Each of these was temporary apparatus added to chase one bug, and each was
supposed to be deleted in the commit that fixed it. Fourteen closed issues
later they were still here: #337's support/wire-mesh trio, #171's sticky
timeline, #119's viewer and entity dumps, #113's phantom probe, and a dozen
more. 3,493 lines removed; the client now reads 144 environment variables
instead of 161, and 47 temporary probes remain instead of 64.
This is not only tidying. Every probe leaves a branch on its hot path when
unset, several re-read the environment per call rather than caching, and
the volume buries the diagnostics that are actually load-bearing. It is
also a headless correctness matter: HeadlessStaticStateAudit reflects over
PhysicsDiagnostics' flags to refuse a multi-session host when any is set,
and cannot see probes that live outside that owner.
Four files went entirely — WalkMissDiagnostic.cs, CollisionMeshWireframe.cs
and two test files whose only subject was a deleted probe.
TransitionTypes.SetContactPlane also sheds its CallerMemberName /
CallerLineNumber parameters, which existed solely for #337's cpSrc=
attribution and carried the instruction to strip them with the probe
family; no call site passed them, so no behavior changes. F2's collision
overlay survives and reverts to its proxy-cylinder form, which is what
removing the ACDREAM_WIRE_MESH upgrade means.
LaunchOptionsDocumentationTests earned its keep here: it refused the
deletion until docs/launch-options.md moved the 17 rows into Retired and
the frozen direct-read counts came down (PhysicsEngine.cs to zero,
TransitionTypes.cs 3 to 2). The documentation could not drift during a
cleanup this wide.
The 14 probes that name no owning issue are deliberately NOT deleted.
Nothing records when they became safe to remove, and guessing is how a
future investigation loses apparatus it needed; #435 stays open for their
attribution.
Full hermetic suite 15,321 passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The client reads 161 ACDREAM_* environment variables across 79 files. Only
about 25 were written down, and the audit found the documentation drifting
in both directions: CLAUDE.md still advertised ACDREAM_RUN_SKILL /
ACDREAM_JUMP_SKILL (deleted; skills are server-authoritative now, and the
jump fallback is 300, not the documented 200), while flags with real
side effects had no description at all.
docs/launch-options.md documents every one by lifecycle — production,
command line, measurement, automation, permanent diagnostics, temporary
probes, deprecated, retired — with a mandatory side-effects column. That
column is the point: #432 cost three days of taxed measurements because
ACDREAM_AUTOMATION_ARTIFACT_DIR reads like an output path and also builds
a per-frame diagnostics referee, and ACDREAM_STREAM_RADIUS silently
measures a streaming window production never uses. Rows now say so. Other
surprises the audit surfaced and recorded: ACDREAM_DUMP_SCENERY_Z swaps in
a duplicate scenery-placement path rather than only logging,
ACDREAM_PROBE_VIS silently also enables ACDREAM_PROBE_ENVCELL, and
ACDREAM_DUMP_ENTITY's id list doubles as an unrelated probe's watchlist.
LaunchOptionsDocumentationTests enforces it, because a hand-maintained list
of 161 flags is stale within a week: an undocumented flag fails, and so
does a documented row whose read site was deleted. It scans string literals
rather than GetEnvironmentVariable call shapes — the startup path reads
through an injected delegate, so a call-shaped pattern silently missed
ACDREAM_LIVE, ACDREAM_PAK_PATH and every other production flag. A third
test freezes per-file direct-read debt by exact count (20 files outside the
owner classes), so structure rules 4 and 5 can be paid down but not
regressed.
CLAUDE.md's 94-line env-var section becomes a 16-line pointer, and its
stale test-character paragraph is corrected.
Also fixed, all doc-vs-code mismatches the audit proved:
- RenderingDiagnostics.FrameProfEnabled described a GPU-query self-disable
that Campaign V slice V11 deleted.
- Two comments named ACDREAM_RENDER_BACKEND as a live co-requisite; it died
with the OpenGL backend.
- EnvCellRenderer.CollectCellAuditLines and its ACDREAM_A8_AUDIT doc: the
method had no caller anywhere and its documented caller never existed.
Filed rather than fixed, to keep this a documentation change: #434 (the
DebugPanel/DebugVM surface is never constructed, so ~40 "runtime-toggleable"
comments are false and 35 env reads are unreachable) and #435 (17 temporary
probes outlived their closed investigations; 14 more name no owner).
Full hermetic suite 12,202 passed / 0 failed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Live Holtburg at 2560x1440: the Low preset needed 67,368,164 resident bytes
(screen-sized HDR/depth/ray targets are 44 MB of that) against an absolute
64 MiB ceiling that had only been validated at 1080p, so Options -> Apply
fell back to the default path; every later Apply was then refused by the
controller's failure memo, which treated the user's deliberate choice like
automatic re-activation.
RenderPackResidentBudget.Effective scales the declared 1080p figure by the
viewport's pixel-count ratio (never below 1), still capped by the hardware
MaxPackResidentBytes; both pack graphs use it and the performance-matrix
tool judges its resident column by the same rule (contract test updated).
RenderPackController.Request gains explicitUserChoice, which clears the
memo for that selection; RenderPackSelectionBinding passes it on every
display edge (Apply, including resolution changes) and keeps the memo for
the startup request.
Tests: RenderPackResidentBudgetTests (1080p/720p keep the declared
ceiling, 1440p = 16/9x, 4K = 4x, hardware cap wins, zero extent rejected);
controller explicit-retry; the binding test now proves the user's next
Apply activates once the cause is gone. App hermetic lane 6,068/0 (Release).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Imported from the codex worktree's uncommitted work
(C:/Users/erikn/.codex/worktrees/bd98/acdream on
codex/atmospheric-rendering-campaign), on top of its 8b7b601b.
The remote-player shadow gate ran two clients on fixed sleeps, so the
observer could move before the primary had taken its 'before' screenshot,
or the primary could take its 'after' shot before the observer had moved.
Now the harness publishes named signal files into each client's artifact
directory and the routes block on them:
- IRetailUiAutomationRuntime.TryIsAutomationSignalPublished, implemented
by WorldLifecycleAutomationController over <artifactDir>/signals/
<name>.signal (names validated by AutomationArtifactName, so no path
escape).
- 'wait signal <name> [timeoutMs]' in RetailUiAutomationScriptRunner.
- run-connected-render-pack-remote-player-gate.ps1 publishes
'primary-before' to the observer after the primary's before-shot
completes, then 'observer-moved' to the primary after the observer's
remote-observer-moved checkpoint.
- Both routes teleport with an explicit heading and wait 12 s to settle.
Build green; the three touched App test classes pass 58/58 including the
two new signal tests.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Every automated pixel gate and every blank-world verdict in Campaign V is
produced by FrameScreenshotController reading the default framebuffer with
glReadPixels. The window is created with the quality preset's MSAA sample
count, so that framebuffer is normally 4x multisampled -- and glReadPixels
against a multisampled read framebuffer is undefined per the GL spec. The
instrument the campaign has been using to decide "did the world render?"
rested on an operation with no specified result.
That is not a theoretical complaint. The blank-world investigation spent
several rounds unable to tell "the renderer drew nothing" apart from "the
readback did not return what the renderer drew", and it took an out-of-process
desktop grab to separate them. A gate cannot arbitrate a rendering defect
while its own read is unspecified.
So the capture resolves first: when the default framebuffer is multisampled
it blits the whole colour buffer into a single-sampled RGBA8 framebuffer with
identical rectangles and GL_NEAREST -- the defined resolve -- and reads that.
A single-sampled default framebuffer keeps the original direct read, so
non-MSAA captures stay byte-for-byte what they were. The blit disables and
restores the scissor test, because a blit is subject to it and a frame that
left a rectangle armed would otherwise resolve only part of the image; that
is the same self-contained-GL-state rule the render passes follow. The
resolve target is created and destroyed per capture -- captures are rare, and
a cache would have to track resize and context teardown for no gain.
GlGpuDevice.CaptureBackbuffer had the identical undefined read. It now routes
through the same path rather than being a second instrument to keep sound.
The IDefaultFramebufferSurface seam grows the draw binding, the sample count,
and the resolve operations, so the bind/query/blit/read/restore order stays
assertable without a GL context; two new tests pin the resolve order and the
resolve target's release on a failing read.
Gates: Release build green. App tests 3,866 passed / 3 skipped. Offline pixel
gate against fed636b9 passes at a differing fraction of 4.08e-05 against the
0.001 threshold -- which is exactly the same-commit control pair measured at
this commit, i.e. indistinguishable from ambient noise. Same-commit controls
re-measured at 17 px (fed636b9) and 23 px (here) out of 563,200; the recorded
band in plan section 5.1 widens to 15-23 px, fraction <= 4.1e-05.
Plan section 5.5.1 records what the connected investigation established: the
interleaved A/B attribution (4/5 vs 0/5, p ~ 0.024), the desktop witness
showing every depth-tested draw missing while the atmosphere clear and the
complete retained UI present, the probe evidence that the CPU dispatched
3,331 statics with no GL error, and the falsification list -- including the
ring glBufferSubData hazard, which condition 1 shipped against and did not
fix.
Section 5.5.2 records this session's second investigation, run against a
staged (never committed) V4c with log-only glGet* probes, and it closes the
shared-3-D-state hypothesis. The depth plane is bit-identical on blank and
rendered frames -- test on, write mask on, GL_LESS, clear value 1.0, range
[0,1], full viewport, full colour mask, no clip distances. The camera
constants are sane and advancing. Forcing gl_ClipDistance off left the blank
rate unchanged at 3/5. glGetGraphicsResetStatus returned NO_ERROR in all
1,814 samples across four blank runs, which also retires the "GPU-side fault"
reading in its context-reset form.
Two sharper facts replace it. Replacing only the frame clear colour with
magenta makes a blank frame come back uniformly magenta under the complete
retained UI, so no 3-D fragment is rasterized at all -- the world is not
drawn-then-hidden, fogged, or overdrawn. And on a blank run the client's own
capture of framebuffer 0 is RGBA(0,0,0,0) in every pixel, including pixels
where the UI is visibly on screen at that moment. That survives this commit's
resolve fix, so it is a second, independent instrument fault: the screenshot-
byte verdict used by the repeat and A/B gates measures the readback, not the
renderer, and those gates need to assert on the desktop witness instead.
V4c is NOT re-landed. No fix was attempted, because the mechanism is not
renderer state and does not sit in V4c's surface as this hypothesis predicted.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
FrameScreenshotController.ReadDefaultFramebuffer called glReadPixels
without binding a read framebuffer, so it captured whatever was bound to
GL_READ_FRAMEBUFFER at that moment rather than the default framebuffer
its name promises. The capture runs at the end of
PrivatePresentationRenderer.Render, after PrivateEntityViewportRenderer
has drawn the paperdoll and appraisal views into its own FBO — an FBO it
clears to exactly RGBA(0,0,0,0). A capture that inherits that binding
writes a fully transparent PNG, which the repeat-run connected gate
scores as BLANK even though the backbuffer on screen was correct.
This was latent for as long as something rebound framebuffer 0 often
enough to mask it. Before Campaign V slice V4c, GL BeginPass bound
framebuffer 0 on every pass with a null colour target; V4c deliberately
stopped doing that (plan §5.4) so the offscreen viewport renderers could
keep their own target across a dispatcher draw. Removing the wide path
exposed the narrow bug underneath it — the same latent-bug-masked-by-a-
fallback class the project recorded for #98.
The read now binds framebuffer 0 to GL_READ_FRAMEBUFFER, reads, and
restores the caller's binding, so a diagnostic capture states its own
source and cannot perturb the frame it observes. The GL calls move behind
IDefaultFramebufferSurface so the bind/read/restore order is assertable
without a GL context; two tests cover the ordering and the restore on a
throwing read.
Gates: Release build green; App tests 3,864 passed / 3 skipped (3,862
baseline plus the two new tests), no #250 flakes; offline pixel gate
against 8dec163f PASS at a differing fraction of 4.26e-05 against the
0.001 threshold, inside the documented same-commit noise band.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Carry one immutable prepared collision closure with each accepted near-tier generation and install graph plus flat views through the same retained publication receipt. Apply the same strict package-only rule to live entities, add exact sampled graph-authoritative comparison artifacts and lifecycle counters, and prove cancellation, demotion, rehydrate, revisit, teardown, reconnect, and the nine-stop route with 14,064 zero-mismatch samples.
Co-authored-by: OpenAI Codex <codex@openai.com>
Build a same-frame candidate product from incremental render-scene indices and compare its exact PView routes against the accepted current path without changing the production draw source.
Construct Slice F's non-drawing scene only for lifecycle automation, drain accepted static and live deltas at the final update boundary, and compare exact current-path fingerprints at cadence and checkpoints. Publish bounded mismatch, journal, index, digest, and memory evidence without changing normal launches or draw submission.
Release: 8,211 passed, 5 skipped.
Extend the non-drawing oracle through ordered PView routes, dispatcher visibility and final instance payloads, and accepted retail selection parts. Lifecycle artifacts can now referee the later shadow scene without influencing production visibility or draw decisions.
Capture deterministic, landblock-aware fingerprints from the accepted partition only during lifecycle automation. This gives the F/G shadow scene an independent oracle without changing normal draw decisions or adding disabled-path per-entity cost.
Record route-lifetime streaming overruns and maximum operation costs, require canonical reveal/resource convergence at every connected checkpoint, and keep recenter semantics independent of the production wall-clock budget.
Co-authored-by: Erik Nilsson <erikn@users.noreply.github.com>
Complete Slice D4 by adding aggregate lifecycle occupancy and traffic facts, validating physical source reports, and including decoded audio under the typed startup budget. Exercise every domain under forced pressure and retain the real cache/fence convergence gates.
Starting dotnet-counters in the process-creation race can suspend the CLR before the graphical host creates its window. Wait for the guaranteed in-world boundary, then attach counters and contention tracing during the route's warm-up interval.
Co-Authored-By: Codex <noreply@openai.com>
Free-running sleeps could label a sample as the old scene exactly when automation began the next teleport. Give every route stop an explicit post-input liveness dwell and post-checkpoint hold, then capture process and frame facts only after its named canonical barrier.
Co-Authored-By: Codex <noreply@openai.com>
The first-player-position line is emitted only when initial streaming must recenter. Treating it as a required login event made a valid same-landblock login time out while route automation was already running. Gate on in-world and the route's authoritative materialization barriers instead.
Co-Authored-By: Codex <noreply@openai.com>
Correct whole-frame GPU timestamps so they bracket only the accepted render transaction and associate delayed query results with the owning CPU frame. Add route-wide frame-history summaries, fixed-camera screenshot comparison, process counters, contention traces, a pinned Arwic workload, and credential-safe launch disclosure.
The reference hardware/display contract now keeps local and RDP populations separate and defines the screenshot and re-baseline rules needed by later prepared-content gates.
Co-Authored-By: Codex <noreply@openai.com>
An adversarial performance review found our own instruments cannot
measure the project's own performance gates:
- FrameProfiler aggregated CPU/GPU/alloc/stage samples into ~5-second
windows and reset the ring buffers after each report, so route-wide
p50/p95/p99 distributions across a whole soak could not be
reconstructed after the fact. ACDREAM_FRAME_HISTORY=<path> now opts
into a separate per-frame history (one record per frame, ~72
bytes/record, accumulated in memory with zero frame-thread I/O) that
a shutdown-only Dispose() writes as CSV. The aggregated [frame-prof]
report format and its existing metrics are unchanged.
- The canonical checkpoint JSON tracked cache residency (entry/byte
counts) but never LOH size/fragmentation, process-wide allocated
bytes, or cache hit/miss/eviction traffic — a committed audit JSON
showed 65% LOH fragmentation that no tracked instrument recorded,
and "does a revisit portal hit or miss the caches" was unanswerable
from an artifact alone. WorldLifecycleResourceSnapshot now carries
loh_size_bytes/loh_fragmentation_bytes (GCMemoryInfo.GenerationInfo
index 3), process_total_allocated_bytes (GC.GetTotalAllocatedBytes),
and Interlocked hit/miss/eviction counters for the CPU mesh cache,
decoded-texture cache, and the four bounded DAT-object caches
(portal/cell/highRes/language, aggregated).
- run-connected-r6-soak.ps1 unconditionally forced
ACDREAM_UNCAPPED_RENDER=1 with no capped mode, while its sibling
lifecycle-gate script correctly gated it behind a switch. Added
-Uncapped (default capped, matching the sibling script's pattern),
fixed the stationary dwell (12s -> 26s, past the 25s
LiveEntityLivenessController deadline the adjacent comment already
cited), and now write an env-disclosure.json into the automation
artifact directory before every launch listing every ACDREAM_* var
the script sets plus -Uncapped, since the prior audit could only see
ACDREAM_DUMP_MOVE_TRUTH and nothing else was ever recorded anywhere.
Cache counters are wired via the existing composition path
(ObjectMeshManager already owns the CPU mesh cache and the mesh
extractor directly; content.Dats is threaded into
WorldLifecycleResourceSnapshotSource the same way every other
composition consumer receives it). The DAT-object cache lives behind
IDatReaderWriter, a third-party interface from the DatReaderWriter
package that cannot be extended; RuntimeDatCollection (the one
production implementation) exposes the aggregate stats directly and a
pattern match reads them, degrading to zero for any test double —
no new static registry was introduced (GpuMemoryTracker remains the
one precedented process-wide static).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
(cherry picked from commit 1da2c33c875b41fa383dd79694ee2765f0e21896)
Pin every ACE teleloc to an explicit starting quaternion and distinguish stable-workload owner growth from legitimate authoritative population, visibility, and cache-retirement changes. Preserve the original process residency and frame-cost limits while making canonical deltas actionable.
Co-authored-by: Codex <codex@openai.com>
Make scripted lifecycle checkpoints acknowledged post-diagnostics render barriers, capture the exact frame outcome beside canonical resource ownership, and harden the nine-stop route with ordered same-location cache and lifetime gates without weakening process residency thresholds.
Co-authored-by: Codex <codex@openai.com>
Move the live-session reset and routing graph, combat and diagnostic command targets, and the sole gameplay input subscriber into Phase 7 before frame publication. Add exact retryable ownership for late bindings so partial startup cannot strand session or component teardown edges.
Co-authored-by: Codex <codex@openai.com>
Move the accepted draw transaction and failure recovery behind typed frame-phase owners so GameWindow only supplies immutable frame input. Preserve retail draw order, make ImGui/bootstrap shutdown ownership explicit, and restore exact text-render GL state on failures.\n\nCo-authored-by: Codex <codex@openai.com>
