* fix: preserve debouncing settings for post-preprocessing arg accumulation
After preprocessing completes, store the flow's debouncing settings in
runnable_settings_handle on v2_job_queue so that maybe_apply_debouncing
can find them when the surviving job is pulled. Without this, the handle
is NULL and arg accumulation silently does nothing for flows with
preprocessors.
Also adds a debouncing badge in flow settings and 4 focused accumulation
tests covering scripts, flows without preprocessor, flows with
preprocessor (with and without the fix).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* chore: update sqlx prepared query for worker_flow.rs change
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* fix(frontend): preserve keycloak realm url between instance settings saves
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix(backend): preserve provider-specific oauth fields through round-trip
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add missing grant_types field to instance config OAuth structs
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add missing display_name and tenant fields to instance config OAuthClient
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
- Show recurring toast every 15s (8s duration) when loading takes long, with stop button
- When streaming by batches of 25 and a batch takes >4s, offer to stream 1 by 1
- Expose batch size in progress bar with editable input to customize on the fly
- Make stop button more prominent (destructive Button component)
- Fix list_jobs UNION: exclude queue jobs when filtering by result or is_skipped=true
- Add "Show skipped" preset to runs filter
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
When failures are sparse (<1%), filtering by failure status on the runs
page required scanning millions of success rows. Add a partial index on
v2_job_completed (workspace_id, completed_at DESC) WHERE status IN
('failure', 'canceled') and switch ORDER BY to completed_at when
filtering failures, so Postgres walks the small partial index directly.
Benchmarked at 5.2M rows / 1% failure rate:
- LIMIT 30: 800ms -> 0.4ms (2000x faster)
- LIMIT 1000: 550ms -> 21ms (26x faster)
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* fix: copy deps and remove user auto-add on workspace fork
Clone workspace_dependencies to forked workspaces and remove
automatic workspace_invite creation for parent workspace users.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* chore: update sqlx offline cache
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* Support arg type decl in postgres
* Python datatable client no longer requires explicit arg typing
* compilation fix
* Set correct type in statement exec
* reset to main
* Explicit pg arg types
* remove code duplication
* update parser js
* FLOAT8 doesn't have space
---------
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
When connecting Nextcloud via workspace integration OAuth, the resource
now includes baseUrl (from OAuth config) and userId (fetched from
Nextcloud OCS API) alongside the token, making it immediately usable
by scripts. Falls back to token + baseUrl if user info fetch fails.
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* refactor: slim down claude instructions for lean context and fast iteration
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: add private and license feature flags to enterprise validation docs
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat: add /refine skill for end-of-session doc evolution
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* refactor: remove architecture.md overview doc per research findings
General codebase overviews distract agents and trigger unnecessary
exploration. Keep only operational docs (validation, enterprise).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat: add autonomous mode doc for bypass permission workflows
Covers: plan-first requirement, tmux pane usage for checking
backend/frontend logs, manual testing via Playwright MCP,
Playwright gotchas, and end-of-task summary expectations.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* feat: add mermaid, playwright, and asciinema tools to autonomous mode doc
Claude should use mmdc for diagrams during planning, playwright CLI for
screenshots of frontend changes, and asciinema for terminal recordings
of CLI changes. All attached to the PR.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: use pastebin for screenshot/recording uploads
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: address PR review findings
- Remove stale docs/architecture.md reference from /refine skill
- Fix script name: ./update-sqlx -> ./update_sqlx.sh
- Remove .claude/settings.local.json mention from enterprise doc
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* fix: replace SELECT * with explicit columns in teams command query
- Update sqlx offline cache for the changed query
- Fix write_latest_ee_ref.sh to prefer matching EE worktree branch
- Update ee-repo-ref.txt
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* chore: update ee-repo-ref to 8ffae1f43b31dc8136714fa612d22b6301773e27
This commit updates the EE repository reference after PR #434 was merged in windmill-ee-private.
Previous ee-repo-ref: da1f8bf8676f85cac2b6fa2705246e1819d4b6f0
New ee-repo-ref: 8ffae1f43b31dc8136714fa612d22b6301773e27
Automated by sync-ee-ref workflow.
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
* feat: add force branch param to git sync settings
* update hub path
* chore: update ee-repo-ref to a797dd4d619cdab737e133ce593f2f8582ba21de
This commit updates the EE repository reference after PR #430 was merged in windmill-ee-private.
Previous ee-repo-ref: 373c5870ae5499c8c8a22cb92b2fd3a64a651183
New ee-repo-ref: a797dd4d619cdab737e133ce593f2f8582ba21de
Automated by sync-ee-ref workflow.
* UI nits
* nit
---------
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
When a flow has a flow-level `skip_expr` (or `no_flow_overlap`) and the first
module has `skip_if` defined, the flow-level condition returns `UpdateFlow`
before any identity job is created. The `UpdateFlow` path passes `Uuid::nil()`
as `job_id_for_status`, causing `fetch_one` to fail with "no rows returned".
- Change `fetch_one` to `fetch_optional` so a missing row returns false
- Short-circuit the DB query with `stop_early && skip_if_stop_early` so both
skip mechanisms (identity job check and early-stop skip flag) are considered
- Also fixes the logical gap where a module with both `skip_if` and
`stop_after_if` would only check the identity job, ignoring the early-stop
skip signal
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
The num_columns i16 was written twice, causing the parser to read
the second copy as column data and misparse the column name.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: optimize slow list_assets query with covering index and v2_job join fix
Add a covering index on asset(workspace_id, path, kind, created_at DESC, id DESC)
with INCLUDE(usage_kind, usage_path) to enable index-only scans for the CTE aggregation.
Fix v2_job join to cast asset.usage_path::uuid instead of job.id::text, allowing
PostgreSQL to use the job_pkey primary key index instead of seq scanning the entire table.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix: drop redundant asset indexes subsumed by new covering index
idx_asset_workspace_created_id and idx_asset_kind_path are fully covered
by the new idx_asset_ws_path_kind_recent + the primary key. Verified all
asset table queries still have optimal index coverage. Reduces write
amplification on inserts.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* fix: use @-prefixed LIKE pattern for email domain matching in auto-invite
The SQL queries for auto-add and auto-invite used `LIKE CONCAT('%', domain)`
which could match emails from unrelated domains (e.g., `user@barfoo.com`
matching domain `foo.com`). Changed to `LIKE CONCAT('%@', domain)` so only
exact domain suffixes match.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* chore: update ee-repo-ref.txt
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
* fix(backend): pass parent_path for trigger renames in git sync
When renaming/moving a trigger path, the old path was not included in
the deployment metadata, so git sync never deleted the old file. This
adds parent_path to all 9 trigger DeployedObject variants and computes
it in update_trigger when the path changes.
Fixes#8014
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* fix path change with common prefix issue
* update ref
* chore: update ee-repo-ref to cb25312072c15c0e9cc375ebc824d41995a52898
This commit updates the EE repository reference after PR #428 was merged in windmill-ee-private.
Previous ee-repo-ref: 7225f7423311f58015a2fab61248c9d89888aef6
New ee-repo-ref: cb25312072c15c0e9cc375ebc824d41995a52898
Automated by sync-ee-ref workflow.
---------
Co-authored-by: Claude Opus 4.6 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
- Remove bidirectional $effect sync in RawAppSidebar; bind FileExplorer
directly to files prop with {} default
- Avoid creating new files object on every keystroke in FilesetEditor;
merge editContent → args in a single effect without intermediate spread
- Simplify no-op `?? undefined` in addResourceType
- Add backend validation: reject create_resource_type when both
is_fileset and format_extension are set
- Fix fileset alert title showing undefined format extension
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Use sqlb.set() instead of set_str() for boolean is_fileset field
to avoid quoting (SET is_fileset = TRUE not 'TRUE')
- Add JSDoc comment to isFilesetResource explaining it matches
children inside .fileset/ directories, not the directory itself
- Update OpenAPI spec for file_resource_type_to_file_ext_map endpoint
to document the new response schema with format_extension and
is_fileset fields
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a new "fileset" resource type that represents a collection of files
stored as a relpath→content map. This enables resource types to manage
multiple files (e.g., config directories, template sets) instead of just
a single file.
Backend:
- Add is_fileset column to resource_type table
- Update CRUD operations and workspace duplication to handle is_fileset
- Add integration tests for fileset resource types
Frontend:
- Add FilesetEditor component with file explorer + Monaco editor
- Extract shared FileExplorer component from RawAppSidebar (dedup)
- Add fileset toggle to EditableSchemaWrapper
- Show fileset editor in ResourceEditor and ApiConnectForm
- Show folder icon for fileset resource types in IconedResourceType
CLI:
- Support fileset resources in sync pull (expand to .fileset/ directory)
- Support fileset resources in sync push (reconstruct from directory)
- Handle !inline_fileset YAML tag in resource resolution
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
