A padding changed on both sides of a branch gave a conflict for every size and place auto
layout worked out from it: the component's, each instance's and their hugging parents' width
and height, each instance's layout, and every child's position (10 in a file with one instance,
138 with 15). The merge was refused until each had a side, yet the merged file laid itself out
again from the padding whatever they got. A `layout` conflict also took all auto layout fields
together, so a gap changed only on the branch was lost when main won the padding.
`merge::plan` now leaves out properties both sides changed to values both files work out for
themselves (`Plan::derived`): a hugging auto layout frame's or a filling child's size, text
sizing to its characters, an auto layout child's place (when both sides turn and scale it the
same way), and a value a layer in an instance has as its main's layer has it. Merging keeps the
file merged into's own value and normalising lays it out from the merged inputs.
`layout` and `childLayout` merge field by field: a field changed on one side (or the same way on
both) merges, and each field changed both ways is a conflict of its own (`Key::Field`,
`<id>/layout.padding`), written whole from the merged fields and each picked side's value. The
review shows it as "auto layout padding" with both sides' values ("8, 8, 8, 8").
Tests: `branch_tests::one_conflicting_auto_layout_edit_is_one_conflict` (the issue's file: one
conflict, either side laid out as if set directly), `auto_layout_fields_changed_one_way_survive_a_conflict`
(merge and update from main), `sizes_auto_layout_works_out_never_conflict`; the merge fuzz test
leaves worked-out values out of what each side must show. Docs: collaboration-parity.md, mcp.md,
ROADMAP.md.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LvG6E2LxvJ6MuFmiLA6Zqm
* perf(collab): memoise layer depths in diff and skip the sibling scan for a new child
* address review: bound the depth walk by the document's node count
* fix(collab): place keyed-list puts through an index within a batch
Each Put of a comment, style, variable or collection scanned the whole
list for its key (cloning String keys on the way) before appending, so a
batch of N new entries cost N squared comparisons on the hub every
session shares. Removing an image also rebuilt the document's image-use
set before checking the image existed.
Replica::apply now keeps a per-batch key index for each keyed list; it
is rebuilt when the list's length or the entry at a key no longer
matches, and dropped on whole-list sets, reorders and real removals.
Removing an absent entry or image returns without scanning.
* fix(collab): keep the keyed-list index current through removals
Build the index without dead code outside tests, drop the wall-clock
assertion from the large-batch test, and shift the index down after a
single-entry removal instead of discarding it, so a run of removals in
one batch does not index the list again each time. Staleness checks for
puts and removals share one helper. Tests cover removals, duplicate keys
from a whole-list set, and absent keys.
* M3: library analytics list how each style and variable is used
A file reports how many of its layers use each of a team library's styles and
variables, and Show usage lists them by a Type choice with their layers and
files. Layers in instances count for their component. Also ignores plugins/target.
* M3: the style analytics test waits for the instance report too
Gap 10: a file counts the library instances it places, creates, pastes or duplicates and reports them like its detaches. The server counts what a file's totals grow past their highest in the current week and keeps a year of weeks (actions.json). library.usage {days} and Show usage give insertions and detachments over the past 30 days, 90 days or year. New variants are appended; an older server refuses them and Show usage says so.
* M12: an ended vote's results say who voted for each object
* M12: a board keeps earlier votes' results, listed in Results and deleted one at a time
* M12: the vote Results popover is drawn by small helpers, and its stepper, rows and Clear are tested
* M3: team library analytics count detached instances
Detaching an instance of a library component counts it in the file (libraries.detached, saved with it, capped). The background library check reports the counts after its fetch (ReportDetached); the server keeps them per file beside the file's usage report in usage.json (same caps) and answers Detached with them. Show usage and library.usage list detaches per component and per file (detached, detachesKnown). The new request and reply variants are appended, so every older library message keeps its bytes; an older server refuses the new requests, which the app ignores. Gap 10.
* M3: the detach test waits for the instance count too (the two arrive in separate reports)
* M8: cap the people a client keeps track of and the deleted ids a session remembers (#1121)
A hostile host or server could grow a client's presence, seen-peers and who-is-here lists without bound, and a session's deleted-id set only grew. Clients keep at most MAX_PEERS (4,096) others, leaving newcomers out with a warning; max_people is at most that; a session remembers at most 262,144 deleted ids, the lowest forgotten first.
* M8: a max_people above what clients track is lowered with a warning, not refused (#1121)
Refusing it would stop a server whose existing config says more from starting after an upgrade. Config::people() lowers max_people to MAX_PEERS for the core's limits, and the startup warnings name the configured and the used value.
* M8: the live welcome's handling in its own function (Sonar S3776)
Someone who joins with a comment link (or its join code) sees the file live and adds, replies to, reacts to and resolves comment threads, but can't change the design: their app refuses other edits and the server accepts a comment link's batch only when every change in it touches comment threads. Sharing online makes a comment link and code beside the edit and view ones; the Share popover offers Can edit, Can comment and Can view, and the commenter's toolbar shows a Can comment badge with Move, Hand and Comment.
Backward compatible: clients announce the comments word; one that doesn't is welcomed with a comment link as a viewer and gets comment links listed as view links, a level this build doesn't know reads as view, and an older server that can't make a comment link leaves the share without one.
Properties were listed by name, so a rename moved one. Components now keep prop_order (the names as they were made; a rename keeps the place) and every listing follows it. Make slot opens a name field under the frame's Slot row; Enter makes the slot, Escape cancels, a taken name keeps the field open.
* M9: one name for comments and live sessions, shown in the Join dialog, kept in the browser, renamed live (#1380)
* M9: a reconnecting session's reset is its own helper (#1380)
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
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
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
- An image a layer or style still shows can't be removed: removing it and
putting other bytes under its key would swap the picture that layer
shows, which giving the key other bytes directly can't. A batch made by
the diff stops using an image (node changes, then styles) before it
removes it; an undo keeps the changes that apply, so its image may stay.
- Restore and replay leave out the image entries history written under
older rules can't apply, using one method shared with the undo history
(Replica::apply_kept, extracted from history.rs). It is exact for every
image rule, including removals in use and replacements, and leaves out
only image entries: any other change that doesn't apply is still an
error. The image-only filter it replaces is gone.
- A server whose restore or replay left image entries out, or whose saved
log puts images, catches every resuming client up with its snapshot,
including one at the same sequence number.
- Tests: an_image_in_use_cannot_be_removed,
an_image_no_longer_in_use_can_be_removed, undoing_an_image_layer_still_works,
a_restart_replays_past_a_logged_removal_of_an_image_in_use,
a_resume_after_history_was_left_out_catches_up_with_the_snapshot,
a_resume_after_a_clean_restart_catches_up_with_the_log,
a_snapshot_with_a_broken_layer_change_still_fails_to_open.
- Docs: collab-protocol.md, collaboration-parity.md.
- without_refused_images judges each image entry against the document it
applies to and the batch's earlier entries (image_for, shared with
apply): a logged put of other bytes under a key the replay state holds
is left out, so a restart no longer stops there and drops the rest of
the log.
- Server::restore keeps its saved log only when the log puts no images:
a resuming client may already hold other bytes under a logged key, which
the restored state can't check, so clients catch up with the checked
snapshot instead.
- Tests: a_restart_replays_past_a_logged_image_replacement,
a_client_resuming_past_a_refused_image_in_the_saved_log_catches_up,
a_client_resuming_past_a_replaced_image_in_the_saved_log_catches_up.
- Docs: collab-protocol.md, collaboration-parity.md.
- The entry size cap is figcraft_doc::ImageData::MAX_BYTES (256 MiB), the
cap the .fig, .sketch and REST importers already read images with (now
one constant), instead of 32 MiB: an imported image larger than 32 MiB
made the document unjoinable and its server snapshot unrestorable.
- History written before the checks (a restored or joined snapshot, a
replayed server log, a file published to a server) applies without the
image entries they refuse (docprops::without_refused_images), instead of
the file not opening or a restart dropping every edit after the image.
- A key that already holds an image can't be given other bytes: producers
key images by their bytes (the REST importer's placeholder keys always
hold the same picture), and merge treats a key's image as never
changing.
- REST import: an unresolved imageRef shaped like a content hash shares
the rest-image-missing placeholder (the placeholder's bytes don't hash to
it); ImageData::claims_content_hash is shared with collab.
- Tests: collab image_tests (4 new, size test uses a small cap), format
image_tests (1 new).
- Docs: collab-protocol.md, collaboration-parity.md, file-format-parity.md.
Image entries applied bytes to any key with no checks (audit COL-9): a
collaborator could poison the shared content-addressed image store, claim
absurd dimensions, balloon memory, or name an image `../../x`, which a
saved file then wrote as the archive member `images/../../x`.
- Every key must be 1–128 ASCII letters, digits, `-` or `_`:
`figcraft_doc::ImageData::is_valid_key`, promoted from the format
crate's private `is_valid_hash` so the file reader and collab share one
rule. The reader now also takes `-` and `_`, so the REST importer's
`rest-image-missing` placeholder survives a save and load.
- The REST importer's side of the same attack: an unresolved `imageRef`
that can't legitimately be an archive member name (path text, an
over-long ref) shares the `rest-image-missing` placeholder instead of
being keyed by its raw name
(`image_tests::a_path_like_rest_image_ref_never_names_an_archive_member`).
- A key with the shape of a content hash (64 hex chars) must be exactly
the lowercase sha256 of the bytes, the form `image_hash` writes; an
uppercase look-alike is refused. Other keys (svg FNV-128, REST
placeholders) claim no content address and only meet the rules above.
- A side is at most 65 535 px; decoded bytes at most 32 MiB, with the
base64 text capped before decoding. 0×0 stays valid: the format crate
stores it for sizes it can't probe.
- Refusals are Error::Invalid and reject their batch, so every replica's
store stays clean; catch-up snapshots apply through the same check.
- Docs: collab-protocol.md, collaboration-parity.md and
file-format-parity.md (rows, status lines, histories).
- Tests: collab image_tests (8), format image_tests: round-trip and the
REST placeholder gate.
* M12: board timer and voting
A countdown everyone on a board sees (board.timer, board.timerState), synced as its end time or the
time left while paused, and voting sessions (board.vote, board.voteResults): votes per person, on
everything, stickies or the selection; clicks vote, your badge takes a vote back; your own votes
show while it runs, everyone's with a ranked list once it ends. Timer and vote sync as document
properties and ballots as one entry per voter; neither is an undo step. A pill at the top right
of the board, new Timer, Pause and Vote icons, labels and strings in every catalog, docs.
* M12: board timer and voting on egui 0.37
Id::new is deprecated in egui 0.37: the timer and voting area and its memory ids use Id::unique.
* M12: board.vote's start and ballot changes in functions of their own (Sonar)
* M6: Dev Mode measurements kept on the canvas, a read-only layers panel
Measurements placed between two layers, along a side or to the layer
around one (devMode.addMeasurement / removeMeasurement / measurements),
stored on the layer they start from (saved, synced, undoable; old files
open without them), drawn in Dev Mode as labelled redlines that follow
the layers and listed, placed and removed in the Inspect tab. In Dev Mode
the layers panel only selects and expands.
* M6: Dev Mode measurements: explicit imports, a measurement row of its own (Sonar)
Board files (File > New board) with sticky notes, shapes with text, connectors that follow what they join (straight or elbow, arrowheads, labels), marker, highlighter and eraser, and stamps; a board toolbar and a contextual bar; the Design panel hidden on boards, Prototype and Dev Mode refused there. Boards are ordinary FigCraft files built from ordinary layers, so saving, undo, export, MCP and live sessions work unchanged. File tabs, recent files and Open files show a pen nib for design files and a note for boards (original icons drawn in code).
Validated formatting, workspace Clippy, workspace tests excluding the two pre-existing localized startup assertion failures tracked in #1017, layers, assets, WASM and plugins. The startup tests are unchanged in this independent fix.
`utc_date` (collab) and `date` (ui-egui) took `u64` seconds, which the old
hand-rolled civil-from-days code converted for the whole `u64` range. The
chrono rewrite used `from_timestamp(..).unwrap_or_default()`, so any instant
past chrono's own year limit (262143) silently became the epoch — changing
existing output. `u64::MAX` seconds, for example, used to read
`584554051223-11-09 07:00 UTC` / `Nov 9, 584554051223`, not 1970.
Fix: add a `utc_parts(days, sod, nanos)` helper per crate that moves the
instant into chrono's range by whole 400-year Gregorian cycles (146_097 days),
lets chrono do the month/day/time, and restores the real year afterwards.
chrono stays the only calendar logic; the cycle count just shifts the year, and
every `u64` day count fits in `i64` so the conversion is always exact (no epoch
fallback). Split seconds with `u64` arithmetic first to avoid the `i64` seconds
overflow near `u64::MAX`, matching the old code.
Tests now pin the old output instead of the epoch:
- collab utc_date(u64::MAX) == "584554051223-11-09 07:00 UTC"
(and utc_date(253_402_300_800) == "10000-01-01 00:00 UTC", already pinned)
- ui-egui date(253_402_300_800) == "Jan 1, 10000"
date(u64::MAX) == "Nov 9, 584554051223"
chrono's %Y writes a '+' before years past 9999, whereas the old hand-rolled
{y:04} did not. utc_date now writes the year with {:04} to keep the exact old
output. from_timestamp already reads a clock past chrono's range as the epoch
rather than panicking. Tests cover the 5-digit-year and out-of-range cases.
Replaces the two hand-rolled Howard Hinnant days-from-civil copies with chrono
as a workspace dependency:
- crates/collab/src/message.rs: utc_date (YYYY-MM-DD HH:MM UTC)
- crates/ui-egui/src/home_recents.rs: date (Mon D, YYYY, UTC; keeps the
translated month names and trf formatting)
Output is unchanged. Neither function had a test, so each gains one pinning the
exact strings the old code produced (leap days included). chrono uses only std,
so wasm stays safe. AGENTS.md/CLAUDE.md now say time goes through chrono.
Follow-up to storytold/cadcraft#1785
Found while testing crates/collab/src/code.rs:~150 (Local::from_link, u64::from_str_radix on the token)
Co-Authored-By: Claude Haiku 5.5 <noreply@anthropic.com>
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Fills, strokes and effects already merged item by item, so both sides
changing different items of one of those lists was no conflict. Every other
list conflicted as a whole: both sides editing different export settings, or
different guides, or different prototype interactions, or different
accessibility annotations or Dev Mode links, each had to be resolved by hand
for the whole list, though the items stand on their own.
Export settings, guides, prototype interactions, accessibility annotations and
the Dev Mode code links and resources now merge item by item, with the same
rule: the same item changed both ways, an item removed on one side, or
different items added on both sides still conflict.
Properties that are keyed maps rather than item lists (bound, varModes,
overrides, propRefs, pluginData) are deliberately left out: their items line up
by key, not by position. Text runs are left whole too: they live in the node's
text data, so merging them item by item means splitting a property the change
format keeps whole.
Replaces the test that documented the gap (other_lists_still_conflict_as_a_whole)
with one that checks each new list merges, and adds one for the same item
changed both ways still conflicting.
The hub kept the addresses removed from every file it ever had open. The
entry now goes when the file is closed (everyone left) or removed from the
server, so a long-running public server doesn't keep collecting them.
The welcome's seq and processed.last_batch are the host's word and were used
unchecked: accepted computed self.seq + 1 and queue bumped next_batch, so a
host sending u64::MAX overflowed the client (panic in debug, wrap in release).
Both counters now step with checked_add/saturating_add, and a sequence that
cannot advance is refused as a protocol error (the established rejoin path).
Test: hostile_sequence_numbers_do_not_overflow_the_client.
Right-click a person's avatar ▸ "Remove <name> from session", or
`collab.removePerson {client}`, when you host the session or own the file
on a server. The hub closes their connection as the end of the session for
them (close 1001: they keep their copy and don't reconnect) and bars their
address from that file while it runs, so the code or link they hold no
longer gets them in; everyone else stays connected.
Protocol addition within version 2: `Admin::RemovePeer {file, client}` on
an open hosting connection with the file's owner token (or the server
key). Older servers say they can't.
Tests: loopback (removed, can't rejoin, others stay), engine (host can,
guest can't, bad ids refused), codec round trip; the collab junk sweep
covers the new command.
The recovery copy of one of your online files is no longer discarded when
the file reopens from its server. It is kept until the outcome is known:
- the server has everything in it: the copy goes;
- the copy has changes that never reached the server and nothing else
changed there: they are applied on top as one edit and sent to everyone;
- anything else (others changed the file, the server forgot the client):
the copy opens beside the file as "<name> (recovered copy)", with a toast;
- the server can't be reached: the copy opens as an unsaved document.
The note beside a copy is written after it and records who the app was in
the file and the server's last change it had. The app joins again as that
client (the first hello resumes; a new sync client numbers its batches
after the ones the server processed), so the layers it made are its own to
send. The old copy is removed only once this run has copied what it became.
collab.recoverFile; a tooltip on the tab mark; docs.
Filip StrajnarandClaude Opus 5.5
Dylan M. Taylor
BFlat
Patrick Wozniak
Christopher Linnett
NetterSilverandClaude Opus 5.5
F3R9
Ashsajal
midasmatterhorn
HaSh3003