* fix: scope cloned app policy and custom path to the fork's creator
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor: share the app custom-path scoping rule across its call sites
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: tighten the cloned-app-policy comments
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: correct the execution_mode and custom-path scoping rationale
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
The consumer name field in the NATS trigger config read "Required is using
JetStream" instead of "Required if using JetStream", matching the wording
already used by the sibling stream name field.
Fixes WIN-2335
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(cli): sync package-lock.json with package.json
`npm ci` fails in cli/ because the lockfile predates two manifest changes:
windmill-parser-wasm-yaml was bumped to 1.770.0 and windmill-yaml-validator
1.1.1 was added, neither of which reached the lockfile.
Regenerated with `npm install --package-lock-only`; the only entries touched
are those two packages and windmill-yaml-validator's transitive deps.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(cli): drop the unread npm lockfile, bun.lock is the CLI's
Every install path in cli/ runs `bun install`: cli-tests.yml, git-sync-test.yml,
backend-test.yml, build.sh, and install_dev.sh (its --node branch installs the
generated npm/ bundle, which carries its own manifest). build-npm.ts synthesises
the published package.json from scratch, and change-versions.sh regenerates the
frontend and yaml-validator lockfiles but not this one.
So package-lock.json was read by nothing and verified by nothing, and drifted out
of sync with package.json unnoticed until `npm ci` refused to install. Deleting it
removes the second source of truth rather than hand-repairing it again on the next
bun-driven dependency change.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: allow custom dev workspace environment labels
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: reject dev labels that shadow a tracked branch's namespace
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: guard dev labels against a repo's assumed default branch
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* chore: state the badge-cap rationale once and drop unenforceable openapi constraints
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor: offer a fixed list of environment labels instead of free text
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: match the accepted label set to the openapi enum exactly
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: stop describing the label set as dev/staging only
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(frontend): hide the fork workspace banner from operators
* fix(frontend): scope the operator gate to the workspace its role was fetched for
* fix(frontend): drop superseded whoami responses instead of writing a stale role
* fix(frontend): guard the remaining workspace-switch userStore writers
* feat: open a session edit in the preview panel from the edits list
* refactor: move the deploy-kind preview mapping next to its siblings
* feat: make preview the primary action in the session edits list
Clicking a row in the Edits popover now opens the item in the session
preview panel; kinds the panel cannot host fall back to their diff. Each
row gains an explicit Diff button for that item, and a pinned footer row
opens Review & deploy for the whole set.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor: fold the open-and-flash rule into SessionPreviewTabs
The rule for when a preview open should flash the tab was written out at
three call sites, one of them with a looser condition. Move it onto the
tab owner as openAndPulse so the three agree.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: preview data-pipeline edits from the edits list
A pipeline bundle is stored at `f/<folder>/data_pipeline` while its editor
is the folder's pipeline view, so the deploy-kind mapping has to route on
the folder. Without it the one remaining previewable kind fell through to
the drawer. Also pin the open-and-flash rule with tests.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor: state the pipeline bundle path layout once
`f/<folder>/data_pipeline` was parsed independently in the compare page,
the home list and the session preview mapping — two of them disagreeing on
whether the trailing segment is checked — and built by hand in the editor.
Route all four through $lib/pipelinePaths.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor: route the pipeline page through the shared bundle path
The route built the bundle path by hand and passed the draft kind as a
literal, the two halves of the editor disagreeing on where the layout is
stated.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: name a pipeline edit by its folder in the edits list
The bundle path is an implementation detail and the row's click lands on
the folder's editor, so showing `f/<folder>/data_pipeline` named something
the click doesn't open. Also derive the bundle regex from the draft-kind
const it has to agree with.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* chore(cli): drop unrelated package-lock.json change
The lockfile diff was an incidental regeneration from an older manifest
(it downgraded the locked svelte range below what cli/package.json
requires) and had nothing to do with this PR. Reset to main's version.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Co-authored-by: Ruben Fiszel <ruben@windmill.dev>
* refactor(recordings): build recordings from the completed run instead of live event capture
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(recordings): budget and isolate replay synthesis against hostile recordings
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(recordings): capture full logs and the run-time flow, upgrade v1 files
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* feat(recordings): pick an existing run for the hub recording
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(recordings): cap recorded logs under the replay loader's text budget
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(recordings): pin picked runs to their executed version, keep v1 streamed logs
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(recordings): bound pipeline finalize fan-out, pin flow schema to run version
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix(recordings): warn on mixed-version fallback, bound code fetches
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* feat: add public sharing option for job pages
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: gate public run sharing and address review findings
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: address review nits on public run sharing
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: key public run view on workspace, job and token
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: keep the mermaid fullscreen dialog inside its pane and its emoji vector
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: load only the fonts a diagram needs and size chrome per breakpoint
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: cover every emoji class in the font preload without restyling diagrams
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: keep the dialog chrome allowance in rem so it scales with the root font
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: decode mermaid entity codes when sampling text for the font preload
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: match mermaid's decimal-only entity codes and decode the Inter sample too
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: drop emoji format characters from the font sample
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: state the emoji subset spread and modifier exclusions precisely
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor: re-render once fonts settle instead of hand-picking emoji subsets
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor: give Modal a fill-height mode instead of measuring its chrome
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: guard the post-fonts re-render against a newer render
The re-render after document.fonts.ready assigned svg without re-checking
renderSeq after its own await. renderedCode is set before that await, so a
stale re-render landing last leaves svg holding the previous diagram while
renderedCode names the current source — showSvg stays true and paints the
wrong diagram.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: compact ai chat context for models with unknown context windows
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* docs: correct stale comment on unknown-model context window handling
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* fix: surface assumed context window in usage indicator for unlisted models
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
* [ee] feat: improve duckdb isolation
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
* chore: update ee-repo-ref to c4e6cbc1a1efeca5b71920c7903db2347d0eeda0
This commit updates the EE repository reference after PR #713 was merged in windmill-ee-private.
Previous ee-repo-ref: f630f7e73cb863e312430738d81d802a3971f7cd
New ee-repo-ref: c4e6cbc1a1efeca5b71920c7903db2347d0eeda0
Automated by sync-ee-ref workflow.
---------
Co-authored-by: Claude Opus 4.8 <noreply@anthropic.com>
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
* chore: track origin EE branch in worktree setup when absent locally
* chore: pass --track to worktree add so upstream is set regardless of git config
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
* chore: guard first worktree arm on local branch to avoid remote-only DWIM
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---------
Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
#10560 changed the script editor's change handler to read `e.detail` instead of
`editorCode`, on the reasoning that the two only diverge while the editor is
being torn down. They also diverge in a live editor: `setCode` dispatches the
string it was handed, but `alignCodeWithEditor` applies that string to Monaco
first, and the resulting `onDidChangeModelContent` runs `updateCode`
re-entrantly — so if the model normalized the text (EOL is the reachable case;
`ScriptBuilder` builds template content with a `\r\n` join), `editorCode`
already holds the buffer's version and the payload is the pre-normalization
one. Taking the payload leaves `code` disagreeing with what the editor shows,
which the external-write effect then tries to reconcile on every change.
The language-switch fix in that PR is `alignCodeWithEditor` clearing the timer
its own write armed; that part stands and is unaffected. This restores the
handler to the buffer-true read.
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
Picking a new language seeds the new template into `script.content`, which
`ScriptEditor` writes straight into Monaco. That write goes through
`onDidChangeModelContent`, which arms the keystroke debounce as if the user had
typed — so when the `{#key effectiveLang}` block then tears the editor down, the
unmount flush sees a pending timer, reads a stale `code`, and dispatches a change
that puts the previous language's template back. The editor kept showing the old
content under the new language.
`alignCodeWithEditor` now clears the timer its own write armed, restoring the
premise the unmount flush is guarded on: a pending timer means Monaco holds
something newer than `code`.
`ScriptEditor` also takes the change payload instead of re-reading `editorCode`.
A destroyed component's `bind:` writes no longer reach the parent, so the
re-read returned the value from before the change — which is what actually
wrote the old template back, and would equally have made the unmount flush
save stale text after real typing.
Fixes WIN-2330
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat(flow-editor): surface an agent's tool-call status without moving the graph
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: count only an agent's tool calls and key them in one place
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: report an agent's replies alongside its tool calls
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: break the agent summary down by action kind
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: key agent tool nodes by kind and keep the summary clear of the tool row
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor: read agent action status from the run's success array
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* test: pin the tool joins a local run cannot reach
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: place the agent summary beside the step and match MCP paths bare
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: feed a single-step agent test's calls into the graph status
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* perf: keep run status out of flow graph node and edge data
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: key the ai tool node memo on the agent's actions
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: state the edge-data and memo-key constraints as invariants
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: restore selection clearing and pin the zoom bar's border colour
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: show an agent's tool calls as they arrive instead of at step end
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* perf: stop editor runs from rebuilding on agent tool calls
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: keep the same_worker pin when a suspend ends without approval
A disapproved or timed-out approval gate hands the flow back through the
UpdateFlow channel with unrecoverable = true. That flag means "the previous
step's worker died", and it is read by six sites. Five of them happen to want
what it does here, but continue_on_same_worker and continue_with_runners do
not: the worker that ran the approval step is alive, so unpinning the error
handler and routing it by tag breaks the ./shared contract of a same_worker
flow and can land it on a worker group that cannot run it — the same defect
#10551 fixed for the three producers that hand back a live flow.
Replace the boolean with StepFailureKind so the suspend producer can say
"worker alive, but this failure is not the module's to handle" instead of
overstating a worker death. The failed module's error policy is deliberately
still bypassed: the failure is recorded against the step the gate was holding
back, which never ran, so its retry would re-open the gate and its
continue_on_error would skip it outright (verified: the gated step is marked
Failure with a nil job id and the flow jumps past it). suspend.
continue_on_disapprove_timeout remains the way to continue past a gate.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat(flow-editor): flag that continue on error does not cover the approval gate
A resolved approval is recorded against the step the gate holds back, not
the step carrying the suspend, so continue_on_error never sees it: the flow
still stops on a disapproval or timeout. Point users at
suspend.continue_on_disapprove_timeout, which is what actually continues past
a gate, whenever both settings are on and that one is not.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: point re-opened previews at the tab already showing them
* fix: judge composed preview mutations as one change
* fix: treat a fullscreen preview as displayed when deciding to flash
* fix(frontend): draw the tab strip's scroll bar instead of the native one
The strip sizes its scroll row to the tabs, but a native horizontal
scrollbar claims layout height on top of that: Firefox spends 11px on
`scrollbar-width: thin` — `--wm-scrollbar-size` is WebKit-only, so the
4px it asks for is ignored there — which clipped the tabs at the top of
the 32px sessions strip and left a wide gutter under them.
Hide the native bar and draw a 4px thumb from `scrollLeft`/`scrollWidth`
instead: it costs no layout height, is the same size in every engine, and
sits on the strip's bottom edge, flush under the tabs. Tabs drop to `h-6`
so they clear it, and the strip's default height matches the sessions
caller's `h-8` so every strip has the same geometry.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(frontend): clamp the tab strip thumb at both ends of its track
WebKit's elastic overscroll drives `scrollLeft` negative, which slid the
thumb out of the track's left edge and into the strip's padding.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat(debugger): install debug session deps from the instance registry settings
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(debugger): keep install-time registry credentials out of the session-visible tree
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: drop em dashes from the debugger registry docs and comments
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(debugger): stop installing for a session that went away during the settings fetch
Also serves nativets sessions the npm settings their installer reads.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
When a module completes without spawning a job — an empty branch, an
empty for-loop, or a module already marked Success — the flow hands
itself back through the UpdateFlow channel, and the result processor
resumed it with unrecoverable = true regardless of what sent it. That
flag means "the previous step's worker died", which holds for none of
the three producers except a suspend that ended without approval.
The stale argument was inert until continue_on_same_worker and
continue_with_runners started reading it, since when the step after such
a module is pushed as an ordinary queued job. It is then routed by tag
and can land on any worker in the pool, breaking both the ./shared
directory contract and the guarantee that a same_worker flow stays on a
worker able to run it — a step whose tag resolves to a worker group that
cannot execute its language fails instantly, taking the flow with it.
Carry the flag on the UpdateFlow message so each producer states its own
case, rather than having the shared receiver assume the worst. The three
that hand back a live flow forward whatever their caller reported, so a
genuinely unrecoverable failure still crosses the hop unchanged.
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: register mounted CA certificates in windmill_extra at startup
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: only claim a CA update when update-ca-certificates can read the mount
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: detect mounted CA certificates the way update-ca-certificates finds them
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
`windmill prepare-deps` was spawned with Bun's raw `spawn` in both the Python and
the TypeScript session, bypassing the nsjail wrapping the debugged script itself
gets. With ENABLE_NSJAIL=true, `uv pip install` (source distributions run their
build backend) and `bun install` (postinstall scripts) therefore executed
package-supplied code unconfined, next to the LSP, multiplayer and gateway
services in the windmill-extra container.
Both installers now go through the same nsjail wrapper as the debuggee, which the
two files no longer build separately. The jail keeps the environment
(`keep_env`), which is what carries the registry credentials and CA settings into
the installer; the debugged script's environment is unchanged and still holds
neither.
Killing the installer also did not reap the `uv` or `bun` it had spawned: those
were reparented to init and kept downloading, so both the timeout and the
cancel-on-disconnect only half-worked. The installer now runs in its own process
group and is signalled as a group, reading the group id back from /proc rather
than assuming it, since a group kill aimed at the service's own group would take
down every service in the container.
Two things that cancellation exposed: a kill was reported to the client as an
install failure, since it ends the read with nothing to parse - blaming the user
for their own Stop; and the standalone Bun server's close handler only dropped
the session from its map, so nothing there was ever cleaned up. The teardown flag
is also scoped to a launch rather than the session, because cleanup() runs when a
program finishes normally too.
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: allow a dev workspace to have its own dev workspace
Fixes WIN-2324
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: keep every dev workspace in a chain on a distinct deploy branch
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: count a dev workspace the caller has no seat in as holding its label
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor: keep the attach form standing when a candidate takes the last label
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: keep the label toggle visible when a candidate's dev workspace clashes
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: describe the cycle guard by what holds, not by what changed
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: refuse to archive a fork-backed dev workspace that owns a nested dev
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: put the deploy target and item filters under the pairing they configure
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: refuse to archive any dev workspace that owns a nested dev
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: fix the fixture family count
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: put the deploy target with the pairing line it restates, above protections
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: name the same family head in the workspace menu and the scope picker
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: stop offering to delete a dev workspace from the sidebar settings menu
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: state the visibility boundary the lineage root actually resolves to
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: serialize dev-pairing creation against teardown of the same workspace
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: lock both sides of an attach so adjacent pairings cannot share a label
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: serialize dev pairings on one key, the invariant being chain-wide
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: scope the pairing lock to the chains an operation reads
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: hold the pairing lock across renames and re-check the cycle under it
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: hide the fork-delete action until the workspace entry has loaded
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: lock archive before it reads the pairing state it acts on
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: describe the archive lock test by what it pins, and drop an unused fixture row
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: add a load all next to load more in the tree view folder pager
* fix: bound tree node rendering and make a long load resumable
* fix: resume a failed first load from its saved cursor
* fix: size the show-more step by what a node holds, not what it renders
* fix: keep the pager visible mid-run and spin only the clicked button
* fix: forward proxy and TLS settings to debugger subprocesses
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: reach uv and the bun debugger with the forwarded network settings
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: map every CA variable spelling onto the one uv reads
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: keep package-index credentials out of debugged user code
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: install debugger dependencies outside the interpreter running user code
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: sandbox and bound the debugger dependency installer
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: correct the installer timeout rationale
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: scope the uv --cert note to the commands prepare-deps runs
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: build the debug venv against the interpreter that runs the script
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: do not start the debuggee for a session that already went away
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: remove the debug script when the session is gone before it starts
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor(ai): render thinking blocks with the shared tool-call card
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat(ai): collapse thinking to a status row with a thought-for duration
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* refactor(ai): separate reasoning-timing reset from duration read
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(ai): render expanded thinking in the body font, not mono
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat(ai): mark in-progress chat rows with a shimmer sweep
Thinking and tool calls both announced themselves with a spinner, which
carried no more information than the row already did and read as visual
noise once several tools ran in sequence.
A white copy of the label now sits over the coloured one and is revealed
through a travelling band, so a running row is marked by motion across
its own text rather than by a separate glyph. Both spinners and the brain
icon are gone, leaving the card with no icon slot at all, and every header
label settles on text-secondary.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(ai): keep a running row marked under reduced motion
The shimmer is the only thing distinguishing a running tool row from a
settled one, and the reduced-motion rule removed it outright, so the two
became identical for those users. The band now degrades to a flat wash
instead of disappearing.
Also covers the reasoning-duration state machine: that thinking stops at
the first answer token rather than at the end of the turn, and that each
reasoning pass of a tool-using turn is timed from scratch.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* test(ai): restore the clock spy after the reasoning-duration tests
The file-level hook only clears call records, so the Date.now spy stayed
installed and would freeze time for anything appended after this block.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(debugger): return PrepareResult when the service prepared the venv
`prepare_dependencies()` returns `PrepareResult` on every path except the
service-prepared short-circuit, which returned the venv path as a bare `str`.
`handle_launch` reads `prepared.error` on it, so every Python session whose
script has a third-party import raised `AttributeError`, hung, and failed at
180s with `Debugpy command timeout: launch`.
The two consumers of a prepare-deps response also read a `stderr` key the CLI
does not emit; the field is `install_stderr`, and it carries the same text
`error` already wraps in a sentence, so take one rather than joining both.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(debugger): keep the failing step in the launch message, drop the dead branch
Preferring the raw `install_stderr` made `_first_line` pick uv's opening
progress line, so a refused launch reported "Using Python 3.12.13 environment
at: venv" — which reads like success. `error` is the same text prefixed with the
step that failed, so it is the better of the two to condense.
The installer-diagnostics pass over a `success: true` response is unreachable:
every `success: true` site in prepare_deps.rs sets `install_stderr: None`, and
its comment claimed the opposite of what that file documents. It existed to work
around a producer that warned and returned success on a failed `uv pip install`;
that producer now returns `success: false`.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: stop the homepage edit-in-fork button from wrapping
* fix: show the full edit-in-fork label anywhere on the button
* fix: thread showEditButton through the homepage tree view
* fix: match the edit-in-fork button styling to the normal edit button
* fix: edit in dev workspace dead-ends on items the dev workspace lacks
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: pull the item's folder before copying it into the dev workspace
* fix: speak the compare page's update vocabulary in the dev-workspace prompt
* fix: raw app with no stylesheet was undeployable across workspaces
* fix: drop the raw-app stylesheet workaround now that the backend serves one
The frontend wrapped `getRawAppData` to report a missing `.css` as empty,
because a raw app with no stylesheet stores no css blob and the shared deploy
treats the resulting 404 as fatal. #10364 fixed that at the source: the backend
now serves an empty body for a missing stylesheet, so the wrapper guards a 404
that no longer happens.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: drop the fork icon from the edit-in-dev-workspace affordances
The row button carried both a pen and a fork, and the menu entries and detail
page buttons carried a fork alone — where the menus already used that same icon
for Duplicate/Fork, so the two entries were indistinguishable. The action is an
edit, so it takes the pen everywhere, matching the ordinary Edit button.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: send edit in dev workspace to the item's editor
The affordance landed on the item's page in the dev workspace and left the user
to open the editor from there. It says "Edit", so it goes to the editor:
`/scripts/edit/...?workspace=<dev>` and the equivalent for flows and both app
kinds. `?workspace=` still does the workspace switch, which the logged layout
applies on any route.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat: choose the on-behalf-of user when updating the dev workspace
The prompt deployed the item with no say over the identity it would run under,
so an item that ran on behalf of someone in prod silently became the deploying
user's in the dev workspace. It now offers the same choice the compare page
does, under the same rules: shown only when the source item carries an
on_behalf_of, picking anyone but yourself gated on admin/wm_deployers in the
target, and confirming blocked until a choice is made — including while the
lookup that decides whether one is needed is still in flight.
The prompt also stops offering the compare page inline; the confirm button
still leads there when the user can't deploy into the dev workspace.
Two fixes the reused selector needed to work inside a dialog:
- ConfirmationModal takes `confirmDisabled`, which also blocks the Enter binding.
- The popover's z-index is now overridable, and the user picker is portalled.
A ConfirmationModal renders above the popover layer, and its card is
transformed for the open transition, which makes it the containing block for
the picker's `fixed` positioning — so both opened behind, and the picker was
laid out inside the card instead of the viewport.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: check deploy rights per item before prompting to update the dev workspace
* fix: read the compare page link before closing the dev-workspace prompt
The link is derived from the request the prompt is answering, so closing first
left an empty string to navigate to: refusing users saw the dialog dismiss and
stay put, with no way through to the compare page.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: keep the new-tab promise and speak up when a popup is blocked
Three defects found by successive cold reviews of the click-time resolution
added earlier in this branch, each one only reachable once the previous fix
existed:
- Safari refuses `window.open` from any promise continuation however fast it
resolves, so the tab the editor dropdown opens after its existence probe
never appeared there. `claimTab` takes the tab inside the click's own
transient activation and points it at the answer once it lands, releasing it
when there is nothing to show.
- That left the two halves of the same action disagreeing: the entry promises
never to navigate the editor away, but when the item turned out to be missing
the prompt took over and navigated in place. The request now carries
`openInNewTab`, and every destination the prompt can reach honours it.
- With popups blocked the fallback called `window.open` without checking, so a
successful deploy closed the prompt and did nothing, silently. It now names
what it could not open.
`openEditInFork` also takes the workspace explicitly. The four editor dropdowns
computed their label from `opWorkspace` but resolved the action from the
navigation store, and `prodWorkspaceId` feeds `deployItem({ workspaceFrom })` —
so a session pane would have deployed from the wrong workspace.
`checkPathWritePermission` is exported with an injectable folder probe and
covered by table-driven cases, chiefly to pin its two fail-open branches, which
otherwise read as dead code inviting deletion.
The two unrelated whitespace hunks in ScriptBuilder.svelte are the repo's
format-on-save hook fixing pre-existing violations in a file this touched.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: create the dev workspace's missing folder without overwriting it
`ensureFolder` delegated to the shared `deployItem`, which re-probes and
switches to `updateFolder` when the folder turns out to exist. Nobody asked for
that folder to be deployed — it is created only so the item has somewhere to
land — so a folder created between the two probes had its owners, ACL, summary
and labels silently replaced with the source workspace's. Creating is now
create-only, and losing that race counts as success: the folder exists, which is
all the caller needed.
The same delegation dropped `default_permissioned_as` and `labels`, which the
shared folder deploy does not send. A folder copied without its create-time
identity rules applies none, so an item landing inside it with no on_behalf_of
of its own resolves to whoever deployed it rather than to the principal the
source folder would have chosen — the exact substitution the rest of this branch
exists to prevent. Both are now carried across.
Also check `window.open` in the no-dev-workspace branch of `openEditInFork`. The
branch beneath it already reported a blocked popup; this one returned as if it
had opened something.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: translate copied folder identity rules into the target workspace
A `u/<username>` names a workspace-local account, so copying a folder's
`default_permissioned_as` verbatim was wrong in two directions: the same
username in the dev workspace can be a different person, who would then be
granted the item; and a username with no account there at all passes the
folder-create check, which is structural, only to fail every subsequent item
deploy on the existence check, including the retry — the folder now exists, so
`ensureFolder` short-circuits and the deploy fails identically, with no way out
of the prompt.
Rules are now resolved source username -> email -> target username, since email
is the only identifier stable across workspaces, and a rule whose principal has
no account in the target is dropped rather than carried. Dropping one makes the
copied folder less restrictive than its source, which is not something to
discover later from an item running as the wrong user, so it is reported.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: refuse to overwrite a concurrent item, and translate every folder principal
Four findings from CI review, all on the implicit half of this flow — the writes
the user did not explicitly ask for.
The item write is now create-only. The shared `deployItem` re-probes and silently
switches to an update, so the caller that acts on an item being *absent* could
still overwrite whoever landed it between the two probes. Rather than
reimplementing the per-kind deploys, the frontend's own provider refuses exactly
the three writes that branch reaches for — `updateFlow`, `updateApp`/
`updateAppRaw`, and a `createScript` carrying a `parent_hash`, which is what
makes an otherwise identical create an update. A refusal reports `conflict`, and
the prompt opens their version instead of replacing it.
Folder principals are translated rather than copied. `u/<username>` is
workspace-local, so a verbatim copy either names nobody or names a different
account that happens to share the username. Users now resolve source username ->
email -> target username, and the two kinds of unresolvable principal are
separated because they fail differently: an owner or ACL entry is dropped, which
can only narrow the folder and leaves the creator owning it; an identity rule
refuses the copy outright, because dropping it runs the item as the deployer and
keeping it creates a folder the server then rejects every deploy into.
Groups resolve against `listGroups` rather than `listGroupNames`, which unions in
instance groups that folder rules do not resolve against — a same-named instance
group would otherwise let an unusable rule through.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: read every page of workspace groups before judging a folder principal
`listGroups` paginates, and the `perPage: 100` it was called with is narrower
than the server's own default of 1000 — so a group past the first page read as
having no account in the target. Since an unresolvable identity rule now refuses
the whole folder copy, that turned into a refusal naming a group that does
exist, and an owner or ACL entry on a later page was dropped silently. Read
until a page comes back short, with a size check as the backstop for a server
that ignores `page`.
`list_users` is unpaginated, so the user half of the same lookup was never
affected.
Also move `makeProvider`'s doc block back onto `makeProvider`; adding
`DeployConflict` had left it documenting the type instead.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* docs: reattach principalTranslator's doc block to principalTranslator
Adding `workspaceGroupNames` above it left the block documenting the helper,
the same way adding `DeployConflict` had displaced `makeProvider`'s.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* feat(mcp): serve the 2026-07-28 spec alongside the legacy protocol
* fix(mcp): keep oauth discovery strict and preserve request limits
* fix(mcp): allow the protocol's own headers through CORS
* fix(mcp): expose the auth challenge to browser clients
* chore: update ee-repo-ref to c1665a881b61616f96ffe7702b44840905304660
This commit updates the EE repository reference after PR #711 was merged in windmill-ee-private.
Previous ee-repo-ref: bc1c001e3e386342415dfb8ac31c6b97f6629320
New ee-repo-ref: c1665a881b61616f96ffe7702b44840905304660
Automated by sync-ee-ref workflow.
---------
Co-authored-by: windmill-internal-app[bot] <windmill-internal-app[bot]@users.noreply.github.com>
* fix(frontend): stop swallowing Ctrl/Cmd+Shift+S in the editors
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(frontend): make Ctrl/Cmd+S from a focused Monaco flush the draft
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix(frontend): broadcast the Monaco save shortcut after the effect flush
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: report python debugger dependency install failures instead of timing out
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: surface swallowed installer errors and stream debugger prepare progress
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: reap the python debugger on a failed launch and bound prepare-deps
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: match uv failure output by stripping progress instead of matching errors
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: treat uv build, download and warning lines as install progress
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
---------
Co-authored-by: Claude Opus 5 (1M context) <noreply@anthropic.com>
* fix: honor python index settings in prepare-deps and report install failures
* fix: forward python registry env to the debugger's prepare-deps
* fix: scope registry credentials to the prepare-deps subprocess
* fix: install python debug dependencies from the service, not the session
* fix: bound the debugger dependency install and keep the proxy bypass default
* docs: name the nsjail config that isolates debug sessions
Ruben Fiszel
hugocasa
Guilhem
AlexRV12