Two live-integration gaps the ef96c554 unit tests could not see:
1. World box stayed empty against ACE: ServerName (0xF7E1) arrives in the
SAME connect batch as CharacterList, so ServerNameReceived fires during
the handshake pump BEFORE the controller binding subscribes - the
event-only wiring proved the state and controller but never the live
ordering. StartCore now reads the durable WorldSession.ServerInfo after
connect exactly like the roster (ILiveSessionOperations.GetServerInfo,
default interface method so no fake breaks); the event remains for
post-connect updates. Pinned by a Start-level test.
2. The exit confirmation rendered far right of the screen: all three
retail dialog views centered against the raw window size while the
active screen lays out in the fixed 800x600 canvas - center-of-1920
is canvas-760, which the stretch pushes off-center. Views now center
against UiRoot.EffectiveCanvasSize (canvas while a pre-world screen is
active, window otherwise). Pinned by growing the window over the fixed
canvas in the exit-dialog test and asserting the scrim spans the canvas
with the popup centered at 400.
Runtime 1666, App 5100+6 skips, green.
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>
Agent was stopped for token budget. Landed here: probe flag through
LiveSessionConnectOptions + the StartCore short-circuit, the mode field
with JsonRequired-to-semantic-validation move, host exit-code mapping,
and 34 passing tests including 3 new probe tests (agent last reported
green before the stop). NOT DONE: the idle-policy unit tests (next
step), full-suite verification, and the WSL run.
Build/test state UNVERIFIED at this commit. Next session: finish idle
policy tests, run Runtime+Headless Release suites Windows and WSL, then
dispatch the Opus dual-lens review.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The LA3 Opus review process note was right: the contract both sides
implement lived only in orchestrator prompts, which is exactly the drift
mode the pin exists to prevent (and it produced the paths-key CRITICAL).
The schema, field rules, probe-mode discriminator, and status vocabulary
are now a binding plan section; amendments change this text first,
implementations second. Ledger: LA3 fix round dispatched.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Closes the two mechanism-lens and blast-lens dual reviews of Campaign OP
slice OP1 (86c0a7e0): docs/research/2026-08-10-op1-review-mechanism.md and
docs/research/2026-08-10-op1-review-blast.md.
MUST-FIX M1 (blast): RuntimeCharacterOptionsState gains a HasServerSeed
latch, set by Replace (the PlayerDescription seed) and cleared by
ResetSession. TryFlush/TryFlushIfAutoSaveDue now refuse before the seed
arrives — closing the window where a bot (or, after this commit, the
timer/logout triggers) could flush client-default option words over a
character's real server-side options before any PlayerDescription ever
landed.
MUST-FIX 1 (mechanism): the 480 s auto-save timer and the pre-logoff
flush are now wired into production, closing TS-71 (retired). Both ride
LiveSessionController's own tick/stop transaction via two new hooks
(ConfigureAutoSaveTick/ConfigurePreLogoffFlush), wired once by
GameRuntime's constructor — a Runtime-internal change requiring zero
host edits, exactly as the review identified. The flush body talks to
WorldSession directly rather than through App's LiveSessionCommandRouter,
which is what keeps this off the S2 lock-order hazard (below). Filed
TS-73 for the two OnChanged side-effect cases (weather/day/combat-
target/fog) TrySetOption still doesn't model — pre-anchored to OP4's
Group B consumer binds.
SHOULD-FIX S2 (blast, prerequisite for MUST-FIX 1): TryFlush/
TryFlushIfAutoSaveDue no longer invoke the flush callback while holding
_dirtyGate — the decision is made and cleared under the lock, but the
callback itself runs outside it, closing the lock-inversion hazard the
natural timer wiring would have hit (Runtime tick's _dirtyGate-then-
_gate vs the router's _gate-then-_dirtyGate).
SHOULD-FIX MF-2 (mechanism): TrySetOption now ports the two
PlayerModule-state-mutating cases of CPlayerModule::OnChanged's local
side-effect switch — turning ON IgnoreFellowshipRequests or
FellowshipAutoAcceptRequests clears the other through a real recursive
TrySetOption call, reproducing retail's second 0x0005 (the clear's send
reaches the wire before the primary option's own send, matching the
nested-call order in the decomp). The signature widened from
Action sendAutoSave to Action<uint,bool> so the recursion can send a
different (id, value) than the caller's own; every production call site
now passes WorldSession.SendSetSingleCharacterOption directly.
SHOULD-FIX MF-3 (mechanism): a hand-transcribed 53-row (id, isOptions1,
mask) theory in CharacterOptionTableTests, independently re-derived from
acclient.h's PlayerOption/CharacterOption/CharacterOptions2 enums rather
than copied from CharacterOptionTable.cs — closes the one column with no
id-by-id pin. Also added the pairwise-distinctness check blast NOTE N7
named.
SHOULD-FIX S1 (blast): LiveSessionCommandRouterTests' CH3/CH4 regression
test now drives the REAL TrySetOption binding instead of a hand-rolled
SetOptionBit substitute that had silently drifted from production after
OP1.
SHOULD-FIX S3 (blast): RuntimeCharacterOwnershipSnapshot gains
OptionsAreClean (!Options.IsDirty), included in IsConverged — a module
whose two words happen to cycle back to their default bit pattern while
still dirty is now caught by the combined ownership ledger, not just by
OptionsAreDefaults.
SHOULD-FIX S4 (blast): SaveOptions no longer encodes "did it actually
flush" as PrimaryObjectId 1u/0u (which read as object guid 0x00000001 in
the K2 event stream). Both host adapters now report the identical shape
(Accepted, objectId 0) — the graphical host never could report this
anyway (LiveCommandBus.Publish has no return channel).
SHOULD-FIX S5 (blast): Replace (the server-seed arrival) now also clears
IsDirty/FirstDirtiedAt — a wholesale re-seed supersedes any pending
batched-but-unflushed local intent (retail's own PlayerModule has no
partial-merge path either), documented at the member.
SHOULD-FIX S6 (blast): a cross-check theory asserting CharacterOptionTable's
masks equal PlayerDescriptionParser.CharacterOptions1/2's independently
(the write path vs the read path TurbineChatMembershipGate/
RuntimeSettingsController consume) — guards the exact CH3 failure class.
Also fixed a real allocation regression found while landing MUST-FIX 1:
the naive per-tick flush closure would have allocated on EVERY
LiveSessionController.Tick() call regardless of dirty state, which broke
the K4 headless 30-session resource-envelope gate. GameRuntime.
FlushCharacterOptions now pre-checks Options.IsDirty (itself retail-
faithful — CPlayerModule::UseTime opens with the identical m_bDirty byte
compare) before allocating the flush closure, so the allocation only
happens on the rare tick that might actually flush.
Dispositions on findings not changed this round:
- Mechanism NOTE 6 / not independently re-flagged: a re-entrant MarkDirty
from inside a flush callback can still be erased by the trailing
"_isDirty = false" — pre-existing, unchanged by the S2 lock restructure
(same outcome whether the callback runs inside or outside the lock),
not reachable from any current caller, not a one-liner to close
correctly (needs a per-dirty-period generation token). Left as documented
in the review; worth closing before the Options panel ever flushes from
inside a change handler.
- Mechanism NOTE 9, blast N2/N3/N4/N5/N6/N8: informational or require
touching files this round doesn't otherwise edit (SocialActions.cs,
CharacterOptionsBlobSource.cs, GameRuntimeContractTests.cs) — left per
the "one-liner in a file already being edited" instruction.
Register: TS-71 retired (both remaining SetCharacterOptions flush
triggers now production-wired); TS-73 filed (the two unmodeled OnChanged
presentation-binding cases, pre-anchored to OP4).
Quality bar: Release build green; full solution suite 12,853 passed / 4
skipped / 0 failed (baseline 12,770/4/0 post-OP2 — 83 new tests added,
zero regressions).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Move canonical per-session teardown into one retryable Runtime transaction, reduce App reset to projection acknowledgements, and prove the same GameRuntime graph through deterministic no-window lifecycle, gameplay, portal, fault, reconnect, and isolation gates.\n\nCo-authored-by: Codex <noreply@openai.com>
Move the canonical WorldSession generation, connect/enter/tick/stop transaction, inbound subscription owner, and retryable teardown acknowledgements into AcDream.Runtime. Keep App as a borrowing graphical host with a single inertable command projection and no mirrored session state.
Validated by 79 Runtime tests, 3,776 App tests with three existing skips, the Release solution build, and 8,428 complete Release tests with five existing skips.
Co-authored-by: Codex <codex@openai.com>
2026-07-25 19:39:24 +02:00
Renamed from src/AcDream.App/Net/LiveSessionController.cs (Browse further)
