Repository navigation
Conversation
This was referenced Oct 28, 2026
yohamta0
force-pushed
the
stack/24-ui20
branch
from
October 29, 2026 13:21
f7cfa1a to
b73208c
Compare
yohamta0
force-pushed
the
stack/23-ui19
branch
2 times, most recently
from
October 30, 2026 06:44
7f1089c to
2e9eeaa
Compare
yohamta0
force-pushed
the
stack/24-ui20
branch
from
October 30, 2026 06:44
b73208c to
b30096f
Compare
yohamta0
force-pushed
the
stack/23-ui19
branch
from
October 30, 2026 08:07
2e9eeaa to
6391cc8
Compare
yohamta0
force-pushed
the
stack/24-ui20
branch
2 times, most recently
from
October 30, 2026 08:43
388aceb to
f6538d0
Compare
yohamta0
force-pushed
the
stack/23-ui19
branch
from
October 30, 2026 11:03
c62562a to
c69c6d8
Compare
yohamta0
force-pushed
the
stack/24-ui20
branch
3 times, most recently
from
October 31, 2026 04:50
72ff7b4 to
f2b5b0d
Compare
yohamta0
force-pushed
the
stack/23-ui19
branch
from
October 31, 2026 05:37
17f31b9 to
7b0a8a1
Compare
yohamta0
force-pushed
the
stack/24-ui20
branch
2 times, most recently
from
October 31, 2026 06:28
765d09b to
f142e71
Compare
yohamta0
force-pushed
the
stack/23-ui19
branch
from
October 31, 2026 06:28
7b0a8a1 to
c4ff36b
Compare
yohamta0
force-pushed
the
stack/24-ui20
branch
from
October 31, 2026 09:21
f142e71 to
3631c39
Compare
yohamta0
force-pushed
the
stack/23-ui19
branch
2 times, most recently
from
October 31, 2026 10:44
0d357a0 to
f52574e
Compare
…gucloud#3403) * UI: dags/components, workflow-workspace/dock, workflow-workspace/step * Expect runs to open on the workflow page in E2E * Open run links of DAGs named apart from files * Expect a sub-run link to drop its step * Give sub-runs an address and show file-less runs * Test the summary link of a waiting agent * Read step reuse from the run record in step facts * Fix draft step runs, foreach items and run links
The workflow page now shows runs: an old ?dagRunId link and a progress card open the run there, and its run bar stops the run. The run page stays until a later change retires it.
A run is selected in the URL after its start resolves, so checking the URL right after the start call or the followed run raced the navigation and failed now and then in CI.
Two fixes of an earlier PR change behavior of the run bar parts that this PR adds tests for: - Under ?def=run the canvas shows the run definition, so the stale dot of the run chip names no count of steps missing from the canvas. - When the history does not hold the source run of reused inputs, the details take the case of its ID from the record of the reused inputs instead of the lower-case label, so "Show run" selects it. useElementWidth now reads the border box of the element on each resize, not the content rect of the observer entry. The run timeline resize test changed only the entry, so it now changes the measured box too.
Several tests passed whatever the code did: - The test-run notice was checked against a mock that kept calls from earlier tests. The mock is now cleared before each test. - The failed-start test swallowed the error in the canvas stub. The stub now records it, and the test checks the message. - The test of a failed editor run checked the followed runs before the start resolved. It now waits for the start. - keepTab was only tested where it falls back to the log tab. A test now keeps the agent tab of a waiting step. - A run seen while ?step is set was not checked after the link is cleared. The test now clears it and checks that the panel stays shut. - The click on the disabled test-run toggle checked nothing. The test now checks that the callback is not called. The render cost of the timeline was measured by counting barTone calls, so memoizing that helper failed the test with no change in behavior. The test now counts the renders of each bar and extent. The step navigation tests now find rows by role and read the selection from aria-current. Tests that render twice call cleanup() instead of clearing the body, which left the first React root mounted.
A run URL now opens a sub-run with ?subRun, so the new-tab and run page tests of sub-runs expect it in the URL. New tests cover: - A ?subRun link opens the sub-run on the stack with its ?step and leaves the URL; a link to a sub-run the run lacks is dropped. - The step sub-run list keeps a child that the sub-run list does not have yet, after the others. - The run outputs stop polling 15 s after the run ended.
The refresh button resets its spinner 500 ms after a refresh with a timer that nothing cleared. When the button went away first, as at the end of a workspace test, the timer updated an unmounted component, and after the test environment closed it threw "window is not defined", failing the frontend run although every test passed.
The progress bar of the run details names the succeeded steps of the total. The test file for the popover starts here.
Step facts label a step reused only from the run record, and read no outputs while a step runs or waits. The test file starts here.
A step run of an unsaved draft takes its previous and upstream steps from the draft, and a foreach step opens on its log with the item list. The mocks of the dirty editor gain getDag. The test files start here.
yohamta0
force-pushed
the
stack/24-ui20
branch
from
October 31, 2026 10:45
3631c39 to
e6ff48c
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
All PRs of the stack
Summary
This PR adds 29 test files for the workspace code of PR 23. It changes no production code. The tests cover the shell, the bar, the dock, the step run view, the commands and the keys.
Why
PR 23 lands a large amount of new UI at once. The tests are in a separate PR so that review of PR 23 stays on the code. They pass on top of PR 23. PR 25 reworks the surrounding pages and adds more tests, and PR 26 tests the editor side.
Changes
29 files, all new, all tests. No existing file is modified. No production file is touched.
Groups of tests:
WorkflowWorkspace.test.tsx,WorkflowWorkspace.keys.test.tsx,selection.test.tsx(which step or run is selected),layout/WorkspaceLayout.test.tsx,hosts/WorkspaceRunHosts.test.tsx.singleTab.parity.test.tsxchecks that the run history loads on demand, that the timeline draws one bar per run, and the trigger of a run.WorkflowBar,RunChip,RunDetailsPopover,RunSplitButton.DockContext,LogsDock,RunStepsNav,RunTimeline,useDockAutoOpen,dock/tabs/dockTabBodies, andrenderCost.test.tsx, which counts renders to keep the dock cheap.StepRunView(and an accessibility test),StepFacts,StepManualActions,StepRunTabs,StepSubRuns,useStepRunActions.WorkflowCommands,useWorkspaceKeys,useWorkspaceNotices,useStepCommandHost,useOpenStepDetails.How to review
WorkflowWorkspace.test.tsxfor the test setup that the others reuse.dock/__tests__/DockContext.test.tsxanduseDockAutoOpen.test.tsx.step/__tests__/StepRunView.test.tsxanduseStepRunActions.test.tsx.renderCost.test.tsx.Most attention: tests that assert render counts or exact text may break on harmless changes. Check that they test behaviour.
Tests
All files listed above run with
pnpm testinui/. Each file sits next to the module it checks, in a__tests__folder.Screenshots
No visible UI change in this PR. The behaviour it tests is visible in the workspace page that PR 23 adds, for example the Activity panel and the step view.