score_search_hit: settlements are near-immune to distance ("brussels"
from Amsterdam means Bruxelles, not the closest Brusselsestraat),
POIs/streets keep strong local bias, and a number token signals
address intent. nav-build --places-only scans a pbf for settlement
nodes into a compact index (Europe: 1.31M places, 67MB, 60s); the app
merges it with the regional full index, deduping same-name near hits.
Route overlay decimation now measures against the last drawn point,
bounding the error at ~1.5px — pairwise skipping compounded and
visibly reshaped the route when zoomed out.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Search results render as two-line cards (name, then category · address
· distance) instead of one cramped line. Picking a result clears the
input — the next query no longer appends to the old text, which broke
every follow-up search. The nav worker restarts on demand if a script
hot-reload wiped the channel state.
Long-press never synthesizes from a held mouse button on desktop, so
double-click now triggers the same action (set position, then route
here); hints updated accordingly.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Flat mode returns to the classic building style (base fills +
outlines) instead of a top-down view of the extruded roofs: crossing
between flat and tilted bumps the style epoch and sentinels every
resident tile's bucket, so the normal stale-bucket restyle path
rebuilds them while the previous geometry stays visible and
cross-fades.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Tilted rendering now resolves occlusion in the depth buffer: map
geometry lives in a negative depth domain (UI/labels/overlay drawn
later always win by call order), view-ground y dominates so walls and
roofs occlude anything behind them under any rotation, and a baked
sort-rank micro-depth in param5 (4 depth quanta per rank step) orders
overlapping layers at equal ground without quantization flicker.
Bridges bias their micro-depth like their sort rank, the dark bridge
edge sits one step under the road fill (exact ties noise-win — black
bridges), and nested greens get distinct sub-ranks (park < garden <
grass < pitch) so parks stop shimmering. Flat mode stays on the
classic call-order path, byte-identical.
Camera ergonomics: right-mouse-drag (or Option-drag) rotates the
camera horizontally and tilts vertically, emitting ViewportChanged on
release.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Axonometric tilt as camera uniforms: screen y compresses about the
view center and building vertices carry their height in meters in
param4, lifted toward screen-top by height * px_per_m * sin(tilt) —
tilt animates freely with zero tile rebuilds. All CPU projections
(gestures, zoom anchor, screen<->lonlat, tile-selection AABB, label
placement + cache, overlay) undo the tilt alongside the rotation.
In 3D mode building footprints come from the all-tag detail archive
(real height / building:levels tags, 8m default) replacing the base
building fills: per-edge flat-shaded wall quads (NW light) painted
north-first as the painter's occlusion approximation, then the lifted
roof. The example app gets a 3D toggle with an eased tilt.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The detail archive's micro-POIs never reached the collector: the
decode-time point filter only admitted pois/addresses/road-label
layers. Bellamyplein now renders its trees, benches, statues, bins,
recycling and bicycle parking like the osm.org reference.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
CAMetalLayer with display sync throttles nextDrawable when the
compositor consumes frames unevenly (hardware-mirrored / scaled
displays): the main thread blocks 10-25ms in phases, felt as hiccups
in any sustained-animation app. Track in-flight presents per window
via addPresentedHandler; when two of the three pool drawables have
not reached glass, skip the paint beat and keep the pass dirty — the
next timer beat retries with the pool drained and event handling
never stalls behind vsync.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Heading-up camera: MapView rotation (bearing-up degrees) as uniforms in
the map shader — geometry and map-aligned glyphs (oneway arrows) rotate
about the view center, POI symbols stay upright billboards. Rotation-
aware tile selection AABB, pan/zoom-anchor/screen<->lonlat math, overlay
projection and puck heading, label placement (street labels follow the
rotated baseline, point labels stay horizontal) and the label pan-shift
cache. The nav app eases the camera onto the travel bearing with a
shortest-arc exponential and resets to north when navigation ends.
Micro-POIs: an optional all-tag detail archive (detail_mbtiles_path)
composes over the shortbread base at icon zooms — trees (canopy discs),
benches, waste baskets, recycling, bicycle parking, playgrounds and
artwork/memorial statues, matching osm.org's park furniture. Icons
only; base labels are never duplicated.
Simulated drive now ticks on NextFrame instead of a 20 Hz timer, so
the follow camera moves once per rendered frame.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Stroke geometry now bakes the centerline anchor per vertex (svg
tessellator emits anchors; offsets ride in param1/2, width-growth
class in param3, shape_id+100 marks expand mode — the shared vertex
format is unchanged). The map vertex shader re-expands each stroke
with a per-class width correction (regular roads, thin paths/rails,
waterways, constant-px building outlines) computed from the tile's
styled bucket vs the live fractional view zoom.
Tiles rendered at a stale zoom bucket keep the exact widths a fresh
restyle would produce, through the whole gesture: no fat roads while
zooming in, no width snap when the rebuild lands, dash dots and
building outlines stay crisp. Fading outgoing generations carry their
own bucket so they correct too.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Per-rule min_zoom on road/waterway rules: footway/path/steps hidden
below z15, cycleway/track below z14, streams/ditches below z13 and
canal lines below z12 — carto hides these long before roads; at city
scale they read as dot confetti over every block.
Waterway centerlines shrink faster than roads below z14
(zoom_mult^1.6): linear scaling kept Amsterdam's canal grid at near
full width at z12 and the whole city read as water; the water
polygons carry the visual weight at those zooms.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Conversion tooling (tools/map_tiles): curated Shortbread base pyramid
from a VersaTiles planet archive (bbox extract, brotli->gzip transcode
parallelized per 256x256 block), all-tag native OSM pbf detail
converter, pbf audit, and nav-build/nav-probe producing the routing
graph + search index artifacts. download_map.sh orchestrates pinned
downloads and conversions. mbtile_reader writer now skips SQLite's
lock-byte page at byte offset 1 GiB — allocating through it corrupted
every database over 1 GiB ("2nd reference to page 16385"); the 31 GB
Europe conversion passes integrity checks.
Navigation (libs/map_nav, new): region.search place/POI/street/address
index (prefix autocomplete, NL/EN category synonyms, proximity
ranking) and region.graph routing graph (CSR directed edges,
car/bike/foot speeds, oneway, turn restrictions, snap grid, A*),
maneuver generation and the NavSession map-matching state machine;
26 unit tests.
MapView interaction layer: MapViewAction, camera API + animated
fly_to, overlay.rs (route polyline with traveled dimming, markers,
position puck), per-widget mbtiles_path override, archive
minzoom/maxzoom honored for tile requests, zoom-level cross-fade over
the previous level's imagery, floating panels win hit-testing.
examples/map is the navigator app: worker-thread search, Drive/Bike/
Walk routing, simulated turn-by-turn with banner, follow camera.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- bridge roads get a dark casing edge (carto bridge style); global
casing pass keeps it under all road fills
- named parks/green areas emit centroid labels in carto green (z15+),
new LABEL_CLASS_GREEN; health-class label color wired in the view
- icon-vs-icon collision at bake time: first symbol wins within an
icon-sized neighborhood, so dense shopping streets show distinct
symbols instead of a stacked carpet
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Zoom-constant arrow glyphs (shaft+head) spaced every ~130 screen px
along oneway highways, offsets pre-rotated into the travel direction
(oneway_reverse honored), emitted into the symbol buffer so they draw
above road centers. Same anchor+screen-offset encoding as POI symbols.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Local tiles decode the MVT protobuf straight into tile-local f32
features (MvtSink trait + MvtLocalCollector with scale-aware thinning);
the MVT -> generated-Overpass-JSON -> parse detour is gone (a ~1-3MB
binary tile inflated to ~10MB of JSON re-parsed on every restyle, most
of the ~1s per-tile rebuild). The shared feature builder now takes
tile-local ways/points; the Overpass network path projects into the
same structures. MvtTileJsonBuilder deleted.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Direct (zoom, column, row) tile queries so the map no longer scans an
entire zoom level per batch, plus a writer for producing mbtiles
archives.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
New tile geometry fades in over 250ms via a tile_fade uniform
(multiplies the premultiplied output); the replaced generation's
geometry stays drawable underneath until the fade completes, so bucket
restyles and fresh tiles blend instead of flickering in. A 16ms timer
keeps frames coming while any fade is active (the paint beat idles
without input).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Cached label placements were redrawn shifted by the raw map-offset
delta; during a zoom the offsets change by world-scale amounts, flinging
the cache thousands of px off-screen (labels vanished just off max zoom
and stayed gone while idle because the re-place rate limit counted
frames). Now: cached glyphs draw through the exact affine screen
transform s*k + (off_new - off_old*k) with glyph size scaled, so labels
track the map mid-gesture; the re-place limit is wall-time based; and
the settle timer re-arms until a budget-truncated label pass completes.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The settle window counted FRAMES, but the paint beat idles when input
stops: after a zoom gesture no frames occur, the frame delta never
reaches the threshold, and stale-bucket restyles only fired when the
user wiggled the map. Settle checks now use wall time and the zoom
gesture schedules a 150ms timeout so a wake is guaranteed.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- in-flight tile requests time out after 600 frames and clear stuck
Loading placeholders; a leaked key would otherwise choke the 12-slot
dispatch cap (zoom-only restyles appeared to never regenerate)
- label placement budget only applies during gestures; at rest a full
pass runs, otherwise the lowest-priority tail (house numbers) was
permanently truncated at the same point every re-place
- per-tile label cap 32768: a city-sized overzoomed tile can exceed 16K
labels and the priority sort dropped addresses first
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Single-batch dispatch meant one tile in flight at a time after a zoom
(requests only advance when a draw happens, draws only happen when a
tile lands) — restyles trickled sequentially. Now every missing/stale
tile dispatches its own worker job, up to 12 in flight, so a viewport
restyle overlaps across the pool.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Gap-distribution instrumentation on the 7680x2160@120 display showed the
renderer can hold 120fps (gap_avg 8.4ms in calm stretches) but dropped
frames wherever labels are dense: full label re-placement (5-22ms) ran
almost every other frame during pan (tile arrivals + pan distance
invalidated the cache) and EVERY frame during zoom (exact zoom equality
in the cache check).
- full re-places rate-limited to ~8/s; between them the cached placement
draws shifted (~2ms), including during zoom gestures where labels stay
screen-pinned briefly (pinch behavior a la Google Maps)
- each placement pass gets a hard 7ms budget; highest-scored labels
place first, the tail defers to the next pass
- gap avg/max and >12ms counts added to the perf log
- tile restyle after zoom: one tile per worker job (10-tile batches ran
sequentially in one closure, 2-3s per batch), settle delay 150->80ms
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Instrumentation showed draw_walk at ~1ms and CPU tile insert at ~2-3ms
while inter-frame gaps hit 200-1000ms during interaction: the stall is
GPU buffer upload/pressure from oversized tiles (a bucket-17 z14 tile
holds ~6.8K outlined buildings -> tens of MB of 19-float vertices).
- scale-aware vertex thinning: drop points closer than ~0.35 screen px
at the styled zoom (invisible detail dominated vertex volume)
- building outlines use miter joins (round-join bevel expansion on ~60K
ring points per tile was the largest single buffer)
- tile GPU uploads throttled to 2 per frame via a pending queue
- resident tile cache capped at 48 (was 640) with faster idle eviction
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- perf log: draw_walk aggregates total/geometry/label ms, max frame,
inter-frame gap and full-placement count per 240 frames into
local/map_perf.log
- labels precompute name_key + tile-local bbox at tile build; the
per-frame scan does no allocation and rejects offscreen labels with
two point transforms before any path work (an overzoomed tile holds
~16K labels, mostly offscreen)
- batched glyph path no longer runs update_draw_vars per call (the open
batch already did); halo underdraw 8 -> 4 diagonal offsets, halving
glyph volume
- fill clip overlap shrunk to ~1 screen px: the wider double-drawn
strip let the later tile's land paint over the earlier tile's
buildings (pale band at tile edges at high zoom)
- point local mbtiles back at noord-holland-shortbread-1.0.mbtiles
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- one glyph instance batch per frame (begin/end_glyph_batch on
DrawRotatedText): every glyph previously opened and closed its own
aligned-instance batch, x9 with the halo underdraw
- label placement cache: pan-only frames redraw the previous placement
shifted by the pan delta (like the tile geometry) instead of
re-scanning up to 16K labels and re-shaping/re-colliding ~1-2K; full
re-placement only after 48px of pan, zoom change, or tile updates
- shaped-run cache keyed by (text, quantized font_scale) so periodic
re-placements stop re-shaping every string every time
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- zoom-bucket restyles wait ~0.3s after the last zoom change so line
widths scale geometrically (smooth) during the gesture and re-style
once, instead of flickering tile-batch by tile-batch mid-zoom
- per-tile label cap 16384 and larger z16.5+ budgets: an overzoomed z14
tile holds a whole city's addresses and lowest-priority house numbers
were truncated before the viewport filter ever saw them
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- openstreetmap-carto symbol SVGs (CC0) tessellated once at final screen
size and baked into a 4th tile-cached vector buffer; vertices encode
anchor + screen-px offset, the map vertex shader (shape id 20) adds the
offset after the map transform so symbols keep constant pixel size at
every zoom — pure GPU vector, no raster atlas
- POI symbols colored by class (orange food, purple shops, brown culture,
red health incl. pharmacies), labels shifted below the symbol
- symbols gated to z17+ and hidden immediately on zoom-out even while
stale higher-bucket tiles are still displayed
- green areas (leisure/grass/forest/gardens) rank above generic landuse:
they share the shortbread land layer with huge residential polygons and
lost to protobuf feature order (Bellamyplein rendered gray)
- cache-only fallback when the local mbtiles source is missing (e.g.
while a larger extract downloads)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
- per-end line caps (Tessellator::stroke_ends): butt at tile-clip cuts,
round at true ends — road cap discs no longer stamp over neighbor
tiles' tram tracks and roads at seams
- fill polygons clipped to their own tile square (Sutherland-Hodgman);
a tile's MVT buffer fragments no longer overpaint the neighbor;
building outlines skip segments running along the tile cut
- stroke clip padding 3px (under the generator buffer) so boundary
cuts are detectable
- white label halos (8-offset underdraw, theme label_halo color)
- carto POI label colors (orange food, purple shops, brown culture,
muted house numbers) from shortbread poi attributes
- label placement hysteresis across frames (no flicker while panning)
- street_polygons fill ranked below sites/parks (Bellamyplein-class
plazas no longer cover the park inside them)
- max zoom 19 with width stops for z18/19
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Replace the abandoned arc2 colour-stub with a real path primitive: the
distance to a bare circular-arc CENTERLINE (no baked thickness), so it
chains after move_to/line_to and is covered by a single stroke(w) -- unlike
arc_round_caps/arc_flat_caps which bake their own width. Angles in radians,
0 = +x axis, counter-clockwise positive; last_pos advances to the arc end so
a following line_to connects.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
feat(draw): add sdf.arc_to circular-arc path segment
A `Fill` child under Flow::Right (or `height: Fill` under Flow::Down) is
deferred and given the row's full slack up front, so end_turtle's align
step takes the deferred branch, which distributes width/height to fills
and drops main-axis align entirely. That's correct when a fill consumes
its whole slot, but a fill that draws narrower than its slot (e.g. an
Image that aspect-fits, or `Fill{max: N}` capped below the container)
leaves real slack that then anchors to the start regardless of align.
Add row_align_x_shift / col_align_y_shift: reclaim align * (inner -
actually_drawn) in each deferred branch. They early-return 0 when align
is 0, the inner size is unknown, or the content fills, so the only
behavior change is deferred-flow containers that set main-axis align and
hold an under-filling Fill child.
Add examples to the uizoo "Layout Demos" tab that proves they boht work.
redrawing is significantly more expensive than text comparisons,
and doing this outside of each widget is difficult and less efficient.
this helped avoid a bunch of redraw cycles in robrix,
so it's probably worth doing for all widgets, especially since
setting text or another basic property now does an auto-redraw
(it didn't used to be like that)
* Fix rendering, gradient, sampling, etc issues on older GPUs
* Fix bug in `box_y` sdf function, which caused gradients to be split
into two bands incorrectly. Mostly a problem on lower-res screens.
* Use per-texture filtering instead of GL sampler objects on Linux,
especially for Mesa drivers that ignore min filter samplers.
This should help prevent blocky/pixellated things like emoji/avatars
* Improve rendering sharpness on low-DPI screens (icons, emoji, images)
On 1.0-DPI screens, emoji/SVG-icons/avatars were minified without
adequate sampling and SVG AA was sub-pixel, producing blocky/aliased
output. This reworks each path and adds optional full-window SSAA.
SVG icons (device-aware AA + round caps):
- draw_svg.rs/draw_vector.rs/render.rs: size the fill & stroke AA
fringe and the curve-flatten tolerance in DEVICE pixels (≈constant
regardless of icon size), so edges resolve via the analytic
d/fwidth coverage and curves stay smooth at any scale; re-tessellate
on scale change.
- triangulate.rs: thread the flatten tolerance through path fill/stroke.
- tessellate.rs: emit round caps as a solid disc (u=0.5) instead of a
radial fade that collapsed to a square at small sizes.
- widgets/icon.rs: don't clip the Icon to its Fit bounds, so round
caps that extend past the box render fully instead of being sheared.
Emoji:
- glyph_raster_image.rs/rasterizer.rs: rasterize color emoji near the
on-screen size with an alpha-weighted box downscale (geometric-mean
scale factor) instead of the font's native PNG strike.
Images / avatars (mipmaps):
- image_cache.rs/texture.rs/draw_list.rs/lib.rs: optionally emit a CPU
mipmap chain (VecMipBGRAu8_32) for non-animated images so minified
avatars/thumbnails sample cleanly. Env-gated MAKEPAD_IMAGE_MIPMAPS;
default on for GL on Linux.
- metal.rs: real per-level mip upload. d3d11.rs/vulkan.rs/web_gl.rs:
safe single-level fallback (no crash; real mips TODO).
Full-window supersampling (optional):
- window.rs: render the whole UI into an offscreen target at
MAKEPAD_SUPERSAMPLE× device resolution and downscale-resolve into the
window. Default 2×, env-tunable, 1× disables. Modeled on the existing
GaussStack render-to-texture path.
* don't unconditionally enable supersampling SSAA of 2x by default
it's too expensive and too slow for most older devices
* cleanup, reduce comment verbosity
* improve SVG anti aliasing
* Windows: fix laggy/juddery scroll performance
- Pace the render loop to the display refresh using a DXGI frame-latency
waitable object and present with vsync, replacing the free-spinning,
uncapped Poll loop that caused uneven scroll cadence.
- Coalesce consecutive WM_MOUSEMOVE messages and paint once per loop pass
to stop the judder when moving the mouse during fling deceleration.
- Cache get_dpi_factor() and the WM_NCHITTEST WindowDragQuery result to
avoid per-mouse-move GetDeviceCaps syscalls and widget-tree hit-tests.
- Throttle XInput/DirectInput polling of empty/disconnected controller
slots, which was stalling the UI thread.
- Rework the momentum fling to a native exponential model with a
frame-interval EMA, and stop the tail auto-scroll from fighting an
active fling/drag.
- D3D11: update the glyph atlas and image textures in place via
UpdateSubresource instead of recreating them on every change, and
spread D3D11 shader-object creation across frames.
- Slug atlas: only force a full re-layout on a width change; append rows
on height growth.
* Windows: correctness fixes from review (off the scroll hot path)
- d3d11: close the DXGI frame-latency waitable HANDLE in Drop (it was
leaked once per main-window lifecycle and the field comment was wrong);
keep popup swap chains at frame-latency 1; track the waitable-swapchain
flag for ResizeBuffers instead of inferring it from the handle; present
without the vsync interval during a live resize.
- windows.rs: poll game input on the idle signal tick so a gamepad button
can be serviced while the app is otherwise idle.
- win32_window / window: invalidate the WM_NCHITTEST / WindowDragQuery
caches on window move and on a caption relayout, with a generation
counter guarding against a reentrant invalidation being clobbered.
- windows_game_input: detect a controller already plugged in at launch via
a one-shot full scan on the first poll, probe slot 0 (Player 1) first,
and offset the DirectInput enumeration so it never stacks with the
XInput probe.
- comment/doc corrections.
* Image cache: accept any Arc<D: AsRef<[u8]> + ?Sized> for async image data
The load_image_from_data_async family required Arc<Vec<u8>>, forcing callers
that already hold the bytes as Arc<[u8]> (e.g. a content-addressed media cache)
to copy the whole buffer via .to_vec() just to satisfy the type. Generalize the
data parameter to Arc<D> where D: AsRef<[u8]> + ?Sized, so those callers can pass
their existing Arc by refcount-clone with no byte copy. The decode path only ever
borrowed the bytes (&[u8]), so this is purely a signature relaxation; existing
Arc<Vec<u8>> callers are unaffected (D = Vec<u8>).
* Linux: GL glyph-atlas in-place texture update + X11/Wayland mouse-move coalescing
* Scroll: unified fling model + native trackpad momentum deceleration
Share one kinetic-scroll model between PortalList and ScrollBar (and thus
ScrollXView/ScrollYView/ScrollXYView) via a new widgets/src/scroll_motion.rs:
- Touch-drag flicks use an iOS-style exponential self-decay, frame-rate
independent via a per-frame integrator with dt smoothing.
- Trackpad scrolling applies the OS momentum directly while fast (responsive,
full native speed), then hands off to a gentler self-decaying tail once it
slows past a threshold, so the deceleration is longer and smoother than the
OS's short, choppy tail. Handoff is seeded at the current speed for a
continuous transition; the seed is clamped against degenerate event timing.
- Add ScrollPhase to scroll events, mapped from NSEventPhase/momentumPhase on
macOS and wl_pointer AxisStop on Wayland; None elsewhere (wheels/X11/Windows
behave as before). MAKEPAD_RAW_TRACKPAD_MOMENTUM=1 bypasses the smoothed tail.
- A press catches an in-progress fling (stops the scroll, consumes the press so
it doesn't also activate a child), matching iOS/Android/macOS.
* Shader codegen: prefix Metal/WGSL locals to avoid reserved-word collisions
The Metal/WGSL backends emitted user-declared shader locals verbatim, so a
local named after a reserved type keyword (e.g. `half`) produced invalid
shader source and failed to compile at runtime. Prefix them with `l_` like the
HLSL/GLSL backends already do.
* TextFlow: don't panic on unbalanced HTML close tags
end_code/end_quote unwrapped the area stack, so a stray `</pre>` or
`</blockquote>` in untrusted content (e.g. a chat message) panicked. Return
early instead, and drop a vestigial per-list-item area-stack push that leaked
an entry and could hand a stray close tag the wrong block's area.
* Scroll: expose fling decel, handoff threshold, & tail-decel as `#[live]` fields
This allows app devs to override the scroll feel per-widget in the DSL,
or globally by overriding the base widget's defaults — verified that a DSL
override takes effect.
* Windows: fix frame pacing, paste crash, and wheel input backlog
- Wait on the frame-latency waitable right before each window's vsync
present instead of on every Paint, so input no longer stalls behind
waits that have no matching present. Drain leftover credits after a
live resize.
- Pasting when the clipboard has no text no longer panics.
- The poll loop now handles up to 32 messages (2 ms) per frame and
merges consecutive mouse-wheel messages, so fast wheels can't build a
backlog that keeps scrolling after the gesture ends. Sleep 1 ms when a
frame presents nothing so animation polling doesn't spin a core.
* Image: don't build unused mip chains; fix stale images in recycled widgets
- Only build the CPU mip chain when a backend actually uploads it
(Metal, behind its env var), and build it on the decode thread instead
of the UI thread. Linux GL still gets its mipmaps via glGenerateMipmap
and now retains less CPU memory per image.
- Recycled Image widgets no longer show the previous item's image or
apply an old decode result. Placeholder textures set via set_texture
(like blurhashes) stay visible while the real image decodes, and a
failed load clears the widget instead of leaving old content up.
* Widgets: avoid needless caption redraws; cheaper PortalList height tracking
- Label::set_text does nothing when the text hasn't changed, and the
window caption title is only synced when it actually changes, so mouse
moves and animation ticks no longer redraw the whole window every
event. The caption centering padding requests its own redraw now.
- PortalList records item heights only when new or changed, and only
re-applies the default height after it drifts by half a pixel, so big
lists don't walk every unmeasured item on every scroll frame.
* Text: cache layouts of long texts; stop cloning glyph outlines every frame
- The layout cache now accepts texts of any length (long messages and
code blocks used to re-layout on every scroll frame). It is a real LRU
with a byte budget on top of the entry cap, and texts drawn in the
current frame are never evicted, so one heavy frame can't thrash the
cache into a permanent miss cycle. The shaper cache is LRU now too.
- Glyph outlines are shared via Rc, so drawing a cached glyph no longer
copies its command list, and outline complexity is computed once when
the outline is built instead of every frame.
* Html: fix stale links/spans in recycled widgets and <details> renumbering
- set_text only rebuilds when the content actually changed, so a
recycled link can't open the previous message's URL, and re-setting
identical content keeps the user's <details> open/closed state.
- Custom widgets and <details> are keyed by their node index instead of
a visit-order counter, so toggling a collapsed <details> can't
renumber the widgets after it and rebind them to the wrong nodes.
item_with_scope also recreates its widget when the template changes.
* Linux: fixed-distance wheel scrolling; Wayland frame-callback pacing
- Wheel scrolling moves a fixed 60 px per detent on X11 and Wayland
instead of a timing-based guess that flipped between 12 px and 240 px
depending on how events batched. Wayland reads real detent counts via
AxisValue120 (wl_seat v9, with AxisDiscrete as the older fallback) and
maps keymaps MAP_PRIVATE as v7+ requires. Touchpads are unchanged.
- Wayland frames are paced with wl_surface frame callbacks and swap
interval 0, so redrawing a hidden or minimized window can no longer
hang the whole app inside eglSwapBuffers (compositors withhold frame
callbacks for hidden windows). Windows with a callback in flight skip
presenting and stay dirty; X11 keeps vsync exactly as before.
* Text: bigger layout cache budget, reclaimed at the end of each frame
A maximal ~60 KB message lays out to roughly 4 MB of glyphs, so the 4 MB
budget couldn't hold even one alongside a normal screen. Raise it to
16 MB, and run eviction at the end of every frame so memory over the
budget is freed one frame after its content leaves the screen, instead
of lingering until some later layout happens to insert a new entry.
* Fix oversized uniform slices: the array lengths were in bytes, not f32 elements
* Wayland: flush buffered mouse motion before scroll events, and drop motion for closed windows
* GL: fall back to non-mipmapped filtering when glGenerateMipmap fails on strict GLES3 drivers
* Text: bucket emoji raster scales so zooming reuses atlas slots instead of re-decoding every step
* Image: add has_content() and record texture provenance on cache-hit loads too
* Scroll: native trackpad momentum, Chrome-model bounce; presses catch motion, never click children
* Dock: redraw the newly selected tab immediately when the active tab is closed
* Html: standard link colors with pressed precedence; skip re-parsing unchanged text; color setters
* Image: don't redraw on cache-hit loads of the already-bound image (per-draw reloaders looped forever)
* Scroll: log macOS momentum-end phase bits to check Cancelled (touch-cut) vs Ended (natural fade)
* Scroll: momentum state machine; flicks survive pagination; edge sentinels & once-per-frame actions
* Scroll: remove the MAKEPAD_SCROLL_DEBUG diagnostics
* Scroll: time-based fling velocity window; pointer fan-out guard; parked flings survive pagination
* PortalList: optional reached-start/end margins (Some(0) default); repositioning re-announces the edges
* Linux/Wayland: honor UI zoom across window resizes; scale caption bar with zoom
Keep the wayland-side window geom in native units so the zoomed dpi isn't
read back as "native" on the next Configure — UI zoom no longer resets or
flickers on maximize/tile. Also let the caption bar height scale with the
zoom (pin to native only on macOS, where the buttons are OS traffic lights).
* Windows: D3D11 fixes for drawlist mgmt
trying to help with the `new_batch` bugs
* draw_list.rs — `set_zbias` now returns whether it changed
* d3d11.rs — uploads draw_call_uniforms on the given condition:
`uniforms_dirty || zbias_changed || buffer.is_none()`
and the zbias advance is hoisted above the early-continues
* Scroll: hard flicks carry farther and boost on re-flick; iOS-style touch rubber band with per-edge bounce gating; Android fling spline behind a flag
* Splitter: redraw both pane subtrees on drag; cached views keep fresh fixed sizes in the dirty check
* iOS: re-deliver window-geometry changes dropped by re-entrant UIKit callbacks; Init only ever from the first draw
* final cleanup fixes. Ensure all examples, experiments, `studio` all work
* Image: skip re-parsing an SVG that is already shown, keyed on the caller's shared bytes
* CachedView: fix upside-down offscreen texture on GL/GL-ES
GL/GL-ES store offscreen FBOs bottom-up, unlike Metal/D3D/WGSL.
The DSL->script-shader migration switched the CachedView composite to plain
.sample() (non-flipping sample2d on GL), so cached views rendered
vertically mirrored on GL/GL-ES; Metal was fine on macOS.
Add a `sample_rt` script sampler (emits the V-flipping sample2d_rt on
GLSL, plain no-flip sample elsewhere) and use it in the CachedView and
CachedRoundedView composites.
* Fix `AdaptiveView::redraw()` to actually do something
Deadzone-rescaled right stick feeds the exact mouse-drag pipeline: same
0.01 rad/px orbit through pseudo-pixels (~2.6 rad/s full deflection, stick
up = look up), same look_dx/look_dy for scripts, same chase-rig authority
(stick held = kid owns the camera, recenters after release) and
cam_dragging visibility. Applied before script camera writes each tick,
like real mouse events, so set_cam_yaw still wins its tick. Zeroed under
tape tests for determinism. Camera-only pads now count in device selection.
splashgame.md: right stick documented; new rule — every new ability must
also be reachable from the gamepad (bind to the named actions).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
VoiceWave grows an opt-in ptt_use_escape flag (Escape doubles as cancel/
dismiss elsewhere, so hosts choose); both keys drive the same logical talk
button. Gamemaker opts in via the caption_bar's hidden voice_wave and the
hints now read 'hold Esc' — the big friendly key for kids. Verified the
nested named-child merge lands on the live widget.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Verdict on the gamemaker judder (measured, 25s A/B runs, same window size):
gauss OFF -> unchanged (25-33ms gaps) [pyramid exonerated]
empty game world -> unchanged [game exonerated]
vsync OFF -> FIXED (119.8fps steady, worst 9ms; slow callbacks 124->2)
The stall is CAMetalLayer display-sync throttling: the compositor (hardware-
mirrored + SwitchResX-scaled 7680x2160) returns drawables unevenly, and
nextDrawable blocks the main thread 10-25ms in phases. The durable fix is
present-gated pacing: track in-flight presents via addPresentedHandler and
skip the paint when the pool is busy instead of blocking the event loop.
MAKEPAD_TIMER_TRACE=1 logs paint-clock fire-to-fire gaps >20ms and slow
callbacks >10ms with a per-step breakdown (live_edit/net/pad/paint/gc).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The fixed 8ms timer0 presented ~125Hz into a 120Hz vsync queue: the drawable
pool drifted full and nextDrawable blocked the main thread in a ~25-frame
sawtooth (PerfGraph 'wait' ramps). Timer now arms at 1.002/max_fps across
attached NSScreens — the +0.2% makes NSTimer lateness drain the queue instead
of accumulating. Sawtooth verified gone.
perf_monitor: 'gpu' channel — presented-frame GPU interval tapped from the
existing command-buffer completion aggregation (atomic hand-off, folded at
frame_boundary; concurrent with CPU channels, plotted violet).
metal: MAKEPAD_GPU_PASS_TRACE=1 logs per-pass GPU time ([gpu-pass] name ms)
for frame-budget hunts.
Measured (gamemaker, empty world vs full racing game — identical ~4.7ms GPU):
the frame cost is the UI glass pipeline, not the game — scene capture +
6 mip downsamples + 3 smooth upsamples re-run every frame at 120Hz because
the game pane dirties the window; at low GPU clocks the 11-pass chain's
latency swings past the 8.3ms budget in phases.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
platform: Cx.perf_monitor — per-frame ring (240) of paint-to-paint gap +
per-channel CPU us. Built-in channels: event dispatch (outermost, minus
app-attributed time), script exec, GC, pass encode, nextDrawable wait.
Apps register custom channels: cx.perf_monitor.channel("physics", rgb).
Off until enabled; hooks in event dispatch, macos repaint, metal draw_pass.
widgets: PerfGraph — corner-pinned live panel (DrawVector strips): frame-gap
bars colored against 120/60Hz budgets with guide lines, stacked per-channel
CPU, legend with averages. Self-positions bottom-right (DrawVector geometry
+ deferred turtle alignment don't mix — no aligning parent).
gamemaker: PerfGraph hovers the game pane (F3 toggles), engine feeds script
+ physics channels (incl. hot-reload evals); engine text overlay moved to
F4; per-phase engine window kept for ag perf / AIGAME_PERF=1; new 'ag perf'
harness verb + template guidance (template CLAUDE.md force-added: runtime
resource, blanket CLAUDE.md gitignore had kept it untracked).
Measured on my-game-5: engine frame CPU ~0.3ms; the hiccup is frame pacing —
the 8ms NSTimer paint clock beats against the 120Hz display, the drawable
pool drifts full and nextDrawable blocks the main thread in a ~25-frame
sawtooth (avg 2-3.4ms, spikes 20-30ms). The graph shows it as red ramps.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
libs/tts was never tracked despite being a build dependency of the gamemaker
example. tools/download_tts.sh fetches the public upstream weights (HuggingFace
Kokoro-82M + whisper.cpp) and converts them locally with the in-repo stdlib-only
converter; model artifacts are gitignored.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
The parallel batch machinery from be21d627a only ever existed to chase a
default-on multi-threaded win that never came (washer w8 still +54% with
the hybrid on). At w1 — the only configuration this opt-in flag is for —
parallel_for runs inline, so the serial path gives the identical result
for ~340 fewer lines. Wire up the previously-dead
dynamic_tree_self_pairs/cross_pairs into a serial collect_batch_candidates
(three BVTT self/cross traversals -> canonical (a,b,child) sort -> serial
filter into move_results[0]) and delete BatchWork, BatchCtx, batch_drain_*,
BatchFilterCtx, batch_filter_*, bvtt_step, dynamic_tree_bvtt_drain/expand,
and the batch_frontier/worker_* scratch fields.
Determinism preserved exactly: OFF 0x61E35C31/step314 bit-identical, ON
0xBE99C5F7/step313 identical across workers 1/2/4. The debug SET-equality
oracle and the determinism_broad_phase_hybrid_across_worker_counts test
are unchanged and still pass; zero warnings.
PGO: pgo.sh never trained -bp=1, so an off-path-only profile laid the
hybrid branch out cold and collapsed the win to ~-5%. Add one -b=8 -bp=1
training run (neutral for the default path — counts merge, the OFF branch
stays hot) and retrain.
Corrected README numbers to measured values (hybrid-trained profile,
paired -bp toggle, washer w1): pair-finding stage 7.6k -> 3.6k ms/1000
(-52%); total ~-17% (~19.0k vs ~22.6k), which beats C (20661) and narrows
Rapier's lead from ~23% to ~12% — not the "17.7k / within 5% / -19%"
be21d627a claimed. Still default-off (w8 regresses ~+54%), opt-in
single-threaded accelerator for churn-heavy scenes.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Broad-phase-hybrid subsection (why washer loses, the batch design, the
determinism proof, the single-thread win / multi-thread floor tradeoff,
and the default-off rationale). Washer row in the single-thread rapier
table annotated with the opt-in number (~17.7s, beats C, ~5% behind
rapier → box3d ahead-or-even on all nine single-threaded).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
WorldDef.enable_broad_phase_hybrid (default false; -bp=0/1 bench toggle):
an adaptive batch broad phase for high-churn scenes. When
move_count*4 > proxy_count, replaces the per-moved-proxy tree queries
(8k proxies x 3 root-descents on washer) with three BVTT self/cross
traversals (dynamic self + dynamic x static + dynamic x kinematic) that
share the upper-tree descent, plus an O(n) bottom-up refit instead of
the median rebuild. Both traversal and candidate-filter are parallelized
across the task system (per-worker buffers → merge → canonical sort by
(shape_a,shape_b,child) → deterministic contact creation).
Correctness: a #[cfg(debug_assertions)] SET-equality assertion (batch
candidate set == per-mover set) runs in every test and never fires — the
proof the BVTT finds identical contacts (the hash can't prove it since
creation order legitimately re-baselines). New test
determinism_broad_phase_hybrid_across_worker_counts. OFF hash 0x61E35C31
bit-identical; ON hash 0xBE99C5F7 identical across workers 1/2/4 +
external tasks. 180/186/180/180 tests, zero warnings, profile retrained.
Single-threaded washer -18.7% (17715 vs 21780, broad phase -51%) — beats
C (20661), within ~6% of Rapier (16844). DEFAULT OFF because it regresses
multi-threaded (washer w8 +52%): the batch materializes ~40-50k
candidates/step and serially merges+sorts them (a fundamental floor the
inline per-mover path avoids by filtering in the query callback), so at
w8 the parallel per-mover queries win. Cannot be worker-gated (would
break cross-worker determinism). Correct, deterministic, zero-cost when
off — an opt-in single-threaded accelerator for churn-heavy scenes.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Adds the C Box3D column to the nine-scene rapier table, all from one
same-window interleaved single-thread run (2026-07-06). Percentages vs
the box3d Rust column. box3d Rust beats Rapier on 8/9 (junkyard flipped
to a +5% win post-tier-2; only washer lost, Rapier's incremental-BVH
broad phase). vs C: within ~7% geomean, ahead on both pyramid scenes,
worst is junkyard +18%. Replaces the derived † junkyard cell with a
direct measurement. Headline updated 7/9→8/9.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
