* fix(producer): pass variables to duration probe
* fix(producer): tolerate rounded frame-boundary durations
* fix(cli): resolve relative data-start references in composition duration
`compositions --json` computed each timed child's start with a bare
parseFloat(data-start ?? "0") in parseCompositions (host duration) and
parseSubComposition (sub-comp duration). A relative reference like
data-start="s1" ("start when clip s1 ends") is not numeric, so parseFloat
returned NaN and that clip's contribution to the max-end was silently
dropped — a host with two 3s clips (2nd data-start="s1") reported duration 3
instead of 6, breaking compositions/inspect/snapshot for composition-clip
relative timing.
Resolve relative references the same way the extractor does (parseStartExpression
from @hyperframes/core + a findReferenceTargetEl/resolveReferencedStart port,
since the engine's referenceResolver isn't a public export across the package
boundary). Verified: host duration now 6; 3 tests pass.
(Implemented via Codex; verified independently.)
* feat(studio): mirror canvas z-order actions into timeline lanes, badge z overrides
Track order = default paint order; authored z = advanced override.
- timelineZMirror.ts: pure resolver mapping a successful z-menu action to a
timeline lane move — closest track in the action's direction that is free
over the clip's whole span, else a new lane adjacent to the crossed
neighbor; temporal-overlap scope (default pending product sign-off, see
module doc); visual zone only; same-file reference scoping; persistTrack
via the shared authored-space rules. null for non-clips (menu stays
z-only) and at-extreme/no-overlap cases.
- useCanvasZOrderTimelineMirror.ts: after the z commit resolves, the mirror
persists the lane move through the same machinery as a timeline lane drag
(optimistic store update, authoredTrack refresh, rollback); inserts reuse
commitTrackInsert's renumber via a shared buildTrackInsertEdits core. Both
writes share one coalesce key (zReorderCoalesceKey) and fold into ONE undo
entry (test proves it over the real history reducer). The mirror never
triggers the lane->z stacking sync, so it cannot fight the z values the
action just set.
- timelineZOverride.ts + TimelineClip badge: clips whose paint order
contradicts lane order among temporally-overlapping same-context visual
neighbors (laneIsAbove XOR paintsAbove, the stacking-sync predicates) show
a 'z' badge — authored z overrides are surfaced instead of silently
disagreeing with the timeline.
- Timeline.tsx track derivations extracted to useTimelineTrackDerivations
(600-line cap).
* fix(studio): fold mirrored z-order gestures into one undo entry across slow persists
Live verification caught the z write and the mirrored lane write splitting
into two undo entries: the mirror runs after the z persist's server round
trip, which exceeds editHistory's default 300ms coalesce window under real
latency (the unit test's deterministic clock sat inside it).
zReorderCoalesceKey now mints a per-gesture-unique key (monotonic seq, the
laneChangeGestureSeq precedent) and both records carry coalesceMs Infinity —
distinct gestures can never merge, and one gesture always folds regardless
of write latency. coalesceMs threaded through the persist chain alongside
coalesceKey. Also hardens the existing lane-drag move->z fold, which had the
same latent split. Fold test now simulates a 400ms gap (failed before the
fix, passes after); a two-separate-gestures test asserts two entries.
* feat(studio): flashless lane mirror, z-order menu icons, close-gap track menu
- Track-only batch moves (the z-mirror's lane hop and the insert renumber)
skip the GSAP fallback round-trip and the preview reload entirely — the
renderer never reads data-track-index, and the live DOM patch + optimistic
store update cover the UI. Mixed batches keep current behavior. Kills the
canvas blink on mirrored Bring/Send actions (live-verified: an
iframe-scoped marker survives the whole gesture).
- The four z-order menu items get 16px stroke icons (single layer diamond +
directional arrow for Forward/Backward; pierced two-layer stack for
Front/Back); labels unchanged — they are the industry-standard names.
- New track context menu on empty lane space: 'Close gap' (shifts the next
clip and every clip after it on that lane left by the clicked gap's width;
leading gaps count, so a single clip with empty space before it compacts
to 0) and 'Close all gaps' (whole lane contiguous from 0). Pure gap math
in timelineGaps.ts; persists through the drag path's atomic batch move
(one undo per action); refuses when a clip that must shift is locked;
items disable when there is nothing to close.
* fix(studio): rebind-only preview sync for unmutated timing edits, classical z-menu order
Timing edits that rewrote NO GSAP positions (gap closes and moves of
selector-addressed caption clips, zero-delta batches, comps without a
rewritable script) full-reloaded the preview — and the rerun-current-scripts
attempt was wrong for real compositions: re-executing init-style scripts
(three.js scenes, caption engines) is exactly the unsafe case, verified live
by doubled init warnings and a fallback reload anyway.
The correct observation: when mutated === false the existing __timelines are
still valid — only the runtime's clip visibility windows are stale, and the
live DOM timing attributes were already patched. So the no-mutation path now
runs applySoftReloadFinalization only (seek + __hfForceTimelineRebind +
manual-edits reapply), extracted from the soft-reload machinery — zero
script execution. This also un-blinks comps with no GSAP script at all,
which previously always remounted. Rewritten-script soft reloads,
cannot-soft-reload, otherFileChanged, and mutation failures keep their
existing behavior. gsapSoftReload's undo/redo restore section moved verbatim
to gsapUndoRestore.ts for the 600-line cap.
Also: z-order menu items reordered to the classical arrangement (Bring to
Front, Bring Forward, Send Backward, Send to Back).
Live-verified on a three.js-heavy composition: Close-all-gaps shifted 4
caption clips with correct cumulative amounts, the preview iframe was never
remounted (marker survived), and one undo reverted everything.
* fix(studio): bound forward/backward mirror to a one-element step
User-specified semantic: Bring Forward / Send Backward move the clip past
EXACTLY ONE element. The mirror's lane target is now bounded by the next
temporally-overlapping element beyond the crossed neighbor: a free lane
strictly between the two is taken (closest to the neighbor), and when they
are back-to-back a new track is inserted immediately beyond the crossed
element — never past the second one. Previously the resolver took the
closest free lane anywhere beyond the neighbor, which could carry the track
past a second element while the z action only stepped past one — a
track/paint contradiction our own zOverride badge would flag. Front/back
keep whole-set semantics (past everything; back stays above the audio
zone). End-to-end test pins the 3-stacked case through commitZMirrorLaneMove
to the persisted renumbered tracks.
* feat(studio): permanent gap-menu rows with hover and click-select gap highlights
- TrackGapContextMenu always renders both rows; an inapplicable action dims
with a tooltip ("No gap here" / lock reason / "No gaps on this track")
instead of vanishing into a one-item menu. Width badge only when a gap
exists under the pointer.
- Hovering an ACTIONABLE row highlights the strip(s) it would close in the
timeline: the single gap for Close gap, every current gap (leading included)
for Close all gaps. New resolveAllGapIntervals in timelineGaps.ts reports
present-state intervals (epsilon-tolerant, overlap-safe), distinct from
resolveAllTrackGaps' post-compaction starts.
- Click-selecting a single clip paints a quieter tint over its lane's gaps
(suppressed for marquee multi-selection and during drags; the gap-menu hover
wins on its own lane). Derivation lives in useTimelineGapHighlights with the
pure buildTimelineGapStrips exported and unit-tested.
- Strips render in TimelineCanvas with the drop-placeholder geometry (row top
+ clip inset), dashed accent for hover, faint tint for selection.
- Timeline.tsx stayed under the 600-line cap by extracting the scroll-viewport
plumbing (ResizeObserver width + shortcut-hint sync) into
useTimelineScrollViewport, behavior unchanged.
* feat(studio): stronger capcut-style timeline zoom steps
One button press / pinch gesture now moves the zoom meaningfully: step
factors 1.25x/0.8x -> 1.5x/(2/3) (kept reciprocal so in+out round-trips) and
pinch sensitivity 0.0035 -> 0.007. Addresses "zooming several times to get
anywhere" feedback; cursor anchoring unchanged.
* feat(studio): three-way z sync — layers drags mirror timeline lanes, panel tracks live z edits
Completes the layers/canvas/timeline sync triangle: the Layers panel was the
one surface whose reorders never reached the timeline, and the one that went
stale when the other two wrote z flashlessly.
- Layers drag -> minimal z + equal-jump lane mirror. handleReorder now uses
the canvas menu's realization core via resolveZOrderReposition (one
between-z write when a strict gap exists, band-safe scoped renumber
otherwise) instead of computeReorderZValues' all-sibling stamp — that
helper is deleted, completing the #2347 unification follow-up. The drop
then mirrors into a timeline lane move through the same machinery as the
canvas menu (new resolveRepositionLaneMove: the clip lands on a free lane
strictly between its NEW paint neighbors' lanes — nearest clip siblings in
the desired render order, decorations skipped — else a track insert at
that boundary; audio zone never crossed). Both writes share one
per-gesture zReorderCoalesceKey with an unbounded fold window, so a drag
is exactly ONE undo entry; useCanvasZOrderTimelineMirror's plumbing is
factored into useMirrorLaneMoveCommit and reused by the new
useLayerReorderTimelineMirror. A same-slot drop is a hard no-op (new
order-equality guard in resolveZOrderReposition).
- Panel staleness fix: flashless z commits (skipReload) reload nothing and
bump no refreshKey, so the panel's z-sorted order went stale while paused.
handleDomZIndexReorderCommit now bumps a store zEditVersion on apply AND
rollback; the panel re-collects on it. Verified live: the panel re-sorts
the instant a drag commits and again on undo.
- Layer click reveal (useLayerRevealOverride): clicking a layer that stays
hidden at the current frame (animation-parked opacity, non-clip
display/visibility hides, hidden ancestors) temporarily forces the chain
visible with live inline styles — exact priors restored on deselect, on
another reveal, on play, and on unmount; never persisted (file diff == 0
verified live). Clips keep the existing seek-into-window behavior; the
override applies on a short defer so a seek-revealed clip needs none.
- layerOrdering's unused hasExplicitZIndex probe (zero callers) removed.
Live-verified on a bed copy: a 2-position layers drag wrote exactly one
element (z 6->23 + data-track-index 15->2), the timeline lane moved without
a reload, and a single Cmd+Z restored the file byte-identically.
* feat(studio): full-track selection highlight, borderless gap hover strips
- Click-selecting a clip now lights the WHOLE lane minus its clips — leading
gap, inter-clip gaps, and the open space after the last clip to the rendered
end (new resolveLaneEmptyIntervals; displayDuration threaded into the strip
derivation). Still click-only: any drag/resize suppresses the strips, and a
marquee multi-select never shows them.
- The gap-menu hover strips drop the dashed border (user feedback) — fill only,
nudged to 0.18 alpha to keep the same visual weight.
* feat(studio): selected layer paints on top via a reader-transparent z lift
Clicking a layer in the Layers tab now shows the element as if it were at the
very top of the stack while selected — whatever its authored z or panel
position — extending the reveal override (which already forced hidden chains
visible) with a temporary inline z lift:
- liftElementToTop parks the TRUE effective z in data-hf-reveal-prior-z and
writes a far-top inline z; a static element gets a layout-preserving
position:relative with its prior parked in data-hf-reveal-prior-pos. Only
the RENDERER sees the lift: all three studio z readers
(readTimelineElementZIndex, getElementZIndex, readEffectiveZIndex) return
the parked prior while the attribute is present, so the canvas z-menu, the
zOverride badge, the lane mirror, the stacking sync, and the panel sort
keep reasoning on the element's real z.
- Strictly ephemeral: exact priors restored on deselect / another reveal /
play / unmount, each property only while it still holds the value the
override wrote (a later real edit is never clobbered). File diff == 0
verified live across a full lift/restore cycle.
- A z-reorder commit CONSUMES an active lift (handleDomZIndexReorderCommit
reads the parked position for its persist-position:relative static check,
then drops the attributes) — the committed z becomes the truth and the
later restore is a guarded no-op.
* fix(studio): flashless undo/redo — three full-reload causes in the soft-restore path
Cmd+Z blinked the canvas on essentially every undo. Three independent causes
in applyUndoRestoreToPreview, each sufficient on its own:
1. Master-view path gate: activeCompPath is NULL at the master view, so the
'paths[0] === activeCompPath' eligibility check could never match the
index.html restore and every default-view undo full-reloaded at the first
gate. Normalized to the codebase-wide 'activeCompPath ?? "index.html"'.
2. Nested identity innerHTML check: the diff compared each identified
element's innerHTML, but the composition root wraps every clip — any child
change re-detected at the root rejected the restore. Change detection now
compares only each element's OWN attribute surface; structure/text
integrity is still guaranteed by the normalize-residual whole-doc pass
(text nodes, added/removed elements, and un-identified attrs all remain
after normalization and force the full reload).
3. id-only identity: elements addressed by data-hf-id / selector (no DOM id)
fell outside the diff entirely. Identity is now id OR data-hf-id, with the
live sync resolving either.
Also stop re-running an UNCHANGED GSAP script: attribute-only restores (z,
lane, timing, style — the overwhelmingly common undo) now use the rebind-only
finalization (seek + __hfForceTimelineRebind + manual reapply, zero script
execution — the same path as flashless timing edits), instead of tearing down
and rebuilding live timelines or full-reloading when the script can't be
scoped. A restore whose script text genuinely changed still re-runs it via
applySoftReload, and structural restores (split/delete) still full-reload.
Live-verified on the bed (iframe marker): gap-close undo AND redo both keep
the iframe mounted, live DOM lands on the restored values, disk restored
byte-identically.
* feat(studio): left breathing pad before t=0, double zoom sensitivity again
TRACKS_LEFT_PAD (48px) — the horizontal sibling of TRACKS_TOP_PAD: empty lane
surface between the sticky gutter and the ruler's 00:00 / the first clips,
scrolling WITH the content.
- The lanes and the ruler realize it as a plain flow spacer between the
sticky gutter cell and the time-mapped content div, so every
content-relative computation (clip left = t*pps, beat lines, lane-menu
time, clip drag deltas) is untouched by construction.
- Canvas-space overlays shift by the pad: playhead (getTimelinePlayheadLeft),
gap strips, drop placeholder, snap guide, range highlight, marquee clip
rects, beat SVG; the insert line spans the pad.
- Every pointer->time inverse subtracts it symmetrically: seekFromX, razor,
range/marquee anchors, asset drops, and the zoom-anchor gutter basis; fit
pps and the display width account for the consumed viewport width.
- Live-verified: t=0 clip edge, the 00:00 tick, and the playhead line center
all sit at GUTTER + TRACKS_LEFT_PAD, and a ruler click lands the playhead
center exactly under the pointer.
Also doubles the timeline zoom sensitivity again (user feedback after
feel-testing the first bump): button steps 1.5x/(2/3) -> 2x/0.5, pinch
0.007 -> 0.014.
* fix(studio): left pad renders as true empty space, not lane surface
The pad before t=0 inherited each row's background and bottom border from the
row wrapper, so it read as track lanes. Lane visuals now live on the cells:
the sticky gutter keeps its own separator (header column stays delineated),
the time-mapped content div carries the row background + separator, and the
pad spacer stays transparent — bare shell background, no lines. The
new-track insertion line also starts at the pad's end instead of crossing it.
* fix(studio): no vertical line in the ruler band before 00:00
The ruler corner's right border drew the header-boundary line through the
ruler strip, so the band didn't read as starting at 00:00. Dropped it — the
boundary line belongs to the track rows below; the ruler stays completely
clean from the panel edge to the first tick, matching the empty left pad.
* refactor(studio): remove the timeline z-override badge
User decision: the "z" chip on clips never earned its place — dropped
entirely (timelineZOverride.ts + test deleted, TimelineClip badge rendering
and the zOverrideKeys derivation/threading removed). This also eliminates the
review's D2 finding at the root: the badge's cross-document comparison
(stackingContextId ?? null collides across source files in the expanded view)
produced false positives, and there is no longer a detector to mis-fire.
overlapsInTime/paintsAbove lose their export (the badge was their only
external consumer); the paint-order predicate itself is unchanged.
* fix(studio): collision-free expanded child lanes and host-window gap floors
Review findings D1 (blocker) and 4.
- D1: buildChildElements assigned expanded children synthetic display rows as
`host.track + index` — integers that can EQUAL a real clip's lane in another
file (host on 0 with two children puts child #2 on 1). Lane grouping merges
purely by track number, so the collision fused clips from different source
files into one display lane, and lane-scoped actions (the gap menu) then
batch-persisted a foreign file's clip. Children now take FRACTIONS strictly
between the host's lane and the next integer — structurally unable to
collide with any normalized lane, while still rendering as ordered rows
under the host. Regression test pins the reviewer's exact two-file scenario.
- Finding 4: gap math compacted toward absolute 0, but an expanded child's
display time is host-anchored — close/compact could drag it before its host
window and persist a wrong (even negative) local time. All gap functions
now take a lane FLOOR (laneGapFloor: 0 for ordinary lanes, the children's
expandedParentStart for child lanes — single-origin per lane post-D1),
threaded through the menu model, hover highlights, selected-lane strips,
and both commits. Close-gap shifts clamp at the gap's own left edge.
* fix(studio): scope mirror references, insert writes, and crossed-neighbor identity
Review findings 1, 2, and 3.
- Finding 1: buildTrackInsertEdits normalized the FULL display set and
persisted every shifted clip — writing host-lane numbers into OTHER
composition files when expanded children were showing. The renumber write
set is now the edited element's own source file (the sanctioned multi-write
converges one FILE to lane space, never neighbors' files); foreign clips
keep their authored tracks and re-derive display lanes. The locked-clip
refusal scopes the same way. Expanded-origin elements refuse the insert
outright (a new lane is a host-space renumber, meaningless in the child's
file), and the mirrors restrict an expanded child's lane candidates to its
own siblings' lanes — a sub-comp child still mirrors WITHIN its sub-comp
(persisting the sibling's authored track) but can never land on a host lane
with no same-file occupant. authoredTrackForLane's offset fallback rounds:
fractional synthetic rows must never leak fractions into data-track-index.
- Finding 2: the mirror comparison sets required only sameSourceFile, but a
file can contain several CSS stacking contexts and leaf z is only
comparable within one. Both resolvers now scope by samePaintScope — same
source file AND same stackingContextId (the file check also stops null root
contexts of different files from comparing equal in the expanded view).
- Finding 3: the crossed-neighbor key was derived without selectorIndex, so
duplicate class selectors (.sub) resolved to occurrence 0 — a different
clip. The key now carries getSelectorIndex, matching how z-reorder entries
derive theirs.
* fix(studio): z-to-lane gestures are one serialized transaction gated on durable persists
Review findings 5 and 7.
- Finding 5: commitDomEditPatchBatches resolved successfully even when the
server matched NO patch target — the z write never reached disk (the
preview reloads to reconverge) yet the lane mirror still ran, desyncing
track order from what actually paints. The commit now resolves a durability
report ({allMatched, changed}; the save queue and commit types are generic
over the result), and the mirror phase is skipped on allMatched === false.
- Finding 7: the z persist rides the DOM-edit save queue while the lane move
rides the timeline/SDK path — two queues, so a second rapid gesture's z
write could land BETWEEN the first gesture's z and lane phases. Every
z-to-lane gesture (canvas z-order menu AND Layers-panel drag) now runs
through runZLaneGesture: a single module-level tail that serializes the
COMPLETE two-phase transaction, with unit tests for ordering, the
durability gate, and queue resilience to failed gestures. The timeline
lane-drag's inverse (move-then-z-sync) shares its phases' await ordering
already; cross-gesture serialization for that path is noted as follow-up.
- LayersPanel's pure sort helpers moved to layersPanelSort.ts (600-line cap).
* fix(studio): multi-clip GSAP batch mutations roll back on late failure
Review finding 6. finishGroupTimingGsapFallback mutates files sequentially
per clip; a late per-clip failure left the earlier rewrites on disk with no
aggregate history entry — unreachable by undo. foldGsapMutationIntoHistory
already snapshots every touched path before mutating; on a mutation failure
it now restores each path whose disk content changed (all-or-nothing batch),
reports restore errors without masking the original failure, and rethrows.
Regression test drives a two-clip batch whose second rewrite fails and
asserts the first clip's write is restored byte-identically.
* fix(studio): scope mirror inserts to their lane zone
* fix(studio): unify source-scoped clip identity
* fix(studio): isolate track insert topology
* fix(studio): harden timeline paint synchronization
---------
Co-authored-by: Miguel Angel Simon Sierra <miguel.sierra@heygen.com>
Add a `core-skills` entry to .claude-plugin/marketplace.json declaring
the core skill set (the /hyperframes router, the hyperframes-* domain
skills, and media-use). The upstream vercel-labs/skills CLI reads that
entry's `skills` array for its interactive picker: the core set renders
under a "Core Skills" group and everything else falls into "Other", so
a human running `npx skills add heygen-com/hyperframes --full-depth`
can tell the always-needed core set apart from the on-demand creation
workflows — mirroring the core/on-demand tiers `hyperframes skills
update` already enforces (isCoreSkill in skillsManifest.ts).
The array deliberately lives on a separate marketplace entry, NOT on
plugin.json or the `hyperframes` entry: Claude Code treats a manifest
`skills` array as that plugin's skill allowlist (verified against
claude CLI), so attaching it to the full plugin would narrow it from
all skills to the core 8. As a side effect the new entry is itself a
coherent Claude Code plugin — `core-skills@hyperframes` installs just
the core set — while `hyperframes@hyperframes` keeps auto-discovering
everything.
A new pin test keeps the marketplace list in lockstep with isCoreSkill
and the skills/ tree (alongside the existing FALLBACK_CORE_SKILLS pin),
and asserts the full plugin carries no allowlist. Agent installs are
unaffected — the upstream CLI detects agent environments and installs
non-interactively, and the hyperframes CLI always passes explicit
--skill flags.
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
Miguel R4 blocker on #2359: my R3 fix at renderOrchestrator only updated
the observability copy, leaving the authoritative captureForceScreenshot
local at compileResult.forceScreenshot (false for auto→software). The
frameCapture side clamped its own local and correctly routed screenshot,
but downstream orchestrator code overwrote observability back to
beginframe from the still-false local at two sites:
- Parallel-stream label at renderOrchestrator.ts:2293 mis-labelled the
stream as 'beginframe' when actual capture was 'screenshot'.
- capture_strategy telemetry at renderOrchestrator.ts:2440-2450
overwrote the earlier observability correction, so the final
captureMode observation flipped back to 'beginframe' while the
engine actually captured screenshot.
Fix: extract the clamp into a caller-facing helper
applyConcreteGpuScreenshotClamp(current, resolvedGpuMode, cfg) that
returns the (possibly-promoted) new boolean. Callers assign it back to
their authoritative local, so routing + telemetry + strategy code read
one value.
Changes:
- packages/engine/src/config.ts: new exported
applyConcreteGpuScreenshotClamp; delegates to
shouldClampToScreenshotForConcreteGpu but computes the caller's
final value, not just the clamp decision. Reads the programmatic
opt-out from cfg.forceScreenshotExplicitlyOptedOut. Idempotent on
already-true input.
- packages/engine/src/index.ts: export the new helper.
- packages/engine/src/services/frameCapture.ts: replace the inline
OR expression with applyConcreteGpuScreenshotClamp.
- packages/producer/src/services/renderOrchestrator.ts: assign result
into the AUTHORITATIVE captureForceScreenshot local (was updating
only observability). Downstream parallel-stream label at :2293 and
capture_strategy telemetry at :2440-2450 now read the corrected
value.
Tests: 6 new caller-level cases for applyConcreteGpuScreenshotClamp
covering the exact matrix Miguel called out:
- resolved software + default false → promotes to true (screenshot)
- resolved software + programmatic opt-out → stays false (BeginFrame)
- resolved hardware + default false → stays false
- resolved software + already-true → stays true (idempotent)
- resolved software + env PRODUCER_FORCE_SCREENSHOT=false → stays false
- resolved software + undefined cfg → promotes to true (frameCapture path)
Local: 67/67 engine config tests pass (was 61). oxfmt clean.
Miguel R3 blocker on #2359: the runtime helper only checked the env opt-out
(PRODUCER_FORCE_SCREENSHOT=false), silently defeating the documented
programmatic escape hatch (overrides.forceScreenshot === false) on the
browserGpuMode:'auto' → software probe path. At the concrete-resolution
site the boolean forceScreenshot === false is ambiguous between default
and explicit opt-out — resolveConfig sees the provenance but the runtime
helper does not.
Fix: persist provenance on the resolved config.
- New INTERNAL EngineConfig field forceScreenshotExplicitlyOptedOut, set
by resolveConfig when EITHER env or programmatic explicit-false is
present. Purpose-documented in the type as 'not intended to be set by
callers'.
- shouldClampToScreenshotForConcreteGpu gains an opts.programmaticOptOut
parameter; returns false early when set. Env stays as the third arg
(backward compatibility with existing tests).
- frameCapture.ts and renderOrchestrator.ts pass
config.forceScreenshotExplicitlyOptedOut through at both call sites, so
the auto→software probe path preserves the same escape hatches as
literal browserGpuMode:'software'.
New tests: 5 additional cases across the helper (programmatic opt-out
alone; programmatic beats missing env) and resolveConfig provenance
(programmatic sets flag; env sets flag; neither leaves it undefined).
Local: 61/61 engine config tests pass (was 56).
* feat(cli): stable-URL re-publish (send owned id, honest UX, --update, team id file)
Resolve a stable project id (committed team id > machine store > mint), send it
when authenticated so an owned re-publish updates the same URL in place, and
persist the server id+url. Report updated-vs-created honestly, add --update to
target a project explicitly, write a committable .hyperframes/project.json so a
team shares one link, and show the prior URL on re-publish.
* test(cli): env-gated E2E round-trip for stable-URL re-publish
Publish -> edit -> re-publish asserts one URL with updated content against a
live EF (HYPERFRAMES_E2E_API_URL + an authenticated runner); skipped otherwise.
* fix(cli): real team space, auth-gate --update/--space, safe team-file write
- --space + committed .hyperframes/project.json (projectId+spaceId) send X-Space-Id so a
team converges on one link; personal space stays the default for solo users
- --update/--space error when unauthenticated, and warn loudly when a resolved-but-invalid
token silently downgrades to a new anonymous URL (no more generic-tip-only)
- team-file write in its own try/catch so a read-only dir can't fake 'Publish failed'
- parseUpdateTarget handles scheme-less URLs + query/hash; X-Space-Id on metadata only (not S3 PUT)
- share readJsonRecord across the local + team descriptors
* test(cli): e2e team-space convergence + cross-space hijack guard
* test(cli): unit-test parseUpdateTarget url shapes (export for test)
* fix(cli): warn on committed-team miss too, not just --update
The SDR-to-HDR extraction tests synthesize their HDR fixture with
-color_trc/-color_primaries flags and rely on the encoder propagating
them into the bitstream. The pinned Windows CI ffmpeg build drops the
transfer on that path, so after #2377 narrowed HDR detection to the
transfer function, the fixture probes as SDR on Windows and both tests
fail (hdrPreflightCount 0). Write the VUI directly with the
h264_metadata bitstream filter so the tag survives on every build.
Sub-compositions are required to wrap markup, script, and style in <template>, but template content is an inert DocumentFragment that document-level querySelectorAll does not traverse — so 'hyperframes keyframes' silently surfaced zero tweens and zero CSS keyframes for every spec-conformant sub-composition. Extract from the document and every template content fragment (nested templates included, walked iteratively). Documents the extraction ordering contract and the deliberately document-root-only sub-composition discovery scan; pins template, multi-template, nested-template, and mixed-script cases with tests.
Review 1 (restore commit):
- asset reveal now clears any open preview overlay (stuck-overlay repro:
preview on A, click already-added B — A stayed open over the reveal)
- duration readout rolls back on failed persist: captureDurationRollback
snapshots store + live root data-duration before the optimistic sync and
restores both in every move/resize/delete/group catch (golden's
previousDuration pattern)
- asset preview opened during running playback dismisses immediately (the
RAF loop bypasses the store, so the subscription alone never fired)
- persistTimelineBatchEdit resolves the target (findTagByTarget) before
treating identical output as a no-op — a mistargeted member now throws
like the single-element path instead of being silently dropped
- a post-mutation history-fold failure no longer suppresses the preview
sync: fold errors are surfaced separately and the rewritten script still
syncs (previously the preview kept stale GSAP positions with no recovery)
- timelineRevealScroll guards degenerate viewports (windowSize <= 0)
- CodeQL: encodeURIComponent(projectId) at all timelineTimingSync fetches
Review 2 (single-source-of-truth pass):
- createTimelineElementFromManifestClip — the one manifest->element
boundary — now carries authoredTrack and stackingContextId; expanded
sub-comp children preserve both (authoredTrack in their OWN file's space)
- authoredTrackForLane scopes occupants to the dragged clip's sourceFile
(a foreign file's authored values are a different coordinate space);
nearest-same-file-lane offset fallback
- optimistic store updates mirror the persisted track into authoredTrack
(and roll it back on failure), so consecutive drags before a reload
resolve from fresh data
- spill sub-lanes: documented decision — dropping onto a spill lane is a
legitimate same-track join (occupants share the authored track by
construction); false 'never a lane-move target' docstring rewritten
- single-element fallback persists vertical-only moves (early return now
requires neither start nor track changed; live DOM patch includes
data-track-index)
- canonical contextKey helper for stacking-context normalization
- new pipeline test crosses the REAL factory boundary (sparse authored
tracks -> factory -> expansion -> normalize -> drag commit -> persisted
attribute), no injected fields
Two legibility fixes for the canvas z-order menu, from user feel-testing:
- z-only commits no longer remount the preview iframe. The commit hook
already applies the inline z (+ injected position) to the live elements and
updates the store synchronously; the post-commit reloadPreview() was a
redundant full remount that read as a canvas 'blink' on every action.
commitDomEditPatchBatches gains skipReload, engaged only when provably
safe: every op is an inline-style patch AND the server reports every patch
matched — anything else falls back to the reload so the preview reconverges
with disk. The file-watcher's own reload stays suppressed by the existing
domEditSaveTimestampRef window, so the skip is real.
- Bring Forward / Send Backward step over the next VISIBLY overlapping
sibling. The nearest z-neighbor in a composition is often invisible at the
current frame (runtime hides time-inactive clips with inline
visibility/display; GSAP parks elements at opacity 0), so the step crossed
something the user couldn't see — 'enabled but nothing happens'. The
forward/backward set now filters on element-level computed visibility
(display/visibility/opacity, injectable for tests); enable/disable shares
the resolver so the menu is honest: actions disable when no visible
neighbor exists. Front/back keep the full painting family.
- The neighbor that was stepped over gets a 600ms accent flash, drawn in the
studio overlay layer (never in the iframe DOM), so the action shows its
work.
An adversarial review of the canvas context-menu z-order pipeline (Bring to
Front / Forward / Backward / Send to Back) found the resolver math sound but
the glue between the menu and the commit hook broken:
- The menu optimistically wrote style.zIndex AND position: relative to the
live elements BEFORE the commit hook ran. The hook decides whether to
persist position by checking getComputedStyle(el).position === 'static' —
always false after the pre-apply — so the position patch was never
persisted on the menu path and the reorder silently reverted at the
post-commit reload for any nested/static element (root clips survive only
because the runtime forces position:absolute). The same pre-apply made the
failure rollback capture the already-mutated values, restoring the broken
state on persist errors. The menu no longer pre-applies; the hook owns the
live writes (it already applied both synchronously) and now sees true
priors. Siblings without a persistable identity still get their z applied
live-only so a renumber stays visually coherent.
- The commit hook's entry.key store-sync plumbing had zero production
callers; the store zIndex went stale until full reload. All three callers
(canvas menu via PreviewOverlays, timeline lane z-sync, LayersPanel) now
derive and pass the timeline store key (new deriveTimelineStoreKey helper).
- patchElementBatch discarded the server's per-patch matched[]; unresolvable
siblings persisted partially and silently. Unmatched targets now warn and
report save-failure telemetry (z-reorder-unmatched) without rolling back
the matched subset.
- template/noscript elements counted as painting siblings, so renumber
fallbacks wrote z-index/position into <template> tags in the source file.
Excluded from the sibling family.
- The default undo coalesce key merged DISTINCT z actions within 300ms into
one undo entry; the action kind is now part of the key (LayersPanel drags
keep coalescing within a drag; explicit lane-move gesture keys untouched).
- rectsIntersect comment claimed touching rects intersect; the strict
inequalities say otherwise — comment fixed.
Vertical clip moves committed in the store but never survived: two persist
bugs plus a runtime renumber all fought the stable-track-lanes model.
- timelineMoveAdapter deliberately stripped the track from lane-reorder
persists ('z-only reorder path' — the old z-driven lane model). Lane =
authored data-track-index now: lane-reorder and track-insert both persist
the track; plain timing moves omit it to stay SDK-fast-path eligible.
- Display lanes and file tracks are different coordinate spaces:
normalizeToZones packs sparse authored tracks (1,2,... or gaps, or DOM-index
fallbacks) onto contiguous display lanes, and lane edits persisted the LANE
number — silently re-targeting the wrong row in any non-0-contiguous file.
Elements now record their authoredTrack when remapped; a lane change
persists the target lane's authored track (store stays in lane space).
- The runtime split same-track clips of different kinds (video vs caption
div) onto separate renumbered tracks at discovery, so authored indices
never round-tripped ('drop onto an existing track' bounced back). Removed:
data-track-index is honored verbatim (render never reads it); kind-based
row presentation belongs in the display layer if ever wanted.
Adversarial review fixes on the same pipeline:
- runtime: parseInt(attr) || fallback dropped authored track 0 for GSAP and
overlay clips (parseAuthoredTrack helper honors 0)
- single-clip move fallback persisted only data-start — lane changes snapped
back on reload (now passes the track to the patch builder)
- lane-change z-sync candidate ignored a multi-selection's time shift, so
patches were computed against stale overlap sets
- track insert around a locked clip persisted a colliding renumber (the next
normalize merged lanes); the insert is now refused with a warning
- computeStackingPatches compared leaf z across CSS stacking contexts, where
ancestor z decides paint order; the sync now partitions by
stackingContextId and never patches across contexts
Timeline geometry (user-reported):
- fit zoom leaves 20% trailing headroom (FIT_ZOOM_HEADROOM in
timelineLayout.ts; single fit-pps source, so ruler/lanes/playhead/drag all
inherit it)
- playhead line center now sits exactly on GUTTER + t*pps at every zoom
(wrapper had shrink-wrapped to the 9px diamond, off-centering the line);
ruler ticks center on their timestamp
- ruler: frame-mode steps snap to whole frames (no duplicate labels), hour
steps added for far zoom-out, tick positions computed as exact multiples
(no float drift)
The Studio stack rebuild (#2291) landed the remaining NLE layers but dropped
or regressed several final-wave behaviors from the reviewed studio-dnd stack,
and never repaired the stale timelineZones.ts that #2279 introduced. Restores:
- TimelineRuler: sticky under vertical scroll, full-height gridlines removed
(beat lines only), frame-number tick labels via a persisted timeDisplayMode
store preference (PlayerControls toggle now store-backed)
- timelineZones: stable track lanes — lane = authored data-track-index
ascending; z is paint order only (replaces the stale z-driven lane pack,
which broke track insert-band commits that contractually depend on it)
- persistTimelineBatchEdit: a batch member whose patch is a no-op (attributes
already at target values, e.g. in a track-insert renumber) is skipped
instead of aborting and rolling back the whole batch — this alone made
new-track creation (incl. the top insert band) fail silently
- useTimelineStackingSync: unresolvable clips read as NaN again so
timelineStackingSync's Number.isFinite exclusion contract holds (z=0
fabrications skewed stacking boundaries)
- timelineAssetDrop: drops land on the drop track (no overlap bump to
max-track+1), data-hf-id stamped, audio gets data-volume
- timing edits: soft-reload the server's rewritten GSAP script instead of a
full iframe remount (no all-clips flash on move/resize); full reload only
when no scriptText or the soft path can't apply, and one full reload when a
group edit touches non-active files (new hooks/timelineTimingSync.ts)
- duration: content-driven grow-AND-shrink on move/resize/delete, synced
optimistically to the store and the live root data-duration at release
(was a grow-only ratchet; shrink never updated the readout)
New UX: sidebar asset click opens a compact non-modal preview over the canvas
(dismiss on outside click, Escape, playback, or seek), and clicking an
already-added asset reveals its clip in the timeline (smooth minimal scroll
to its time and lane; vertical-only in fit zoom).
Verified by pointer-driving a real project: sticky ruler + gridline removal,
no iframe remount on move/resize (marker survives, GSAP tween positions
rewritten in place), duration readout 40->37->40 on shrink/stretch, and
top-insert-band track creation renumbering lanes correctly on disk.
Addresses Miguel's R1 blockers:
1. `browserGpuMode: "auto"` that runtime-probes to software slipped past the
`resolveConfig` clamp — that clamp only sees the pre-resolve string. Add
`shouldClampToScreenshotForConcreteGpu(resolvedGpuMode, currentForceScreenshot, env)`
in `packages/engine/src/config.ts` and apply it at BOTH concrete-resolution
sites:
- `packages/engine/src/services/frameCapture.ts`: downgrades `preMode`
from "beginframe" to "screenshot" when resolved GPU is software (respects
`PRODUCER_FORCE_SCREENSHOT=false` env opt-out), fixing the routing.
- `packages/producer/src/services/renderOrchestrator.ts`: updates
`captureObservability.forceScreenshot` (and thus `captureMode`) at the
same call site, fixing the observability truth on the auto → software
case.
2. New unit tests in `config.test.ts`:
- Documents the auto-branch gap (resolveConfig leaves auto as
forceScreenshot=false — the runtime companion closes it).
- 5 branch tests on `shouldClampToScreenshotForConcreteGpu` covering
software / hardware / already-forced / env-opt-out / non-"false" env
values.
Full suite: 56/56 pass.
Scope narrowing on Blocker 2: the distributed rendering path at
`packages/producer/src/services/distributed/plan.ts:753-754` and
`renderChunk.ts:462-466` explicitly hardcodes `browserGpuMode:"software",
forceScreenshot:false` post-resolveConfig and stays outside this PR's
invariant boundary. `compileStage` may still flip it to true for alpha
formats, but generic MP4 distributed renders on SwiftShader hosts remain
BeginFrame. That's a separate architectural cleanup (needs its own
behavior-change trace); the PR body now scopes the invariant to the
in-process CLI/orchestrator path.
Send `X-HeyGen-Client-Source: hyperframes` from buildAuthHeaders on every
HeyGen API call (both OAuth and API-key), so backend billing meta can
isolate hyperframes CLI usage. The single buildAuthHeaders chokepoint covers
the core CLI + cloud client. Mirrors the media-use tagging; the OAuth-only
X-HeyGen-Source cli free-gate header is unchanged.
* feat(cli): coordinate-frame layout findings in check
Four production compositions shipped with 100-600px layout drift, each a
different coordinate-frame confusion the check graded info or missed
entirely: viewport pixels written as container left/top, gsap x/y
treated as absolute position, a -350px margin fighting flex centering,
and stage-relative path coords drawn into a nested SVG.
Three new layout findings close the class:
- positioned_out_of_parent: an absolute/fixed element rendering mostly
outside its positioning ancestor (warning) — the parent needs no
overflow clipping, which is what let container_overflow miss it.
- box_out_of_canvas: a painted panel breaching the canvas (warning) —
text is canvas_overflow's, media is frame_out_of_frame's, painted
boxes were nobody's.
- connector_detached: a connector path whose endpoints land far from
every anchorable element (warning) — measured coordinates drawn into
an SVG with a different origin.
canvas_overflow additionally promotes from info to warning when held
across samples AND the breach exceeds 5% of the canvas.
All three are persistence-tiered and respect data-layout-allow-overflow.
Verified against the four incident compositions: every one now surfaces
its drift as held warnings (previously: info or silence).
* fix(cli): harden coordinate-frame findings against review false positives
Reworks all three findings after two-lens review (adversarial FP hunt in
real Chrome + maintainer pass):
- escaped_container (was positioned_out_of_parent): uses offsetParent
(transform-aware, skips fixed-as-canvas), exempts fully-detached
callouts within an attachment allowance while still flagging
touching-but-mostly-outside drift.
- panel_out_of_canvas (was box_out_of_canvas): paint alone qualifies
(flat solid panels were a false negative), fully off-canvas rects are
parked entrances and stay silent, pointer-events:none marks decorative
layers, hero-sized breaches warn while small bleeds stay info.
- connector_detached: endpoints via getPointAtLength + getScreenCTM
(viewBox, preserveAspectRatio, group transforms, every command type),
defs/marker/clipPath subtrees skipped, word-boundary connector naming,
containment tier limited to opaque non-ancestor targets (a text-bearing
wrapper contains its own diagram's endpoints).
- canvas_overflow promotion requires partial visibility — a fully
off-canvas rect is a parked entrance, not drift.
Verified: the four incident compositions still surface their drift as
held warnings; the review's false-positive repros (fixed HUD, callout,
parked entrance, corner bleed, marker arrowheads, g-transform and
viewBox-scaled connectors) are clean at warning level. Docs and the CLI
skill reference now describe the coordinate-frame findings.
* fix(cli): panel ownership is geometric — direct-text panels were a silent false negative
A painted panel whose direct text stays in-bounds while its box breaches
the canvas produced neither finding: canvas_overflow measures the text
range and panel_out_of_canvas skipped every own-text element. Skip the
panel finding only when the element's own text ALSO breaches (that
geometry belongs to canvas_overflow); pin the message/fixHint wording of
all three findings with positive assertions; document the SVG-internal
anchor blind spot.
* fix(cli): classify panel decoration by paint kind, not pointer-events
pointer-events:none exempted the framed-painting incident's gold frame
layers — hero content that happens to disable hit-testing. Decoration is
now gradient-only paint (spotlights, textures, vignettes); url() images,
solid fills and borders are content regardless of pointer-events.
* fix(cli): add fixHint to the test-local AuditIssue shape
* fix(cli): gradient stops decide content vs decoration; ownership matches canvas_overflow's tolerance
A gradient with any solid stop (alpha >= 0.6) is content — heroes and
cards painted with linear-gradient were invisible under the blanket
gradient exemption; all-translucent stops (spotlights, vignettes) stay
decoration. The text-ownership check now uses the audit tolerance that
canvas_overflow itself fires at, making the contract strict-mutex: any
text breach past that tolerance cedes the element, so a shallow 20px
text breach no longer double-reports.
Address Miguel's R1 blocker + Rames/Miga's testability nit:
- The `-frames:v 3` decode samples AT MOST 3 frames; a legitimate 1- or
2-frame WebM (256 or 512 bytes) was returned as `undefined` (probe
failure), silently skipping the advisory even when every available
pixel was opaque. Accept any positive whole-frame byte count ≤ 768
(multiples of 256), distinguishing successful short-EOF from partial/
malformed decode.
- Export `sampledAlphaIsFullyOpaque` and add 11 direct tests covering:
3/2/1-frame opaque decodes → true; transparent pixel at pos 0 or
final byte → false (guards the alpha-byte stride); non-frame-multiple
/ over-3-frame / zero byte counts → undefined; execFileSync throw →
undefined; findFFmpeg missing → undefined; and one args-shape guard
pinning the load-bearing `-c:v libvpx-vp9` before `-i` (without which
the default decoder silently discards VP9 alpha and the whole check
would false-positive on genuinely-transparent WebMs).
- Update advisory wording from "3 sampled decoded frames" to "every
sampled decoded pixel" so the message is honest for short WebMs.
18/18 tests pass locally under `vitest run`.
Review feedback on #2358: the batch-miss RENDER_FAILED in
runAssetImportMany (asset.ts) throws the same typed error as
client.ts's single-node renderNode, but wasn't labeled — so
cli_error.endpoint would silently come back undefined for the
flow that most heavily exercises /v1/images.
cli_error had no way to tell which figma REST call (images, files_nodes,
variables_local, styles, ...) actually hit RATE_LIMITED/FORBIDDEN/etc, so
the dashboard could see failures spike but not which call caused them.
FigmaClientError now carries a low-cardinality endpoint label (never the
raw fileKey/nodeId), threaded through to cli_error's endpoint property.
Extends webmAlphaCheck.ts (from #2044) with a pixel-level decode probe.
After the tag check passes, decodes 3 sampled frames via
`ffmpeg -c:v libvpx-vp9 -pix_fmt rgba -f rawvideo` at 8x8 and emits a
distinct advisory if every alpha byte reads 255.
#2044 detects the "tag absent" failure mode (ffprobe shows no
`alpha_mode` in stream tags). It doesn't catch a stricter case reported
on CLI 0.7.56 / Windows 11: ALPHA_MODE=1 present but BlockAdditional
alpha side data empty. Under current logic webmAlphaAdvisory sees
`alphaMode: true` and stays silent, so the render ships as opaque
without any signal.
The -metadata:s:v:0 alpha_mode=1 push is a muxer directive that some
ffmpeg builds write unconditionally, independent of whether libvpx-vp9
emitted the alpha plane. Tag presence is necessary but not sufficient
evidence of preserved alpha.
Advisory text names both possibilities (opaque composition OR silent
alpha drop) plus the concrete workaround (png-sequence + prores repack).
Fast path (no tag or missing tag) is unchanged. Probe adds ~1s per
WebM render only when the tag says alpha.
When `browserGpuMode === "software"`, set `forceScreenshot = true` in
`resolveConfig`. Explicit opt-outs (`PRODUCER_FORCE_SCREENSHOT=false`
or `overrides.forceScreenshot === false`) are honored.
This is defense-in-depth on top of the existing platform gates:
1. Linux + software (SwiftShader host) skips BeginFrame, avoiding the
compositor stall on shader-heavy frames under CPU raster (same
motivation as the closed PR #822).
2. `renderOrchestrator`'s reported `captureMode` field is derived from
`cfg.forceScreenshot ? "screenshot" : "beginframe"` — without this
clamp it misreports `"beginframe"` for the actual screenshot capture
on darwin + software.
3. Any new BeginFrame or drawElement entry point that forgets to gate
on GPU mode still routes to screenshot here.
Does NOT fix SwiftShader-on-darwin text-rasterization artifacts (an
ANGLE-SwiftShader issue on macOS text — the fix there is to use
`--browser-gpu`, which routes to `--use-angle=metal`).
Adopt the fake-timer pattern the sibling "recovers when a crashed reclaimer
leaves both lock directories" test in the same file already uses. Without
fake timers, if the dynamic `import("./manager.js")` beat between
`installFsMocks({ initialMtimeMs: Date.now() })` and the `withInstallLock`
call exceeds `staleMs` (50 ms on a busy shared runner with `vi.resetModules()`
per beforeEach), the new immediate-stale short-circuit added in #2328 fires
on iteration 1, breaks out before `waitedMs` reaches `waitNoticeMs=20`, and
no "Waiting for another hyperframes process" warn ever emits — the assertion
at `manager.test.ts:428` (`expected false to be true`) then fails.
Under fake timers, `Date.now()` is frozen at the mtime seed, so the lock
stays non-stale across the dynamic import and the wait-notice branch
observes real polling; `vi.advanceTimersByTimeAsync(staleMs + pollMs * 5)`
then drives the loop past both the wait-notice threshold and the stale
deadline so the reclaim + acquisition still resolves.
Test-only change; no production-code diff. Verified 27/27 in
`packages/cli/src/browser/manager.test.ts` under `vitest run`.
Miguel Ángel