ci(deps): bump storytold/craft-common-actions/.github/workflows/windows-arm64.yml from 8bb4e5ab95443171104ad88113fe1ef9b0e5ce3f to fa5ef18780cf5e7d5d3b8de068de56cffc83cac5 in the actions group
* M9: a per-address budget for server admin requests, and join codes found by hash (#1075)
* M9: the server tests share one snapshot and address helper (#1075)
* M4: a letter and its combining marks shape with one font (#1478)
* M4: the stacked-mark test checks positions only in fonts whose anchors stack them (#1478)
* M4: the cluster face rule is tested with stand-in faces, so it runs without system fonts (#1478)
* M4: the #1478 tests use stand-in faces only, so they run everywhere
Every message (the first and each reply) gets `mentions`: {name, known,
you} for each @mention, by the rule the thread view highlights with
(figcraft_doc::mentions, now shared with mention_ranges). Names resolve
against the file's people and you; unknown names are kept with
known: false; a@b.com mentions nobody.
A mask group's pixmaps cover only where its mask or content can paint within the view, rounded out to whole pixels, drawing the same pixels (a 40x40 mask on a 3000x2000 view: 96 MB -> 61.5 kB). Isolated layers share a 2 GiB per-thread budget; a layer past it is left out with a warning.
From the review of #1496: three background blurs stacked on each other across the band
boundaries draw the same in bands as in one piece (within anti-aliasing), so the bands'
overlap, the reach of the largest blur, is enough, as it is for the desktop's parallel bands;
pinned by a test. The docs and the `MAX_BANDS` comment no longer put the longest frame at "an
eighth of the render": a frame draws at least a band, a part of the render rather than the
whole, and the numbers measured for #1496 are cited instead.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* M2: components can't contain their own instances (#1084)
Edits that would make a component hold an instance of itself, directly, through other components or through another variant of its set, are refused with the loop named. Files holding such a loop open with one instance per loop detached and a note; a rebuild never expands an instance inside a copy of itself, so a loop merged in a live session stops growing.
* M2: a variant may hold a sibling variant's instance; only true loops are refused (#1084)
Each variant counts as its own component: refusals and open-time repairs apply only to a component leading back to itself through instances (an instance-swap value included), so files with nested sibling variants open unchanged. sync_instance_depth is split into match_layers and copy_layers.
Found while testing crates/engine/src/cmd/props.rs:614 (also :708, :717, :718, :826 and :1007)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Once a zoom had paused, the browser drew its full render in one frame. On a heavy page that
frame took seconds, and no input was handled meanwhile: the wheel steps and drags that came
were seen only when it was over.
The renderer's `Retained::begin` hands a full render back as `Steps`, drawn `band_h` rows at a
time with the same `render_band` and `put_band` the desktop's parallel bands use (bands overlap
by what background blurs read and keep their own rows; a blur reaching past a band makes the
bands taller rather than one piece), and `finish` keeps the image as a full render. A render
that is reused or patched is done at once, as before.
The browser's `Inline` worker refines through `Renders::refine_for`, which draws at least a
band a frame, more while they fit in `STEP_MS` (24 ms), and asks for the next frame while bands
remain, so input is handled between them; a view has at most eight bands, so on a page whose
render takes seconds the longest frame is about an eighth of it. A render drawn for the very
request is done once drawn. A pan or an edit meanwhile at the same zoom keeps the render going
and is patched in once it is drawn; a pan too far to patch or an edit the patch can't take draws
it again in steps, the image just drawn showing till then (`Refined::Shown`); a zoom, another
page or file, another size or other render options drop the render for one of the new view,
and what its bands took stands as the measure the next zoom's preview and wait go by. A render
in steps counts as one render (glass keeps the shapes a later render asks for again). The
desktop, with its render thread, is unchanged.
Measured in Chrome on two synthetic pages (20 images and 60 shadowed rectangles, with and
without two background-blur panels) whose full render takes 3.6–4.0 s at a deep zoom: the
longest frame while that render runs went from 3.7–4.0 s to 0.9–1.4 s (blur) and 0.69–0.80 s
(no blur); the bands together take longer than one render (the refined image comes after
6.7–8.1 s and 5.3–5.5 s instead of 3.8–4.1 s), each band rendering its overlap again and
repeating some per-band work, shadows mostly. The zoom steps are as before.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
With map: true, edit.duplicate also returns map: each original layer's id, descendants included (nested frames, groups, instances and their sublayers), paired with its copy's id, so a batch can edit inside a copy through a step reference instead of inspecting it. Without map the result is unchanged.
The batches a session keeps for apps that reconnect were bounded only by count on a desktop host
and a private figcraft-server (`max_log_bytes` defaulted to usize::MAX). It now defaults to
64 MiB, and figcraft-server takes it from a new `max_log_mb` setting (default 64; open to the
public, also at most a quarter of `max_file_mb` as before). An app behind the trimmed log catches
up with a snapshot, as it already did past the count limits. The protocol is unchanged.
Nine 8 MiB renames with default limits: all 75.5 MB kept before, 58.7 MB after.
Fixes#1069
Tools that act on a document, and the control methods engine.execute,
document.inspect, node.get and ui.render, take an optional `document`
(an open file's id from doc.list / document.inspect, or its path). The
call acts on that file and the tab in front stays in front, so an
attached agent keeps editing its own file when the user switches tabs.
Every replica placed a created, moved or reordered layer by counting each sibling that sorts
before it, so a batch of many layers under one parent took quadratic time, on the server's
single hub thread too. `Replica::slot` now binary-searches the `(key, id)` order on each side of
the layer's old place (a position change has already given it its new key). The order is
unchanged, ties and fractional positions included.
20,000 creations and 10,000 moves under two parents: 20.05 s before, 0.33 s after (debug build).
Fixes#1065
Found while testing crates/engine/src/cmd/variables_io.rs:~92 (css_safe, used by css for String values and mode names)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Every command an MCP client runs acts as the client's name from initialize (clientInfo.title, else .name) with ' (agent)' after it, or 'AI agent' without one, carried to the app as engine.execute {agent}. Its messages, reactions and read state are its own, and its messages are marked (agent: true in the file and comment.list, an Agent label in the thread). The user's name and comments made in the app are unchanged; in live sessions the name and marker travel with the thread.
The server's fixes for a refused batch walked every deleted layer's whole subtree again and
compared each document-level fix with all the earlier ones, so a batch deleting a deep chain took
quadratic time on the hub thread. Each subtree walk now stops at a layer already restored, and
document-level fixes are looked up and sent once per target. The fixes are unchanged.
A refused batch deleting a 20,000-deep chain plus 40,000 comment puts: 27.2 s before, 0.14 s after
(debug build).
Fixes#1063
* M8: large images on the canvas draw from a smaller copy (#1483)
An image shown smaller than its pixels draws on the canvas from a copy scaled down to 4096 or 8192 px on its longer side, cached beside the image in the same budget. Two 16000 x 10000 images (640 MB each) evicted each other on every frame that drew both; as copies they take 80 MB together zoomed out. Moving one over the other: about 2.1 s a frame before, 11 ms at fit and 15 ms at 30% after. Exports, thumbnails and views at 100% and closer draw from the image itself. Tests in large_image_tests; timings in large_image_bench_tests (ignored).
* M8: drop the ignored large-image benchmark; its numbers are in the PR (#1483)
BFlat