4ff9cba76d2d3d126830514e7e297c5a82a9daf5
14026 Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 Thomas Trompette
|
4ff9cba76d |
fix(server): stop the global catch-all filter from shadowing typed GraphQL exception filters (#23508)
## Context Sentry [TWENTY-SERVER-60Y](https://twenty-v7.sentry.io/issues/6633503406) ("Permission Denied: Entity performing the request does not have permission") is still firing at full rate on `v2.25.0`: ~10.8k events in the last 7 days, 24k total. #23104 tried to fix it by registering `PermissionsGraphqlApiExceptionFilter` globally via `APP_FILTER`. That registration is correct but **inert in production**, and the integration test added alongside it passes for a reason unrelated to prod behaviour. ## Root cause `main.ts` registered a catch-all filter after bootstrap: ```ts app.useGlobalFilters(new UnhandledExceptionFilter()); ``` Nest builds each resolver's filter list as `[...global, ...class, ...method]`, reverses it, and selects **exactly one** matching filter — there is no chaining. `APP_FILTER` providers are collected during module scan; `useGlobalFilters` appends after that, so the catch-all ended up at the head of the list: ``` 1. UnhandledExceptionFilter @Catch() <- matches everything, wins 2. PermissionsGraphqlApiExceptionFilter <- never reached 3. BillingGraphqlApiExceptionFilter <- never reached ``` On a GraphQL host `UnhandledExceptionFilter` then no-ops: `host.switchToHttp().getResponse()` returns the GraphQL args object, `response.header` is undefined, so it hits `return;`. Nest treats a falsy return as unhandled and rethrows the original `PermissionsException`, which reaches the Yoga error hook as a non-`BaseGraphQLError`, is serialized `INTERNAL_SERVER_ERROR`, and is reported by `shouldCaptureException`. The 28 resolvers carrying `@UseFilters(PermissionsGraphqlApiExceptionFilter)` were unaffected — method-level filters are evaluated before globals. Only the resolvers relying on the global registration leaked, which is exactly the set showing up in Sentry (`findOneApplication`, `uploadFilesFieldFileByUniversalIdentifier`, `UpdatePageLayoutWithTabsAndWidgets`, ...). Two other global filters were shadowed the same way and have never run: `BillingGraphqlApiExceptionFilter` and `FlatEntityMapsGraphqlApiExceptionFilter`. ## Why the existing test did not catch it `test/integration/utils/create-app.ts` builds the app from `AppModule` directly and never executes `main.ts`, so `useGlobalFilters` does not exist in the test process. It registered `MockedUnhandledExceptionFilter` as an `APP_FILTER` on the root testing module, which is collected *first* and therefore evaluated *last* — the exact inverse of production precedence. The `findOneApplication` denial test passed while the same query kept reporting to Sentry. ## Fix Register `UnhandledExceptionFilter` through `APP_FILTER` on `AppModule`. Root-module providers are scanned first, so it is collected first and evaluated last. The filter stays global, stays catch-all, and keeps its CORS-header role for HTTP; it simply no longer cuts in front of the typed filters. Un-shadowing the other two global filters means they now actually run, so `FileStorageExceptionFilter` and `FlatEntityMapsGraphqlApiExceptionFilter` get the `host.getType() !== 'graphql'` rethrow that `Billing` and `Permissions` already had. Without it they would start throwing GraphQL error objects into the REST pipeline. `MockedUnhandledExceptionFilter` is removed: `AppModule` now supplies the real filter in the same position, so the mock was dead weight. ## Test Verified against a real server (not the integration harness), calling the exact document from Sentry event `8d19eb7c` as a member with no permission flags: ``` query ($v1:UUID){findOneApplication(id:$v1){applicationVariables{key,value}}} ``` | | response code | exceptions captured | |---|---|---| | before | `INTERNAL_SERVER_ERROR` | 1 | | after | `FORBIDDEN` | 0 | Capture count measured through the console exception-handler driver, i.e. the same `captureExceptions` call site that is the Sentry driver in production. New unit spec `src/filters/__tests__/unhandled-exception.filter.spec.ts` boots a Nest + Yoga app both ways: it asserts `FORBIDDEN` with the `APP_FILTER` registration, and pins the shadowing behaviour of `app.useGlobalFilters` so the pattern cannot come back unnoticed. `granular-settings-permissions.integration-spec.ts` passes (10/10). Note it also passes *without* this fix — the harness cannot observe bootstrap-only configuration, which is the underlying reason #23104 shipped green. Closing that gap properly means sharing the post-`create` bootstrap between `main.ts` and `create-app.ts`; left as a follow-up. `file-storage-exception-filter.spec.ts` extended with a non-GraphQL host case. ## CI follow-up `failing-file-by-id-download.integration-spec.ts` snapshots were updated. That REST endpoint's 403 body changed in tests from `{}` to `{"statusCode":403,"error":"Forbidden","message":"Forbidden resource"}`. The old `{}` was an artifact of the mock: `MockedUnhandledExceptionFilter` rethrew, the exception escaped Nest's handler into Express's default error handler, and supertest saw an empty body. Production has always run the real `UnhandledExceptionFilter`, which writes `response.status(status).json(exception.response)` — the new snapshot. Production HTTP behaviour is unchanged by this PR: no other global filter matches an `HttpException` (the typed ones rethrow outside GraphQL), so the same filter handles it whether it is evaluated first or last. |
||
|
Félix Malfait
|
b643ddf119 |
Keep fullWidth buttons at full width while loading (#23536)
When a `Button` enters its loading state, the wrapper gets
`.wrapperLoading { max-width: calc(100% - 32px) }` to make room for the
spinner on auto-width buttons. `.fullWidth` only set `width: 100%`
without a max-width, so any full-width button shrank by 32px for as long
as its loader was visible, leaving a visible gap on the right. Most
noticeable on the "Add credit card" button in the billing modal while
the card is being validated.
Fix: `.fullWidth` now also pins `max-width: 100%`. It is declared after
`.wrapperLoading`, so it wins the cascade at equal specificity and the
loading shrink keeps applying only to auto-width buttons.
---
_Generated by [Claude
Code](https://claude.ai/code/session_01W1J7cW2MhaqXGfdjvHeFoX)_
<!-- This is an auto-generated description by cubic. -->
<a
href="https://cubic.dev/pr/twentyhq/twenty/pull/23536?utm_source=github"
target="_blank" rel="noopener noreferrer"
data-no-image-dialog="true"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img
alt="Review in cubic"
src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a>
<!-- End of auto-generated description by cubic. -->
|
||
|
github-actions[bot]
|
33fb57d128 |
i18n - translations (#23525)
Created by Github action Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
Thomas des Francs
|
030ee2c7cc |
Move settings tabs below page headers (#23519)
## Summary - Position the Admin Panel tabs directly below the settings header. - Position the Communication tabs directly below the settings header. - Reuse the shared settings tab bar while preserving permissions, disabled states, and hash navigation. - Keep the responsive tab overflow menu available on narrow settings cards. ## After <img width="3456" height="2008" alt="Admin Panel tabs positioned below the settings header" src="https://github.com/user-attachments/assets/c8ff965c-d4b2-4700-85e1-0763f9f0852d" /> |
||
|
github-actions[bot]
|
58ebbe0394 |
i18n - docs translations (#23523)
Created by Github action <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23523?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
Etienne
|
a276f3277f |
feat(workflow): pin concrete model on AI agent node creation and exclude interactive tools from workflow runs (#23447)
## Context The AI Agent workflow node's model dropdown could show a model that was not the one used at run time (e.g. the node displayed "Claude Haiku 4.5" while the run log showed `openai/gpt-5.6-sol`). Root cause: workflow agents were created with `modelId: AUTO_SELECT_SMART_MODEL_ID`. The builder's model `Select` cannot represent that value — auto-select ids are filtered out of the options (`useWorkspaceAiModelAvailability`) and the pinned "default" option remaps its value to the resolved concrete model id (`useAiModelOptions`) — so `Select` silently fell back to `options[0]`, the alphabetically first enabled model. Meanwhile the runtime correctly resolved auto-select to the instance's default smart model. ## What this PR does ### 1. New workflow agents store a concrete model id `WorkflowVersionStepOperationsWorkspaceService` now reads the workspace's `fastModel` setting, expands it through `AiModelRegistryService.getEffectiveModelConfig`, validates it with `validateModelAvailability`, and stores the concrete model id — so the dropdown displays the model that will actually run, and workflow agents default to the cheaper fast tier instead of the smart one. Falls back to `AUTO_SELECT_FAST_MODEL_ID` if the lookup or validation fails (workspace missing, no AI provider configured, model disabled), so node creation never breaks. ### 2. Exclude `search_help_center` and `navigate_app` from workflow agent runs `ActionToolProvider` adds both tools unconditionally, but they only make sense in an interactive chat session (navigation targets the user's browser; help-center search is a support tool). They are now excluded via `WORKFLOW_AGENT_EXCLUDED_TOOL_NAMES` in `AgentAsyncExecutorService`, alongside the existing output-navigation exclusions. Chat agents are unaffected. ## Test coverage - Existing specs for `WorkflowVersionStepOperationsWorkspaceService` and `AgentAsyncExecutorService` updated/passing (new constructor deps mocked). - `nx typecheck twenty-server` and `lint:diff-with-main` pass. <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23447?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
martmull
|
c4a79c50c3 |
Install pre-installed apps in a dedicated job after the workspace upgrade cursor is written (#23517)
## Problem Application registrations flagged `isPreInstalled: true` were not installed on newly created workspaces. The call was wired in, but it ran too early. `activateWorkspace` invoked `preInstalledAppsService.installOnWorkspace` from inside `prefillCreatedWorkspaceRecords`, which runs **before** `activateAndInitializeUpgradeState`. The install path validates app/workspace version compatibility: - `ApplicationInstallService.runInstall` reads `engines.twenty` from the app's `package.json` and calls `validateWorkspaceCompatibility` - `ApplicationVersionValidationService.validateWorkspaceCompatibility` resolves the workspace version through `UpgradeStatusService.getWorkspaceCompletedVersion` - that reads the workspace's upgrade-migration cursor, which is only written by `markAsWorkspaceInitial` inside `activateAndInitializeUpgradeState` During creation the workspace has no cursor row yet, so `getWorkspaceCompletedVersion` returns `null`, the install throws `INVALID_WORKSPACE_VERSION`, and the failure is swallowed twice over: `PreInstalledAppsService` logs per-app failures without rethrowing, and `activateWorkspace` wraps the whole call in non-critical error handling. The workspace comes up silently missing its apps. This affects most real apps, since they pin `engines.twenty`: `fireflies`, `last-contact`, `people-data-labs`, `call-recorder`, `postcard`, `self-hosting`, `twenty-partners` (`>=2.23.0`) and `exa`, `real-estate` (`>=2.19.0`). Only apps with no `engines.twenty` installed successfully. The same interaction is already documented in `2-23-workspace-command-1784565137000-upgrade-people-data-labs-application.command.ts`, which works around it with `skipWorkspaceCompatibilityCheck: true`. ## Changes - Added `InstallPreInstalledAppsJob` on the workspace queue, mirroring the existing `InstallOnboardingAppsJob`. - `activateWorkspace` now enqueues that job instead of installing synchronously, so workspace creation no longer blocks on package fetching and manifest application. - The enqueue happens after `activateAndInitializeUpgradeState` writes the upgrade cursor, so the compatibility check has a workspace version to resolve by the time the worker picks the job up. ## Notes Workspaces created before this fix can be repaired with the existing `install-pre-installed-apps` backfill command, which is idempotent. |
||
|
Raphaël Bosi
|
ada7eb1d88 |
Restore the Figma halftone shapes and calm the welcome animation (#23495)
https://github.com/user-attachments/assets/0c06d190-977e-4ad6-8a02-251f341ee310 The welcome overlay settled into circles instead of the halftone from Figma: the densify step that took the dot set from 681 to 2897 emitted points, so every dash had zero length. All 681 dashes in the Figma export (`public/images/onboarding/welcome-halftone.svg`) share a constant length / stroke width ratio of 1.452, so the shape is restored by deriving the length from the stroke width, with no data regeneration. Dashes now stretch into shape while they are still flying in, rather than popping once they have landed. The rest of the pass makes the animation quieter. The shine sweep is gone, along with the highlight colour that only fed it. Particles approach from much closer on a gentler ease, the idle drift is roughly halved, and the exit is a soft outward drift instead of a burst that threw everything off screen. The white pill behind the title and the person chip's surface are removed too, so the title reads directly against the halftone. <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23495?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Raphaël Bosi
|
684259fd5e |
Make onboarding cards contrast with the page background (#23516)
Onboarding card surfaces used `background.secondary`, the same token as the onboarding page background, so they blended in (visible on the workspace selection step). Switched them to `background.primary`, matching the other onboarding cards (plan card, install apps, trust badges). <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23516?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Thomas Trompette
|
58b8e8afee |
Fix clipped New chat button label in localized navigation drawer (#23511)
Fixes #23503 <img width="161" height="76" alt="Capture d’écran 2026-07-29 à 17 20 25" src="https://github.com/user-attachments/assets/a6ea0a12-b91d-419f-8023-8e65d5dce65b" /> The expanded "New chat" button had a hardcoded `width: 103px`, sized for the English label. Localized labels ("Neuer Chat", "Nouveau chat") were clipped, since `OverflowingTextWithTooltip` can only truncate inside a parent it cannot resize. The expanded wrapper is now `width: max-content` with `max-width: 100%`, so it grows with the label and only truncates (with tooltip) when the sidebar has no space left. `min-width` keeps the pill from collapsing below icon size. Collapsed state is unchanged. ### Verified locally (German locale) | | wrapper width | label | |---|---|---| | before | 103px (fixed) | `Neuer C...` clipped | | after | 109px (max-content) | `Neuer Chat` in full | - English: 103px -> 98px, visually identical. - Collapsed drawer: still exactly 24x24, unchanged. - Very long label (Vietnamese-length): wrapper stops at the sidebar edge, no overflow, label truncates with tooltip. |
||
|
github-actions[bot]
|
8707ebb7ac |
i18n - docs translations (#23515)
Created by Github action Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
nitin
|
e5c6cbcf80 |
Resolve route-trigger workspace from bearer token on bare hosts (#23490)
When a request reaches the `/s` route on a host that names no workspace
(bare `SERVER_URL` on a multiworkspace instance), resolve the workspace
from the bearer token — the same source `/graphql` uses — instead of
failing with `WORKSPACE_NOT_FOUND`. Hosts that do name a workspace keep
host resolution unchanged, and requests without a token are unaffected.
This makes the client SDK's same-site `${apiBase}/s` fallback work on
multiworkspace instances without a configured public domain: app logic
functions calling their own HTTP routes (e.g. call-recorder artifact
import) currently 404 there, because `TWENTY_FUNCTIONS_URL` is injected
empty and the bare server host carries no workspace identity. Cloud
(workspace public origin injected) and single-workspace self-host (host
resolves the default workspace) never hit this path.
Note: this also allows public routes to be reached through a bare host
when a valid token identifies the workspace. It does not change route
authorization; the token is used only for workspace resolution.
|
||
|
github-actions[bot]
|
cedb3768ec |
i18n - translations (#23513)
Created by Github action <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23513?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> --------- Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
Weiko
|
adfb96c7ce |
Keep email participant avatar colors consistent (#23444)
## Summary - Fixes issues where the same record didn't share the same avatar in multiple places of the inbox. - Add a shared avatar color seed resolver for email participants. - Apply consistent placeholder colors across participant chips and email thread previews. <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23444?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Paul Rastoin
|
133b3375b6 |
[Upgrade] Stop command gracefully on SIGINT/SIGTERM (#23481)
Ctrl+C on `upgrade` used to kill the process wherever it happened to be, potentially in the middle of a workspace command. It now stops at the next iteration boundary instead. ## Behavior - **First SIGINT/SIGTERM** — the runner finishes what it started, then stops instead of starting new work. Exits with `130` (SIGINT) or `143` (SIGTERM), following the 128+signal convention, so orchestrators can tell an interruption apart from a failure. - **Second signal** — immediate exit, leaving the command in progress unfinished. - **SIGKILL** — untrappable, same outcome as a second signal. Nothing is rolled back on stop: the run resumes from the last command recorded in `upgradeMigration`. ## Opt-in per command Registering a `SIGINT` listener removes Node's default kill-on-signal behavior, so a command that installs a handler without honoring the flag would ignore the first Ctrl+C entirely. Handlers are therefore opt-in via `CommandShutdownService.listenToShutdownSignals()`, called by the two commands that stop at a boundary: - `UpgradeCommand` - `WorkspaceCommandRunner`, the base for standalone workspace commands Everything else keeps today's behavior and dies on the first signal, `run-instance-commands` included: instance commands are transactional and cursor-guarded, so a hard kill rolls back and a rerun skips what completed. `install-application`, `rebuild-application-default-deps` and `install-pre-installed-apps` iterate over workspaces without going through `WorkspaceCommandRunner`, so they are not armed either; they are one call away if we want them. The server and worker processes share these services and never arm anything, so their shutdown semantics are unchanged. ## Where the flag is checked `CommandShutdownService` exposes a single boolean, `isShutdownRequested()`, read only by the iteration runners: - `UpgradeSequenceRunnerService.runInner` — before each sequence step - `WorkspaceIteratorService.iterate` — before each workspace There is deliberately no `AbortSignal`: in-flight work is never cancelled, it is allowed to finish. Individual commands know nothing about shutdown, so a workspace that has started runs its whole pending segment before the run stops. Each workspace ends up either fully done with the segment or untouched, never scattered at some cursor inside it. That keeps resume state coarse and the change out of the command layer, at the cost of a longer stop latency, which the second Ctrl+C covers. `WorkspaceIteratorReport` gained an `interrupted` flag. The sequence runner needs it: stopping partway through the workspace list and then advancing the cursor would run an instance step against workspaces that are not aligned yet, so it returns instead. ## Deployment note Under Kubernetes, `terminationGracePeriodSeconds` must exceed the time for one workspace to finish its segment, otherwise the SIGTERM path degrades into a SIGKILL. Documented in `docs/UPGRADE_COMMANDS.md`. ## Testing - New unit test for `CommandShutdownService` (7 cases); 293 tests pass across `database/commands` and `core-modules/upgrade` - `tsgo -p tsconfig.json` clean - oxlint and oxfmt clean on all touched files |
||
|
nitin
|
742f76e318 |
[BREAKING-CHANGE] Add NOT_RECORDED call recording status (#23478)
Adds NOT_RECORDED to the CallRecording status select, for meetings where nothing was captured (bot never admitted, meeting not started, nobody joined). First part of twentyhq/core-team-issues#2706, split out so existing workspaces are upgraded before the call-recorder app starts writing the new status. - NOT_RECORDED enum value + standard select option - 2-26 workspace upgrade command adding the option to existing workspaces (idempotent, same option id as the standard definition) App-side classification from Recall sub codes comes in a follow-up PR. <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23478?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Paul Rastoin
|
4ec65ed08d |
System view tooling explicit params key naming (#23506)
# Introduction View field system always result from a field existence, the application universal identifier should be the related field one Same but for views and object <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23506?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Félix Malfait
|
33970dd45d |
Align campaign view column positions with the standard layout (#23496)
Follow-up to #23493 (merged). Now based on `main`. ## Problem `AddMessageCampaignNameFieldCommand` places the new `name` column below the lowest existing position when it is the label identifier (`min - 1`). That satisfies `isViewFieldInLowestPosition` once, but does not hold: `viewField.position` is compared by the standard-application sync (`position: { toCompare: true }`), so the stored position is pulled back to the standard one and re-fails `validateLabelIdentifierFieldMetadataIdFlatViewField`, throwing `WorkspaceMigrationBuilderException` on every sync. That sync runs during `WorkspaceManagerService.init`, which is what `ActivateWorkspace` calls. cubic flagged this on the original PR: [discussion_r3653452416](https://github.com/twentyhq/twenty/pull/23188#discussion_r3653452416). ## Confirmed against prod Queried the workspace failing in Sentry ([TWENTY-SERVER-JJ2](https://twenty-v7.sentry.io/issues/7639768673/)): the `allMessageCampaigns` view has no `name` column at all, `subject` at position `0` (still the label identifier), and every column at the old layout. The activation sync tries to create `name` at standard position `0` while repointing the label identifier in the same batch, ties with `subject` at `0`, and throws. Exactly the mechanism this PR fixes. ## Change A `2-25` workspace command (`upgrade:2-25:align-message-campaign-view-field-positions`, timestamp `1785332560000`) that aligns the `allMessageCampaigns` columns to the standard layout, so the sync has nothing left to change: - Standard columns take their standard positions (`subject 0→1`, `status 1→2`, …), freeing slot `0` for `name`. - Columns the standard application does not know about keep their relative order and move above the standard ones. - Updates run in two migration passes: everything else first, then the lowest-target column alone. The migration builder validates updates one at a time against optimistic maps it mutates as it goes, so a single batch is order-dependent (caught by Greptile below); two passes keep every intermediate state valid. Placed in `2-25/` rather than `2-26/` so it can ship in a 2.25.x patch — the failures are on `v2.25.0` instances. This is why `server-previous-version-upgrade-mutation-guard` is red: it needs the `ci:allow-previous-version-upgrade-mutation` label (same as #23493). The position arithmetic and the pass-splitting are pure utils with 12 tests. ## Traced against the confirmed prod state Pass 1 moves `status`…`createdAt` up while `subject` (label identifier) stays lowest at `0`; pass 2 moves `subject` to `1`, still strictly lowest. The subsequent activation sync then creates `name` at `0` below `subject` at `1` and repoints the label identifier. Converges. Run order matters only relative to the sibling command from #23493: removal (`…550000`) sorts before alignment (`…560000`), which is correct. |
||
|
Félix Malfait
|
e79424ddb0 |
Stop building the Campaigns navigation menu item (#23493)
Campaigns show up in the sidebar of every workspace, including brand new ones, while the feature is still behind `IS_EMAIL_GROUP_ENABLED`. ## Why it leaked `navigationMenuItem` has no `conditionalAvailabilityExpression` column. Only two entities do: - `page-layout-widget.entity.ts` - `command-menu-item.entity.ts` So there was no flag to attach. #23188 gated every surface that supports gating — the campaign command menu items carry `featureFlags.IS_EMAIL_GROUP_ENABLED` (`standard-command-menu-item.constant.ts:720` and `:735`) and the page layout widgets are gated the same way — but `allMessageCampaigns` was added to `FLAT_NAVIGATION_MENU_ITEM_NAMES`, which builds unconditionally. `messageCampaign` is `isSystem: true`, so the navigation item was the only thing exposing the feature. ## Change - Drop `allMessageCampaigns` from `FLAT_NAVIGATION_MENU_ITEM_NAMES`. Its definition stays in `STANDARD_NAVIGATION_MENU_ITEMS` so the identifier remains reserved and re-enabling is a one-line change. - Add workspace command `1785324390000` to delete the rows from workspaces already provisioned with the item. It collects every matching row rather than looking the identifier up once, because each user workspace gets its own. ## Trade-off Campaigns become unreachable from the sidebar even with the flag on, which restores the pre-#23188 state. The durable fix is adding `conditionalAvailabilityExpression` to `navigationMenuItem` so the item can be gated like the command menu items — a schema change, deliberately not done here. ## Verification Reproduced on a fresh `database:reset` against `main`: two `20202020-b00b-4b0b-8b0b-c0aba11c000b` rows (type `OBJECT`, position 7) for the single seeded workspace, structurally identical to the other nav items. Typecheck, `oxfmt --check src/` and `oxlint --type-aware src/` clean on this branch. The post-fix database check did not complete — local Postgres died mid-reset — so the removal is verified by construction and by the build-list change, not yet by a second reset. --- _Generated by [Claude Code](https://claude.ai/code/session_01NZKeMSUBG3VQPuoy8i5Awe)_ <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23493?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
github-actions[bot]
|
63d34e0133 |
i18n - translations (#23509)
Created by Github action <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23509?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> --------- Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
Weiko
|
e2e2142998 |
Add calendar date field submenu with grouped field options (#23445)
## Summary - Add a submenu for selecting calendar date fields as it was not clear enough that you could pick 2 date fields. - Group date and datetime fields with non-selectable headers ### Before <img width="240" height="252" alt="Screenshot 2026-07-29 at 11 32 54" src="https://github.com/user-attachments/assets/169bd965-676f-4ef9-8ba6-9ecac80547c6" /> ### After <img width="260" height="246" alt="Screenshot 2026-07-29 at 11 31 58" src="https://github.com/user-attachments/assets/a4471bb0-5c1f-4f5e-bddb-165a5c9da4c7" /> |
||
|
Paul Rastoin
|
e99452e00d |
feat(server): attach app attributes to application rate-limited metric (#23500)
## What
`MetricsKeys.CommonApiApplicationQueryRateLimited`
(`common-api-query/application-rate-limited`) is emitted from the
per-application throttler in `common-base-query-runner.service.ts`
**without any attributes**. Downstream that means a single
undifferentiated counter — there is no way to tell *which* application
is hitting `APPLICATION_API_RATE_LIMITING_LIMIT`.
This attaches the app dimension:
```ts
await this.metricsService.incrementCounterForEvent({
key: MetricsKeys.CommonApiApplicationQueryRateLimited,
shouldStoreInCache: false,
attributes: {
universal_identifier: authContext.application.universalIdentifier,
app_name: authContext.application.name,
source_type: authContext.application.sourceType,
},
});
```
## Why these three attributes
Same trio already emitted by `application-registration.service.ts`,
`application-install.service.ts` and `application-gauge.service.ts`, so
per-app API rate limiting joins cleanly against the existing app
lifecycle metrics rather than introducing a second naming scheme.
## Notes
- **No new data fetching.** `authContext` is already narrowed to
`ApplicationWorkspaceAuthContext` by the `isApplicationAuthContext`
guard at the top of the method, and the throttler key a few lines above
already reads `authContext.application.universalIdentifier`. `name` and
`sourceType` are plain columns on `FlatApplication`, present on the same
in-memory object.
- **Bounded cardinality.** `universalIdentifier` is stable for an
application across workspaces (see the unique index on
`(universalIdentifier, workspaceId)`), so the label set is bounded by
the number of distinct applications, not by installs.
- **`source_type` typing.** `ApplicationRegistrationSourceType` is a
string enum, assignable to OTel's `AttributeValue`.
## Context
The consuming dashboard panels are already merged-pending in
`twentyhq/twenty-infra` ([PR
#835](https://github.com/twentyhq/twenty-infra/pull/835)) and written
with a `coalesce(nullIf(Attributes['app_name'], ''),
nullIf(Attributes['universal_identifier'], ''), 'unknown')` fallback, so
they render today as a single `unknown` series and start splitting per
app automatically once this ships — no dashboard change needed on either
side of the deploy.
## Test plan
- Behaviour is unchanged: the counter still fires once per
`ThrottlerException`, and the error is still rethrown. Only the
attribute bag is new.
- I did not run the twenty-server test suite or typecheck locally — this
was authored against a shallow clone without a monorepo install, so I'm
relying on CI for both.
---
_Generated by [Claude
Code](https://claude.ai/code/session_01VbFtaCS5RkxDhLJApNSteV)_
<!-- This is an auto-generated description by cubic. -->
<a
href="https://cubic.dev/pr/twentyhq/twenty/pull/23500?utm_source=github"
target="_blank" rel="noopener noreferrer"
data-no-image-dialog="true"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img
alt="Review in cubic"
src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a>
<!-- End of auto-generated description by cubic. -->
|
||
|
github-actions[bot]
|
6db019686e |
i18n - translations (#23507)
Created by Github action --------- Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
Paul Rastoin
|
0c545bcdeb |
[BREAKING-CHANGE] Centralize system View viewField side effect (#23081)
# Introduction Closes https://github.com/twentyhq/core-team-issues/issues/2669 Part of the `isSystemSideEffect` engine-ownership effort. Until now, a custom object's default **INDEX** table view (`All {objectLabelPlural}`) and its view fields were built imperatively in `ObjectMetadataService` with random `v4()` identifiers, while `twenty-standard` authored its own copies with hardcoded literals. The two never converged, an object rename could drift the view, and nothing marked these rows as engine-owned. This PR makes the metadata side-effect engine the **single owner** of the INDEX view and its view fields, on name-free deterministic identifiers, for custom and standard objects alike. ## Core design - **Name-free deterministic identity.** The INDEX view identifier derives from `object identifier + ViewKey.INDEX` (`getSystemViewUniversalIdentifier`); each view-field identifier derives from `view identifier + field identifier` (`getViewFieldUniversalIdentifier`). An object rename (with a pinned object identifier) keeps the same view, losslessly. - **`isSystemSideEffect: true` is provenance.** Every INDEX view / view field the engine emits is flagged system-owned, so manifest deletion inference never drops it. The flag follows the view: a view field inherits its parent view's flag. - **The engine is the sole owner of the INDEX view.** It always emits it; a caller providing one with the same derived identifier is a genuine conflict surfaced by the engine's reserved-identifier collision, not silently deferred. ## Changes ### Shared (`twenty-shared`) - `getIndexViewUniversalIdentifier` → `getSystemViewUniversalIdentifier`, now taking a `viewKey` (generalizes to any singleton engine-owned view). - Standard field identifiers extracted into a new `STANDARD_OBJECT_FIELDS` constant, so both an object's `fields` and its INDEX view read the same field identifiers. - `buildStandardObjectIndexView` derives the standard INDEX view + view-field identifiers from `STANDARD_OBJECT_FIELDS`, replacing the hardcoded literals in `standard-object.constant.ts`. ### Metadata side-effect engine (custom objects) - **`objectSystemFieldsAndIndexViewOnCreate`** (replaces `objectSystemFieldsOnCreate`): on object creation, provisions the 7 reserved system fields **and** the INDEX view with one view field per displayable system field, all `isSystemSideEffect: true`. - **`fieldIndexViewFieldOnCreate`** (new): on field creation, provisions the field's INDEX view field. Object created in the same batch → visible, positioned before the system view fields; pre-existing object → hidden, appended (preserving the historical `createOneField` behavior). Both branches resolve the INDEX view by its derived identifier (single map access, never a scan). - **`fieldSystemViewFieldsOnDelete`** (new): on field deletion, cascade-deletes every engine-owned view field displaying it. - **`objectSystemSideEffectsOnDelete`** (extended): now also cascade-deletes the object's engine-owned views and their view fields (in addition to system fields, indexes, searchFieldMetadata). Every lookup walks a foreign-key aggregator down from the deleted object, so the work is proportional to what the object owns, never to workspace size. - Object-create and field-create positions are derived from the same caller-input field list, so the INDEX view layout is contiguous with no handler-ordering dependency. - `view` / `viewField` added to the side-effect companion metadata names for `fieldMetadata` and `objectMetadata`. ### Reserved-identifier invariant A caller can never define an entity whose identifier collides with one a system side effect produces: caller inputs are forced `isSystemSideEffect: false` at every entry point (API and app-manifest transpilers), and the engine raises `RESERVED_SYSTEM_UNIVERSAL_IDENTIFIER`, aborting the operation, when a system emission lands on a caller-claimed identifier. Covered by a new engine-level test. ### Caller-side provisioning removed The imperative INDEX view + view-field provisioning is removed from `ObjectMetadataService.createOneObject`. The record-page `FIELDS_WIDGET` view is intentionally left caller-side and deferred to the follow-up (see below). ### `twenty-standard` convergence Standard INDEX views and their view fields converge on the same derived-identifier + `isSystemSideEffect: true` scheme as the engine. `twenty-standard` syncs through the from/to migration path (which never runs the side-effect engine), so it authors this INDEX surface itself, matching what the engine produces for custom objects. ## Rollout Two `2.26.0` workspace commands, running after the `2.25` messageCampaign commands: - `upgrade:2-26:reconcile-index-view-universal-identifier` re-owns the INDEX views of the **twenty-standard and workspace-custom applications** and all their view fields to the derived identifiers with `isSystemSideEffect: true`, in a single per-workspace transaction. Each view field identifier is keyed on the application of the **displayed field** (an app or user column on a standard INDEX view converges too). Soft-deleted views and view fields are skipped: one can coexist with an active successor on the same derivation inputs and both would derive the same identifier. Children reference the view by primary key, so the re-own is lossless. - `upgrade:2-26:demote-and-backfill-application-index-view` handles **manifest-installed applications**, which never had their INDEX view auto-provisioned: every caller-authored INDEX view of another application is demoted to `key: null` (a plain additional view under its manifest identifier), then every application object gets the engine-owned INDEX view and its full view-field layout backfilled through the migration pipeline's legacy path (no side-effect expansion), views committed before view fields across applications since a view field belongs to the application owning its field. Idempotent and retry-safe: engine-owned INDEX views are neither demoted nor re-backfilled, and view creation and view-field creation are gated independently, so a retry after a partial failure still backfills the missing view fields of an already-committed view. Both support `--dry-run` and invalidate the full flat-maps closure (parents aggregate the re-owned identifiers, children resolve them as universal foreign keys, and page-layout widget universal configurations resolve view PKs at cache-build time). The `2.25` `upgrade:2-25:add-message-campaign-name-field` command is adapted to resolve the campaign INDEX view by its INDEX key on the object instead of by universal identifier: it now runs before the reconcile, on workspaces still holding legacy identifiers. ## ⚠️ Breaking change This PR **mutates 187 previously hardcoded universal identifiers** — the standard objects' INDEX views and their view fields (the literals removed from `standard-object.constant.ts`), now derived. - **Handled by the `2.26` commands above** for all existing workspaces. - **The INDEX key is now engine-reserved.** The flat view validator rejects caller-created INDEX views (API and manifest inputs are forced `isSystemSideEffect: false`) and enforces a single non-deleted INDEX view per object; `view.key` is no longer a comparable/updatable property, so no writer can promote or demote a view after creation. `ViewManifest.key` is deprecated and ignored (manifest views are always additional views, so old apps keep syncing and demoted views are not promoted back); the REST/GraphQL create path now rejects `key: INDEX`. In-repo example apps (`hello-world`, `document-generator`) no longer declare it. - **12 declared-but-never-seeded standard INDEX view field identifiers deleted** (the former `preservedViewFields` on `timelineActivity`, `workflowRun` and `workspaceMember`): after the reconcile, no workspace row references them. - **`computeFlatViewFieldsToCreate` now derives view field identifiers** instead of drawing `v4()` ones, which also changes what the committed `1-23` record-page backfill produces going forward (deliberate, documented in-code). - **Record-page views and view fields are not affected** (identifiers unchanged). - **In-repo apps: `twenty-last-contact` updated.** It was the only app declaring explicit INDEX view fields (10 columns across `allPeople` / `allCompanies` / `allOpportunities`) through manifest `viewFields`. Those target identifiers are now engine-owned and derived, so the manifest inputs no longer resolve and install failed with `View not found`. The app now declares only its fields; the engine's `fieldIndexViewFieldOnCreate` provisions the matching INDEX view field automatically. No other app under `packages/twenty-apps` references any of the 187 mutated identifiers, and apps that target standard views point at record-page views (e.g. `real-estate` → `opportunityRecordPageFields`) or their own objects (`twenty-partners`), all unchanged. ### Loss of granularity for app maintainers The engine now owns the INDEX view field of every field a caller adds to an object, so app maintainers lose direct control over those columns. Previously an app could target the engine-owned INDEX view with an explicit manifest `viewField` and set its `position` and `isVisible`. Now `fieldIndexViewFieldOnCreate` appends a **hidden** view field in caller-input order on field creation, so: - Columns an app previously showed at a **dedicated position** and **visible** (e.g. `twenty-last-contact`'s last-contact columns) become **hidden** and **appended in input order** after install. - There is currently **no manifest way to override** the engine-provisioned INDEX view field's position, visibility, or size. This is a deliberate regression accepted for the sake of single-ownership, and app maintainers should expect their INDEX columns to move/hide after upgrading. A follow-up override API will let maintainers reclaim per-field control over the engine-provisioned INDEX view field. ## Testing - Unit specs for each handler: object create (system fields + INDEX view/view fields, override, position offset), field create (same-batch vs existing-object, non-displayable noop, no-INDEX-view noop), field delete, object delete (fields/indexes/searchFieldMetadata/views/view fields cascade, reverse-relation view field on another object). - Engine-level test for the reserved-identifier collision. - `twenty-standard` guard test that its INDEX views/view fields stay on the derived scheme and stay system-owned. - Integration test: full engine provisioning of the INDEX view/view fields on object creation, same view id preserved across an object rename, and cascade delete on object deletion. ## Follow-up The full record-page stack (record-page view, its view fields, view field groups, page layout / tab / widget) is still built imperatively and moves into the engine in https://github.com/twentyhq/core-team-issues/issues/2721. |
||
|
Thomas Trompette
|
a0281f635b |
Restore soft-deleted junction record when re-adding a junction relation (#23371)
Fixes #23305. Removing a junction relation soft-deletes the intermediate junction record. Re-adding the same relation created a brand new record, which the composite unique index on the junction object (e.g. `personId` + `companyId`) rejected with "This record already exists". `useUpdateJunctionRelationFromCell` now creates the junction record through `useCreateManyRecords` with `upsert: true`. The server matches on the unique index with `withDeleted()` and clears `deletedAt` on the matched row instead of inserting. ## This revives, it does not create Worth being explicit, because it is a deliberate trade and not obvious from the diff. When a soft-deleted row exists for the pair, the user gets that row back. Same id, same `createdAt`, same `createdBy`, and anything attached to it (notes, files, and any fields the app added to the junction object). Verified on a dev instance: after re-adding a link through the picker, the row still reported a creation date from days earlier. That is fine for a pure link. It is a lie for a junction that carries data, for instance a `PersonCompanyRelationship` with a role and dates. The alternative designs are hard-deleting on detach (needs `canDestroyObjectRecords` on the junction object, which most roles do not grant) and partial unique indexes (needs `indexWhereClause` exposed to app-declared indexes, which the SDK does not support today). Both are larger changes. This one unblocks affected apps without requiring anything from them. Note that upsert conflict detection ignores `indexWhereClause`, so shipping partial indexes later would not change this behaviour on its own. ## Why the id handling changed The hook no longer sends a client-generated id. Under upsert the server decides which row you get, so the id in the input would be discarded. The optimistic store entry still uses a local id so the chip appears immediately, then adopts the persisted id once the mutation resolves. Without that, a revive left an id in the store that exists nowhere on the server, and the next removal failed with "This record does not exist or has been deleted". Supersedes #23365, which took the same approach but added `$upsert` to the shared `createOne` mutation. That capability already exists per call on `createMany`, and widening the shared document broke the `useCreateOneRecordMutation` and `useCreateOneRecord` tests. ## Testing Verified end to end on a dev instance with a junction object carrying a non-partial composite unique index, since the dev seed does not create one and the bug cannot reproduce without it: - before: re-adding after a removal fails with "This record already exists" - after: the original row is restored, `deletedAt` cleared, one row throughout, and removing again in the same session works, with no GraphQL errors Added an integration test for the path this depends on: a soft-deleted record matched on its unique fields alone, with no id in the input. The existing coverage only exercised upsert by explicit id. ## Known gaps, deliberately not addressed here **Toggling a link off while its creation is still in flight.** The store id is provisional until the mutation returns, so a removal issued inside that window deletes an id the server never received. Reproduced by stalling the create and clicking remove during it: `CombinedGraphQLErrors: Record not found`, and the link stays active despite the user removing it. The window is one mutation round trip. Left for a follow-up; the likely fix is a per-record operation queue so adds and removes on a field run in click order. **Junction objects with no unique index.** Remove then re-add still creates a duplicate row there, on this branch as on main, because conflict detection is driven by unique indexes. --------- Co-authored-by: Shinu Cherian <129690295+Shinu-Cherian@users.noreply.github.com> |
||
|
Abdul Rahman
|
ea7863dc4e |
feat(ai-agent): lazy tool loading for the open-ended runAgent path (#23454)
## Context `AgentAsyncExecutorService.executeAgent` is shared by workflow agent nodes and the `runAgent` mutation (Slack assistant, apps). Workflow nodes run a scoped task, so pre-loading the few explicitly-granted object tools is fast and skips the `learn_tools` round trip (the behavior settled in #23400 / #23358). `runAgent` is open-ended: its role grants broad object access, so pre-loading inlines every CRUD/action schema on every step. That is what makes the Slack assistant take 3-4 minutes for a prompt Ask AI answers in seconds. Confirmed still slow with #23400 merged, so this is the payload, not object scoping. ## What Add a `toolLoadingStrategy` to `executeAgent` (default `'preload'`, so workflow nodes and evals are unchanged). `AgentRunService.run` opts into `'lazy'`, which exposes a compact tool catalog in the system prompt plus the `learn_tools` / `execute_tool` meta-tools, using composed role permissions rather than explicit grants only, so the agent keeps broad access without the full-payload latency. ## How - Split tool provisioning into two focused methods on the executor; a 3-line dispatch chooses per strategy. The pre-load path is unchanged. - `buildLazyRegistryToolset`: one reusable definition of lazy registry provisioning (catalog + meta-tools), so the chat and agent executors can share it. - Extract `buildToolCatalogSection` out of `SystemPromptBuilderService` into a `tool-provider` util so both paths format the catalog identically (no dup). - Replace the meta-tools' `excludeTools` denylist with a single `isToolAllowed` predicate: the agent path passes an allowlist closed over the shown catalog (enforced at call time), MCP passes its existing deny predicate. Workflow node and eval behavior is unchanged. <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23454?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Thomas Trompette
|
9dd9709e97 |
test(workflow): cover the workflow core mirror end to end (#23434)
## What Adds the integration coverage `core.workflow` mirroring never had: create, rename, delete, restore and destroy, asserting the core row appears, follows the rename, disappears and comes back. No production code changes. The async `WorkflowCoreDualWriteListener` stays as the mirror for the workflow entity. ## Why this PR changed shape It originally replaced the listener with query hooks, to remove the last async best-effort write path. That was the wrong call, for three reasons found while reviewing it: 1. **It would have introduced drift, not removed it.** `workflow-trigger.workspace-service.ts` updates `lastPublishedVersionId` via `workflowRepository.update(...)` when a version is activated. That is a repository write, not an API mutation, so query hooks never fire for it and `core.workflow.lastPublishedVersionId` would have gone stale on **every workflow activation**. The consistency cron compares that field, so it would surface as `fieldMismatch` drift. The listener catches it because events are emitted at the ORM layer. 2. **Hooks are no more atomic than the listener here.** Workflow CRUD goes through the generic API, so there is no dedicated mutation to wrap and the generic runner holds no transaction: it commits, then runs post-hooks. Both approaches are post-commit, with the same drift guarantees. 3. **Events carry the full record; hook payloads do not.** `workspace-insert-query-builder` passes the full formatted result to the event while the client response is filtered by `returning`. That partial payload is what forced the re-fetches, the `coreWorkflowId` pre-hook injection, and the destroy pre-commit special case. All three were workarounds for a problem the listener does not have. The principle we settled on: **use the transaction where one exists, use the listener where there isn't one.** Post-hooks were the worst of both here. Version *content* writes keep their transactional mirror, since those go through dedicated mutations that own a transaction. ## Note on the test The mirror is asynchronous, so the assertions poll (20 attempts, 250ms) rather than reading core immediately after the mutation returns. Without that it would be racy. ## Verification - `nx typecheck twenty-server` green - `oxfmt` + `oxlint --type-aware` green |
||
|
Etienne
|
5ebcce0a51 |
feat(ai-tool): resolve and default icons in AI metadata tools (#23480)
## Context Objects and fields created through the AI chat / MCP metadata tools almost never get an icon, so they all render with the meaningless `123` fallback icon. Two causes: - The `icon` tool input was described only as `"Icon name"`, so the model had no idea what the value space is and mostly skipped an optional field it couldn't fill confidently. - Any invalid name is silently swapped for `Icon123` by `useIcons.getIcon` on the frontend, so near-misses were indistinguishable from unset. ## What this PR does **Guide the model** (icon names are Tabler names, which LLMs know well): - `icon` / `targetFieldIcon` schema descriptions now state the convention with examples (`IconBuildingSkyscraper`, `IconPaw`, …) and ask for one to always be set - The `metadata-building` skill gains an "Icons" section; the MCP server instructions gain a one-line reminder **Normalize server-side** (new `resolveIconName` util, used by all create/update/batch metadata tool executes incl. `relationCreationPayload.targetFieldIcon`): - Fixes shape mistakes: raw tabler slugs (`"building-skyscraper"`), separators, missing or lowercased `Icon` prefix - Deliberately does NOT validate existence against the full ~4.2k icon registry — an unknown name is harmless since the frontend falls back to its default icon, exactly as for icons stored via the API today - Unusable input (empty/garbage) resolves to nothing: creates fall back to a default, updates keep the existing icon **Fall back sensibly for fields**: - New `FIELD_TYPE_DEFAULT_ICONS` in `twenty-shared/constants` maps every `FieldMetadataType` to a sensible icon (mirroring the settings UI type illustrations), applied when the model provides no usable icon — an AI-created field always gets a meaningful icon - Lives in twenty-shared so the frontend can reuse it later (e.g. as `getIcon`'s custom default for fields) The REST/GraphQL metadata APIs are untouched — this only affects the AI tool layer. ## Test plan - `resolve-icon-name.util.spec.ts` — canonical pass-through, slug/prefix/separator fixes, unknown-name pass-through (FE fallback contract), unusable inputs, icon-key dropping on updates - `FieldTypeDefaultIcons.test.ts` — every field type mapped, all values canonically shaped (values hand-checked against the twenty-ui `ALL_ICONS` registry) - `nx typecheck` twenty-server + twenty-shared, oxlint/oxfmt clean on changed files <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23480?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
github-actions[bot]
|
2304d19c8e |
i18n - docs translations (#23483)
Created by Github action <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23483?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
BOHEUS
|
a3b54e834c |
PDF upload fix (#23473)
Sometimes uploading PDF files resulted in "Non-whitespace before first tag." error, updating parsing library fixes the error <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23473?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Paul Rastoin
|
25d20731ac |
Include root cause errors in workspace migration runner exception message (#23416)
closes https://github.com/twentyhq/core-team-issues/issues/2733 ```diff - "message": "Migration action 'create' for 'index' (universalIdentifier: 67c6811e-...) failed" + "message": "Migration action 'create' for 'index' (universalIdentifier: 67c6811e-...) failed: [workspaceSchema] could not create unique index "IDX_UNIQUE_85951922a2..." (pg code: 23505, detail: Key ("externalId")=(DUPLICATED-VALUE) is duplicated.)" ``` ## Problem When a workspace migration action fails, the `EXECUTION_FAILED` exception message only states which action failed: ``` Migration action 'create' for 'index' (universalIdentifier: 9e20a0f6-7a18-51c5-a422-5dc4dbd1d972) failed ``` The underlying errors (`metadata`, `workspaceSchema`, `actionTranspilation`) are attached to the exception instance but dropped by most surfaces: the SDK CLI only prints `errors[0].message`, server logs only log `error.message`, and the REST/GraphQL paths lose the postgres driver details. Debugging a failed app install (like the stale index universalIdentifier case in the issue) requires guessing. ## Change - New `formatWorkspaceMigrationRunnerExecutionErrors` util that builds a compact one-line summary of the underlying execution errors, including the postgres error code and `detail` for `QueryFailedError`, capped at 1500 chars. - The `EXECUTION_FAILED` exception message now appends that summary: ``` Migration action 'create' for 'index' (universalIdentifier: 9e20a0f6-...) failed: [workspaceSchema] relation "IDX_..." already exists (pg code: 42P07) ``` Since every surface (CLI, server logs, Sentry, REST, GraphQL) shows `error.message`, the root cause now propagates everywhere without touching those surfaces. - Since actions run inside a single transaction, when one branch fails with `25P02 current transaction is aborted` (collateral of the other branch's statement aborting the transaction), the summary keeps only the real root cause. A lone 25P02 error is still shown. ## Notes - The SDK's `getSyncErrorRecoveryHint` matching (`/migration action .* failed/`) still works with the suffixed format. - Commit-time failures from `DEFERRABLE INITIALLY DEFERRED` FK constraints still bypass this path (wrapped as `INTERNAL_SERVER_ERROR` with no action attribution) and are left as a follow-up. ## Tests - New spec for the formatter util (labels, pg code/detail, 25P02 demotion, truncation). - Extended `workspace-migration-runner.exception.spec.ts` with a root-cause message assertion. - Updated `format-upgrade-error-for-storage` snapshots (first line now carries the enriched message). --- _Generated by [Claude Code](https://claude.ai/code/session_01Mu8t74X14oYVkLrFBZHs2o)_ <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23416?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
BOHEUS
|
6d60fac92f |
Dashboard doc fix (#23475)
<!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23475?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Thomas Trompette
|
198e3969df |
feat(workflow): read workflow version content from core in the engine, behind a flag (#23403)
## Scope: engine only Behind the existing `IS_WORKFLOW_VERSION_IN_CORE_ENABLED` flag (**off by default**), `getWorkflowVersionOrFail` sources a version's `trigger` / `steps` / `status` from `core.workflowVersion` instead of the workspace row. That covers its 11 call sites: workflow build, validation, schema, run, and trigger dispatch. **The UI is not switched here.** `useWorkflowVersion` and the content fetch in `useWorkflowWithCurrentVersion` still go through generic object CRUD against the workspace columns. That work needs the client to stop treating the workspace record as the home for content, and is planned separately around `flowComponentState` (the jotai state the builder already reads from). It is the read that must land before the workspace `trigger` / `steps` columns can be dropped. ## Why an overlay, not a repository swap `core.workflowVersion` is a content-only projection: its own `id` (not the workspace version id), `workflowId`, `triggers[]`, `steps[]`, `status`. No `name`, no `position`. So core cannot fully back the entity. The flip is therefore an overlay: identity, name and position stay from the workspace row, and only content comes from core (`triggers[0] -> trigger`). ## Reversibility Falls back to workspace content when the flag is off (default), when the version has no `coreWorkflowVersionId` soft-ref, or when the core row is missing. Merging changes nothing until the flag is enabled per workspace, and it can be flipped back at any time. ## Verification - `nx typecheck twenty-server` green, `oxfmt` + `oxlint --type-aware` green - Unit test covering four branches: flag off, flag on with the core row present (overlays content **and** preserves `id` / `name` from the workspace row), flag on with the core row missing (falls back on `trigger`, `steps` and `status`), flag on with no soft-ref (skips the core read) ## Enablement gate Do not enable the flag in any workspace until the drift dashboard reports zero drift for `core.workflowVersion` and legacy drift is repaired. Enabling before that turns latent drift into live dispatch behaviour. |
||
|
Félix Malfait
|
b602294f1d |
Hide the command menu button while the mobile side panel is open (#23471)
On mobile the side panel covers the page, but the page header stays mounted underneath. Its command menu button (`⌘K`, the `⋮` icon) sits at the same coordinates as the panel's own close button, so the two icons render on top of each other. Measured on a 390x844 viewport with the AI chat open: - `Command Menu` button at `x=346, y=8, 32x32` - `Close side panel` button at `x=358, y=14, 24x24` `SidePanelToggleButton` already hid itself for the command menu and search pages, but the AI chat pages (`AskAI`, `ViewPreviousAiChats`) are not in `COMMAND_MENU_SIDE_PANEL_PAGES`, so the button stayed and overlapped. ## Change Hide the button on mobile whenever the side panel is open, rather than enumerating pages — the header is not reachable behind a full-screen panel either way. Layout customization mode is the exception and keeps it: `alignWithSidePanelTopBar` deliberately repositions the button into the side panel top bar there, so that path is preserved. Desktop is unaffected. ## Testing Three cases added to `SidePanelToggleButton.test.tsx` (hidden on mobile with the panel open, kept on mobile in layout customization mode, kept on desktop with the AI chat open); the `useIsMobile` mock is now switchable per test. All 10 tests pass. Verified in the browser at 390x844: with the AI chat open only `Close side panel` remains in the top bar, and the button reappears once the panel is closed. --- _Generated by [Claude Code](https://claude.ai/code/session_018gcsCQbuTMsyFWv874p25Q)_ <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23471?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Félix Malfait
|
0859133774 |
Disable double-tap zoom while keeping pinch zoom (#23476)
Adds `touch-action: manipulation` on `body`. Context: #8477 disabled auto-zoom on iOS only, via `maximum-scale=1` behind a UA check. That leaves double-tap-to-zoom active everywhere, which is what makes taps feel laggy on mobile (the browser waits ~300ms to see if a second tap is coming) and what causes accidental zooms when tapping small targets twice in a row. `touch-action: manipulation` removes double-tap-to-zoom and the associated tap delay, and leaves pinch-to-zoom fully intact. So the page still zooms the way a website should, it just stops zooming when you didn't ask it to. This is deliberately not a revert of #8477 and not an extension of `maximum-scale` to Android: blocking pinch zoom fails WCAG 1.4.4, and being able to zoom is part of what makes this feel like a website rather than a native app. --- _Generated by [Claude Code](https://claude.ai/code/session_018gcsCQbuTMsyFWv874p25Q)_ <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23476?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Thomas Trompette
|
933ae9c20b |
fix(page-layout): render standalone rich text widget on record pages (#23435)
Fixes #21093 ## Problem A `STANDALONE_RICH_TEXT` widget can be created on a record page layout through the metadata API, and `getPageLayoutWidgets` returns it, but the record page renders nothing for it. `StandaloneRichTextWidget` only resolved a target id when `layoutType === PageLayoutType.DASHBOARD`, then bailed out with `return null` whenever that id was undefined. On a record page it never rendered. ## Why the guard was there It was correct when it was written. In #16437 the widget used the full `BLOCK_SCHEMA` and uploaded files: ```ts return await uploadAttachmentFile(file, { id: dashboardId, targetObjectNameSingular: CoreObjectNameSingular.Dashboard, }); ``` Without a dashboard id there was nowhere to attach an upload, and dashboards were the only layout type in play, so refusing to render was a coherent stance. #17934 then disabled file upload because the urls were not signed. It swapped in `DASHBOARD_BLOCK_SCHEMA`, dropped `useUploadAttachmentFile` and `prepareBodyWithSignedUrls`, and added `filterSupportedBlocks` to strip file blocks out of previously saved bodies. It left behind the attachments query, the `attachments` prop and the `useAttachmentSync` call. After that, `dashboardId` had one consumer left: a filter that could no longer match anything. The `return null` underneath it was guarding nothing. Record page layouts then made the widget reachable outside dashboards, and the stale guard blanked it. ## Fix The body lives on the widget configuration, not on the target record, so the widget needs no record id to display. The leftover attachment fetch has nothing to act on: - `DASHBOARD_BLOCK_SCHEMA` declares only paragraph, heading, lists, checklist, codeBlock, table and quote. No image, file, video or audio block. - The three sync utils all key off `ATTACHMENT_BLOCK_TYPES = ['image', 'file', 'video', 'audio']`, so they return empty for any body this editor can produce. - No `uploadFile` option, no `onPaste` handler, and `filterSupportedBlocks` strips unsupported blocks on load, so such a block cannot get in. So rather than generalise the attachment filter to every object type, this removes it: the `useFindManyRecords` call, the `attachments` prop, and the `useAttachmentSync` call in `StandaloneRichTextEditorContent`. It finishes the cleanup #17934 started. `useAttachmentSync` is untouched and still used by `RichTextFieldEditor`, which does support file blocks. Net result is a pure deletion, and the widget renders on every layout type. If image blocks are ever added back to `DASHBOARD_BLOCK_SCHEMA`, attachment sync will need to come back with them. ## Testing Local instance, widget created through `createPageLayoutWidget` on the default Company record page layout. - On the unpatched component the widget is absent from the page. - With the fix it renders read-only in the record page column. - Also verified with the payload shape from the issue (`markdown` set, `blocknote: null`); the server converts it to blocknote on write, so it renders too. - Verified on `calendarEvent`, an object with no `attachments` relation. Renders clean, no console or GraphQL errors. Generalising the old filter instead would have sent `targetCalendarEventId` and hit `Object attachment doesn't have any "targetCalendarEventId" field.` - Dashboard rendering unchanged, and editing still round-trips: typed into the widget in dashboard edit mode, hit Save, confirmed the new body in `core.pageLayoutWidget`. |
||
|
Félix Malfait
|
5d90fb33c0 |
Open records on a full page instead of a side panel on mobile (#23474)
On mobile the side panel covers the whole screen, so a record opened in
it arrives cramped behind an "Open" button offering the full page it
should have gone to in the first place.
`useResolveOpenRecordIn` already forces `RECORD_PAGE` on mobile via
`canDisplaySidePanel: !isMobile`, but it is a resolver callers have to
opt into, and only five do. Thirteen other call sites reach
`useOpenRecordInSidePanel` directly and get a panel on every device,
including:
- `TaskRow` and `NoteTile`, the activity lists inside a record's tabs
- `EventRowActivity`, `EventCardMessage`, `EventRowGenericLinked` on the
timeline
- `SidePanelSearchRecordsPage`, `EmailThreadPreview`,
`useOpenCreateActivityDrawer`, `useAddNewRecordAndOpenSidePanel`
## Change
Decide it inside `useOpenRecordInSidePanel` rather than at each call
site, so no caller can wedge a record into a panel by forgetting to ask.
On mobile it closes the panel and navigates to `AppPath.RecordShowPage`,
then returns before any of the side-panel setup runs.
Two details carried over so the redirect is not lossy:
- `setRecordPageActiveTabId` still runs first, so a caller passing `tab`
lands on the right tab.
- `isNewRecord` forwards `{ isNewRecord, objectRecordId,
labelIdentifierFieldName }` as navigation state, mirroring what
`useCreateNewIndexRecord` already does on its `RECORD_PAGE` branch, so a
freshly created record still opens its title for naming instead of
arriving untitled.
Side-panel-only effects are skipped rather than lost.
`runWorkflowRunOpeningInSidePanelEffects` ends in
`openWorkflowRunViewStepInSidePanel`, which auto-opens a step *in the
panel*; with no panel there is nothing for it to do, and the workflow
run's record page renders its own diagram.
The two hooks that already branch on `useResolveOpenRecordIn`
(`useOpenRecordFromIndexView`, `useCreateNewIndexRecord`) never call
into this path on mobile, so this is a no-op for them rather than a
double navigation.
Uses `useIsMobile` rather than `useIsTouchDevice`, matching
`useResolveOpenRecordIn`: this is a question of whether there is room
for a panel, not of how the user points.
## Testing
At 390x844, opening the search side panel and tapping a result now
navigates to `/object/person/<id>` with the panel closed, where it
previously stayed in the panel. Typecheck and lint clean.
---
_Generated by [Claude
Code](https://claude.ai/code/session_018gcsCQbuTMsyFWv874p25Q)_
<!-- This is an auto-generated description by cubic. -->
<a
href="https://cubic.dev/pr/twentyhq/twenty/pull/23474?utm_source=github"
target="_blank" rel="noopener noreferrer"
data-no-image-dialog="true"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img
alt="Review in cubic"
src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a>
<!-- End of auto-generated description by cubic. -->
|
||
|
github-actions[bot]
|
3e9e75e774 |
i18n - docs translations (#23466)
Created by Github action Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
github-actions[bot]
|
75047f3237 |
i18n - translations (#23463)
Created by Github action --------- Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
martmull
|
00e418a039 |
Show call recorders as calendar event participants (#23380)
Closes twentyhq/core-team-issues#2729 Call recordings attached to a calendar event are now displayed next to the human participants, in the timeline event card (`EventCardCalendarEvent`). They are rendered as the source app's `AppChip`, rounded so it sits in the participant avatar group, with the recording status in tooltip https://github.com/user-attachments/assets/e0c393c8-fd65-4468-8c4f-503dd22c13d5 ## Before No call recorder chip displayed TODO: add this in the calendar views (`CalendarEventRow`) |
||
|
github-actions[bot]
|
f84f242f9d |
i18n - docs translations (#23460)
Created by Github action <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23460?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
Félix Malfait
|
840c6d0129 |
Take openRecordIn from the view in scope instead of a global atom (#23422)
Stacked on #23424 (mobile chip navigation). Review that one first; the diff shown here is only the delta. ## Problem `recordIndexOpenRecordInState` was a global atom mirroring `view.openRecordIn`. It was written whenever any index view loaded and never reset, so a record chip behaved according to whichever view had been browsed last: - Companies view set to "record page". Open a Company, tap a related Opportunity chip. The Opportunities view says "side panel", but the chip reads the leftover Companies setting and opens a full page. - Visit the Opportunities index first, then the same Company page, and that same chip now opens a side panel. The setting is per view in the database, but the frontend kept it in one slot as though it were a user preference. ## Approach The value already lives on the view, so the mirror is deleted rather than scoped: - `useResolveOpenRecordIn` reads the current view of the surrounding context store. On a record index that is the view being displayed. On a record show page `MainContextStoreProvider` resolves a view for the object in the URL — the last visited view for that object, falling back to its index view — so chips there follow a view belonging to the object they sit on, rather than whatever was loaded last. - Where no context store is mounted at all (a mention inside a note, for instance) there is no view to take a setting from, so the hook falls back to `DEFAULT_VIEW_OPEN_RECORD_IN`. The instance lookup is non-throwing on purpose: `RecordChip` renders in a lot of places, and an existing test caught this crashing when the read was strict. - The options dropdown now reads and writes `currentView.openRecordIn` directly, the same way `isCompact` beside it already works, so `setAndPersistOpenRecordIn` only has to persist. - `useGetOpenRecordIn` is gone; every call site had the object name available at render, so the reactive hook covers all of them. ## Behaviour change A chip whose behaviour previously came from an unrelated view now follows the view in scope. That is the point of the change, but it does mean some chips will open somewhere different from before, always in the direction of "what this list is configured to do" rather than "what the last list was configured to do". ## Testing - New `useResolveOpenRecordIn` tests: falls back to the default with no context store, follows the context store's view when there is one. - Full frontend suite: 951 suites, 5598 tests passing. Typecheck and lint clean. - Not exercised in a running app: no database in this environment. The dropdown's optimistic behaviour in particular relies on the same view store refresh that `isCompact` already depends on, so it is worth a click-through before merge. |
||
|
github-actions[bot]
|
057468343f |
i18n - translations (#23459)
Created by Github action Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
Félix Malfait
|
8990334cad |
Make record chips open records natively on touch devices (#23424)
Two mobile problems in the record table: tapping a chip in the first column takes two taps, and chips in every other column open a side panel where a full page is wanted. #23422 is stacked on this branch. ## Two taps to open a record The table's interactive layer lives in a hover portal mounted from `onMouseMove` on the table wrapper. Touch has no hover, so the browser fakes one, and the synthesised `mousemove` arrives *before* the `mousedown`. React commits the portal in the microtask between them, so the whole tap is hit-tested against a subtree that did not exist when the user aimed. Confirmed in Chromium with real touch input (`page.tap`, Pixel 5 emulation), mounting an overlay from the mousemove handler: ``` mousemove target=chip >>> overlay mounted <- the hover portal mousedown target=portalChip <- a node that did not exist when the finger went down mouseup target=portalChip click target=portalChip ``` The same test also ruled out `preventDefault` on the compat `mousedown` as a cause, and showed a `setTimeout`-deferred mount does *not* retarget — it is specifically React's sync flush timing that does. So hover state is now only tracked on hover-capable pointers. `useMoveHoverToCurrentCell` becomes the single writer and absorbs the deduplication `RecordTableContent` was duplicating inline. The interaction/layout split matters here: `useIsMobile` is a 768px width query, which answers "how much room is there to lay out", not "how does this person point". The new `useIsTouchDevice` uses `(hover: none) and (pointer: coarse)`. Layout keeps using width; interaction uses capability. ## Side panel on mobile "Where does a record open" was computed independently in six places and only `useOpenRecordFromIndexView` knew about mobile. `RecordChip` — every chip outside the first column, plus board cards and relation fields — had its own copy without that check. On mobile the side panel animates to `fullScreen`, so it is a full-page view with no URL and no back button. That decision now lives in one `resolveOpenRecordIn`: the view setting is an intent, and the side panel is only a real destination when there is room for it and the object supports it. Also here: `MOUSE_DOWN` navigation downgrades to `CLICK` on touch. It only buys a frame on a real pointer, since a tap synthesises its mouse events after the finger is already gone. ## Hover styling Separate layer, same root cause. A tap leaves CSS `:hover` applied until the next tap lands elsewhere, so a row you came back from keeps reading as selected. Nine `:hover` blocks across the record table, `Chip` and `Avatar` are now fenced behind `(hover: hover)` — the same media feature `useIsTouchDevice` branches on, via a new `hover-capable` SCSS mixin on the twenty-ui side and inline media queries in the Linaria components. Desktop rendering is unchanged, since Chrome matches `hover: hover`. Verified the built CSS emits the wrapper correctly, and checked the nested form through stylis directly for the Linaria side. ## Testing - New unit tests for `resolveOpenRecordIn` and for hover not being tracked on touch devices. - Full frontend suite: 951 suites, 5598 tests passing. Typecheck and lint clean. - Not observed end to end in a running app: no database in this environment, and the `RecordIndexPage` story renders an empty table under its msw mocks. The browser-level mechanism is verified and the fix removes the mid-gesture DOM change, but it is worth one pass on a real device before merge. ## Follow-ups not in this PR - The whole first cell navigates but only the chip-sized part of it gives tap feedback, and `isRecordTableRowActive` is only set on the side panel path — setting it on the navigate path too would keep the row lit while the page loads. - Rows are 32px against a 44px minimum touch target. - Giving the side panel a URL would make "panel vs page" a rendering decision on the same location, rather than something each call site has to branch on. --- _Generated by [Claude Code](https://claude.ai/code/session_019cDWPgWESbdRUhGxGb66j8)_ <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23424?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
Félix Malfait
|
2d9582117b |
feat(front): preview highlighted record in the search command bar (#23413)
Previewing the highlighted record while searching. The preview is a card anchored to the highlighted result, and its fields come from the object's **index view** (the base list view), so what you see while searching matches the list you came from. ## Screenshots > Note: these were taken before the latest design pass (smaller header, field cap). The layout is otherwise unchanged. Highlighted result, showing the index view's columns:  Overflow and hidden columns sit behind the `More (N)` expander:  Arrowing to a Person re-anchors the card and renders that object's own index view fields:  ## Changes - `SidePanelSearchRecordPreviewCard.tsx` — the card: the shared `SidePanelPageInfoLayout` header (avatar + name + created-at), then read-only `FieldDisplay` rows, then the `More (N)` expander. - `useSidePanelSearchRecordPreviewFields.ts` — resolves the object's index view through `useViewOrDefaultView` and splits its `viewFields` into visible and hidden, sorted by position, dropping the label identifier since the card already shows the record name. - `useSidePanelSearchRecordPreviewItem.ts` — resolves the highlighted item from the selectable list, following the selection immediately. - `useSidePanelSearchRecordPreviewRecord.ts` — hydrates the record into the record store so field displays can read their values. The fetch is debounced 200ms so holding an arrow key doesn't fire a `findOne` per row crossed, and it reports whether the record is hydrated yet. - `SidePanelSearchRecordsPage.tsx` — anchors the card with `AppTooltip` (`place="left-start"`, controlled `isOpen`, `clickable` so the expander is reachable) against a per-result anchor id. ## How many fields show Collapsed, the card shows at most seven of the index view's visible columns. Everything past that, plus the columns hidden in that view, sits behind `More (N)`. So the expander appears whenever there are more fields than fit, not only when the view happens to have hidden columns. ## Keeping the card stable while it loads The first cut jumped: measuring it over time gave `288px → unmounted for ~200ms → 232px`. Two separate causes, both fixed. It was **unmounting between records** because the previewed item was debounced and briefly resolved to `null`. The selection is now followed immediately and the *fetch* is what's debounced instead, so the card is reused across records rather than remounted. Its **size was derived from the data**, so every value that arrived nudged the layout. The header, rows (24px) and width are fixed, with skeleton placeholders for values until the record is hydrated. The field list comes from view metadata, which is available synchronously, so the card is its final size on first paint. Measured after that change: a constant `360x328` across 13 consecutive records spanning Person and Workspace Member, and no unmount. The skeleton and loaded states are the same height, so values just fade in. The card still disappears briefly on a brand-new search. That tracks the results list turning over — the anchor row it attaches to is genuinely removed from the DOM — so following the list is the correct behaviour there rather than holding a stale card against a deleted anchor. ## Notes The preview is read-only (`FieldDisplay`, not `RecordInlineCell`) — a floating preview isn't the right place to start an inline edit, and it keeps the card out of the field hover/edit portal machinery. The `More` button is wrapped in a container that prevents the default mousedown focus shift. Without it, clicking the expander moved focus out of the search input and arrow keys started driving the record table behind the panel instead of the results list. The section heading still reads `Results`; the design says `Records`. Left as-is since it is out of scope here. ## Testing - `nx typecheck twenty-front` passes. - `oxlint --type-aware` and `oxfmt` clean on the changed files. - Verified manually against a seeded dev workspace: anchoring and re-anchoring on arrow navigation, Company vs Person rendering their own index view fields, the `More` expander, arrow keys still driving the results list after clicking it, and the card holding a constant size through load. |
||
|
Abdul Rahman
|
965e033f3f |
[breaking-change] fix(server): return runAgent execution failures as result errors (#23390)
## Summary
- When `executeAgent` throws, `runAgent` now returns `{ success: false,
error }` instead of a GraphQL exception
- Callers (workflows, Slack assistant, etc.) can surface the failure to
users instead of hanging or failing opaquely
Run agent exception response error format breaking-change, errors moved
from the GraphQL error channel into the response payload.
<!-- This is an auto-generated description by cubic. -->
<a
href="https://cubic.dev/pr/twentyhq/twenty/pull/23390?utm_source=github"
target="_blank" rel="noopener noreferrer"
data-no-image-dialog="true"><picture><source
media="(prefers-color-scheme: dark)"
srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source
media="(prefers-color-scheme: light)"
srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img
alt="Review in cubic"
src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a>
<!-- End of auto-generated description by cubic. -->
|
||
|
github-actions[bot]
|
5b10869a2b |
i18n - docs translations (#23455)
Created by Github action Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
Thomas Trompette
|
b6a4c635ee |
fix(workflow): rename the trigger step through the dedicated mutation (#23450)
## Bug Renaming the **trigger** step from the workflow side panel fails with: > Updating a workflowVersion through the generic mutation is restricted. steps, trigger, status, position, workflowId and coreWorkflowVersionId cannot be changed... Renaming a **regular** step works, which is why this is easy to miss: only the trigger branch is broken. ## Cause This is a regression from #23207. That PR added the server-side denylist on `updateOneWorkflowVersion` and switched `useUpdateWorkflowVersionTrigger` to the dedicated `updateWorkflowVersionTrigger` mutation, but missed the call site in `SidePanelWorkflowStepInfo`, which still did: ```ts if (isTrigger) { await updateOneWorkflowVersion({ // generic mutation, sends `trigger` updateOneRecordInput: { trigger: { ...stepDefinition.definition, name: title } }, }); } else { await updateWorkflowVersionStep({ ... }); // dedicated, unaffected } ``` The observed request confirms it: `UpdateOneWorkflowVersion` with `input.trigger`. ## Fix Route the trigger branch through `updateTrigger`, which already resolves the draft version, calls the dedicated mutation, marks the step for recomputation and updates the cache. `useUpdateWorkflowVersionTrigger` now accepts an **optional** `instanceId`. This matters here: the side panel computes the visualizer instance id explicitly (it already passes it to `useGetUpdatableWorkflowVersionOrThrow`), and without it the hook would resolve the updatable version from a different component instance. Being optional, the four existing callers are unaffected. Also removes the now-redundant `getUpdatableWorkflowVersion()` call on the trigger path, so a rename no longer risks resolving the draft twice. ## Verification - `nx typecheck twenty-front` green - `oxfmt` + `oxlint --type-aware` green on both changed files - `useUpdateWorkflowVersionTrigger` unit tests green (2/2) - Not yet clicked through locally; the reporter hit this on a dev instance and can confirm the rename now succeeds <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23450?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> |
||
|
github-actions[bot]
|
9acc5c192f |
i18n - docs translations (#23452)
Created by Github action <!-- This is an auto-generated description by cubic. --> <a href="https://cubic.dev/pr/twentyhq/twenty/pull/23452?utm_source=github" target="_blank" rel="noopener noreferrer" data-no-image-dialog="true"><picture><source media="(prefers-color-scheme: dark)" srcset="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"><source media="(prefers-color-scheme: light)" srcset="https://www.cubic.dev/buttons/review-in-cubic-light.svg"><img alt="Review in cubic" src="https://www.cubic.dev/buttons/review-in-cubic-dark.svg"></picture></a> <!-- End of auto-generated description by cubic. --> Co-authored-by: github-actions <github-actions@twenty.com> |
||
|
Charles Bochet
|
ea2de2dc2b |
ci(pr-review): label-only manual trigger, thin dispatcher (#23449)
Addresses the review feedback on #23418 (Paul + Copilot + cubic) and switches the manual trigger from comments to **labels**, consistent with the e2e labels. ## What changed - **Manual reviews are label-driven** — add `pr-review-security` / `pr-review-triage` / `pr-review-standard`. Labeling requires write access (team-only). The **`/pr-review` comment trigger is removed.** - Like the e2e labels, a check runs on every push **while its label is present** (the orchestrator reads the PR's current labels each run) — so `pr-review-standard` keeps the deep review current until removed. - **Dispatcher is now dumb** — it forwards only `pr_number`. All resolution + validation lives in the privileged orchestrator (Paul's suggestion: it fetches PR metadata, incl. labels, there anyway). This fixes the bot findings (regex allowlist bypass, `/pr-review`→standard default, delimiter edge cases) at the source. - `cancel-in-progress: true` (latest-push-wins, matching the previous dispatcher). Fires on non-draft PR events (the auto `security,triage` gate) and on `pr-review-*` label adds. ## Depends on A companion change to the privileged CI (reads labels + resolves/validates checks) — merge that first; it's backward-compatible, so nothing breaks in between. |
