note): the latch is retail's OWN
discriminator one layer down — 0x0055E8B0 case 1 branches on
GetVerificationState()==PENDING (create) vs not (restore) — now cited in
both the latch doc and CharGenVerificationResponse.cs.
Core.Net 994, Runtime 1667, Launcher.Core 324, all green Release.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Wire (Core.Net):
- CharacterCreate.cs: outbound 0xF656 builder, byte-exact port of
Proto_UI::SendCharGenResult@0x00546a70 -> ACCharGenResult::Pack@0x005c7570
-> CG_Pack@0x005c7200. Account String16L first (packed outside CG_Pack),
then the constant-1 u32, heritage/gender, 14 appearance strip/style/color
u32s, 6 f64 shades (skin/hair/headgear/shirt/trousers/footwear, retail
order), template, 6 attributes, slot, classId, numSkills + exactly 55
u32 skill-advancement classes (ReadOnlySpan validated ==55, throws
ArgumentException otherwise — ACE terminates the session on any other
count via PlayerFactory.CreateResult.ClientServerSkillsMismatch), name
String16L, startArea, isAdmin, isEnvoy, and a trailing checksum whose
exact 19-term accumulation set (heritage+gender+3 strips+hairColor+
eyeColor+hairStyle+headgearStyle+shirtStyle+trousersStyle+footwearStyle+
template+6 attributes) is read byte-for-byte off CG_Pack's decompiled
accumulator (0x005c7213-0x005c74c3) — headgearColor/shirtColor/
trousersColor/footwearColor/shades/slot/classId are deliberately absent
from the sum despite sitting adjacent on the wire. Cross-checked against
ACE's CharacterCreateInfo.Unpack/Appearance.Unpack and holtburger's
CharacterCreateRequestData (types.rs:236-369), which agree on every
field and order. Retail routes via SendToLogon — the same queue
CharacterDelete already uses.
- CharGenVerificationResponse.cs (new): promotes the shared 0xF643 parse
out of CharacterRestore — full Code enum (Undef..AdminPrivilegeDenied=7,
ACE's CharacterGenerationVerificationResponse) plus the conditional
Ok-only identity payload (guid/String16L name/u32 secondsGreyedOut).
CharacterRestore.Parse now delegates to it; CharacterRestore's public
Parsed shape, Parse signature, and every existing test expectation are
UNCHANGED.
- PacketWriter.WriteDouble: f64 little-endian helper for the shade fields.
WorldSession dispatch (Core.Net):
- Added an awaiting-request latch (None/Restore/Create), armed by
SendRestoreCharacter/the new SendCharacterCreation immediately before
each send (SendCharacterCreation builds the body first so a skill-count
throw never arms the latch for a request that was never sent), cleared
the instant a matching 0xF643 is dispatched (success OR parse failure —
a malformed reply must never wedge the latch open) and on Dispose.
0xF643 now routes to CharacterRestoreReceived or the new
CharacterCreateResponseReceived (Action<CharGenVerificationResponse.Parsed>)
by that latch; an unexpected 0xF643 with nothing outstanding logs once
and is dropped, never misattributed. Fixed
WorldSessionCharacterSelectionTests' restore-dispatch test, which
previously fed a bare CharacterRestore response with no preceding
SendRestoreCharacter — that shape is now the "no outstanding request"
drop path by design.
Status events (Runtime + Launcher.Core, contract first):
- Amended docs/plans/2026-08-14-launcher-campaign.md §LA1's pinned status
vocabulary to add characterCreated{guid,name} (Ok reply identity, named
to mirror CharGenVerificationResponse's own fields and to read distinct
from enteredWorld — retail logs a freshly created character straight in
without a fresh characterList) and creationFailed{code,name} (raw Code
value + its enum member name).
- SessionStatusWriter.CharacterCreated/CreationFailed implement that
contract.
- Launcher.Core: CharacterCreatedStatusEvent/CreationFailedStatusEvent +
StatusEventParser cases, in lockstep.
Tests: CharacterCreateTests (byte-exact layout incl. checksum term-set,
55-slot fixture, wrong-count throws), CharGenVerificationResponseTests
(every Code value), WorldSessionCharacterCreationTests (create-then-
response routes correctly, restore unaffected, no-outstanding drop,
second-response-after-consumed drop, Dispose clears the latch, a builder
throw never arms it), SessionStatusWriterTests + Launcher.Core
StatusEventParserTests/StatusFileTailerTests (pinned shape + tailer
round-trip) for the two new events.
Verified: dotnet build AcDream.slnx -c Release — 0 errors. Full solution
test run green (Core.Net.Tests 993/993, Runtime.Tests 1667/1667,
Launcher.Core.Tests 323/323, plus every other project in the solution).
WSL Ubuntu: Core.Net.Tests 993/993, Runtime.Tests 1667/1667.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Recon-grounded (three parallel sweeps 2026-08-15): full retail flow
(gmCharGenMainUI mode 0x1000000b, six ECGProgress pages with every widget
id), byte-exact 0xF656 layout incl. the trailing checksum ACE never reads,
0xF643 create semantics (local roster append + retail logs straight in; no
fresh CharacterList), the chargen DAT table shape, the preview rig
(gmCG3DView over CreatureMode - Appearance/Summary pages only), and every
acdream seam to reuse (generic layout resolver, fixed canvas, offscreen
viewport pipeline, RuntimeCharacterSelectionState as the J-owner template,
the 0xF643 correlation landmine). Slice order CC1 data / CC2 wire (parallel)
then CC3 Runtime owner, CC4/CC5 form pages, CC6a/b appearance+preview
staged, CC7 end-to-end + user gate.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
F1 (MUST-FIX): RetailWaitDialogView was the ONE dialog view the 0a7dc7d6
EffectiveCanvasSize sweep missed - the Entering World wait dialog (fires
on ENTER, the char screen primary action) still centered against the raw
window and landed off the visible canvas. Same three-line fix as its three
siblings; the enter-wait test now grows the window over the fixed canvas
and asserts canvas-space centering.
F3/F4: two stale assertions about the DELETED first AD-98 substitution
(the register section-2 header line and the live-DAT oracle test doc) now
describe the completed FixedCanvasSize mechanism - the C4-closeout failure
mode, caught before it cost anything.
F5: RetailDialogData.Confirmation sets ElementAttribute40 itself (retail
MakeConfirmExitDialog writes 0x8E=1, 0xAC=1, 0xC5); the manual set in
GameplayConfirmationController is gone.
F6: MapWindowToCanvas truncates instead of rounding - rounding mapped the
window far edge one past the canvas last valid coordinate, a 1px dead
hit-test band; test updated to truncation semantics + far-edge case.
F7: AD-98 records that the no-letterbox aspect claim has no decomp
citation and is confirmed by the user live gate pass 2026-08-15.
F8: the durable world-name read in StartCore is IsCurrent-gated like every
neighbouring step.
F2 filed as #401 (invert RetailUi to opt-out - product-default decision,
not a gate fix).
App 5100+6 skips, Runtime 1666, green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The ef96c554 batch ghosted Credits with a comment claiming it was filed
as #397 - that number is the Windows graceful-stop issue and no Credits
entry existed. #400 now records the gap properly.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Finding 1 (Exit button dead): retail's gmCharacterManagementUI Exit
button (element 0x100003A4, offset 7 from the listbox base in
ListenToElementMessage@0x004ed5a0) opens MakeConfirmExitDialog
(0x004ed250), whose exact ID_CharacterManagement_ConfirmExit text
(table 0x23000002) and m_confirmExitDialogContext re-entry guard are
now ported. On confirm (matching RecvNotice_CloseDialog@0x004ed760
case 1's ConfirmationResult check) the client exits through the
EXISTING graceful window-close seam (CharacterSelectionRuntimeBindings
.RequestExit -> d.Window.Close, the same delegate
GameplayInputCommandController's Escape fallback already uses) so
disconnected/exited status events still fire via GameWindow.OnClosing
-> CompleteShutdown. Retail's real post-confirm destination is
QueueUIMode(0x10000009) -> gmEpilogueUI, an epilogue screen this round
does not port — recorded as AD-99. Credits (element 0x100003A3,
QueueUIMode(0x10000005) -> gmCreditsUI) stays visibly ghosted like
Create, same treatment, out of scope this round.
Finding 2 (row names center-aligned, retail is left): the character
row template (LayoutDesc 0x21000004, element 0x100003A5, live-DAT
confirmed HJustify=Left with three stateful Type-3 highlight-art
children and no Type-12 caption child) authors its OWN justify
directly, with no separate text child to lift a label from.
DatWidgetFactory.BuildButton's Left-justify branch required
!ReferenceEquals(labelInfo, info) — true only when a label was LIFTED
from a distinct child — so a button's own direct HJustify=Left was
silently dropped to UiButton's Center default. Widened the branch to
also honor the direct case, preserving the existing lifted-child
LabelOffsetX behavior and leaving genuinely-centered buttons
(CREATE/ENTER/DELETE/RESTORE) untouched.
Finding 3 (World box empty): parsed ACE's GameMessageServerName
(opcode 0xF7E1, ACE.Server/Network/GameMessages/Messages/
GameMessageServerName.cs; retail CM_Login::DispatchUI_WorldInfo
@0x006ad860 -> ClientUISystem::Handle_Login__WorldInfo@0x005641a0 ->
ECM_Login::SendNotice_WorldName@0x00692b10, notice 0x186a2, consumed
by gmCharacterManagementUI::UpdateWorldName@0x004ec120 /
RecvNotice_WorldName@0x004ec360 onto element 0x1000039B) as
src/AcDream.Core.Net/Messages/ServerName.cs, cross-checked against
holtburger's ServerNameData. WorldSession.ServerNameReceived fires
alongside CharacterListReceived (ACE sends both in one
SendConnectResponse batch); RuntimeCharacterSelectionState.
ApplyWorldName is the new J-owner field (ungated by lifecycle, since
either message can arrive first); CharacterManagementUiController
binds it onto the WorldTextElementId UiText. Per the LA1 status
vocabulary, the characterList STATUS event's worldName field is
intentionally NOT added this round (kept bounded to the client-side
fix) — a follow-up if the launcher UI wants it.
Also corrects AD-44, discovered stale while filing AD-99: its opening
claim ("acdream has no retained character-management screen") was
false as of this session — LA7/LA8 shipped the screen in earlier
commits without updating this row.
Tests: exit-confirm open/cancel/confirm/re-entry-guard flow;
DatWidgetFactory own-HJustify-Left/Center regression tests plus the
live-DAT pinned row-justify assertion; ServerName parse round-trip
(byte-exact vs ACE's AceWireWriter fixture, truncation/wrong-opcode
cases); WorldSession dispatch test (roster+world in one wire batch);
RuntimeCharacterSelectionState.ApplyWorldName tests (order-independent
of ApplyRoster, unchanged-value no-op, Reset clears); controller test
binding the World text element to the live snapshot. Extended the
shared RetailDialogFactoryTests.BuildDialogLayout test fixture with a
Confirmation-type branch (Accept/Reject buttons) since this is its
first RetailDialogType.Confirmation consumer.
Suites: full solution Release build green; AcDream.App.Tests 5100/6
skips, AcDream.Core.Net.Tests 965/0, AcDream.Runtime.Tests 1665/0, all
Release, 0 failures; live-DAT probes (ACDREAM_PROBE_LIVE_MOUNT=1)
green.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
AD-98's fixed-canvas stretch (73041d70) scales every retained-UI quad at
TextRenderer.AppendQuad, but the live gate reported it JAGGED — text
especially. Cause: dat-font glyph atlases and IconComposer's composited
icons upload nearest (TextureCache.UploadUiTexture's UiNearestRepeat
sampler) — correct at the native 1:1 scale (pixel-exact retail art), but
aliased once magnified 2.4x1.8. Chrome/background art was already fine:
it uploads through GpuSamplerDescription.WorldRepeat (linear) by default.
Retail's own fixed-canvas presentation is a single bilinear-filtered
frame blit, never a per-texture stretch — this closes that gap one step
earlier, at the source texture, without adding RHI surface area.
- TextureCache.GetOrCreateLinearUiTwin: lazily registers a SECOND table
slot for a nearest handle's IGpuTexture, sampled WorldRepeat (linear)
instead of nearest — no re-decode, no re-upload, no extra memory-ledger
bytes. Returns the handle unchanged for anything never registered
nearest (chrome, UiTextureTableHandle.None), so it's a cheap
unconditional probe. Twin slots are released in Dispose without
double-disposing the shared texture.
- TextRenderer.LinearTwinResolver + the DrawSprite chokepoint: swaps a
sprite's texture handle through the resolver only while
CanvasScale != One. At CanvasScale == One the resolver is never even
called — zero overhead on the ordinary in-world/UI path.
- InteractionRetainedUiComposition wires the resolver to TextureCache
right after every UiHost acquisition (the lease can hand back a host
from a prior session against a fresh TextureCache).
- AD-98's register row gets one added sentence recording the fix.
Tests: TextRendererLinearTwinTests pins the renderer-side handle-swap
seam GPU-free (segment handle selection); TextureCacheLinearTwinTests
pins twin creation/reuse/dispose against RecordingGpuDevice. App suite
5097/3 skips (Release, ACDREAM_PROBE_LIVE_MOUNT=1 live-DAT probes
included). Full solution builds clean.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Third iteration on the screen, completing AD-98. The previous substitution
stretched only the root BACKGROUND while the child widgets stayed at their
authored 800x600 pixel positions - and the background painting carries
visual anchors (the World/Characters captions are art), so the user gate
showed captions overlapping the listbox and every widget misaligned
against the stretched art.
Retail model (established at 71bf24fb): fixed-canvas pre-world screens
render at authored 800x600 and the whole composed frame stretches once at
presentation; the blitter has no stretch mode. Our equivalent now does the
same one stage earlier:
- UiRoot.FixedCanvasSize: while the char-select screen is active, the
retained tree lays out in its authored canvas and Draw scopes a uniform
scale onto TextRenderer.CanvasScale; the mouse entry points apply the
exact inverse so MouseX/MouseY and every hit test live in canvas space.
- TextRenderer.AppendQuad is the single emission chokepoint - sprites,
rects, AND glyphs scale together, including retail-authentic non-uniform
aspect distortion and stretched text. World-space HUD stays native (the
scale resets outside UiRoot.Draw).
- CharacterManagementUiController stops resizing Root to the viewport;
activate/deactivate/dispose set and clear the host canvas.
- UiDatElement returns to retail-pure copy-or-tile; the interim
StretchOwnBackgroundToFill flag is deleted.
- AD-98 updated to describe the completed substitution.
Tests: canvas-scale quad math, inverse input mapping (window click lands
on the canvas-space widget), degenerate-size guards, controller keeps
authored extent + sets/clears the canvas. App suite 5085/6 skips; live-DAT
char-select probes 3/3.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The LA8 char-select root (0x1000039A) authors LeftEdge=TopEdge=RightEdge=
BottomEdge=0 ("no anchor") in the installed DAT — confirmed via the new
CharacterManagementLiveDatTests.RootAuthorsNoEdgeAnchors_RetailNeverResizesItSelf
gate — so retail's own UIElement::UpdateForParentSizeChange (0x00462640) never
resizes this element; it stays a fixed 800x600 rect in retail's own tree.
Retail's generic UI sprite blit, Graphic::Draw (0x00693b20) dispatching to
Graphic::PutImage (0x00693a30) for an exact/undersized destination or a
modulo-wrapped tile loop otherwise, has no third "stretch" mode — confirmed
against BlitMode (acclient.h ~3135) and MD_Data_Image::m_drawMode/DrawModeType,
both COLOR-blend selectors, not tile-vs-stretch geometry modes. The prior
"Normal -> tile, matching ImgTex::TileCSI" citation in UiDatElement was a
mis-attribution: ImgTex::TileCSI (0x0053e740) is called exclusively from
TexMerge::CopyAndTile/ImgTex::CopyCSI for LAND-SURFACE terrain texture
compositing, never from the UI element system.
Given the dat authors zero resize anchors and the blitter can only copy or
tile, the only way retail's whole pre-world scene (background + buttons +
listbox together) fills an arbitrary window resolution is that these
fixed-canvas "flow" screens render at 800x600 and the WHOLE FRAME is
stretched once at presentation — outside the UI sprite system entirely.
acdream has no offscreen fixed-resolution UI render target / present-time
scale pass; CharacterManagementUiController's constructor instead resizes
the MOUNTED ROOT element itself to the live viewport, which is why its own
background tiled (Width/tw > 1 at any resolution above 800x600, wrapped by
GL_REPEAT).
Fix: UiDatElement gains StretchOwnBackgroundToFill (default false, every
ordinary chrome/container element keeps tiling) — when set, the element's
own DirectState background draws as one UV-0..1 quad instead of the native
tile formula. CharacterManagementUiController sets it on Root right where
Root is resized to the host viewport, reaching the same visual result as
retail's present-time stretch (no tiling, no aspect-preserving letterbox)
through a different mechanism. Divergence register row AD-98 records the
substitution.
Tests: three new UiDatElementTests pin the UV-span mechanism generically
(tile past 1.0 when unset and rect exceeds native size; clamped to 1.0 when
set; byte-identical to the old tile formula when rect equals native size,
so every unaffected panel is untouched). CharacterManagementUiControllerTests
pins Root.StretchOwnBackgroundToFill == true post-construction. The live-DAT
gate confirms the root's zero edge-anchors and Type=3 against the installed
DAT. AcDream.App.Tests: 5084 passed / 3 skipped with ACDREAM_PROBE_LIVE_MOUNT=1
(5081/6 skipped without it — the 3 live-DAT-gated tests skip).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
F1: the crash reporter comment claimed the launcher never holds a password
in any field - false (ProfileEditorDialogViewModel, AccountProfile.Password,
StartRequest.Password). Reworded to the true, narrower invariant (no throw
site interpolates a credential VALUE into an exception message) and pinned
it with CrashReportNeverContainsAStoredPassword: a real STJ failure over a
profiles document containing a known password, corrupted after the
credential, must yield a crash file with the stack and without the value.
F2: the co-deploy Inputs covered only Bake own sources; a Content edit
never refreshed the 83 MB exe. Now the full reference closure. Fixing it
surfaced two more incrementality traps, both fixed and comment-documented:
SkipUnchangedFiles left the output older than the triggering input (target
re-ran forever - added an explicit Touch), and %(Item.Metadata) in a plain
Include does not batch (the literal percent-text became a permanently
out-of-date phantom input - globs are now spelled per project). Verified:
Core edit retriggers, then two consecutive clean incremental builds.
F3: RID publishes ran BOTH co-deploy paths (two self-contained bake
publishes). Build-time target now guarded on _IsPublishing; verified a
real win-x64 publish runs zero build-target co-deploys and still ships
both exes.
F4: comment misattributed PublishBakeTool=false to CI lanes; it is
target-local recursion guarding. F5: the x:Name reflection sweep now walks
the markup as XML and tolerates template-scoped names (no generated field
exists for those). F6: dead using removed. Hardening: the crash reporter
positional --data-dir fallback requires a fully-qualified path so a
relative or flag-shaped value cannot create ./crash-reports at an
arbitrary CWD.
Launcher 67/67, Launcher.Core 317/317.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
#398 was a crash on every modal open/close caused by MainWindow's
constructor calling AvaloniaXamlLoader.Load(this) instead of the
generated InitializeComponent() — only InitializeComponent assigns the
x:Name backing fields, so every named control was null and the first
Dispatcher.UIThread.Post callback in OnViewModelPropertyChanged threw
NullReferenceException, killing the process. It reached the user gate
because no test in tests/AcDream.Launcher.Tests (ViewModel-only) ever
constructed a MainWindow. #399 is the process gap that let that class of
defect through 14,012 green tests.
Adds Avalonia.Headless.XUnit 12.1.1 to the launcher test project. Its
net10.0 dependency group targets xunit v3, so the project migrates
xunit 2.9.3 -> xunit.v3 3.2.2 (drop-in: all 54 pre-existing tests compile
and pass unchanged under dotnet test via xunit.runner.visualstudio 3.1.4,
which already supported v1/v2/v3; two call sites needed
TestContext.Current.CancellationToken per the new xUnit1051 analyzer).
TestAppBuilder.cs wires [assembly: AvaloniaTestApplication] to a headless
AppBuilder.Configure<App>() so the real App.axaml FluentTheme is live in
tests.
MainWindowViewTests.cs adds 12 [AvaloniaFact]/[AvaloniaTheory] tests:
- an explicit non-null + type check of every x:Name field the
code-behind dereferences (ProfilesTree, ServerNameTextBox,
AccountNameTextBox, CharacterNameTextBox, EditorSubmitButton,
FirstRunDatDirectoryTextBox, FirstRunCloseButton, UpdateCloseButton)
- a reflection sweep over every x:Name found in MainWindow.axaml, so a
future named control without a matching non-null field fails loudly
- one open+close round trip per ProfileEditorKind (all seven, including
Remove), plus the first-run wizard and the update prompt, each pumping
Dispatcher.UIThread.RunJobs() so the queued focus callback actually
executes instead of just being asserted vacuously
- a dedicated test for the _focusBeforeModal-restore branch (not just
the ProfilesTree.Focus() fallback), anchored on a real focusable
button since ProfilesTree (TreeView) has Focusable="False" under
FluentTheme — its own tab stops are TreeViewItem rows, so the
close-path assertions check "no exception escaped the dispatcher"
rather than "focus landed on ProfilesTree"
Falsification (required evidence): reverting MainWindow's constructor to
AvaloniaXamlLoader.Load(this) and rerunning gives 12 failed / 0 passed —
10 tests throw NullReferenceException at MainWindow.FocusActiveModal,
propagating cleanly out of Dispatcher.UIThread.RunJobs() (confirming
dispatcher exceptions are not silently swallowed), and the 2 reflection
tests fail on an explicit "x:Name 'ProfilesTree' was null after
construction" message. Restoring InitializeComponent() gives 12 passed /
0 failed. Full launcher suite: 66 passed / 0 failed, reproduced on both
Windows and native Ubuntu (WSL, no display/Xvfb — Avalonia.Headless needs
none). AcDream.Launcher.Core.Tests: 317/317 unaffected.
No CI workflow change needed: .github/workflows/headless-portability.yml's
portable-launcher job already runs dotnet test on the launcher test
project on both windows-latest and ubuntu-latest with no display setup,
which is sufficient for Avalonia.Headless.
Closes#399.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
#399 (HIGH, process class): no test constructs MainWindow — the launcher
test project is ViewModel-only with no Avalonia headless package, which is
how a crash on every modal open/close passed 14,012 green tests and reached
the user gate. Fix direction is Avalonia.Headless.XUnit plus a view test
that drives every modal open/close, catching the class rather than one
spelling.
#398 (MODERATE): the top-level guard prints only ex.Message, so the fatal
NullReferenceException fixed at d54b8a78 surfaced with no file, line, or
frame; diagnosis needed a temporary code edit and rebuild. Fix direction is
a redaction-scanned crash file under the data root.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
