The registry tests mock the RPC and check every call recorded while the
mock is installed. A component sync deletes the thumbnails of the
frames it changes 200 ms later, so a sync at the end of an earlier test
of the suite can call the mock in that window: under the coverage run,
the frontend suite failed in register-test, a different test each time,
with an extra :delete-file-object-thumbnails call. The mock now records
and answers only the plugin profile commands; any other call gets an
empty answer.
Also in this commit:
- Update the component memories for the audit and fix phase. The contract memory
says how the code now holds it: the eager cascade and where it lives, the
automatic text name, the children of a transformed copy shape, measured text
sizes on both renderers, repairs as derived writes, the tolerance of "differs"
and the :set-touched operation. The debugging recipes add the override
scanner; the data model and the composable test notes follow the changes.
AI-assisted-by: claude-opus-5-5
- common/core: the variant logic namespaces now build transform into
a variant, combine as variants and the layers-panel rename of a
member, and the sync walk is two plain functions whose change pairs
go through pcb/add-changes.
- frontend/composable-component-tests: renaming a member in the
layers panel, removing empty properties and setting a member's
error are variant property edits both runners run; case AW joins
the JVM set.
AI-assisted-by: claude-opus-5-5
A profile of the benchmark worlds showed where a sync spends its time.
- The token sync of shapes that apply no tokens took 38% of a sync in
the 1000-copy world: for every synced shape, generate-update-tokens
walked every sync attribute, resolved its group and its token
attributes, and only then found that neither the main shape nor the
copy applies a token. It now returns at once, and the token
attributes of each sync attribute are computed once, at load.
- A sync looks through every object of each page and component to
find the copies it updates, so in a file with many shapes and few
copies the search costs more than the sync itself. The new benchmark
world main-edit-among-many-shapes syncs a main edit into 10 copies in
a page with 20,000 plain rects; its baseline, measured before the
change below, joins the recorded one with a note, and its result
joins the golden master. generate-sync-container called the
uses-assets? multimethod on every object through a lazy filter; the
predicate is now chosen once per container (asset-user?) and the
objects walked with reduce-kv into a vector, in the same order.
- compare-children searched the remaining children (d/seek) for every
child out of place, and dropped a child taken out of order with a
lazy remove, so a reorder took quadratic time and nested one lazy
sequence per child taken. It now looks children up by their
reference or id (index-positions) and skips those taken out of order
when the walk reaches them, with the same callbacks in the same
order. The sync of a reorder stays slow: when a main with 1,000
children moves its first one to the end, each copy still gets 999
moves (F85), and applying them is what costs (4.75 s to 4.26 s for
10 copies). Fewer moves change what a sync writes, so that waits for
Phase 4c.
No behavior change: the golden masters and the suites are unchanged.
Benchmark (lower quantile, ms; baseline / now / ratio):
deepest-edit-through-three-levels 35.2 / 18.6 / 0.53
edit-through-a-chain-of-dependents 68.9 / 36.4 / 0.53
library-update-into-copies 83.1 / 43.0 / 0.52
main-edit-among-many-shapes 4.8 / 2.3 / 0.48
main-edit-into-10-copies 1.8 / 0.9 / 0.48
main-edit-into-100-copies 15.4 / 8.4 / 0.54
main-edit-into-1000-copies 153.1 / 90.1 / 0.59
variant-switch-into-copies 32.1 / 20.0 / 0.63
(main-edit-among-many-shapes: baseline measured before the search
change.)
AI-assisted-by: claude-opus-5-5
Which shapes of a copy a transform leaves untouched (the ignore tree)
was frontend code, written twice: calculate-ignore-tree for the
modifiers of the SVG renderer and calculate-ignore-tree-wasm for the
matrices of the WASM renderer. The second is the single-pass version
of the first: each shape still compares itself with the copy root of
the latest transformed shape above it, as the repeated walks did when
the last one won.
app.common.logic.placement now holds it once: calculate-ignore-tree
takes the transformed ids in order and a function that transforms a
shape, built with modifiers-transform (SVG) or matrices-transform
(WASM), both cached per shape. check-delta and transform-attrs moved
with it. The frontend calls the common version, and its tests moved
to common-tests.logic.placement-test, which runs on the JVM and in JS,
with a rotation through modifiers added.
No behavior change: the golden masters and the suites are unchanged.
The benchmark does not run this code.
AI-assisted-by: claude-opus-5-5
The namespace held 3,300 lines: component sync, instance creation and
detach, swap and the variant switch's attribute carry, duplication,
and the sync of library colors and typographies. Each now has its own
namespace:
- app.common.logic.libraries (cll): component sync, reset, update
main, attribute and placement sync, and the sync of a file with its
libraries;
- app.common.logic.instances (cli): create, instantiate and restore a
component, detach a copy;
- app.common.logic.swap (clsw): swap a copy head, carry overrides
across a variant switch;
- app.common.logic.duplicate (cld): duplicate components and shapes;
- app.common.logic.library-assets (clla): sync of library colors and
typographies, which the asset multimethods delegate to.
Each new namespace may require libraries, never the reverse; the
helpers they share became public there (attr-operations,
reposition-shape, orient-like, swap-content-root-id, ...). Callers use
the new aliases. The functions move as they were.
No behavior change: the golden masters and the suites are unchanged.
Benchmark (lower quantile, ms; baseline / now / ratio):
deepest-edit-through-three-levels 35.2 / 37.3 / 1.06
edit-through-a-chain-of-dependents 68.9 / 69.7 / 1.01
library-update-into-copies 83.1 / 79.8 / 0.96
main-edit-into-10-copies 1.8 / 1.6 / 0.92
main-edit-into-100-copies 15.4 / 15.4 / 1.00
main-edit-into-1000-copies 153.1 / 156.0 / 1.02
variant-switch-into-copies 32.1 / 27.0 / 0.84
AI-assisted-by: claude-opus-5-5
Before refactoring the component sync code, record the exact files the
component tests produce, so a refactor that changes any of them fails.
The JVM test common-tests.logic.golden-master-test runs every variant of
the composable cases the pure runner runs and every benchmark world. It
records the file after each user, undo and redo step (the runner's new
:observe option), the final files with the signatures of the undo/redo
round trip failures, and each world's file after its timed work. The
frontend interpreter records the store's files of every variant of its
cases and of the completion contract the same way.
app.common.test-helpers.golden-master writes a state as the normalized
files in sorted "path value" lines: ids become test labels, or numbers
in tree order when unlabelled; grid cells are named after their shape
and position; timestamps are left out. So a state is the same in every
run whatever ids it draws. Only a digest per state is committed; the
full lines go to tmp/golden-master/ and tmp/golden-master-frontend/
(recorded/, and actual/ for a state that differs), so diff -r shows a
change. PENPOT_GOLDEN_MASTER=record records them again.
Both sides repeat exactly across runs. Dropping :opacity from the
attributes sync copies fails the case C variants that edit it.
The interpreter's fixture also turns cts/wasm-enabled? off for each
composable test and restores it after, as it already restores the
store. A test that starts the WASM renderer left the flag on, so in a
full frontend run the cases ran on WASM shape proxies, which leave out
the attributes that are nil or false, and the frontend golden master
differed from a run of the cases alone (:flip-x and :flip-y).
AI-assisted-by: claude-opus-5-5
A copy text had three override subgroups: its letters, any style, and
its structure (the number of paragraphs or runs). One style override
kept every other style of the main out (F72), letters and a style
together kept everything out (F73), a structure change kept the main's
styles out unless both texts had one style (F74), which the editors'
content never had (F79), and alignment, direction and vertical
alignment overrides were lost at the next sync (F76).
Now a copy text has :text-content-text for its letters and line breaks
and one group per style unit: family, weight (with style and variant),
size, line height, spacing, decoration, case, fills, typography,
alignment, direction and vertical alignment (cttx/style-units). An edit
sets the groups of what it changes: a letter both texts share that
changes its style, or a style that was not in the text before, so typed
letters touch no style. A token in the content sets the groups of the
styles it sets, and token sync of a text follows them.
Sync and the keep-touched variant switch share one merge
(cttx/merge-overrides): the letters of one side, and each unit from its
side placed on them. A unit with one value applies everywhere; else
letters both texts share (stretches of three or more) take the source's
style, and the rest follow the letter before them. The switch carries
an override only when the old and new mains agree on it (F77, F81), and
the copy then holds the groups its result differs in. A typography
reference stays only when every unit it sets comes from its side. The
variant id goes with the weight; where the family comes from the other
side, the frontend has to pick the family's closest variant.
app.common.types.text reads a content as letters, each with its run and
paragraph, and builds the content back from them; it gives the style of
one unit for a letter, compares styles by what they show, and finds the
letters two texts share with a letter diff (Myers). The component
contract memory describes the model.
Files older than this keep :text-content-attribute and
:text-content-structure, read as every group they may cover until the
migration recomputes them.
Tests whose expected values the model changes: text-sync-test (the
copy's weight, letters or new run no longer keep the main's size out;
the letters read whole), text-touched-test (its edits went to the
paragraph node, which shows nothing; they now edit the run),
variants-switch-test (the copies with two runs keep their letters
across a switch, and the size override the mains agree on), the
editor output table, case Y and the notation test (new groups). The 1j
pins of F72-F79 and F81 go; F79's direct test of equal-attrs? becomes a
sync of editor content. A new check holds the production override rule
to the 1j reference rule. New tests: the letter view, and a detached,
duplicated or reset copy text keeping or clearing its override groups.
Benchmark (lower quantile, ms; baseline / now / ratio):
deepest-edit-through-three-levels 35.2 / 36.6 / 1.04
edit-through-a-chain-of-dependents 68.9 / 67.2 / 0.98
library-update-into-copies 83.1 / 72.5 / 0.87
main-edit-into-10-copies 1.8 / 1.5 / 0.82
main-edit-into-100-copies 15.4 / 14.9 / 0.97
main-edit-into-1000-copies 153.1 / 153.4 / 1.00
variant-switch-into-copies 32.1 / 27.1 / 0.85
The worlds hold no text. A main text edit synced into 100 copy texts
that changed their letters takes 40 ms, against 29 ms before: the merge
aligns the letters of each copy with the main's.
AI-assisted-by: claude-opus-5-5
When a component nested in others is deleted (inner in mid in outer,
mid deleted), a change to inner still reaches mid's stored shapes,
mid's copies, outer and outer's copies, and restoring mid brings back
an up-to-date main. Keep this behavior: outer is a live component, and
its users cannot see a deleted mid between it and inner.
Extend the deleted-component propagation test with a copy of the
deleted component and its restore, and reword the contract, which said
copies of a deleted component do not sync.
AI-assisted-by: claude-opus-5-5
When a copy head took the place of a main shape and its component was
the one of that place, it stopped being a swap, but its children kept
referring to the main of its own component instead of to the children
of that place. Validation did not see it, and the next sync deleted
those children and added them again, losing their overrides, or the
head did not follow later changes to that place. It happened in:
- a variant switch that carries a swapped child into a variant whose
child there has the same component;
- update main of a copy that swapped a nested copy back to the
component a deeper main puts in that place, after the main in between
swapped it for another one; the clone in the main also kept
:swap-group;
- the repair of :ref-not-in-near-main, which also never cleared
:swap-group.
The swap in a copy, the switch, update main and the repair now share
ctf/take-position: the head refers to the main shape of its position,
is swapped while its component differs from that shape's, and else its
descendants refer to the descendants of that shape.
The switch also read the swapped child from the objects of the swap,
where it was already gone with its old parent. It now reads it once it
is moved into the new copy, so its placement turns its children too,
and the undo takes it back to the root as the move to the root left it.
The swap finds the new shapes in its own objects, not in its changes.
The component memory says how a head takes a position.
Benchmark (lower quantile, ms; baseline / now / ratio):
deepest-edit-through-three-levels 35.2 / 37.1 / 1.05
edit-through-a-chain-of-dependents 68.9 / 70.5 / 1.02
library-update-into-copies 83.1 / 77.3 / 0.93
main-edit-into-10-copies 1.8 / 1.7 / 0.94
main-edit-into-100-copies 15.4 / 16.0 / 1.04
main-edit-into-1000-copies 153.1 / 160.6 / 1.05
variant-switch-into-copies 32.1 / 28.7 / 0.89
AI-assisted-by: claude-opus-5-5
A swapped copy head followed the position, constraints and layout item
attributes of the main child it replaces, but not its rotation or
flips. When the main rotated or flipped that child, the copies that
kept it turned and the swapped ones did not. Sync, and a variant switch
that carries a swap, now give the head and its subtree the flips and
then the rotation of that child, around the head's own center, unless
the user changed the head's geometry. The swap itself uses the same
step.
Benchmark (lower quantile, ms; baseline / now / ratio):
deepest-edit-through-three-levels 35.2 / 34.1 / 0.97
edit-through-a-chain-of-dependents 68.9 / 75.5 / 1.10
library-update-into-copies 83.1 / 81.3 / 0.98
main-edit-into-10-copies 1.8 / 1.6 / 0.88
main-edit-into-100-copies 15.4 / 15.6 / 1.01
main-edit-into-1000-copies 153.1 / 153.2 / 1.00
variant-switch-into-copies 32.1 / 31.4 / 0.98
AI-assisted-by: claude-opus-5-5
The migration 0033 stores every swap of a file as an override group.
A swap made in a copy is paired with the child of the near main that
holds the same position, as sync paired it; any other slot is dropped.
All pairings are found on the data as it was, so swaps nested in swaps
pair against the content of the outer one, and libraries pair the same
whether they are migrated or not. A head that refers to its own slot
only inherited it, and drops it even when its near main cannot be
reached. Any other slot that matches no child, or whose near main is in
a library that is not loaded, keeps the slot as its ref, so validation
reports it. An import, which migrates its files before it links their
libraries, leaves those slots and converts them once the libraries are
linked. A migrated head that sits elsewhere than the child it replaces
would place it gets its geometry marked as touched, so the first sync
does not move it. Migrated and new files carry the feature
components/swap-group, so a client that still writes slots cannot open
them.
Validation checks the new form: :misplaced-slot reports any slot,
:misplaced-swap-group the group anywhere but on a nested copy head in a
copy, :ref-not-in-near-main a nested copy head whose ref is not a child
of the near main of its parent, and :duplicate-ref two children of a
copy that refer to the same main shape. Each has a repair.
Nothing but the migrations reads slots any more: the slot branches go
from sync, swap, switch, reset, detach, paste, duplicate and move. The
helpers that read and write a slot on a shape move to files/swap-slots,
and the ones that follow refs to files/comp-processors. The graph
projection drops the FillsSwapSlot edge, the components debugger shows
the component a swapped head takes its content from, and the plugin API
documents componentRefShape for a swapped copy. The swap tests are
named after what they check.
Benchmark (lower quantile, ms; baseline / now / ratio):
deepest-edit-through-three-levels 35.2 / 34.6 / 0.98
edit-through-a-chain-of-dependents 68.9 / 63.5 / 0.92
library-update-into-copies 83.1 / 76.4 / 0.92
main-edit-into-10-copies 1.8 / 1.6 / 0.87
main-edit-into-100-copies 15.4 / 15.1 / 0.98
main-edit-into-1000-copies 153.1 / 148.8 / 0.97
variant-switch-into-copies 32.1 / 26.4 / 0.82
AI-assisted-by: claude-opus-5-5
A swapped copy sub-head now keeps the shape-ref of the main child it
replaces (its position), takes its content from its own component, and
carries the touched group :swap-group. It used to point at the root of
the new component and store a swap slot naming the replaced child.
Every path reads and writes that form:
- sync pairs a swapped head with the child it replaces and syncs it
from its own component; it takes from the replaced child only its
placement: position, constraints and layout item attributes;
- reset undoes a swap in place, keeping the head's id and the other
overrides of the copy;
- detach, paste and duplicate make a head that leaves every copy level
a plain copy of its own component, and advance a deeper one by
position; the children of a swapped head never advance;
- a variant switch keeps a swap stored this way, and counts only the
copy's own swap, not one made in a main of the chain;
- update main replaces the main child with a clone of the swapped one,
and from a swapped head updates its own component;
- a swap inside a copy keeps the old shape-ref and adds :swap-group;
swapping back to the component of that position gives a plain nested
copy whose children refer to the position's children; a swap inside
a main is a plain nested copy.
Files that hold swap slots keep working until a migration converts
them. The component memories describe the model.
This fixes what came from the slot: reset of a swapped child cleared
the other overrides of the copy (F12) and gave copies of a main new
ids (F66); update main left a swap mark (F16); a second variant switch
lost the overrides (F29); grid cells did not pair a swapped child
(F44); a swapped child did not follow its position (F65). F49 keeps
failing, in another form.
The tests that asserted slots now assert the override group; the tests
of the slot code build the old form with :as-slot?.
Benchmark (lower quantile, ms; baseline / now / ratio):
deepest-edit-through-three-levels 35.2 / 35.4 / 1.01
edit-through-a-chain-of-dependents 68.9 / 68.7 / 1.00
library-update-into-copies 83.1 / 77.5 / 0.93
main-edit-into-10-copies 1.8 / 1.6 / 0.92
main-edit-into-100-copies 15.4 / 15.4 / 1.00
main-edit-into-1000-copies 153.1 / 151.1 / 0.99
variant-switch-into-copies 32.1 / 32.4 / 1.01
AI-assisted-by: claude-opus-5-5
Add a battery of tests for text overrides in copies.
- A short notation for text content in tests (thtx): paragraphs of
runs with short style names, read back with equal runs joined,
edits in the notation, set-text and set-content through the change
path, and the override rule for texts (override-units), which says
which overrides a copy must hold against its main.
- Sync matrix: every whole-text copy edit against every main edit,
plus the cards of the worked examples, the "Hello world" table, a
token on a copy, typographies and a library color sync. Where the
code differs from decisions (k) to (o), the result is pinned
against F72 to F80.
- Switch matrix: copy edits against variants that differ in letters,
size, color, a new line or a bolded word; F76, F77, F79 and F81
pinned.
- Composable case BA: EditText also edits fills, font family and a
whole text; the copy and the main change different styles, then
reset or update main, in both runners; F72 pinned.
- Editor output: the content the WASM and v1 editors commit when a
copy text is typed into, retyped, split with Enter and has a word
bolded, recorded in the workspace and replayed on a copy. Typing
sets only the text groups and a style change no text override;
F79 pinned on both editors' content.
A review found tests that would pass with what they test broken.
- Runners: both judge one variant at a time, fail a variant with no
assertion and a known-failure mark that applies nowhere, and give
the asserter the state the operations produced. The pure runner
validates the final file and rejects operations of unknown kind.
The frontend interpreter fails a step when the store reports an
event error (Potok only logs it), reports asserter errors against
their case, and stops the watchers that install! starts.
- Nodes drive the real code: a main root rename goes through
end-rename-shape, and the clipboard's advance of nested copies is
one common function, cll/advance-copied-objects, that copy-selected
and both test paths call.
- Cases N, AH, AI, AK, AD, AE, K, C, H, S, AY, I, Y, AX and F check
what they claim; O, R and T get real sizes.
- Stricter common unit tests found F60 (a repair detaches healthy
copies), F61 (a switch drops a fixed-sizing size), F62 (the ref
chain never stops at its target) and F63 (latent); all pinned.
- Frontend library tests: the RPC double records on subscribe, the
dropped commits check that nothing syncs, and the link and share
tests check the events sent.
- Plugin suite: OpSwap and OpDetach throw when the API call did
nothing, and case AB checks that the reset head is an instance of
the original component.
- The composable suite memory records the new rules.
The composable interpreter treated a step as done after a 60 ms gap
with no commits, and treated its 2 s timeout as done too. Slow sync or
reflow work could finish after the step, so a case could assert on a
half-synced file.
Each event step now waits, as the plugin waitForLayoutUpdate does,
until every pending work tracked by the reflow registry has drained:
the watcher's sync wait, the sync it starts, and layout and text
reflow. A step that does not drain within 2 s fails the test and names
the work still pending, and the next variant starts with nothing
pending.
The tests cover an edit with a deliberately slow sync, undo and redo
steps that must wait for the watcher's touch-component, translation
and derived commits, a sync that never finishes, and a restart of the
edit watchers, which must keep the undo entries of an edit the same.
Also in this commit:
- Keep page-less changes from adding a nil page. Applying a commit locally
rebuilds the object indices of every page its changes name. Changes that touch
no page, such as :mod-component, name none, and their nil page id was rebuilt
too: the file data in the store gained a page under a nil key holding
{:objects nil}. Every component touch after an edit to a main added it, and
undo never removed it.
- Check undo and redo of every step in the composable cases. The composable
cases tested undo only where a case asked for it, so a change generator with a
wrong inverse, or a step that splits into several undo entries, went
unnoticed.
- Add operations and cases to the composable component suites. The composable
runners could edit, add, remove and move children, swap and switch, but not
run the other component operations.
- Test component validation, repair and library asset sync.
- Add a component sync benchmark with a recorded baseline. The coming refactors
and fixes change sync code, and a slowdown must be seen and explained in the
commit that causes it.
- Tidy the composable suite and make its cases stricter.
- Test the component sync paths no test reached. A coverage run found sync paths
that no test ran.
- Test variant events, token sync and switch overrides. The variant events had
almost no frontend tests, and several switch, paste, token and restore paths
never ran.
- Test the file and container helpers and the library events. The component
helpers of types/file and types/container and the library events of the
workspace had the lowest coverage left.
The composable test interpreter wired its own copy of the undo stack
subscription and started only the component watcher, so it reacted to
edits differently from the app, and a missing subscription failed
silently.
Gather the subscriptions that react to edits while a file is open
(undo stack, layout updates, text reflow, WASM text measurement) in
initialize-edit-watchers, stopped by finalize-edit-watchers, and the
per-page component and token watchers in watch-library-changes. The
workspace and the interpreter both start them.
Starting the edit watchers again replaces the running ones. A file
reload starts the workspace again without finalizing it, which left two
undo stack watchers running, so an edit added an extra undo entry.
The composable test engine, nodes, setups and cases lived in the
frontend test tree, so only the frontend interpreter could run them.
Move them to CLJC test helpers under app.common.test-helpers.composable
and add a pure runner (comp.runner) that does what the workspace does
around each operation, without the frontend store:
- after a user operation it syncs the changed components until none is
left, as the frontend watcher does with one pass per sync commit;
- the operation and its syncs form one undo group, and the undo node
reverts the latest one;
- it validates the file once the sync or the undo has settled.
Each node declares its kind: assembly, user, undo or frontend-only.
User nodes record the production changes they apply, and the runner
throws when a node breaks that. The swap and switch nodes build their
changes with the production clv/generate-component-swap-in-place.
Rotate and ChangeHeight need the frontend placement step, so the pure
runner rejects them.
The cases are zero-arg functions in comp.cases, shared by the frontend
interpreter and common-tests.logic.composable-sync-test. The latter
runs cases B to M and the new case O, a switch of an overridden copy,
on the JVM and in JS. Case N and the rotation case stay frontend-only.
Write down the intended semantics of copies, :touched and component
sync: what counts as an authored or a derived write, the override
rule, and what each operation (edit, rename, sync, reset, update
main, detach, swap, switch, undo) does to touched groups. It also
records the settled decisions on keep-touched switches, deleted main
children, the eager sync cascade, automatic renames, derived changes
on mains, resized copy roots and layout-owned geometry.
Link it from common/core and common/component-data-model.
The 🔒 section is generated from advisories instead of being written
by hand: an advisory becomes an entry when its patched_versions include
the release version, so unpublished (triage/draft) advisories are valid.
- gh.py advisories --patched-version filters on the exact token of
vulnerabilities[*].patched_versions; the list output keeps the summary
schema.
- changelog.py security renders the entries sorted by GHSA id, links the
merged milestone PRs labelled "pr pt" that reference the GHSA id and
warns when a reference has no matching patched version. The subsection
follows 🚀 (or 💥, or the version header) and is removed
when the version has no advisories.
- Tests cover the filter, the entry rendering, PR discovery and the
subsection placement.
- Memories and the update-changelog skill document the contract.
- CHANGES.md: add 2.19.1 with its security fixes and the remaining bug
fix.
AI-assisted-by: deepseek-v4.1-flash
The changelog generator copies the note lines verbatim, so a note
hard-wrapped at 80 columns ends up wrapped in CHANGES.md. Record
that the breaking-change note must keep one line per paragraph or
bullet, in both the issue and changelog workflow memories.
AI-assisted-by: inkling-2-small-free
* 🐛 Restrict the team invitation token and list to admins
The token of a team invitation is a bearer credential: whoever holds it
accepts the invitation with the role of the invitation row. The RPC that
mints it only required edition permissions, while the rest of the
invitation commands require admin, so a team editor could get the token of
any pending invitation, including the ones that grant the owner role.
The token now requires admin, and the token of an owner invitation is
restricted to the team owner: an admin cannot create nor promote an owner
invitation, so it cannot keep its token either. It reuses the error code
already used for that rule.
Minting the token now leaves an audit event named as an action, carrying
the granted role and the id of the invitation, so the extraction does not
go unnoticed.
The list of pending invitations is admin only as well: it carries the
email of the invited person and the role offered, and the dashboard was
painting those emails to any member. The frontend catches the permission
error and shows the permission notice instead, and the copy action is not
offered for an owner invitation, which the backend rejects for an admin.
The five invitation commands now answer with the same permission error,
so they speak one language.
Relates to GHSA-4mmc-j8rh-74pf
AI-assisted-by: minimax-m3.1-flash
* 🐛 Clear the cached invitations and gate the list on team admin
The pending invitations are a team admin only list. The dashboard still
offered the entry to every member and keyed the section on the invite
policy of the organization, so a team admin blocked by that policy saw
the no-permission notice although the backend lets them list.
- `fetch-invitations` answers a denied refetch with an
`invitations-fetched` event carrying an empty list, so a demoted admin
does not keep the cached emails and the header no longer paints the
invite button from a stale list.
- The header tabs and the team options dropdown hide the invitations
entry without `:is-admin`, which already includes the team owner.
- `invitations-section-content` keys the section on `:is-admin`, and the
invite policy only gates the invite actions: resending keeps the
`can-invite?` check, deleting and row selection only need team admin.
AI-assisted-by: deepseek-v4.1-flash
* 🐛 Apply the share-link page scope to the comment queries and writes
A share link can be restricted to a subset of the pages of a file, and
the comment queries enforced that scope. But the RPC that resolves the
profiles of the commenters authorized the request with
check-comment-permissions! and dropped the returned perms, so its query
filtered only by file id. The holder of a page scoped link got the id,
email, fullname and photo of every profile that commented on any page
of the file, including the pages the owner left out of the link.
The query now takes the pages of the share link and only selects the
authors of the threads on them. The profile of the caller is still
added, so the viewer keeps its own avatar.
The scope was missing as well on the comment writes that take a page id
or a thread id: create-comment-thread accepted a page outside the link,
create-comment accepted a thread of an unshared page, and
update-comment-thread, update-comment-thread-position and
update-comment-thread-frame accepted a thread of an unshared page. They
now raise not-found, the same way the comment queries already did.
Replying to a thread of an unshared page also sent the emails to the
thread participants.
The page scope rule now lives in one place: share-link-pages returns
the pages granted by a share link and check-page-scope raises not-found
when a page is outside of it, so get-comment-threads reuses the helper
instead of checking the pages inline.
The docstring of get-profiles-for-file-comments said that it returns
the participants of all the comment threads of the file, which is not
true anymore for a share link. The docs are public, so the text and
the changes list are updated.
The tests of the share link page scope are hardened: the setup asserts
that the comment threads are created, the membership case has a second
commenter so it can no longer pass for the wrong reason, and the link
without pages, the allowed paths of the thread updates and the
who-comment=team denial are covered.
Relates to GHSA-89cq-3wv7-987r
AI-assisted-by: minimax-m3.1-flash
* 🐛 Enforce the share-link page scope on the remaining comment writes
Check the page scope of the share link before mutating in the four
comment writes that still skipped it: update-comment-thread-status,
update-comment, delete-comment-thread and delete-comment. The check
runs before the ownership checks, so an out-of-scope write fails with
not-found and the response does not reveal that the comment or thread
exists. Load the file in create-comment only after the permissions and
page scope checks, matching create-comment-thread, so a rejected reply
does not load the file data.
Cover the denial and the in-scope paths of the four commands, and the
fail-closed case of a link without pages on the comment reads. Move
the outsider thread creation after the profiles assertion so the
caller-union assertion fails when the union is removed. Document the
page scope invariant in the backend permissions memory.
AI-assisted-by: deepseek-v4.1-flash
* 🐛 Escape SVG values and sanitize every SVG upload
The SVG sanitizer removed dangerous elements and attributes by name and
then wrote the tree back with clojure.xml/emit, which escapes nothing.
A payload hidden in an attribute value or in a text node became live
markup again: an id of "x' onload='alert(1)'" was stored as a real
onload handler, and an entity encoded script tag was stored as a real
script element.
Write the filtered tree with a serializer that escapes text and
attribute values, and route every path that stores an uploaded SVG
through it. The thumbnail writers (create-file-object-thumbnail and
create-file-thumbnail) and the organization logo stored the upload
untouched, so they could plant live markup in a public bucket.
The filter only knows about SVG elements, so it also accepted
documents that are not SVG at all: a page with an iframe or a meta
refresh was stored as an image. The sanitizer now rejects any document
whose root is not an svg element in the SVG namespace, and it compares
the scheme of an url the way the browser does, so a scheme split with
a tab or a line break no longer slips through the javascript: filter.
The attributes are written sorted by name, because the parser hands
them over in the reverse order of the document and the output changed
on every sanitization.
The profile picture of an identity provider is not stored either when
it is an SVG: the avatar is always rendered as an image, so an SVG
would only add the risk of keeping live markup in the storage. The
dimensions of an SVG are now read from its local name, so an SVG with
a namespace prefix can be uploaded like any other.
The logo had no size limit and the sanitizer reads the whole file, so
it now honors the media max file size.
The output shape changes with it: no line breaks, double quotes, and
the single quote of an attribute value escaped as an entity. The hash
of every stored SVG object changes with it.
The media tests use cuerdas.core instead of clojure.string, as the rest
of the codebase does.
Relates to GHSA-qwhj-rchj-58mg
AI-assisted-by: minimax-m3.1-flash
* 🐛 Drop non-SVG namespace subtrees in the SVG sanitizer
The sanitizer only checked the namespace of the root, so a child could
rebind xmlns to XHTML and smuggle live HTML (iframe/srcdoc/meta) past the
blocklist. The serializer escapes attribute values, but the browser decodes
them again, so srcdoc stays live.
Track the namespace bindings through the recursion and drop any element
that is not in the SVG namespace, with its subtree. The root is validated
up front with the same model, so a document that is not an svg document is
rejected with an honest hint instead of a misleading one.
The default namespace is stored under a namespaced key, because a prefix
literally named `default` collided with the internal key and re-opened the
same hole through a prefixed binding. The attribute filter now keeps
`xmlns`/`xmlns:*` declarations, which are not event handlers, so a prefix
named like one does not end up undeclared in the output.
The blocklist also drops HTML-only elements written in the svg namespace
(iframe, meta, embed, link, base).
Add regression tests for the child rebinding, for the prefixed variants,
for the `default` prefix, for the root local name, for the
create-file-media-object-from-url path (mocked download), and for the
HTML-only tags. Assert double-quoted attributes in the clean content test.
Relates to GHSA-qwhj-rchj-58mg
AI-assisted-by: longcat-2.5-free
* ⚡ Keep the translation flag on undo and redo of a move
Previously undos and redos always forced a component sync + thumbnail
rerender for every component touched even if the underlying commits were just
translations. This propagates the translation flag to undos and redos and
propagates it to commit-changes. An entry merged from several undos/redos is
a translation only if all of its undos/redos are translations.
This reduces a UI freeze when undoing a move of a board with 453 components
from 2.42s -> 0.22s.
AI-assisted-by: claude-sonnet-5-5
* :translation? -> :translation
---------
Co-authored-by: Andrey Antukh <niwi@niwi.nz>
* 🔧 Bump Rust version of the devenv image to 1.99
* 🔧 Run the pinned Rust setup from the lint and format scripts
./build and ./test switch to the Rust version in render-wasm/_build_env
through setup, but ./lint did not, and formatting called cargo fmt
directly. In CI both ran on the devenv image's default toolchain.
Make ./lint run setup and add a ./fmt script that does the same, so all
render-wasm entry points use the pinned version. CI and scripts/ci now
call ./fmt instead of cargo fmt.
AI-assisted-by: claude-opus-5-5
The memories claimed link-issue does not close an issue
retroactively when its PR is already merged. Adding 'Closes #12896'
to merged PR #12067 and running link-issue closed the issue
immediately instead (observed 2026-11-09). Update the
mem:workflow/creating-issues, mem:workflow/creating-prs and mem:scripts/gh
entries to describe the observed behavior.
AI-assisted-by: inkling-2-small-free
* 🐛 Return the full profile from the LDAP invitation branch
The invitation branch responded with only {:invitation-token token},
so the session start received a map without id or auth-backend. With
the new auth-origin vocabulary that map fails validation and the login
breaks in development and tests; in production the event would be
emitted without a channel.
Return the profile plus the token, like the regular branch, and cover
both branches with RPC-level tests.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* 🐛 Mint an auth token in the OIDC callback with the invitation inside
The callback used the invitation JWE as the session token, so the
frontend landed directly on token verification and never ran the
regular session start: no signin event, no teams, no host hook.
Always mint an :auth token and carry the invitation in its claims.
Token verification returns the claims verbatim, so login-from-token
attaches the invitation to the profile and logged-in does the rest,
the same pattern the verify-email token already uses.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* ✨ Keep the session origin in the telemetry shadow rows
The anonymisation filter drops every string unless it is on a short
allowlist. Add source and activation next to lang and auth-backend so
the shadow copy keeps them; the email and the full name are still
dropped and the email domain is still derived from the email.
The filter keeps its two bindings: the transducer discards everything
that is not a UUID, a boolean or a number, so the allowlist is applied
on the filtered map, never on the raw props.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* ✨ Describe the signup outcome with activation, not with action
The three success branches of register-profile labelled a new account
as "login", "email-verification" or "accept-invitation", so the event
contradicted its own name and the value only lived in context, where
the analytics consumers never read it.
Add activation to the event props and drop action from those branches:
the name already says it is a signup, and a constant action would only
add noise. The retry branches keep their action in context because
there it distinguishes four different outcomes.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* ✨ Describe the session origin in the signin event
The event carried the profile but never said why the session started:
the only field that tried was frontend metadata that never leaves the
browser, and several flows filled it with the wrong value.
Add a small closed vocabulary (source, auth-backend, activation,
from-invitation) and compose it from the profile when signin is
emitted. The assertion fails in development and tests and is elided in
the release build, so a bad value cannot break a real login.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* ✨ Emit the session origin from every login entry point
The four emitters stamped a free-form string as metadata, so the
session event could not rely on it: two flows that were signups wrote
the same value as a login.
Route them through origin-meta: password and LDAP logins declare
source login, register declares source register with activation
immediate, and login-from-token derives both from the token issuer:
:auth for a login and :verify-email for a signup that needed the
email confirmation.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* 🐛 Send the invitation signups through the regular session start
Registering from a team invitation link navigated by hand to token
verification, so the session start never ran: no signin event, no
teams, no host hook. The same navigation was copied in four places.
Teach login-from-register to carry an optional invitation token,
extract the profile decoration into with-register-origin, and call it
from the four sites instead of the hand-written navigation.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* 🐛 Resolve the invitation without waiting for the teams
Registering from an invitation link now goes through the session
start, and that start only navigates after the teams request
completes. fetch-teams has no error handling, so a network failure
there would leave the invitation unaccepted and the user on the signup
screen with a session already minted.
Move the invitation navigation out of the teams path: it emits as soon
as the profile carries a token, and the teams path emits nothing for
an invited profile so it cannot move the user away from the token
verification screen.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* 📚 Document the session origin vocabulary and the invitation flow
Add a focused memory with the four fields, where each value comes from,
the table of flows, the LDAP exception and the invitation path through
logged-in, and link it from the frontend routing.
Update the audit-log memory: source and activation are now on the
telemetry allowlist, activation lives in props and action only exists
on the retry rows.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* 📎 Cover the invitation session start and fix the auth memories
The invitation flows and the token decoration in login-from-token were
only covered by helper tests, so removing the guard that keeps the
teams path from navigating, or the token attach, broke the behaviour
silently.
Run the events through ptk/watch with a teams-fetched event as input:
an invited profile navigates to the token verification screen and
never to the dashboard, and the signin keeps from-invitation. Add the
login-from-token cases with and without the token, and fix the memory
statements about action, the flow table and from-invitation.
Closes#12849
AI-assisted-by: deepseek-v4.1-flash
* 🐛 Normalize stored SVG attribute keys to kebab-case
The v3 reader already emits kebab-case keys, but birth stored the
camelCase React prop names, so import and export never converged:
fill rules painted wrongly and filter defs lost their blur.
Birth now normalizes every stored attribute key to kebab-case
through the `svg-props` whitelist (`attrs->kebab-props`), applied on
all element paths of the svg importer, on the file builder and on
defs. svg-raw `:content` is kebab-ized in spelling only and keeps
unknown keys, because the sidebar reads them. Non-spec attrs Penpot
must keep live in `penpot-extra-attrs`, seeded with `:stroke-style`.
Render boundaries translate back to camelCase for React and for the
wasm props, and the shape schema declares the closed kebab contract
derived from that same whitelist, so birth and validation cannot
drift.
Migration `0035-fix-shape-svg-attrs-and-defs` converts old files
once with the birth transforms, and carries back the svg-viewbox
rect normalization the withdrawn `0021` had. That old id leaves
`available-migrations`, so files recording it migrate again through
this one, and the binfile cleaner drops its own svg repair.
Review fixes included here: `setup-opacity` and the blend mode read
the stored keys, gradient stops accept `:style` as a CSS string or a
map, `inherit-attributes` reads `:style` so a child's inline style
beats the inherited group attr per CSS, and the svg-attrs menu shows
spec spelling (`viewBox`, `class`) through `stored-attr-display-name`.
Tests cover the migration, the closed svg-attr schemas and the
import/export round trip; the svg-content schema test pins its
sample seed, which dealt no svg-raw shape on unlucky runs.
Closes#11947
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Keep legacy svg node tags and transforms valid in migration 0035
The closed svg node schemas require keyword tags and map attrs, but
stored files hold def and svg-raw content tags as strings, and some
node `:attrs` as nil. Migration 0035 left them as they were, so the
file data check that runs after migrating rejected those files and
they could not be opened. `normalize-node-tree` now turns string tags
into keywords and nil attrs into an empty map, at birth and in 0035.
With keyword tags, the svg defs renderer applies the shape transform
to user-space defs. Legacy `:svg-transform` values are plain maps,
which multiply into NaN matrices. `gmt/coerce` turns them into
matrices, in 0035 and in the binfile cleaner, because every v3 import
decodes the transform as a plain map again. Legacy masks and gradients
now render as designed in the svg renderer.
The binfile test helper calls `export-files!` with the current config
keys.
AI-assisted-by: claude-opus-5-5
---------
A component whose :main-instance-id points to a missing shape fails
validation on its copies with :ref-shape-not-found. A repair scoped to
the copy's page never sees the main's own error, so it detached the
copy, even when the main was still on the component's page.
The repair of :ref-shape-not-found now looks for the component's main
on its page first and points :main-instance-id back to it, which keeps
the copy linked. A copy is still detached when no main is found.
A rollback reverts the local data with a rebase, which is not a commit,
so the parts of the workspace that follow commits missed it:
- Comment pins moved in the reverted action stayed where they were, on
screen and on the server, and pins on a board the rollback removed
kept pointing to it. They go back to the reverted position, and pins
on a removed board move to the page root.
- Focus mode stayed on with no shape left when the rollback removed the
focused ones, and the next shape drawn turned it on again. It ends
when its last shape goes, and the selection drops removed shapes.
- Component thumbnails kept the rejected look. The thumbnails of the
reverted mains are rendered and uploaded again, and an upload whose
shapes changed during the render is skipped.
- Under WASM, an open text editor reset to the reverted content but the
canvas kept the old layout. It is laid out again.
- A library sync whose save failed showed no update notice again once
the release undid it. The notice is checked again after that
rollback.
Before it repairs a damaged file, the backend keeps a copy of it as a
system snapshot, and the repair notice points to the version history to
find it. The History panel listed that copy among the autosaves, so it
could not be told apart.
The History panel lists the copies kept before a repair as their own
entry per day, "Before automatic repair, <date>", above that day's
autosaves, and the preview of such a copy uses the same label.
The notice of a rejected save always said the same thing. It did not
name the file when the user had moved to another one, gave no hint when
the page held data that was already invalid, and replaced the notice of
a backend repair still on screen, since the store holds one toast.
The backend flags a rejection whose first commit touches pages or
components that failed validation before it (:preexisting-damage). The
rejection notice names a file other than the open one, says when the
data was damaged before, and folds into a repair notice still on
screen; a later repair notice folds a rejection notice in the same way.
The notice says that the latest changes were undone, no longer that the
rest of the work is saved.
The answer to a save lists the changes other sessions stored before it,
such as the backend's repair of a damaged file. When the answer came
before the message with such a change, the saved commit was
acknowledged first, and the late change was then applied on top of the
local edits instead of under them, out of the server's order.
A saved commit of the open file now stays pending until this tab has
received every change the answer lists from other sessions, so a late
change is applied under the local edits. If one does not arrive within
10 seconds, the file reloads.
When a peer restores a version, the file reloads to it. Edits made on
the replaced version stayed queued: their save failed for good and
locked the workspace, and a rollback or a rebase replayed them on the
restored data.
Loading another version forgets the queued, buffered and saved commits
made on the old one. A save of such a commit, in flight or answered, is
dropped instead of failing or being rolled back, and one save carries
the edits of one version only.
While saves fail the workspace turns read-only, but only some tools
checked it. Edits from the asset, color and token panels, and a drag or
a drawing in progress when the lock came, still made commits that
queued behind the failed save.
commit-changes refuses commits while the save-failure lock holds,
except those that are not user edits: the undo of rolled-back commits,
text position data and component touches. Undo and redo do nothing
while locked. When the lock comes, moves, resizes, rotations, drawings
and the drags of flex, guide and grid handles in progress are cancelled
and commit nothing, and the design panel shows the Inspect view.
The MCP plugin sends the rollbacks it saw with its next task response
and then forgets them. When that response came after the server had
given up on the task, the server dropped it and the agent never heard
about the rollback. A reload of the tab or of the plugin lost them too.
The server now sends the rollbacks of such a late response back to the
plugin, which reports them again with its next response. The plugin
keeps the rollbacks not reported yet in its local storage, one entry
per file, so a reload does not lose them. Both MCP plugin manifests ask
for the allow:localstorage permission for that.
Plugin writes failed whenever the workspace read-only flag was set. The
Inspect tab sets it too, although it only hides the editing tools, so an
MCP agent could not edit while the user was inspecting. A user who
became a viewer had the flag cleared, so plugin writes did not throw and
silently did nothing.
Writes now fail on what makes them unsafe, not on the flag: failing
saves, a user who cannot edit, a version preview or a lost render
context. The error says which one, and the MCP instructions tell the
agent which of them pass by waiting.
The backend rejects a save after the call that made the changes has
returned, and the workspace then undoes them. A plugin, or the MCP
agent, went on working with shapes that were no longer there and never
learned why.
The rollback event now names the files that lost commits and the pages
and page shapes their changes touched, and plugins get it as the new
`contentrollback` event (`ContentRollback` type) through a subscription
to the store stream, so no rollback is dropped by debouncing.
The MCP plugin keeps the rollbacks it is told about and sends them with
its next task response, failed or not. The server collects them per tool
call and puts a warning first in that tool's response, naming what was
affected; the instructions tell the agent to read the pages again and
redo what is missing.
A read-only workspace (saves failing, a user who cannot edit, a version
preview) only hid the editing tools: plugins, the MCP among them, still
changed shapes, and those edits were never saved. Every plugin write now
throws a read-only error, whatever the throwValidationErrors flag, the
same way a write to a removed shape does.
`check-permission` with a write permission runs the new `check-writable`,
which covers most writers. Writers without a permission check (most
context methods such as create*, group, align or flatten, the z-order
methods, applyTypography on a text range, removing a ruler guide) call it
themselves, and shape setters get it through `check-shape-writable`.
Reads, comments and undo blocks keep working, so a block opened before a
lock still closes.
The MCP instructions tell the agent what the read-only error means.
- Ending a drag inside the frame the shapes already belong to sent an
empty reparent save and bumped the revision. The move commits only
when some shape changes parent.
- When a user restored a version, other sessions reloaded the file with
no notice. They now see a dialog naming who restored which version
(by its name, else its date), saying that later changes are no longer
in the file and, when they had changes not saved yet, that those were
lost.
- Exports drawn in the browser (PDF, SVG, and PNG when the WASM renderer
draws it) and duplicating a file from the dashboard started at once,
so an export could show changes a pending save then rejected, and a
copy could miss queued edits. Both wait for the file's pending saves.
A client export warns that it includes unsaved changes; a copy warns
that it misses them.
A rollback can remove or change shapes the UI still refers to.
- The selection drops shapes that are gone and the color picker that
edited them closes.
- The undo history is cut at the reverted action; it stays when no
entry holds a reverted change. Undo and redo commits carry real undo
changes and the history index they moved from, so a reverted Ctrl+Z
puts the change and the index back.
- An open path edition whose shape the rollback changed closes without
saving the editor's stale copy. The v2 text editor loads the reverted
text again; the WASM editor already showed it.
- The Alt measurements look the hovered shape up in the current data.
- Plugin setters, resize() and rotate() on a removed shape throw
"Shape not found", whatever the plugin's throwValidationErrors flag.
- A save that failed on a short platform failure (a lost or refused
database connection, lack of resources, a server shutting down, a
serialization failure or deadlock, a socket or HTTP timeout, a host
that did not resolve, an interruption) answered 500, which stops
autosave. It now answers 503, which the frontend retries. Other I/O
and database errors keep answering 500, timeouts 504.
- A save answered not-found no longer replaces the workspace with the
access page. It is a terminal save error: the workspace stays open
with the failed save notice, report and reload link. When edit rights
are lost, the notice says how many changes could not be saved.
- A library sync marked the library synced three seconds later,
whatever happened to its save. A rejected sync was undone but stayed
marked: copies stayed out of date and the "updates available" notice
never came back. The mark now waits for the save, a rollback of the
sync brings the notice back, and a failed mark (e.g. the user just
became a viewer) is logged instead of leaving the workspace.
- While saves fail the workspace is read-only. The view-mode banner
says editing is disabled because changes can't be saved, with the
Inspect banner's size and bold lead and no Done button. Users can
still switch pages, pan, zoom and inspect. The lock lifts when a save
lands and brings back the mode the user had or picked meanwhile; the
WebGL context lock keeps the workspace read-only if it lasts longer.
The lock lives in app.main.data.persistence.lock.
- A viewer is never locked. A role change that leaves the user a viewer
keeps the edits that lost permission, queued and on screen, and shows
the count notice again. The role change that brings edit rights back
releases them and rolls them back on screen, as a rejected save would.
Any other terminal failure is released by the next role change or file
opened.
The debug faults `io` (socket timeout, 503) and `eof` (500) cover both
sides of the I/O rule.
A rollback reverts whole undo groups, across files and saves, so that
no user action is left half done. Some actions still split:
- An action that commits several times inside one undo transaction
(e.g. applying a token to several shapes) gave each commit its own
group, so a rejection could revert part of it: shapes ended up linked
to a token whose value was never applied. An undo transaction now
opens with a commit group that every commit made inside it takes.
- The commit that undoes saved commits of a rejected action joined an
open transaction's group, so a later rejection there reverted the
undo too. It gets a group of its own.
- Emptying the open transaction during a rollback dropped its group, so
its later commits were no longer one unit. It keeps its group.
- Waiting for a save ended as soon as the queue emptied, while later
commits could still wait in the buffer. Waiters now also wait for an
empty buffer.
A test covers a token action of three commits in one save whose third
commit is rejected.
The update-file tests each built the same rect change and the same
change that fails while applied. The rect change moves to the test
helpers as th/add-rect-change and the failing one gets a named helper.
The test of a database failure during the retry no longer counts calls
to process-changes: it fails every attempt that leaves the failing
commit out.