Introduce a new OpenCode workflow skill that guides users through
backporting commits by applying diffs instead of using cherry-pick.
This is useful when cherry-pick is undesirable (e.g. divergent
histories, binary conflicts, or partial porting).
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
Two coupled defects made shape.applyToken(), token.applyToShapes() and
token.applyToSelected() silently no-op when invoked from JavaScript with
an array of strings (e.g. token.applyToShapes([rect], ["fill"])):
1. token-attr-plugin->token-attr only consulted its alias map when the
input was already a keyword; string inputs fell through unchanged,
causing downstream token-attr? to return false.
2. The inner schemas used plain [:set ...] which lacks the :decode/json
transformer for JS array -> Clojure set coercion. Switching to
Penpot's custom [::sm/set ...] lets the standard JSON decoder
pipeline handle the conversion automatically.
This is a backport of commit 1eac3e2be5f973359ad2ec9bac4e80a9d5a9e022
which fixes GitHub #9162.
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
Improve MCP instructions on design creation:
* Agents should make use of layouts when appropriate
* Agents should name all elements appropriately
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
* 🐛 Preserve renamed layer name when re-entering edit mode
When a layer was renamed and the user clicked its name again to edit
it, the input opened with the type-based default name instead of the user's saved name. Pressing Enter then
silently overwrote the saved name with the default. Read the current
shape :name when seeding the rename input so the user's previous
rename is preserved.
Signed-off-by: jack-stormentswe <crazycoder131@gmail.com>
* 💄 Remove redundant DOM-refresh effect from layer rename input
Signed-off-by: jack-stormentswe <crazycoder131@gmail.com>
---------
Signed-off-by: jack-stormentswe <crazycoder131@gmail.com>
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
Co-authored-by: Andrey Antukh <niwi@niwi.nz>
* ✨ Show specific error messages for invitation token failures
Surface distinct error messages for the three invitation-token failure
modes that the backend already distinguishes: email mismatch, expired
token, and invalid/corrupted token. Replaces the single generic
could not accept invitation message with actionable text so the
user knows what went wrong and how to recover.
Signed-off-by: jack-stormentswe <crazycoder131@gmail.com>
* 💄 Update CHANGE.md
Signed-off-by: jack-stormentswe <crazycoder131@gmail.com>
* 💄 Address review feedback on invitation-error messages
Signed-off-by: jack-stormentswe <crazycoder131@gmail.com>
---------
Signed-off-by: jack-stormentswe <crazycoder131@gmail.com>
Co-authored-by: Andrey Antukh <niwi@niwi.nz>
Explicitly state that the team is small and reviews may take a
few days as they are handled in dedicated time blocks.
Remove mention of GitHub Discussions since it is not used.
Reword the discussion requirement to manage expectations: do not
expect a PR to be accepted without prior discussion.
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
Rename subscription.settings.sucess.dialog.{title,footer} to
subscription.settings.success.dialog.{title,footer} in en.po and
update the three callsites in subscription.cljs.
Closes#9203
Signed-off-by: jack-stormentswe <crazycoder131@gmail.com>
* 🐛 Fix Plugin API token application for JS array of strings
Plugin code calling `shape.applyToken(token, ["fill"])` or
`token.applyToShapes([rect], ["fill"])` from JavaScript supplies a JS
array of strings. The plugin proxies expected a Clojure set of
keywords, and two coupled defects made the calls silently no-op (or,
with `throwValidationErrors` enabled, throw "check error"):
1. `token-attr-plugin->token-attr` only consulted its alias map when
the input was already a keyword. String inputs like "fill" fell
through to the identity branch, so the downstream
`cto/token-attr?` predicate (which checks against a set of
keywords) returned false for every string. Coerce strings to
keywords first.
2. The `applyToken` / `applyToShapes` / `applyToSelected` schemas
used plain `[:set ...]`, which has no `:decode/json` transformer
for JS array → Clojure set coercion. Switch to the registered
`[::sm/set ...]` (in `app.common.schema`) which provides the
array → set decoder. After the switch, the standard JSON pipeline
converts `["fill"]` to `#{"fill"}`, then the inner
`[:and ::sm/keyword [:fn token-attr?]]` decodes each element to a
keyword and validates it.
Also extends the docstring on `token-attr-plugin->token-attr` to make
the string-friendly contract explicit, and registers a new
`tokens-test` ns under `frontend/test/frontend_tests/plugins/` with
six `deftest` blocks covering:
- known keywords passing through unchanged
- keyword aliases (`:r1` → `:border-radius-top-left`, etc.)
- string inputs coerced to keywords (regression for #9162)
- `token-attr?` accepting both keyword and string inputs
- `token-attr?` rejecting unknown attrs and nil
Closes#9162
* 🐛 Fix wrong direction in plugin-name alias tests
The added tests in tokens_test.cljs and the new docstring in tokens.cljs
described the alias resolution in the wrong direction. The map is
{:r1 :border-radius-top-left, …} then map-invert'd, so
token-attr-plugin->token-attr maps verbose plugin-side names
(:border-radius-top-left) to canonical internal short names (:r1),
not the other way around. Inputs already in canonical form (:r1, :fill,
"fill", …) pass through unchanged. Flipped the alias-resolution test
expectations and the keyword/string-input cases, refreshed the docstring
and the regression-coverage comment to match.
---------
Co-authored-by: Andrey Antukh <niwi@niwi.nz>
Combine margin-block-start and margin-block-end into the margin-block
shorthand to satisfy the declaration-block-no-redundant-longhand-properties
stylelint rule.
Signed-off-by: Andrey Antukh <niwi@niwi.nz>
When working with large asset groups, users asked for a one-click way
to remove every asset under a group path. Multi-select across hundreds
of items is impractical, and ungrouping first and then deleting leaves
the orphaned items in the flat list.
This change adds a "Delete group" option to the assets-panel
context-menu for three asset types that already carry group structure:
- Components (including variants — sibling variants sharing a variant
container are deduplicated, and the container is deleted once via
the same dispatch the per-item delete uses in file_library.cljs).
- Colors.
- Typographies.
A confirmation modal is shown before deletion, with the count of
assets to be removed, so the action is never silent. All deletes run
inside a single undo transaction, so one Cmd+Z restores the whole
group.
Changes
-------
- `assets/groups.cljs` — `asset-group-title*` accepts an optional
`on-delete-group` prop and conditionally adds the menu entry
between "Ungroup" and "Combine as variants". When the callback is
not supplied the option is hidden, so asset sections that do not
implement it stay unaffected.
- `assets/components.cljs` — threads `on-delete-group` through the
recursive `components-group*` and defines the section-level
handler, dispatching to `dwsh/delete-shapes` for variant containers
and `dwl/delete-component` for plain components.
- `assets/colors.cljs` — same threading + a simple `dwl/delete-color`
dispatch per color in the group.
- `assets/typographies.cljs` — same threading + a
`dwl/delete-typography` dispatch per typography in the group.
- `translations/en.po` — three new strings: the menu label
(`workspace.assets.delete-group`) and the modal title/message
(`modals.delete-asset-group.title`/`.message`, plural-aware).
Github #9141
Signed-off-by: FairyPigDev <luislee3108@gmail.com>
Signed-off-by: FairyPiggyDev <luislee3108@gmail.com>
The ::mf/props and ::mf/wrap-props metadata keys are no-ops on modern
components (those defined with mf/defc and the * suffix) since the *
suffix already triggers the props behavior these keys attempt to
configure. This cleanup removes the redundant metadata from modern
components across all UI directories.
Changes:
- comments/: comments
- dashboard/: comments, deleted, files, fonts, grid, import, libraries,
pin_button, projects, search, sidebar, subscription, team, templates
- exports/: files
- modal/: modal
- settings/: subscription
- static/: static
- viewer/: comments, interactions, viewer
- workspace/: context_menu, libraries, sidebar/assets,
viewport/gradients, tokens/settings/menu
Let users pick the pixel grid color from a standard color picker.
The grid color was previously hardcoded, making it invisible on
mid-tone canvases. Choice is stored on the file so it persists
across sessions. Defaults preserve the current appearance when
unset.
Closes#7750
Signed-off-by: jack-stormentswe <crazycoder131@gmail.com>
Closes#7736.
The Layers sidebar offered no way to expand every nested level of a
single subtree at once. Unfolding a layer that wraps a deep tree
required clicking each disclosure indicator one level at a time -
O(siblings * depth) clicks. The asymmetry was particularly visible
next to the existing Shift+click gesture, which collapses every
layer in the panel in a single action via `dwc/collapse-all`, with
no expand counterpart for either a single subtree or the whole
tree.
Add a new `dwc/expand-subtree` event in
`app.main.data.workspace.collapse` that uses
`cfh/get-children-ids-with-self` to gather the shape's id together
with every descendant id, then merges `{descendant-id true}` entries
into `[:workspace-local :expanded]` so the entire subtree opens in
one update. Existing expansion state on unrelated branches is left
untouched (`merge`, not `assoc`), matching the per-key shape used by
`toggle-collapse` and `expand-collapse`.
Wire the gesture into `layer_item.cljs` `toggle-collapse` callback as
a third branch:
- Shift+click while expanded - collapse every layer (existing).
- Alt+click while collapsed - expand the entire subtree (new).
- Otherwise - toggle this single level (existing).
Alt is chosen instead of Shift to avoid the ambiguity the issue
author flagged: "for a layer of middle depth it is unclear whether
[Shift+click] should fold all (up to the topmost parent) or expand
all (only the current subtree)". Alt is a common platform
convention for "do this recursively" (Finder, file managers,
several IDEs), so the asymmetric mapping matches user expectations.
The callback's `mf/deps` vector is extended with `id` and `objects`
so the closure refreshes when the shape tree changes.
CHANGES.md entry added under the 2.17.0 New features section.
