mengw15 opened a new pull request, #7916:
URL: https://github.com/apache/texera/pull/7916
### What changes were proposed in this PR?
Extends `operator-property-edit-frame.component.spec.ts` with 23 tests
covering
the Python-UDF environment-loading branch and the early-return guards.
Measured
locally with `--coverage --coverage-reporters=lcovonly`:
| `operator-property-edit-frame.component.ts` | Before | After |
| --- | --- | --- |
| lines | 350/422 (82.94 %) | **373/422 (88.39 %)** |
| branches | 230/292 | **252/292** |
| functions | 70/92 | **76/92** |
Every line the issue lists is now covered. The remainder of the file is the
Quill
title-editing block, which the issue scopes out.
**Python UDF environment loading.** The branch only runs for a Python UDF
operator and the mock metadata declares none, so the test stubs
`DynamicSchemaService.getDynamicSchema` rather than adding a fixture operator
type; `ComputingUnitStatusService` and `WorkflowPveService` are stubbed with
`of(...)` / `throwError(...)`, so every emission is synchronous.
- `defaultEnv` is seeded when the operator's properties omit it, and an
explicit
`defaultEnv: false` is left alone.
- The unit's `cuid` reaches `fetchPVEs`, and the fetched names arrive as the
`envName` field's `props.options` — that is the observable end of
`patchPythonUdfEnvironmentSchema` → `setFormlyFormBinding`. The two other
arms
(a unit with no `cuid`, and no unit at all) assert the fetch is skipped
and the
field binds with an empty option list.
- Both sides of the error handler's `err instanceof Error` ternary, each
asserting
the exact `notificationService.error` message *and* that the fallback
binding
still ran (`props.options` empty, `expressions.hide` set).
- `patchPythonUdfEnvironmentSchema`'s guard on a schema with no `properties`
and
one where `properties` is `true`; `hideEnvNameWhenDefaultEnvChecked` with
no
`envName` field present.
**Guards.** `ngOnChanges` with no id; the status-update subscription with and
without a selected operator; the ui-parameter subscription for a different
operator; `isHuggingFaceOperator`; `checkOperatorProperty` on an operator
deleted
mid-debounce; both `typeInferenceOnLambdaFunction` early returns; and the
attribute-type validator's `isDefined` chains — no operator selected, a rule
naming a property the schema does not declare, a property with no
`autofillAttributeOnPort`, a `const` rule with no `$data`, and a `$data`
comparison whose two attributes resolve to the same type.
One item from the issue is **not** covered, deliberately: the `if
(this.currentOperatorId)`
guard inside `registerOnFormChangeHandler` (line 738).
`checkOperatorProperty`
filters the stream *after* the debounce and already returns false when no
operator
is selected, so the handler can only ever run with an id set — the false arm
is
unreachable through the public stream. The test kept for that scenario
asserts
what is actually observable: deselecting mid-debounce results in no property
write. No production code was changed.
### Any related issues, documentation, discussions?
Closes #7910.
### How was this PR tested?
`ng test --watch=false --include
src/app/workspace/component/property-editor/operator-property-edit-frame/operator-property-edit-frame.component.spec.ts`
— 219 passed / 1 skipped (196 passed before, 23 new), repeated 3× for
stability;
the whole `property-editor/**` folder stays green at 288 passed. `yarn
format:ci`
clean. Failure path verified by breaking one assertion in each of the 23 new
tests: every one turned red with a non-zero exit, then all were restored to
green.
### Was this PR authored or co-authored using generative AI tooling?
Generated-by: Claude Code (Opus 5)
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
