Shift did nothing with the Pencil: the stroke stayed freehand, while the
Line tool and the Pen keep theirs to 45° steps. pencil_add, shared by the
drag and the release, now keeps one straight segment from where the
stroke started to the pointer at the nearest 45° step while Shift is
held (the Marker and Highlighter too); a Shift-click without a move still
draws nothing.
Publishing keeps the variables its styles bind and where its copies of other libraries' assets came
from; importing files such a copy under that library when the file links it.
From review: the asset list follows the layers placed in the tree, so a
smaller tree (maxNodes, or the MCP tool's retry for maxBytes) means a smaller
list, and `truncated.assetsLeftOut` counts the rest; following aliases reports
`truncated.variables` when it stops at its budget with variables still to
read; and a font is checked with figcraft_text::has_family by its trimmed name,
as the missing-fonts list does, so a family stored with outer spaces isn't
taken for installed. A regression test for each fails without its fix.
From review: the variables a style binds and those an alias or colour
calculation reads are joined too, followed to any depth; fonts come from the
text layers placed in the tree, so a second target's font isn't lost to a
separate walk's budget; and a malformed `id` or `ids` entry is an error
instead of reading the selection or dropping the entry. A regression test for
each fails without its fix.
inspect.designContext and the MCP tool design_context read the selection (or
ids, or a scope) in one call. Each layer keeps apart the values the designer
set (explicit), those layout or the file worked out (derived) and those with no
direct CSS equivalent (unsupported, with the reason). The variables (every
mode's stored and resolved value), styles, components, fonts and assets the
tree uses are joined once, and `missing` says what the design doesn't specify.
A connected instance gives its code snippet instead of its layers. Large trees
are capped and say so, and the MCP tool asks for fewer layers until the result
fits the client.
node.get and document.inspect write `dashCap` on a layer with a dash cap, but
it was missing from NODE_KEYS, so `fields: ["dashCap"]` was refused as an
unknown key (and the MCP tools' `fields` schema left it out). A test now walks
the keys a varied layer tree writes and checks each is one `fields` accepts.
A fill-only text was left as it was: Outline stroke skipped every layer
without a visible stroke. A text now flattens in place into a vector of
its glyphs, filled as the text was and keeping its colour style,
constraints, visibility and lock; a text with no glyphs stays.
The nested frame is 100 x 100 at (180, 180) and its rectangle is 40 x 40 at (200, 200), as the code creates them.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 driver, no subagent
A press on a layer inside a selected one now drags the selection, and a click
without a drag selects that layer (also in Dev Mode and for a single selection),
as in Figma. Cmd/Ctrl, Shift and layers outside the selection are unchanged.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
transform.nudge applied the step in the parent's local space, so Down inside a frame
rotated 90 degrees moved the layer sideways. The step is now turned into the parent's
space first (falling back to the plain step if the parent's transform is not invertible).
Auto layout children are unchanged.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
CodeRabbit's majors on #1735: evicting a file from the capped usage list
dropped asset_peaks, so a later identical report was counted twice.
Retention now includes asset_updated, and the file just heard from is
kept even when timestamps tie. The document keeps up to 1000 asset
counts, but each wire message is still capped at 250, so reports are
sent in batches an older server will accept.
Tests: evicting_capped_usage_rows_keeps_asset_peaks_so_a_repeat_report_is_not_counted_twice
and asset_action_reports_cover_every_count_past_the_message_cap.
Signed-off-by: FredApps <153173854+FredApps@users.noreply.github.com>
counting_links compared library_links, which walked every node, so applying or
detaching a library style or variable on an instance or a layer in it counted,
while duplicating that instance counts only its component. library_links now
walks the pages and skips instances with what they hold, as count_inserted does.
Test: library_asset_counts_tests::applying_and_detaching_inside_an_instance_counts_nothing
(an instance, its layer, a nested instance and its layer). The server test waits
for the detachments too, not only the insertions.
Applying a style or variable, or copying or duplicating content that uses it,
inserts it; the Detach button detaches it; an instance counts as its component
alone. Show usage and library.usage give them over the chosen period.
A property added to a set goes to all its variants (component.addProperty), but
Combine as variants kept each component's own definitions, so a property made
on one component before combining was missing from the others: their layers
couldn't be bound to it ("the component has no property") and an instance
switched to one of them dropped its value. make_variant_set now gives every
variant the properties any of them has (share_props, through the give_prop step
addProperty uses too); a variant with a property of that name keeps its own.
Tests: variant_shared_props_tests.
Cmd/Ctrl+V with one layer inside a frame selected pasted onto the page on top of the frame. It now pastes into the layer's parent just above it, in an auto layout frame the item after it.
Page-level layers, several layers, instance children, board atoms, component sets and copied components keep today's behaviour.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
In vector edit mode, align.* and distribute.* act on the selected
points instead of moving the whole vector. The Design panel shows the
alignment row, and the existing shortcuts apply to those points.
Closes#2109.
In a hugging parent, Fill on the other axis follows the line the way
Figma does, instead of forcing the parent Fixed. This also covers item 2
of #2090 on both axes: a hugging main that becomes Fixed is followed by
its instances' roots.
Fixes#2091 item 3. Fixes#2034. Fixes item 2 of #2090.
A layer stored its vertical text alignment, text case, paragraph spacing and
indent, OpenType features, start and end caps, a frame's variable modes, which
component property controls a field, and whether a nested instance is exposed,
but node.get and document.inspect left them out. They are reported now, under
the names node.set takes so a value read can be written back, and only when
they aren't the default.
Automation roots: same_file doesn't open a path on another computer, native NT link targets and names Windows trims are refused, and the docs say network paths are refused
* M9: page.setOpening sets the page a shared file opens on, for everyone who joins (#1817)
A file on a figcraft-server opened on the page current at publish time, with no way to change it:
`DocProp::CurrentPage` was synced but nothing wrote it (`page.select` is each viewer's own page,
only a local save stamps the document). `page.setOpening {id?}` writes `doc.current_page` as a
document edit: one undo step, a synced change in a live session, so every later join opens on it;
the current page when none is named; a divider or a layer refused; nobody's own page moves. The
pages list's context menu gets "Set as opening page". Docs: mcp.md, collaboration-parity.md.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* M9: the pages menu's new string is in the English catalog (#1817)
The gate's i18n test wants every tr() key in locales/en.toml.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* Opening page: the parity doc states the default when no opening page is set (#1972 issue)
With none set, a file opens on its saved current page, not the first page, and a link cannot name a page, so the row is partial.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* Opening page: one sharing row, and the spotlight exception on the opening-page row (#1972 issue)
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* Opening page: the parity row states where the joiner's page comes from (#1972 issue)
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* Opening page: a no-op set adds no step, deleting the opening page falls back, and the tests cover junk ids, refusals, undo and the menu item (#1817)
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
* M5: the chosen opening page is saved with the file and opens it for everyone (#1817)
The opening-page flag now lives on the page node (NodeKind::Page opening), so it syncs
through the existing node kind. Saving writes the flagged page as current_page instead of
the page the saver was viewing. A duplicated page does not copy the flag, and deleting the
page removes it.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
---------
Co-authored-by: Claude Fable 5.1 <noreply@anthropic.com>
* M8: Duplicate repeats the copy's rotation as well as its move, and forgets the step on a deselect (#2293)
Duplicate remembered only the shift of the copies' bounds corner, so a turned copy repeated an arbitrary offset and no rotation. The step is now the move of the copies' centre in the page frame plus the turn each copy has over its source, applied to the new copies about their own centre (one layer) or their shared centre (several). Size is not stepped; a resized copy repeats its move, not its growth. Measured in Figma's web app on 2026-11-12. The step was already dropped when the selection changes to anything but the copies (clean_selection), and a test now pins that.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 queued; Opus 5.5 briefed; Sonnet 5.5 measured and wrote at medium
* M8: the duplicate step ignores turns in place, size changes and flips (#2293)
The move of the step is now measured per pair, as each copy's own centre minus its source's, less the shift a size change gives that centre (taken in the source's frame), averaged over the pairs. A turn about a layer's own centre, uneven layers and a widened-then-turned copy no longer record a move. A copy whose flip state differs from its source's repeats no turn, since flips were not measured.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 queued; Opus 5.5 briefed; Sonnet 5.5 measured and wrote at medium
* M8: a turn of several copies together is not counted as a move in the duplicate step (#2293)
Duplicated::step took the mean of the per-pair centre moves, so a Rotate on several uneven layers (turned together about the selection centre) read as an extra move and the next Duplicate applied it. The step now also tries the reading where the copies were turned together about the sources' bounds centre, and uses it when every copy fits it and the in-place reading does not fit. Two tests cover a turn together and a move plus turn together.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
* M8: the duplicate step reads a turn about the rotation handle's centre, and a resize after a turn, as no move (#2293)
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
* M8: the duplicate step turns several layers about their handle box's centre, with no pivot inferred (#2293)
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Fable 5.2 briefed; Sonnet 5.5 wrote at medium
* Docs: repair a row the earlier merges garbled (Alt-drag duplicate)
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Model-Effort: Opus 5.5 drove; Opus 5.5 briefed; Sonnet 5.5 wrote at medium
---------
Co-authored-by: Claude Sonnet 5.5 <noreply@anthropic.com>