Both conflicts are tail appends onto regions develop rewrote. In
migrations.clj keep both appended blocks, develop's two entries
first so the branch's four stay the tail of the vector. In
climit.edn take develop's rewritten entries and re-append the
branch's provisional block, moving the map's closing brace to the
last line.
AI-assisted-by: glm-5.3-flash
Signed-off-by: Álvaro Tejero Cantero <alvorithm@teje.ro>
* 🐛 Detect and repair children missing from their parent's list
A shape could point to an existing parent that did not list it in its
:shapes. The validator missed it, since it only walks the tree from
the root plus the shapes whose parent does not exist. Operations that
walk :shapes missed it too: deleting a variant also deleted its
container, which looked empty, and left the other variants without a
parent, so the backend rejected the change and the file stopped
saving.
- get-orphan-shapes also returns shapes not listed by their parent,
so :child-not-in-parent is reported.
- The :child-not-in-parent repair adds every unlisted child of the
parent. All repairs share the same file data and each one sets the
whole :shapes, so with several children only the last one was kept.
- New migration 0032-fix-children-not-in-parent repairs the affected
files when they are opened.
Closes#11758
AI-assisted-by: claude-opus-5-5
* 🐛 List shapes moved to root by the parent-not-found repair
The :parent-not-found repair set the shape's parent to the root but
did not add it to the root's :shapes, so it left the shape in the
child-not-in-parent state that the validator now reports. It now
adds every shape without parent to the root at once, since all
repairs are built from the same file data and each one sets the
whole :shapes.
Also cover the missing parent case in the repair tests.
AI-assisted-by: claude-opus-5-5
* ♻️ Simplify unlisted children validation and repair
Address review comments on the validation and repair of shapes that
their parent does not list.
- get-orphan-shapes no longer rebuilds *children-sets*: both callers
bind it. The lookup moves to a listed-in-parent? helper shared with
check-parent-children, which scans :shapes when the var is unbound.
- The :child-not-in-parent and :parent-not-found repairs no longer
walk every shape of the page. Each one moves only its own shape with
pcb/change-parent. The :mov-objects change adds the shape to the
parent's :shapes when applied, so several repairs on one parent no
longer overwrite each other.
- The :child-not-in-parent repair test uses plain shapes instead of
variants.
AI-assisted-by: claude-opus-5-5
* ♻️ Move some constants out of app.render-wasm.api
- Moved to aapp.render-wasm.wasm where they can be imported
without pulling half the editor with them
- Renamed one for consistency
* 🔥 Delete unused function, update docstring
* ✨ Add stroke side-width helpers to common types
Add set-width-to-all-sides, set-width-to-single-side and side-width to
app.common.types.stroke, mirroring the border-radius helpers in
app.common.types.shape.radius. They back the per-side stroke width API
and are covered by unit tests.
AI-assisted-by: deepseek-v4.1-flash
* ✨ Round-trip per-side stroke widths through the plugin API
Add strokeWidthTop/Right/Bottom/Left to format-stroke and parse-stroke so
reading shape.strokes exposes the per-side widths and writing a stroke
back preserves them instead of collapsing to the uniform width.
parse-stroke prefers a live stroke proxy's hidden raw state when present,
so a commit only rewrites the attributes the plugin touched.
AI-assisted-by: deepseek-v4.1-flash
* ✨ Add per-side stroke widths to the plugin API types
Add the four per-side members to the Stroke interface and to the
borderWidth and dimension token property lists, and refresh the generated
API surface used by the plugin test suite.
AI-assisted-by: deepseek-v4.1-flash
* ✨ Expose per-side stroke width on the stroke proxy
Add strokeWidthTop/Right/Bottom/Left to StrokeProxy. Reads fall back to
the uniform strokeWidth; strokeWidth now writes all four sides, like
borderRadius writes all four corners; each side setter writes only its
own key and the top side also updates strokeWidth.
Reject per-side writes on shapes other than boards and rectangles with a
validation error naming the property, and expose the raw stroke map
through a hidden proxy member so a commit preserves untouched attributes.
AI-assisted-by: deepseek-v4.1-flash
* ✨ Support stroke width tokens per side in the plugin API
Token attribute expansion now accepts the uniform strokeWidth property and
expands it to the four per-side attributes, mirroring the design tab.
shape.tokens exposes the per-side keys and reports strokeWidth when all
four sides share one token, and applyToken routes through the shared
expander.
AI-assisted-by: deepseek-v4.1-flash
* 📚 Document per-side stroke width in the plugins changelog
Note the new Stroke members, the restored strokeWidth token property and
the shape-type validation in plugins/CHANGELOG.md.
AI-assisted-by: deepseek-v4.1-flash
* 🐛 Fix some minor issues
---------
Co-authored-by: alonso.torres <alonso.torres@kaleidos.net>
Setting the local file back as the tokens source, or unlinking the
library that was the source, stored the file's own id in
:tokens-source. ensure-tokens-lib only created a tokens-lib when
:tokens-source was nil, so a file without a tokens-lib never got one,
and every token change was dropped without any error.
Now ensure-tokens-lib also creates the tokens-lib when the source is
the file itself, which fixes files already in that state, and
set-tokens-source removes the attribute instead of storing the file's
own id.
Closes#12180
AI-assisted-by: claude-opus-5-5
* ♻️ Factor GL context initialization out of canvas init
- fixes setting of gl-context-handle to be after _init so an exception
does not leave invalid state behind.
- api.init-canvas-context uses a new webgl.init-context!
- this decouples config from init, so we can reuse it in benchmarks
- some helpers move to the webgl ns
AI-assisted-by: muse-spark-1.3, deepseek-flash, glm-5.3
* 🐛 Release acquired GL context when renderer init fails
- Properly release resources if init fails
- Rethrown exceptions with the original cause
- Fix vacuous check in clear-canvas
AI-assisted-by: muse-spark-1.3, deepseek-flash, glm-5.3
* ✨ Test webgl context init and release
AI-assisted-by: muse-spark, glm-5.3
* 💄 Address review comments
* 💄 Drop `!` from some function names
A plugin with "scope": "global" keeps running across the dashboard
and files and can find and open the user's projects and files. Any
manifest could ask for this without the user being told.
Add the allow:global permission. The plugin registry adds it to
every manifest with global scope, so the install dialog lists it,
and a plugin that becomes global in a later version shows the
permissions update dialog again. The permission is valid in the
shared schema, which the backend checks when saving, and in the
runtime manifest schema. The built-in MCP plugin declares it too.
Document the scope property and the new permission for plugin
developers.
Global plugin lifecycle:
- Add optional manifest scope "global" for plugins that remain active
across dashboard and workspace navigation within one browser tab.
Existing plugins retain workspace scope by default.
- Preserve global sandboxes when a workspace ends or another plugin
loads, including when that other plugin fails to initialize.
- Avoid duplicate instances of an already-running global plugin.
- Close global plugins on explicit closure or logout, with idempotent
cleanup of listeners, timers and UI resources. Global scope does not
imply automatic startup or persistence across a browser reload.
Plugin API and navigation:
- Keep a permanently live penpot facade whose getters follow the current
workspace, while user, theme and UI facilities remain available on
the dashboard.
- Expose penpotMgmt to global plugins with workspace metadata,
none/loading/ready status and workspacechange subscriptions.
- Add awaitable openFile(fileId, {teamId}) navigation in the current tab,
defaulting to the current team and resolving only once the requested
file and its active page are ready.
- Reject invalid identifiers, failed or superseded navigation and waits
exceeding 30 seconds; release navigation watchers on completion.
- Guard workspace-bound operations while no workspace is ready and
return null from root, currentFile and currentPage in that state.
- Require content:read for the initial management API. Callers must
reacquire file, page and shape references after changing files.
MCP lifecycle and connection controls:
- Make MCP a global plugin and initialize its integrated controller per
authenticated application session instead of per workspace.
- Load the enabled integration with a valid token on the dashboard as
well as in a file, retaining its existing hidden-iframe implementation.
- Keep enabling and connecting separate: an enabled plugin starts idle,
and each tab connects explicitly through dashboard or workspace
controls.
- Preserve the iframe, WebSocket, heartbeat and reconnect machinery
during navigation; disconnect on request and close on disable/logout.
MCP sessions and execution:
- Derive the short MCP session ID solely from currentUser.sessionId,
excluding the file ID. Retain the SHA-256-based, 50-bit Base32 encoding
to produce a stable, copyable 10-character identifier.
- Allow sessions without an open file and update workspace metadata over
the existing connection without replacing its routing identity.
- Refresh context after the connection handshake so initialization does
not leave stale file metadata.
- Expose penpotMgmt to execute_code, preserve execution storage across
navigation and skip layout waits when no workspace is ready.
- Update API types, tool guidance, documentation and the changelog.
Validation:
- Cover authenticated startup, workspace survival, logout cleanup,
readiness, navigation failure, duplicate loads and session routing.
- Verify that short IDs remain stable across files and reconnects and
that disconnect cancels pending asynchronous ID initialization.
- Pass 16 focused frontend tests, 73 runtime tests, 30 MCP plugin tests,
builds, TypeScript checks and frontend/runtime lint.
- Pass all 73 MCP server tests with --test-force-exit; the normal runner
still stays alive after assertions finish.
- Verify a live dashboard-to-file-to-dashboard round trip without a
document or iframe reload, preserving facade identity and storage.
AI-assisted-by: gpt-6
Two comments cited an internal work-tracker id that a code reader
cannot resolve and that the tracker may rename. Both now say what the
rule says: one side keeping its value in a merge is a recorded choice,
never a silent win.
AI-assisted-by: glm-5.3-flash
Signed-off-by: Álvaro Tejero Cantero <alvorithm@teje.ro>
Stroke width tokens used to be applied to the single `:stroke-width`
token attribute. Applying tokens per side replaced it with the four
`:stroke-width-*` attributes, but no migration moved the existing
references, so files written before 2.19 look unapplied: the token
pill stays off, the Design tab shows a plain number and propagation
finds no update action.
Add migration 0031, which copies the legacy reference to every side,
keeps any side reference already present and drops the old key. It
covers page and component shapes.
Closes#12016
AI-assisted-by: deepseek-v4.1-flash
Wrap every id prop in the comp-processors warn logs
with str so the log shows plain text instead of
the #uuid reader form. Also use :hint instead of
:msg and lowercase text, as the rest of the system
does. No logic changes, logging only.
AI-assisted-by: muse-spark-1.3
The branch compare panel now offers per-change selection, grouped by
the component each shape belongs to. Every summary entry carries a
stable change key (kind/page/id, the design-branches PoC format), and
the panel's checkboxes key on it.
merge-file-branch and update-branch-from-main take the deselected keys
as an excluded set. An excluded change is dropped on the user's word:
a conflicting entity resolves to the target side (the recorded choice,
per the no-silent-winner decision), and an excluded addition cascades
to its added descendants, so no emitted change references a parent the
merge does not create. A merge whose every change is excluded refuses
with :nothing-to-merge.
The update squash needs no filter: whatever excluded main work is
missing from the branch lands in the squashed net as a branch-side
delta, so the branch state still replays from its repositioned base.
AI-assisted-by: glm-5.3-flash
Signed-off-by: Álvaro Tejero Cantero <alvorithm@teje.ro>
Adding a flex or grid layout to a board with children derives the
paddings from the distance between the board and its children. A child
that overflows the board, or floating point noise after "Resize board
to fit content", gave negative paddings. The shape schema rejects them
since paddings must be non-negative, so update-file failed and autosave
stopped. Clamp the derived paddings to 0, as the derived gaps already
are.
Closes#12107
AI-assisted-by: claude-opus-5-5
* 🐛 Sanitize smuggled SVG on binfile import
Route imported SVG storage objects through the SVG sanitizer on
all three binfile paths (v3, v1, v2), closing the second bypass
of GHSA-ffhp-m958-qxvr. Integrity checks still run on the raw
bundle bytes first; only the persisted copy is the sanitized one.
Adds a shared sanitize-imported-svg helper with unit tests and a
v3 export-tamper-import test proving smuggled scripts no longer
survive the import.
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Normalize SVG content-type on binfile import
Close the residual GHSA-ffhp-m958-qxvr bypass where a crafted
content-type spelling (uppercase, parameters) skipped the import
sanitizer. Detection now uses a shared case-insensitive predicate,
stored values are canonicalized, and the v3 re-read honors the
import size limit.
Extends the tamper tests to obfuscated spellings and adds
exhaustive v1 coverage, including the tempfile branch.
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Gate binfile content-type on known storage types
Define the complete set of storage content-types in
app.common.media and enforce it on the binfile import schema, so
unknown types fail closed instead of passing through. Detection
keeps a single normalization at the boundary with an exact
predicate, and the string helpers use cuerdas.
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Keep only produced types in storage-object-types
Drop apng, avif, penpot and plain text from the set: no flow
stores them, they only exist in the extension mapping table.
Every member must have a producing code path; unproduced types
stay out until some flow actually stores them.
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Reject unknown content-type on v1 binfile import
Enforce the storage content-type allowlist on the v1 import
path, like v3 already does via schema. Bundles declaring types
outside the set, or none at all, fail closed with the same
media-type-not-allowed error as uploads. The sanitize branch is
explicit: SVG bytes are sanitized, anything else passes through.
AI-assisted-by: muse-spark-1.3-contributor
Branches and databases from the preview's earlier shapes are not
supported, and such branches are recreated. The earlier shapes are the
copy model, where a branch file kept its own data payload, and the
tables registered as 0150-add-file-branch-table,
0152-add-file-branch-table and 0153-add-file-pull-request-tables. This
commit removes the code that existed only for them. A fresh database
ends with the same schema as before.
- 0155-add-file-branch-table.sql loses its header about the old
registrations, every IF NOT EXISTS, and the base_branch_revn upgrade
and backfill. CREATE TABLE file_branch already declares the column.
- 0156 and 0157 lose every IF NOT EXISTS.
- 0158-add-file-branch-change-data-version folds into 0157, and its
file and registration go. CREATE TABLE file_branch_change declares
data_version NOT NULL, because both writers of the op log, a branch
save and the squash of an update from main, stamp every row.
- need-migration-to? loses its nil case, which replayed an unstamped
op-log row at the base's version, and the branch derive loses the
comment that described it. Stamped rows replay as before.
- The tests for a re-run of the branch table migration and for the
replay of an unstamped row go, with the require the first one used.
A database that holds the preview's older shapes must drop its
branching tables and the file.is_branch column, and delete the
branching steps from its migrations table, before it boots this code.
A database that ran 0157 without 0158 lacks data_version and counts as
an older shape. A database that already ran 0155 to 0158 under these
names keeps working, because a migration runs once per name.
AI-assisted-by: claude-opus-5-5
develop moved token activation out of the tokens lib into the document's
:tokens-status (penpot/penpot#10293): ctob/get-active-theme-paths is
gone, :set-active-token-themes is no longer a change type, and a set is
activated by id in the status rather than through the hidden theme. The
branch merge still read active theme paths off the lib, which on the JVM
is a compile error in app.common.files.branch-merge and therefore stops
every backend namespace that requires it from loading.
The merge now reads both halves of activation off :tokens-status, as ids,
and keeps the two conflict kinds and resolution keys the UI already
offers (:active-themes, :active-sets). compute-changes settles each half
by the rule it used before and emits one :set-tokens-status, because the
change replaces the whole status. The hidden theme stays out of the
theme diff, as before.
The two activation tests now build the status with develop's own
cfo/activate-set and cfo/activate-theme, and a third pins that a
conflicting activation keeps main's status until a resolution takes the
branch's.
common-tests.files-branch-merge-test: 84 tests, 331 assertions, 0
failures.
AI-assisted-by: claude-opus-5-5
Depth gate: the compare reads the op log and applies the depth gate
before it derives the branch, the same order merge and update from
main use, so a log too deep to replay refuses before the derive runs.
The refusal names the materialise and the delete, the two remedies the
gates do not refuse, in the hint and in the notification copy.
Resolutions: the conflict answers of merge-file-branch and
update-branch-from-main carry the main revision the conflicts were
computed against, and the client sends it as expected-main-revn when
resolutions are applied before the dialog's diff fetch returns.
Libraries: absorbing skips branch files, so unpublishing or deleting a
shared library writes no payload on a branch and the branch keeps its
own link; a soft-deleted library resolves nothing there until the
branch's next update from main brings main's absorbed copies. The
absorbed write passes the file's features to update-file!, which the
hand-built params dropped, so the pointer-map fragments the absorption
regenerates are persisted and no payload references a fragment no
write produced.
Queries: the project file count, the team file count, the list of
files that use a library, and the socket-REPL team duplication skip
branch files.
Snapshots: the branch refusal lives in check-restorable!, which the
restore command calls before its backup snapshot and the two
socket-REPL restore helpers reach through restore!.
Migration 0155: the base_branch_revn backfill repositions only the
rows whose branch file carries a data payload, so a re-run changes
nothing on the branches the op-log model creates.
Tests: the branch merge tests run in JavaScript, which the common test
runner lists; new backend tests pin the gates, the absorption skip,
the query filters, the restore refusals, the migration re-run, and the
conflict answers' revision; a new frontend test pins the token the
senders carry with resolutions.
AI-assisted-by: mixed models
Signed-off-by: Álvaro Tejero Cantero <alvorithm@teje.ro>
The branch comments called the set of ids a document's references are
written in (main's file id and media rows, or the branch's) a frame.
Penpot's code uses frame for a board, through `:frame-id`, so a reader
could not tell which sense a comment meant. The comments, the
docstrings of `get-base-data`, `branch-base-data`, and `remap-changes`,
and one test name now say id space. Every board-sense use of frame is
unchanged.
AI-assisted-by: mixed models
Signed-off-by: Álvaro Tejero Cantero <alvorithm@teje.ro>
Merge engine: a same-parent reorder, an addition under a parent main
deleted, and a branch deleting a container main edited each lost work
silently. Each now follows a policy read from
backend/resources/app/branch-merge-policies.edn, re-read when the file
changes, with the defaults: merge the reorder, and raise a conflict on
the deleted parent or container. Added siblings keep the branch's order.
A :mov-page index counts the target's full page list, a cleared page
background reaches main, a per-attribute resolution can remove an
attribute, and a token set reorder survives a rename on the other side.
Conflict payloads name :main and :branch by document in both
directions. Reorder detection costs nothing measurable on a page of
20,201 objects.
Data versions: each op-log row records the data version it was written
at, the branch replays its log in segments and migrates the document
at each boundary, and the comparison base is migrated like main and the
branch, so an upgrade no longer shows as false conflicts.
Writers: a binfile import, an admin repair, file GC and a snapshot
restore refuse a branch file, whose payload nothing reads, and the
branch base pin cannot be renamed into a user snapshot. A branch save
runs the ordinary media fix-up and logs the rewritten references. An
update from main deletes the branch's copy of media main deleted and
squashes no phantom media changes.
Commands and UI: the pull-request commands publish their messages after
the commit, the listing takes one connection per comparison and shows a
warning instead of "in sync" when a comparison fails, materialize takes
:branch-file-id, and a dashboard fetch that never answers ends in an
error state instead of a spinner, with a refetch on reconnect.
backend/dev/branch_merge_cases.clj builds one example file per merge
policy case for demonstrating the alternatives.
Two models wrote this change, claude-opus-5-5 and mimo-v2.6-pro, and
the work switched between them, so the trailer reads mixed models.
AI-assisted-by: mixed models
The source notes on files_branch.clj listed 32 review findings. This
commit fixes the 21 that are defects of the code; each fix carries a
regression test that failed before it and passes after.
Concurrency: merge, update from main, rename, archive and delete now
re-read the branch row under the advisory lock and work off that row,
so a concurrent merge or update can no longer leave a ten-year base
pin behind or release a superseded base twice. Delete and archive take
the branch file lock, create and reopen of a pull request refuse a
branch that a concurrent merge closed, and create checks the branch
quota under the lock. Merge, update and delete publish their msgbus
messages only after the transaction commits.
Merge, update and materialize re-validate after the repair: merge and
update refuse a result that stays invalid, materialize logs it. The
change log row carries the repair changes, and the three paths drop
the cached file summary under :redis-cache.
Contracts: a :conflict refusal answers 409 instead of 500, resolution
values must be :main or :branch, and update from main accepts the
expected-main-revn token the frontend now sends from a loaded diff.
Smaller fixes: create refuses a blank name, writes no file_data row,
and a rename reaches the branch file's own name. Deleting a kept
merged branch keeps its merged status. get-branch-diff reports the
creation time of the current base snapshot. Rename, archive and delete
report their duration and outcome. An image used only by a library
colour is copied into the target file. Dead fallbacks and dead
deleted-at checks are gone.
AI-assisted-by: claude-opus-5-5
Comments written for the duplicate-file prototype still described a
branch as a copy of its file, a merge as clean-only, token diffing as
missing, a review as a read-only snapshot, and file_change rows as
living for the team deletion delay. Each now states what the code at
this revision does: a branch is a file row derived from a pinned base
plus its op log, every change kind except :page-attrs is translated,
merges apply resolutions, a review opens the live editable branch, an
update squashes the log and repositions the base, and file_change rows
expire after one hour. The audited docstring names the props it sets,
the step 6 comment says each revn counter starts at its own side, and
handle-file-merged names both topics it arrives on.
Comment, docstring, and SQL-comment lines only. clj-kondo reports no
errors on the touched Clojure files. Translations are untouched.
AI-assisted-by: claude-opus-5-5
Two defects with one cause, which is that a comparison and a write want
different identifier frames and the code used one frame for both.
Bringing main into a branch compared in the branch's frame. It remapped
main and the base into that frame, which puts the branch's inherited
content where it cannot agree with the base: everything the branch
inherited still names main, so every reference the remap rewrote read as
a difference. On a demo branch that was 364 of 368 reported conflicts,
and none of them was an edit anybody made. The comparison now runs in
main's frame, exactly as the compare view and the listing summary
already do, and the changes it produces are translated into the branch's
frame before they are applied and logged. That translation is the new
`branch_merge.cljc::remap-changes`, which is a separate function rather
than a call to `remap-refs` for a reason its docstring gives: `relink-refs`
finds a reference where the attribute and the value travel together, and
a change vector carries the identifier alone. The write map is renamed to
`apply-map`, so no later reader can reach for it as the comparison map.
The compare also decoded its merge base twice. The derive resolves and
decodes the pinned snapshot in order to fold the op log over it, and the
caller then decoded the same bytes again for the comparison's third
version, which is about 406 ms on the design-system file and is paid by
every uncached branch summary in a listing as well. The derive now
offers the pre-fold value it already holds, behind an explicit opt-in, so
an ordinary read gains neither a key nor a cost, and a branch read that
forgets the flag raises rather than comparing against nothing. The
invariant a comment now states is that the comparison's base is the
pre-fold value, which is sound because the fold is pure: `process-changes`
threads a value, every method builds persistently, and `PointerMap`'s
`assoc` and `without` return new objects, so the pre-fold base shares
structure with the folded result and costs no extra memory for the pages
the log did not touch.
The new test asserts both halves, because either alone passes a wrong
implementation. A branch carrying only inherited content, updated from a
main that changed elsewhere, reports no conflicts and leaves an empty op
log. Main and the branch repainting the same shape still reports exactly
one conflict, naming that shape.
Verified: the whole backend suite, 919 tests and 5,525 assertions with 0
failures.
A branch copy re-points its local references through the duplication pipeline
and leaves keys like :fill-color-ref-file and :typography-ref-file present with
nil where the source omitted them. Comparing the two forms as different made an
untouched entity look edited on one side, which becomes a modify-modify
conflict as soon as the other side moves the same shape.
The comparison now drops nil-valued keys deeply, for all three sides, in
three-way-entities and so for every pass that goes through it. The new engine
test failed before the change on both orderings and passes after; the three
branch suites report 30 tests, 245 assertions, 0 failures.
A 4-arity that delegates through the var counts as two calls to any test
that redefines it, which made a call-counting assertion measure dispatch
instead of comparisons.
The op log IS the delta between a branch and its merge base, so folding
validate/extract-affected-ids over it yields the exact set of pages the
branch can differ in. compute-merge takes :only-pages and bounds every
page pass to it, the presence pass included, which is where a whole-file
comparison spends its time.
The forward pass is bounded; the reverse one reports main's changes and
stays whole-file. The bound is dropped whenever the log holds a change
the reducer cannot scope to a page or a component (a deletion, page
order, a colour, a token), because for those the affected set is
incomplete and a bounded diff would hide work.
This commit carries the design-branching preview by Juan de la Cruz:
elhombretecla/penpot:branching at 0ac265a3f5, which holds 45 commits
and 3 develop merges and was last synced with develop on 2026-09-16.
The branch is squashed together with a merge of origin/develop at
7c039231f7, so the tree is the merged result.
The merge resolved conflicts in six files:
- backend/src/app/migrations.clj: develop's 0155 and 0156 steps run
first, then the branching steps; the names differ.
- backend/src/app/rpc/commands/viewer.clj: both requires stay.
- common/src/app/common/flags.cljc: both flags stay.
- frontend/src/app/main/ui/dashboard/file_menu.cljs: "Create branch"
becomes a menu-item* with the same id, guard and handler.
- frontend/src/app/main/ui/dashboard/grid.cljs: the portal goes with
develop's menu rewrite; the branches popover stays in the trigger.
- frontend/translations/en.po: both sides' entries stay, with
develop's source references.
Co-authored-by: Juan de la Cruz <delacruzgarciajuan@gmail.com>
AI-assisted-by: mixed models
Drop the equals-sign banners from the CLJS throwable formatter,
so new frontend error reports are born without them. The JVM
branch never printed them. Adds a test asserting full reports
carry sections but no separator runs.
AI-assisted-by: Muse Spark 1.3 Free
* 🐛 Keep interactions of flex children when syncing components
The flex-child sync between main and copy children propagated every
attribute in swap-keep-attrs, which includes :interactions. Since
:interactions has no sync group, the touched check never protected it,
so any change that reflowed a flex component (like toggling a child's
visibility) overwrote the copy children's interactions with the main's.
Closes#11746
AI-assisted-by: claude-opus-5
Signed-off-by: Shreyash Agare <264953665+ShreyashAgare26@users.noreply.github.com>
* 🐛 Add another test and simplify comments
---------
Signed-off-by: Shreyash Agare <264953665+ShreyashAgare26@users.noreply.github.com>
Co-authored-by: Shreyash Agare <264953665+ShreyashAgare26@users.noreply.github.com>
Co-authored-by: alonso.torres <alonso.torres@kaleidos.net>
Kaocha finds every test namespace on the JVM, but the JS runner only
runs the ones listed in runner.cljc. Sixteen namespaces were missing
from both lists, and files.shapes-builder-test was required but not
listed, so 125 tests never ran in JS. Add them to both lists, sorted.
The only one that failed in JS was make-tokens-status-from-tokens-lib:
(.sets base-lib) is a method call in ClojureScript. On the JVM it read
the field, but (:sets ...) on the result gave nil, so the legacy lib
was built without sets or themes. Use (.-sets base-lib) and
(.-themes base-lib), as the first part of the same test does.
AI-assisted-by: claude-opus-5-5
* ✨ Add rich HTML paste to the v3 text editor
Pasting HTML into the v3 text editor keeps bold, italic, decoration and
text-transform, and drops fonts, colors and links. Gated behind
text-editor-wasm/v1-html-paste.
Refs #10465
AI-assisted-by: claude-opus-5-5
* ✨ Paste HTML on the canvas as a styled text shape
With v3 HTML paste on, pasting HTML on the canvas creates a text shape
with the same emphasis. Clipboard options are now read at paste time.
Refs #10465
AI-assisted-by: claude-opus-5-5
Translations that touch no group-like shape or ancestor do not need
the builder objects updated, since only parent resizing reads them.
AI-assisted-by: claude-opus-5-5
Collect children ids into one transient vector and track WASM shape
changes in a JS Map instead of a volatile persistent map.
AI-assisted-by: claude-opus-5-5
* ✨ Gate error reporters behind :error-reporting flag
Error reporters used to start on every boot with no way to
switch them off. They now only start when the new
:error-reporting flag is on (off by default, enable with
PENPOT_FLAGS=enable-error-reporting). The database reporter
needs the flag alone; mattermost needs the flag plus its
webhook url. Shutdown with the flag off is safe, and the
database loop now honors its runtime switch.
AI-assisted-by: muse-spark-1.3-contributor
* 📚 Add audit-log reference memory and update dev tooling
Adds the top-level audit-log reference memory with all event
flows and turns backend/audit-log into a redirect stub.
Also includes the staged dev-tooling updates (playwright
deps, mdts args, gitignore).
AI-assisted-by: muse-spark-1.3-contributor
* ✨ Assign server-side initiator to frontend audit events
Frontend audit events now carry a server-assigned initiator
derived from the x-client header (penpot-frontend goes to
app, penpot-admin-console goes to admin-console, legacy
penpot-nitrate stays on admin-console, all else goes to
app). Client-sent values are always overwritten, and the
key survives on telemetry shadow rows.
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Keep initiator on frontend telemetry context
The strips-pii-keys test still asserted the old contract
(initiator stripped). The initiator only names the sending
app, so it is not personal data and must survive the
telemetry filter like on the backend channel.
AI-assisted-by: muse-spark-1.3-contributor
* 📚 Document initiator on PostHog and browser origins
Complete the audit-log memory for the latest changes: the
initiator property forwarded to PostHog on both channels
and the eventOrigin rule for admin-console browser events.
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Fix unscaped argument in script
---------
Co-authored-by: alonso.torres <alonso.torres@kaleidos.net>
Run the audit archive cron when :nexus or :admin-console is on. Ship
allowlisted events to Admin Console first (with row ids for
idempotency), then the full chunk to Nexus when :nexus is set. Mark
archived_at for the whole chunk on success, including nitrate-only
mode. Rename the gate flag from :audit-log-archive to :nexus.
AI-assisted-by: Composer
Co-authored-by: Cursor <cursoragent@cursor.com>
* ✨ Add size limits to profile props and plugin registry
Bound the total serialized size of profile settings to 2 MiB
(:profile-props-max-size), checked on the merged result before
persisting, with a controlled :props-too-large error. Profiles
that already exceed the limit can still shrink but cannot grow.
Cap plugin registry entries in the shared schema (code 1 MiB, 50
plugins max, bounded name/host/description/icon) and restore rate
limiting on the plugin RPCs (profile-mutations bucket, one write
at a time per profile). The plugin manager now asks for
confirmation before removal and ignores repeated clicks while a
persist request is in flight.
Closes#11592
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Enforce plugin count cap, byte sizes and removal guard
Enforce the declared 50-plugin cap in add-profile-plugin with a
specific :too-many-plugins error (updates of existing entries
still pass); the cap lives in a shared max-plugins constant.
Measure profile props size in UTF-8 bytes instead of chars so
multibyte content cannot slip past the limit.
Cover install/remove persist logic with mocked-RPC frontend tests
(release semantics, in-flight dedupe, validation vs rollback
split) and add the missing boundary tests in common.
Expose the in-flight persist set from the plugin registry and
disable the remove button of entries being saved.
Closes#11592
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Fix rollback loops and restore paths in plugin registry
Restore the previous plugin version instead of dropping the entry
when a validation error rejects an update of an installed plugin.
Make compensating writes one-shot with terminal callbacks so a
persistent failure cannot ping-pong between install and remove.
Restores keep the original list position; the unused public
plugin-persisting? predicate is removed.
Pin count-before-size precedence with a dedicated test and fix
translation source refs to their canonical lines.
Closes#11592
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Guard notifications write and fix restore ordering
Route update-profile-notifications through check-props-size! so
oversized profiles cannot grow through that path; document the
exempt system writers. Remove the duplicated stale entries in
en.po, keeping the canonical translation refs.
Restore rejected plugin updates at their original list position
instead of leaving the optimistic move in place.
Closes#11592
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Skip no-op plugin removal and clarify size comments
Return early from remove-profile-plugin when the id is absent:
no wasted write, no size check, and no manufactured :plugins key
that could spuriously fail on oversized profiles.
Clarify that per-field string caps count chars while the byte
budget is enforced by profile-props-max-size.
Closes#11592
AI-assisted-by: muse-spark-1.3-contributor
* 📎 Fix formatting in rlimit.edn for profile operations
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
* 📎 Fix formatting of import-binfile/global entry
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
* ♻️ Simplify props size check and tighten plugin entry caps
Measure props with transit bytes directly instead of the
PGobject string roundtrip.
Rename check-props-size! to check-props-size: single hard limit
on the merged props, no growth comparison, and return props so
writers thread the check into the update.
Move the 2 MiB default into default-props-max-size on the
profile namespace, still overridable with the optional
:profile-props-max-size config entry.
Tighten registry-entry :code and :icon to 500 chars: they hold
manifest paths, not content.
Closes#11592
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Fix compatibility problems
---------
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
Co-authored-by: alonso.torres <alonso.torres@kaleidos.net>
* 🐛 Gate pastes on page load and harden base-shape lookup
Pasting while the workspace is still opening crashed the
session: the layer-order lookup called rseq on a missing
root children list.
Ignore paste events until the page objects are loaded (the
clipboard keeps its content, so retrying works), return empty
instead of throwing from the shared layer-order helpers, and
fall back to pasting at the pointer position when the selection
is detached from the shape tree. Selecting the page root keeps
working as before through the frame branches.
Closes#11666
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Cover props paste and root-plus-other in paste guards
Address review follow-ups on the paste-before-init fix: gate
props pasting on page readiness like the shape entries, and
route root-plus-other selections without a base shape to the
pointer fallback instead of the unguarded else branch. Pin
single-root selection to the frame path with a regression test.
Closes#11666
AI-assisted-by: muse-spark-1.3-contributor
* ♻️ Hoist page lookup out of paste-shapes gate
Bind page and page-objects once in an outer let instead of
calling lookup-page twice (once for the readiness gate and
once inside the body). No behavior change.
AI-assisted-by: muse-spark-1.3-contributor
* ♻️ Reuse bound ids in paste-shapes page lookup
Bind file-id and page-id once and pass them to the lookup-page
arity that takes both, instead of resolving the page twice and
rebinding file-id in the inner let. No behavior change.
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Stop browser logging from crashing on empty levels
Stop level->int crashes from taking down the dashboard when a
nil or unknown level reaches the browser logger. Invalid levels
now warn and are ignored in enabled?, setup! and the console
handler, which renders unknown records with a neutral fallback.
Alias the schema-legal :fatal level to :error in the browser
mappings and validate the JS-exported debug.set_logging, which
previously threw on missing arguments and wrote unreachable
keyword keys into the loggers map.
Closes#11690
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Guard logger args and strengthen logging tests
Close the residual throw paths next to the empty-level crash:
guard non-string loggers in enabled? and setup!, coerce
set_logging arguments safely, and validate logger keys.
Strengthen the regression tests so the fatal alias cannot
regress silently: enabled-logger filtering, JVM fatal and bogus
cases, setup! skip proof, and invalid-logger cases.
Related to #11690
AI-assisted-by: muse-spark-1.3-contributor
* 🐛 Address low findings from logging review
Validate the logger before the level in console-log-handler,
share a public valid-logger? predicate with debug/set-logging,
and keep warn formatting consistent across boundaries.
Document the fail-soft-FE/strict-BE split on enabled? and the
valid-level? contract on set-level!. Cover safe fallbacks, bad
logger keys, handler logger skips, and loggers-map isolation in
common tests, and add a frontend test for debug/set-logging.
Related to #11690
AI-assisted-by: muse-spark-1.3-contributor
