test(frontend): share workflow-editor TestBed setup, drop double-run

[PG1204]

What changes were proposed in this PR?

Addresses the two review comments on #5318's follow-up implementation:

1. Share a common TestBed setup so the jsdom and browser specs don't drift.

The .browser.spec.ts split created two TestBed configurations for the same component - one in workflow-editor.component.spec.ts (External Module Integration describe), one in workflow-editor.browser.spec.ts. Both configured nearly identical imports/providers arrays, which would inevitably drift over time.
Extracted them into a new sibling file workflow-editor.test-utils.ts exporting two arrays:

export const workflowEditorTestImports = [ ... ];
export const workflowEditorTestProviders: Provider[] = [ ... ];
This follows the existing project convention in frontend/src/app/common/testing/test-utils.ts (which exports commonTestImports and commonTestProviders the same way). Each spec's TestBed now collapses to:

await TestBed.configureTestingModule({
imports: workflowEditorTestImports,
providers: workflowEditorTestProviders,
}).compileComponents();
Adding or removing a service from either spec's setup is now a single edit in one file.

2. Drop the explicit workflow-editor.component.spec.ts entry from the test-browser include in angular.json.

With the six mouse-event tests now living in workflow-editor.browser.spec.ts (picked up by the **/.browser.spec.ts glob), the explicit listing of workflow-editor.component.spec.ts in test-browser's include array was causing the file's 25 jsdom-friendly tests to run twice, once in jsdom (the default test target) and once in real Chrome (the test-browser target). Removed that explicit entry; the test-browser target now picks up only **/.browser.spec.ts files.

Scope note. The pre-existing JointJS Paper describe block at the top of workflow-editor.component.spec.ts has its own deliberately-different TestBed setup (ContextMenuComponent instead of NzModalCommentBoxComponent, Overlay, MockComputingUnitStatusService) and was left untouched, it wasn't part of the duplication introduced by the .browser.spec.ts split.

Any related issues, documentation, discussions?

Addresses review feedback on the follow-up PR for #5318.
Closes #5318
Related to #3614 / PR #5146 (this PR restores tests that were commented out during PR #5146 as collateral from the #3614 fix).

How was this PR tested?

Existing test runs: Verified no test-count regressions in either target, and the double-run is gone:

ng test (jsdom, full suite): 947 pass / 0 fail / 2 skipped / 1 todo across 103 files
ng run gui:test-browser: 13 pass / 0 fail across 2 files (6 from workflow-editor.browser.spec.ts + 7 from the pre-existing code-editor.component.browser.spec.ts)
workflow-editor.component.spec.ts under jsdom: 25/25 pass (unchanged from before this PR)
Browser test count dropped from 38 -> 13, confirming the 25 jsdom-friendly tests in workflow-editor.component.spec.ts no longer double-run in real Chrome.

Static checks: tsc --noEmit and eslint clean on all four changed files.

Was this PR authored or co-authored using generative AI tooling?

Co-authored-by: Claude Code (Anthropic Claude Opus 4.7)

[PG1204]

/request-review @Xiao-zhen-Liu

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 52.61%. Comparing base (7ae9b35) to head (fd8d956).

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #5626      +/-   ##
============================================
- Coverage     53.02%   52.61%   -0.42%     
+ Complexity     2542     2485      -57     
============================================
  Files          1075     1076       +1     
  Lines         41743    41638     -105     
  Branches       4513     4490      -23     
============================================
- Hits          22134    21906     -228     
- Misses        18301    18446     +145     
+ Partials       1308     1286      -22     

Flag	Coverage Δ		*Carryforward flag
access-control-service	64.61% <ø> (-6.82%)	⬇️	Carriedforward from 1abb300
agent-service	34.36% <ø> (ø)		Carriedforward from 1abb300
amber	53.30% <ø> (-0.62%)	⬇️	Carriedforward from 1abb300
computing-unit-managing-service	1.65% <ø> (ø)		Carriedforward from 1abb300
config-service	56.71% <ø> (ø)		Carriedforward from 1abb300
file-service	39.34% <ø> (-17.73%)	⬇️	Carriedforward from 1abb300
frontend	47.43% <100.00%> (+0.05%)	⬆️
pyamber	90.70% <ø> (-0.02%)	⬇️	Carriedforward from 1abb300
python	90.74% <ø> (ø)		Carriedforward from 1abb300
workflow-compiling-service	58.69% <ø> (ø)		Carriedforward from 1abb300

*This pull request uses carry forward flags. Click here to find out more.

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[chenlica]

@Xiao-zhen-Liu Please review the PR.

[Yicong-Huang]

cc @Xiao-zhen-Liu for review.

[Xiao-zhen-Liu]

Clean, well-scoped cleanup. Moving the shared setup into one file means the two test files can't slowly fall out of sync, and routing the mouse tests into a .browser.spec.ts file is the right way to kill the double-run — bonus that six dead commented-out tests now actually run.

One note: the diff also changes real app code in workflow-editor.component.ts (an optional validation param) that isn't in the description. Correct and harmless, but it makes a test-only PR not quite test-only — I'd surface it or split it out. Nothing blocking otherwise.

[Yicong-Huang]

I resolved comments based on discussions. @Xiao-zhen-Liu it might be better for you to check and resolve comments as you posted comments and have context.

Now this PR is good to go!