* feat(core): data-vfx-chain defs and chain parsing (mirrors data-fx-chain)
Mirrors audioFx.ts's shape for the per-pixel VFX chain: a versioned
JSON chain attribute (data-vfx-chain), a def registry with typed
params (number/enum/bool/ref), and parse/serialize/normalize helpers.
HF_VFX holds one def, fractal-noise, with a placeholder frag; the
kernel and its consumers (packages/core/src/runtime/vfx.ts) land in
Task 1.2/1.3 of the vfx-chain-runtime plan, so vfx.ts is added to
.fallowrc.jsonc's entry list for now.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
* feat(runtime): paint data-vfx-chain canvases on every seek
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(vfx): fractal-noise kernel (Basic/fBm approximation) with CPU reference
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(vfx): self capture through layoutsubtree with the page-composite readiness protocol
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(vfx): wave-warp kernel (Sine)
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* feat(vfx): displacement-map kernel (self-referential map)
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* test(vfx): chain order is nesting order; preceding CSS effects are captured
...and following ones apply outside. The plan's one-line message for Task 2.4
is 105 characters; commitlint caps the header at 100, so the tail moved here.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* test(vfx): seek determinism and CPU-reference PSNR for fractal-noise
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
* fix(vfx): size the output canvas from the host's layout box, not its transformed AABB
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
* feat(telemetry): vfx node count, capture class and types on render_complete
Threads vfxNodeCount, vfxCapture and vfxTypes from a static scan of the
compiled composition's data-vfx-chain hosts (mirrors the existing
compositionElementCount/colorGradingCount scan) through
RenderPerfSummary.drawElement into the render_complete event, next to
de_capture_mode. Exports chainCapture/parseVfxChain/HfVfxCapture from
@hyperframes/core's index for the producer to consume. The PostHog
dashboard tile is not created here — left for the coordinator.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
* docs(vfx): data-vfx-chain guide — element shape, capture table, Studio flag, known constraints
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
* fix(vfx): register chain hosts that arrive with a sub-composition mount
`initVfx(document.body, fps)` ran once, synchronously, three lines after
`void loadExternalCompositions(...)` — so every `data-vfx-chain` host that
enters the DOM only when a `data-composition-src` mount resolves was never
registered. An unregistered host paints nothing and reports nothing: no
`[HyperFrames] composition script error`, the `.hf-vfx-out` canvas keeps its
300x150 default, and `paintVfx` never even arms the page-composite flag.
Measured on a real multi-file export served raw: `#main-l5-text`'s output
canvas stayed 300x150 and `window.__timelines` never gained the sub-comp.
Re-scan in the loader's `.finally`, before `maybePublishRenderReady()`, so
registration precedes readiness. `initVfx` now releases the outgoing
registry's programs, capture textures and ping-pong targets first — the
output canvas hands back the same WebGL2 context on a re-scan, so the
previous pass's GL objects would otherwise leak.
Two browser cases in `vfxDeterminism.test.ts`, both against the real runtime
bundle: a host whose `.hf-vfx-in` IS the mount (already worked — the mount
replaces the wrapper's children, not the wrapper, so the registered source
node survives), and a host that arrives INSIDE a mounted sub-composition
(failed with `armed: false` and no error; passes now).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(vfx): runtime tsconfig typecheck passes
Under noUncheckedIndexedAccess, resolveCaptureSource's null return
widened VfxEntry.src past its VfxCaptureSource | undefined type; the
function is internal-only, so its return type moves to
VfxCaptureSource | undefined instead. Indexing PingPong.framebuffers
by a non-literal index produced | undefined against
WebGLFramebuffer | null, coalesced to null. No behavior change.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
* fix(vfx): structural DOM predicates instead of realm-bound instanceof
vfx.ts checked `instanceof HTMLCanvasElement`/`HTMLElement` directly, which
returns false for inner-realm nodes when the Studio previews across an
iframe realm. domRealm.ts had no canvas predicate yet, so this adds
isCanvasElement mirroring the existing isImageElement pattern, then swaps
all three vfx.ts checks to the structural predicates the lint guard requires.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
* fix(vfx): paint self-capture hosts under hyperframes snapshot, not only the engine path
`paintVfx(t, { engineMode: true })` armed the page-composite protocol and
returned, betting that every capturing chain runs under `frameCapture.ts`.
`hyperframes snapshot` (and check/compare/validate/layout, and Studio's
thumbnail capture) seek through the same `renderSeek`, never read
`__hf_page_composite_pending` and never call `__hf_page_composite_resolve` —
so a `self` chain painted nothing there, in silence, because its content
lives in a `layoutsubtree` canvas the browser never paints directly.
Engine mode now arms the preview-side capture as well and paints on whichever
completes first. A seek token stops a superseded async capture from repainting
a newer frame, and a registry-identity check stops it painting an entry
`initVfx` already released. `resolveVfxCapture` stays non-idempotent on
purpose: the engine resolves only after its own before-capture hooks have
mutated the DOM, so its capture is still the authoritative one.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(vfx): an exporter-emitted .hf-vfx-out gets the box the spec fixes for it
findOrCreateOut adopted a canvas the exporter had already emitted without
applying the layout cssText, so it kept the default position:static;
display:inline and wrapped to the line AFTER the layer-wide .hf-vfx-src
canvas — painting a second, misplaced copy of the layer one full layer
height below the host. Measured on retro-wave #main-l6-text: .hf-vfx-src
rect top 315.1 h 93.0, .hf-vfx-out rect top 409.2, both style=null,
position static (ae-mcp findings Task 3.5b). Hosts with no emitted
out-canvas were never affected, which is why only this one host moved.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* refactor(telemetry): vfx_node_count counts hosts, not chain nodes — rename to vfx_host_count
The field increments once per `data-vfx-chain` attribute occurrence, never
per node inside a chain's JSON, so the name promised something the scan does
not measure. Renames the internal `vfxNodeCount` and the wire property
`vfx_node_count` to `vfxHostCount`/`vfx_host_count` across the orchestrator
scan, the perf summary, the CLI render path and the telemetry event. No
dashboard consumes the field yet, so the wire rename costs nothing now and
would cost a migration later.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(vfx): join the preview capture to the seek-completion barrier, not race it
`paintVfx` fired `capturePreviewThenPaint` and forgot it, so nothing the CLI
awaits waited for a `self` chain's paint. Every snapshot-family command
(`snapshot`, `check`, `compare`, `validate`, `layout`) seeks through
`seekCompositionTimeline`, which drains `window.__hfWaitForSeekCompletion`
and then allows one settle race before screenshotting — and only the
`"hf-seek"` event's `waitUntil()` could put anything in that barrier. The
screenshot therefore landed on however many hosts happened to have finished,
which varied with host count and machine speed.
Extracts the "track, untrack on settle, retain the first rejection" block
`dispatch()` built inline into an exported `registerSeekCompletion`, and has
`paintVfx` register the preview capture through it. Also awaits the per-host
paints with `Promise.all` instead of sequentially: N hosts cost about two
animation frames now, not 2N.
Measured on the pre-fix bundle with two `self` hosts, both output canvases
read `[0,0,0,0]` at the moment the barrier returned; they painted only
because the settle race that follows happened to be long enough. Both tests
added here fail on that bundle.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(vfx): report WebGL context loss loudly instead of silently no-op painting
One WebGL2 context per `data-vfx-chain` host, against a browser cap that is
commonly 16 and shared with `colorGrading.ts`'s per-element contexts. A lost
context makes every later call on it do nothing, per spec — so the chain kept
"painting" a blank canvas with no error, which is exactly what the module's
loud-failure contract exists to prevent.
Adds the `webglcontextlost` listener pair's first half, following
`colorGrading.ts`'s precedent: `preventDefault()`, mark the entry, report
once through `reportVfxError`. The entry is then skipped by `paintVfx`,
`capturePreviewThenPaint` and `resolveVfxCapture` — silently, since the one
loud report already happened and repeating it per frame is spam.
Restore is left unhandled on purpose, with the reasoning in the doc comment:
`initVfx`'s existing re-scan already rebuilds every GL object, and a second
narrower rebuild path for `webglcontextrestored` would duplicate it. The
shared-context redesign the review floated is likewise out of scope here.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* perf(vfx): cache uniform locations at pass-build time instead of querying every paint
`setPassUniforms` asked the driver for `u_size`, `u_t`, `u_fps` and every
param's location on every paint of every pass, and `bindPass` did the same
for `u_src` on every bind. `buildPasses` already constructs each `VfxPass`
once at init, right after the program links, and a linked program's
locations never change — so they are resolved there now and read from
`pass.locations` in the hot path.
Same values, same `null`-means-optimised-out semantics (`gl.uniform*(null,
…)` is a defined no-op, which is what the per-frame query produced too), so
the existing vfx.test.ts and the browser determinism suite cover it
unchanged — both still green, the latter against a rebuilt runtime bundle.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* fix(vfx): bound the paint wait per host so a stalled compositor can't hang the seek barrier
`paintVfx` registers `capturePreviewThenPaint` into the shared seek-completion
barrier (48aaa2bbb), and that promise settled only via `awaitCanvasPaint` — a
race between the canvas `paint` event and a double-rAF chain. Under BeginFrame
control (Linux headless-shell, `drawelement` capture) neither can fire before
the barrier is drained: `frameCapture.ts` awaits the barrier in
`prepareFrameForCapture` (~2692), while the frame's
`HeadlessExperimental.beginFrame` is issued later, in the per-frame capture
stage (~3797). Fire-and-forget that was harmless; registered it can stall every
`__hfWaitForSeekCompletion` caller, the engine's render path included.
rAF-under-BeginFrame: CONFIRMED gated, from this repo's own engine notes —
`frameCapture.ts` ~2445 ("waitForFunction uses rAF polling internally, which
won't fire in beginFrame mode") and ~4755 ("headless only fires rAF when a
frame is produced, and nothing produces one until a screenshot asks"). Not
reproduced locally: every environment here is macOS with a free-running
compositor (`beginFrameTimeTicks === 0`). `setTimeout` is not compositor-gated
there — `drawElementService.ts` ~341 describes its 250 ms paint fallback as
burning on every BeginFrame frame, which requires the timer to fire — so a
timer ceiling is a sound escape.
`awaitCanvasPaint` now resolves `"painted" | "timeout"` with a 2000 ms ceiling
(the repo's own `BEGINFRAME_PROBE_TIMEOUT_MS`; ~20x a real two-frame paint,
far short of the ~180 s CDP `protocolTimeout` a hang would burn). The race is
per host, so one stalled compositor no longer delays hosts that would have
painted; a timed-out host is skipped (no capture, no guessed pixels) and
reported loudly. The report is not suppressed in engine mode: `frameCapture.ts`
~2708 skips `__hf_page_composite_resolve` in `drawelement`/`beginframe` modes,
so there is no authoritative second attempt to defer to there.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
* chore(ci): re-pin test-classification.mjs hash after registering vfxDeterminism.test.ts
An earlier commit added vfxDeterminism.test.ts to INTEGRATION_TEST_FILES;
check-test-reachability.mjs pins a sha256 of that file to force a conscious
review whenever it changes. Reviewed: runner[1]'s glob already covers the
file structurally, the classification move (unit -> integration) is correct
for a Chrome-launching test, and the pinned hash is now current.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
* chore(core): regenerate position-edits-render-inline.ts for Linux esbuild output
check:position-edits-render (bun run build:position-edits-render && git diff
--exit-code) failed on both Linux CI runners (Test (runtime), Tests on
windows-latest: studio-core) but passed on macOS: esbuild's minifier named
two identically-scoped WeakMap/function bindings differently on Linux than
on the mac that last committed this artifact, and the divergence appeared
only on this branch (main's own current tip passes CI cleanly), so it is
tied to a change on this branch rather than a random flake, even though it
is not attributable to any specific line of that change.
Confirmed stable across two independent Linux CI runs (byte-identical diff
both times) and reconstructed directly from that diff rather than guessed:
the committed content, byte length (3817), and 3/3 existing artifact tests
all match. A future contributor rebuilding on macOS will see this same
diff again in the other direction until the build script's cross-platform
determinism is fixed at the esbuild-invocation level, which is out of this
PR's scope.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
* fix(engine): classify a single-keyframe video as problematic, not healthy
analyzeKeyframeIntervals short-circuited to isProblematic:false whenever
it found fewer than 2 keyframes, but exactly one keyframe is the worst
case for seek performance, not a healthy result: every seek past frame 0
must decode the entire file.
The single-keyframe branch treats the rest of the video stream's own
duration (not a container duration, which a longer audio track can
inflate) as one uninterrupted GOP. ffprobe's frame timestamps are
absolute presentation timestamps rather than relative to stream start,
so a re-muxed or trimmed clip with a nonzero start time is normalized
the same way extractFinalVideoFrameTimestamp already does for the same
quirk. The cache key now includes both values, matching that sibling
function's convention, since a stale cache entry from a different
duration/start pair would otherwise win silently.
Co-Authored-By: Miguel Ángel <miguel.sierra@heygen.com>
* refactor(engine): share stream-window normalization between keyframe check and final-frame lookup
extractFinalVideoFrameTimestamp and analyzeKeyframeIntervals each
independently computed the same stream-start fallback, finite guard, and
cache-key format from the same metadata shape. Extracted resolveStreamWindow
so the two can't drift out of sync with each other.
Also: match ffprobe's real trailing-comma csv output in one keyframe mock,
add a test at the exact 2s threshold boundary, and comment the
noUncheckedIndexedAccess-only fallback at the single-keyframe branch.
Co-Authored-By: Miguel Ángel <miguel.sierra@heygen.com>
* style(engine): trim long comments to one-to-three lines
---------
Co-authored-by: Miguel Ángel <miguel.sierra@heygen.com>
* fix(producer): bound nested media by a host's data-duration when it has no data-end
The runtime hides a slot's descendants at data-start + data-duration, and
the compiler only stamps data-end on media tags, so a composition host
authored with data-duration was unbounded for the media planner: audio
and video nested in a shortened scene kept playing over the next scene.
The template-wrapped compile test asserted that unbounded end (a 4s clip
in a 2s slot ending at 6); it now expects the slot end.
The host end now comes from the runtime's own resolveAuthoredTimingWindow
rather than a local re-implementation, so precedence matches the runtime
exactly: a positive data-duration wins over a conflicting data-end, and a
zero or negative data-duration is treated as absent (falling back to
data-end, or unbounded) instead of collapsing the host window to its start
and dropping every nested clip. Tests cover the conflicting-attribute case
and the non-positive-duration fallback.
Co-Authored-By: Miguel Ángel <miguel.sierra@heygen.com>
* style(producer): trim long comments to one-to-three lines
---------
Co-authored-by: Val <val@codejam.info>
Co-authored-by: Miguel Ángel <miguel.sierra@heygen.com>
* ci: select the remaining orphan tests in their owning runners
* ci: provide FFmpeg for the catalog preview witness
* ci: pass the installed Chrome path to catalog rendering
* test: align creator capability checks with speed ramp support
#4153 added supported speed ramps as rate lanes in data-automation. The
previously unselected capability tests still required preprocessing.
Require the documented rate-lane representation and its precedence over
the constant playback rate, matching the runtime resolver.
* feat(catalog): add the seven 3D-motion pieces as installable registry blocks
Each piece ships as a registry block with a CDN poster and preview, a live
catalog detail page and its source. Scripts travel inside the catalog payload
(shared vendor libraries fetched once, item scripts inline, the one HDR as a
data URI) because the docs host only serves fonts, images and JSON.
NO_ESSAY_COMMENTS_OK: the long comment blocks are in vendored libraries and
the ported frost source under registry/blocks, not hand-written here.
HF_MAX_NONLFS_KB=6000: four hosted texture PNGs under docs/public/catalog/items
are runtime assets the docs host must serve as files.
fallow skipped locally: it audits against origin/main and this branch is
stacked on an older main.
* chore(catalog): regenerate gallery data and nav after rebasing onto main
* fix(catalog): read the install manifest through the bounded project reader
* fix(catalog): a script the inliner missed fails the payload instead of shipping a dead URL
* refactor(catalog): the leftover-script check reuses the reference finder
* fix(catalog): frost and glass textures load from the CDN and the iteration presets are gone
* chore(catalog): regenerate the frost and glass payloads and nav
* fix(catalog): frost and glass textures are hosted through the registry url field, not hardcoded
* chore(catalog): regenerate the frost and glass payloads against the hosted texture URLs
* fix(catalog): the hosted-reference rewrite refuses a bare name it cannot safely replace
* ci(catalog): verify changed catalog payloads in headless Chrome and check the docs catalog
* feat(catalog): every item page follows the shared layout with a Tune panel and tabs
* fix(catalog): let the texture-mask item host its masks so its tile plays live
* chore(catalog): regenerate item pages, gallery data and the texture-mask payload
* chore(catalog): regenerate item pages after merging main
* fix(catalog): declared-only variables get the Tune panel, Code and Docs tabs capped
* feat(catalog): give the blend-difference caption a demo scene so its tile plays live
* chore(catalog): live payload for the blend-difference caption
* perf(catalog): heavy 3d pieces compile and draw before ready, not on the first visible frame
* feat(catalog): the avatar promo card hosts its videos, audio and fonts on the CDN
* feat(catalog): the avatar promo card tile plays live from its CDN-hosted media
* chore(catalog): the avatar promo card plays live, no allowlist left
* chore(catalog): shorten two comment blocks
* chore(catalog): trim first-party comment blocks to the comment-check limits
* chore(catalog): shorten two frost source comments
* fix(catalog): the Docs tab renders as prose again, and the detail page honours reduced motion
Frame-mode pages carry no prose wrapper, and the detail component sat inside
not-prose, so the Docs slot's headings, paragraphs, links and lists had no
styling at all. The slots now carry their own prose classes and the chrome
around them keeps not-prose.
The recorded clip also gets the reduced-motion guard the docs-snippet check
requires, and the generator reads stability from the source metadata type so
the scripts typecheck passes.
* fix(catalog): docs-tab paragraphs render as blocks
Mintlify emits paragraphs as inline spans, so two paragraphs in the Docs slot ran together.
* fix(ci): drop the duplicate docs_catalog job output
git's line-based auto-merge silently duplicated the outputs key when
main and this branch each added it on non-adjacent lines during the
rebase; oxfmt's yaml parser rejects the repeated map key.
* fix(catalog): close four CodeQL alerts in the frost source
* chore(catalog): allowlist the avatar-promo-card CDN asset deletions, format sync-docs-catalog
check-no-main-deletions.mjs was rejecting 11 local mp4/font files this branch
already deleted in favor of CDN hosting; scripts/sync-docs-catalog.mjs had one
line over the 100-col print width from an earlier edit in this branch.
* fix(catalog): detail tabs use docs prose spacing, Install first; previews skip no-WebGPU runners
* fix(catalog): the docs tab scrolls with the page instead of a capped inner box
* refactor(catalog): split the payload asset and verify scripts into single-purpose helpers
* refactor(catalog): split the preview generator into single-purpose helpers
* ci(catalog): fail when the committed catalog differs from generator output
The docs host serves the committed catalog and nothing regenerates it, so a stale
copy shipped silently. The catalog payloads job now generates into a temp dir and
compares it with the committed tree.
* chore(catalog): regenerate docs/public/catalog with the refactored scripts
One full regeneration from an empty output tree with the current scripts. The tree grows
from 58.4 MB to 65.1 MB. Two hosted assets are about 2 MB each (an avatar png and a wav);
the same bytes are already committed under assets/, so HF_MAX_NONLFS_KB is raised for
this commit only.
* fix(catalog): run the catalog script tests under the runner they are written for
The rebase put three node:test files in the vitest half of test:scripts, where they report
"No test suite found". They now sit in the node --test list.
* fix(catalog): a WebGPU item without an adapter shows its recorded clip and says why
* ci(catalog): the drift check runs from a committed fetch mirror with no network
Generation now serves every font and CDN script fetch from scripts/catalog-fetch-mirror and fails
when a fetch is not in it, even if the caller swallowed the error. CATALOG_FETCH_MIRROR=record
refreshes the mirror from the live network. The docs_catalog filter also watches packages/core,
packages/producer and the mirror, and the 13 re-hashed font files are listed in ALLOWED_DELETIONS.
* chore(catalog): clean up the temp font cache and sort the mirror index without a locale
* fix(catalog): hide Tune when the clip stands in, time out the adapter request, share one check
* fix(catalog): the adapter probe survives a sync throw and a hang; test the clip guard
* fix(catalog): keep the adapter probe inside the component, Mintlify evaluates only exports
* fix(catalog): run the detail and page generator tests in CI; no Tune flash while probing WebGPU
* docs(catalog): credit vendored libraries and make each SKILL.md list what the block installs
CREDITS.md names every vendored dependency with its upstream and licence, and flags the two leaf
textures whose origin is still to be confirmed. Each SKILL.md Files table now comes from the block
manifest and its install and render paths point at compositions/<name>/. A test checks that every
path a SKILL.md names is a manifest file that exists.
* fix(catalog): a declared WebGPU item passes verify by its manifest tag, not by console wording
verify-catalog-payloads forgave a missing adapter only by matching the markup and the message, which
library-reported errors and widgets never satisfy. The webgpu tag in the item manifest is now the
owner: a tagged item ignores its page and console errors, an untagged one gets no forgiveness, and a
marker payload with no html is listed instead of crashing the run.
* docs(catalog): correct the credits per block and guard the mount and render paths
CREDITS.md states the three.js version and vendored-or-CDN status per block, adds d3-delaunay,
Clipper on the CDN and the Google Fonts loads, and lists every texture with no known origin.
The SKILL.md test now also checks the mount and render paths against the manifest composition.
* test(catalog): read each block manifest once in the SKILL.md file test
The two checks shared a copied manifest read, which fallow flagged as a clone with a high score.
* fix(catalog): commit the hosted item files a gitignore rule kept out, and check git not the disk
The compositions/ and *-ad/ ignore rules matched folder names under docs/public/catalog/items,
so 33 hosted fonts, textures and marks were generated but never committed and would 404 on
the docs host. The rules now have a negation for that tree and the files are added.
The drift check read the checkout, where git-ignored files still exist, so it passed locally
and failed only in CI. It now unpacks HEAD with git archive and compares that.
* fix(catalog): a tile keeps its poster until the composition has painted
* fix(catalog): a tagged WebGPU item forgives only the adapter-absent error, and test the wiring
The manifest webgpu tag now decides whether an item may lack an adapter, and the error wording
decides which lines are that absence. Any other page or console error on the item, and every
failed request, still fails. The per-item decision is one exported function the tests drive,
marker payloads are parsed by a tested helper, and the SKILL.md variable counts are checked
against their tables (cuboid-carousel said forty-nine for forty).
* fix(catalog): the pinned group's posters load eagerly at high priority
* fix(catalog): type the tile reveal test so the scripts typecheck passes
* fix(catalog): stop the adapter signature matching an unrelated failed adapter request
failed to request .*adapter also matched errors about other adapters. It now matches only a
request for the WebGPU adapter, so a tagged item cannot hide an unrelated failure behind it.
* test(catalog): a tagged item still fails on an error about some other adapter
* feat(catalog): the install command sits above the preview, as on the live catalog
* fix(catalog): find the install slot through Mintlify's child wrapper
* fix(docs): catalog copy buttons keep their size when the label swaps to Copied
* refactor(docs): name the shown copy label once in the catalog copy button
* fix(docs): primary catalog button keeps a readable label on hover and after copy
* fix(docs): inset the catalog tab strip and tab content to match the boxed panels
* feat(engine): render speed ramps with the same source-time mapping as the preview
The frame extractor, frame lookup table, coverage check, compile-time duration and audio mixer
read a clip's rate lane through sourceTimeAt, so a rendered frame at t shows the source frame
the preview shows. Audio bakes a ramp as consecutive source slices stretched by their mean
rate, keeping pitch. The compile-time attribute reader now handles single-quoted and
entity-encoded JSON attributes. Adds a speed-ramp regression fixture.
* test(producer): regenerate the speed-ramp golden audio with the current engine
The stored baseline predates the per-slice audio fix, so its audio drifted 1 to 3 percent in
pitch from the expected source-time track and the CI render correlated at 0.61. Rendered in the
CI test image, the new baseline tracks the expected pitch to 0.1 percent median, and the compare
now reports visual 0 failed frames and audio correlation 1.000.
* refactor(core): drop data-preserve-pitch, pitch is preserved by construction
The engine always time-stretches audio with atempo, so an opt-out read only by
the preview made preview and render disagree. Remove the attribute and its
reader so both paths keep pitch.
* fix(producer): stop disk capture early when frames cannot fit on disk
* test(producer): a measured disk shortfall is never retried as a browser death
* fix(producer): retry a dead browser tab once on the single-worker capture path
* fix(producer): keep the caller-owned dedup perf list untouched on sequential retry
* test(producer): extend the streaming-stage orchestrator mock with the retry exports
Review follow-ups on the R3 hardening PR:
- A resume that finds every segment complete no longer opens a browser: the
stage closes the probe it was handed, concatenates the plan's segment list
and returns success. A stage test asserts the session factory is never
called on that path.
- retryAfterTargetLoss notes that its recycle also resets the worker's
cadence counter (intended: the fresh browser is as new as a cadence restart).
- The long-form README gains a Known cost section on the narrower retry
surface of segmented capture versus plain streaming, and an operational
note that renders/.hf-segments/<planHash>/ persists after failure by design
and that the plan hash now covers GPU encoder, output resolution and motion
blur.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Four renders on the 60 s fixture, one injected failure each: ffmpeg killed
mid-stream, one worker's Chrome killed, the same under the manual
HF_DE_PARALLEL_STREAM opt-in, and one worker frozen with SIGSTOP under a 5 s
stall window. All four retried on one worker and delivered all 1800 frames;
the worker death surfaced its own error in 0.9 s instead of a 60 s stall
relabel, and the opt-in cohort retried at one worker instead of N. Records
the one caveat: a frozen renderer cannot observe the pool abort, so that
retry waits for the pool to settle.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The streaming stage interleaves on forceParallelStream OR
HF_DE_PARALLEL_STREAM=true, but the plan only carried the two routers'
flags. A manual opt-in below the DE router's frame floor therefore retried
at N workers with forceParallelStream false, which the env var re-interleaved
on the second attempt, contradicting the single-worker guarantee the retry
comment and log line made. isParallelStreamForced folds the env var into the
plan flag at creation, so the plan, the retry target and the
deParallelStreamForced telemetry agree with the distribution the stage
picks. Found by the adversarial review of this branch.
Also from that review: the parallel-stall log and checkpoint no longer claim
a "single-worker" retry, which router-pinned plans do not make (they take
their precomputed routing fallback; the post-replan log carries the real
count); the fallbackReason doc comments list encoder_death and
parallel_stall; the peer-death test names the engine test that pins the
hook contract it replays.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
- tests/long-form/README: the two-clip in-flow fixture, the per-frame luma
check over clip B, the 2026-09-18 matrix (branch -w 1 / -w 2 disk / -w 2
streaming / absolute CSS / published 0.8.44 all 0 of 100 black; 0.8.35 and
0.8.36 99 of 100 as the positive control; fixed in 0.8.37 by #3893), and
why the multi-worker trace's captureMode was not the engine's mode.
- HANDOFF: R3 row now reads "hardening built"; new section 4a lists the
guards, the 694 gate result and what the flip still needs; counts refreshed.
- cli.mdx: the parallel-stream paragraph says the opt-in path now retries a
stalled pool or a mid-stream encoder exit on a fresh single-worker session.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
The capture-stage trace checkpoints read captureMode from the probe session
and, when none is open, assumed "beginframe". The multi-worker path closes
its probe before capture starts, so every macOS and Windows multi-worker
render logged captureMode:"beginframe" while each worker's own
[initSession:screenshot] prefix said otherwise. Found while gating the
PRINFRA-694 fixture on the R3 route. The fallback now goes through the
platform rule (resolveObservedCaptureMode) via a pure helper with tests;
the telemetry field was already derived correctly and is unchanged.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
shouldRetryViaPinnedFallback is the routing-independent retry allowlist;
until now a parallel-capture stall and an encoder dying mid-stream both
fell through it and hard-failed the render on default routing. Two
predicates join the allowlist:
- isParallelCaptureStallError matches the stage's typed stall by name or by
its "[Render] Parallel <label> capture stalled" message.
- isRetryableEncoderDeath matches "Streaming encoder exited before frame N
was written" for N >= 2, so a session that streamed at least one frame is
retried on a fresh encoder, while a death before any frame (an ffmpeg
argument or codec problem that reproduces) and a host-interrupted encoder
(EncoderInterruptedError, which has its own path) are not.
The fallback reason, log line and checkpoint gain matching parallel_stall
and encoder_death branches, and the stage receives the capture label
(screenshot / drawElement / BeginFrame) so the stall names the engine.
Mutation-checked: removing the allowlist line, the frame threshold, or the
EncoderInterruptedError exclusion each fails one of the new tests.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
replanAfterFailure's default-routing fallback kept plan.workerCount for
every non-memory failure. For the parallel-stream router's plan that meant
retrying at N workers with forceParallelStream off, which is contiguous-chunk
streaming: worker k+1's first frame waits for all of worker k's, so
throughput serialises to roughly one worker while N Chrome processes run
and the no-progress watchdog can re-trip on the same render. The retry now
targets a single worker for interleaved plans, the fully hardened path
(typed stall, init-excluded clock, routing-independent retry).
Mutation-checked: dropping the forceParallelStream term fails the new test.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Three of the 5.5 % screenshot-cohort failure modes lived in the parallel
branch of the streaming stage:
- The no-progress watchdog started its clock when the pool was created, so
a fleet whose browsers took longer than HF_DE_STALL_MS to launch was
reported as stalled before any worker had a chance to capture. The timer
now arms on the first frame_capture phase (or first captured frame).
- The stall was a plain Error whose message said "drawElement" regardless
of the capture engine in use. It is now ParallelCaptureStallError with a
variant (drain | pool) and the caller-supplied capture label, so the
orchestrator can route it by name and the log names the engine that
stalled.
- A worker dying mid-render left the reorder buffer parked until the
watchdog fired, which relabelled the real cause as a stall. The stage now
passes an onWorkerFailure hook to executeParallelCapture and aborts the
writer with the worker's own failure the moment the pool declares it
fatal.
Each guard has a test that fails when it is removed (init-hang, hard-coded
label, no-op hook), checked by mutation.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Review notes on the Phase 2d PR: shouldSegmentCapture took workerCount and
never read it (any count segments), so the parameter documented an intent
the body did not have. Removed from the signature, call site and tests, and
the doc comment now says worker count is not a gate. The stage input's
sessionFactory doc says it is shared across workers when no per-worker
factory is given and must return a distinct session per create().
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Three workers produced a byte-identical file to the single-worker segmented
render at 1.6x the speed, which is the property the gate exists to check:
segments finish out of order, so a mismatch would mean the concat list had
picked up completion order instead of the plan.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
shouldSegmentCapture replaces the 2a single-worker gate: the kill switch and
the forced opt-in keep working, and a duration threshold decides the rest.
The threshold ships disabled (the default is Infinity) rather than at the
spec's 600 s. Flipping the route for every long render in the same release
that introduces the route is exactly what the soak gate exists to prevent;
SEGMENTED_MIN_SECONDS_AFTER_SOAK records the intended value, and a fleet can
opt in today with HF_SEGMENTED_MIN_SECONDS=600.
createCapturePlan no longer requires useStreamingEncode for a segmented
plan. That flag answers "one encoder for the whole render", which goes false
for multi-worker — requiring it would have silently dropped every
multi-worker segmented render onto the disk path.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Each worker owns a session (recycled and retried per 2c) and pulls whole
segments in index order. Segments finish out of order, so the concat list is
built from the plan rather than from completion order — ordering never
reaches the encoders.
Three guards, each mutation-checked: every worker settles before any session
is closed (closing one out from under a live worker orphans its ffmpeg and
races the CDP connection); a multi-worker render never takes the
single-encoder fallback, because other workers are already encoding work it
would discard; and the queue is shared, so no frame is captured twice.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
The default deps object captured the engine's closeCaptureSession at module
load, so a test that swaps the engine module after this file is evaluated
never saw the close. The Phase 2a test that checks the probe is closed after
a first-encoder spawn failure failed from this layer upward for that reason.
Calling through the import binding restores the check; production behaviour
is unchanged.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
Recycling the browser on every segment produced a byte-identical file to
the single-session render, which is what makes the target-loss retry safe:
a segment re-captured on a fresh session is indistinguishable from one that
never failed. Five launches cost ~5 s, and each recycle reported a renderer
RSS peak of ~1.4 GB being returned.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
HF_SEGMENT_BROWSER_RECYCLE defaults to 3 segments, which keeps a session
under the 10k-frame band where field crash rates pass 4%. segmentRetries
lands on live capture observability, so a render that retried still reports
it if it later dies.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
A segment boundary is the only safe point to hand capture a fresh browser,
which is what makes the long-render failure class survivable: the session
restarts every N segments, and a Chrome target loss mid-segment re-captures
that segment once on a new session instead of failing the render.
Retry is deliberately narrow. Encoder and authoring errors reproduce on a
fresh browser, so retrying them doubles the time to the same failure — all
three guards (retry exists, retry is classified, cadence boundary) are
mutation-checked. The partial segment file is deleted before the retry so a
later resume cannot see it, and onSegmentComplete still fires only after a
successful close, so the manifest never records a partial segment.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Only the Chrome target-loss class is worth retrying a segment for: encoder
and authoring failures reproduce on a fresh browser, so a retry just doubles
the time to the same failure. Both mutation-checked, including that a
non-Error is never classified as retryable.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Review findings on the Phase 2b PR. Three inputs that change segment bytes
were missing from computeSegmentPlanHash, so --resume could splice
mismatched segments: the hardware encoder (h264_nvenc and libx264 both
report codec "h264"), the device-scaled capture size (the hash used the
composition's CSS size), and the motion-blur options (sample count changes
every averaged frame). All three are hashed now, each with a test that
fails if it is dropped. The hash and the encoder read the image format
from one binding.
The plan -> options -> request -> config chain that once dropped --resume
and --keep-segments had no test. renderOptionsFromPlan is now a pure,
exported hop with a test, renderConfigFromRequest has one for the same
flags, and the batch path, which was still dropping both flags, carries
them too.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
execute.ts is where the plan becomes RenderOptions, and it dropped both new
flags, so --keep-segments deleted the segment directory anyway and --resume
never resumed. Nothing failed: the render just silently re-captured
everything. Caught by the Phase 2b gate, which is the only check that spans
plan -> options -> request -> config; the gate doc now says so.
Gate result recorded: killed at 2 of 6 segments, resume skipped both,
finished in 3m32 against 5m13, and the output was byte-identical to the
uninterrupted render.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>