Fix docs-builder redirect tests

[lcawl]

Fixes #3006

Summary

Implemented fix in tests/authoring/Framework/Setup.fs:

1. RedirectMockTargets module

   • Parses production docs/_redirects.yml with YamlDotNet and collects local targets (same shapes as the real redirect model: scalars, to, many, empty → index.md, anchors-only → from key).
   • For each target, if the real file exists under docs/ and the mock does not already have that path, adds a small stub markdown file (not a full copy of production content).

2. Order of operations

   • Reads _redirects.yml from disk and runs injection before the generated TOC is written so stub paths are included in navigation like other files.

3. Why stubs instead of File.ReadAllText

   • Copying full pages pulled in the real index.md, broke tests that rely on a synthetic index.
   • Full cli/changelog/*.md content also triggered navigation / processing failures (GetNavigationFor) and noisy diagnostics.
   • Stubs keep redirect validation green while keeping the mock docset small and stable. The module comment documents this.

4. Dependencies

   • Added YamlDotNet to tests/authoring/authoring.fsproj (central versions in Directory.Packages.props).

5. Golden test

   • Updated tests/authoring/Generator/LinkReferenceFile.fs so links.json expectations include the new cli/changelog/*.md link entries ({}), the cli/release/* → cli/changelog/* redirects from the current _redirects.yml, and redirect key order matching serialization (YAML order).

Check: dotnet run --project build -c release -- unit-test — passed (all projects including 304 authoring tests).

Tests

I checked out #2980 and cherry-picked this fix this re-ran dotnet run --project build -c release -- unit-test. It now completes successfully.

Generative AI disclosure

1. Did you use a generative AI (GenAI) tool to assist in creating this contribution?

• Yes
• No

2. If you answered "Yes" to the previous question, please specify the tool(s) and model(s) used (e.g., Google Gemini, OpenAI ChatGPT-4, etc.).

Tool(s) and model(s) used: composer-2

[coderabbitai]

📝 Walkthrough

Walkthrough

The changes fix failing redirect validation tests by enhancing the test setup to handle redirects to arbitrary documentation files. The Setup.fs file now parses the production docs/_redirects.yml file, extracts local redirect targets, and creates stub markdown files in the mock test filesystem for any targets that exist in the real docs tree. The LinkReferenceFile.fs test expectations are updated to reflect new changelog-related redirects and their corresponding link entries. A new YamlDotNet dependency is added to enable YAML parsing in the test project.

Suggested labels

dependencies

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title 'Fix docs-builder redirect tests' clearly summarizes the main change: addressing failing redirect validation tests by updating the test setup framework.
Description check	✅ Passed	The description is comprehensive and directly related to the changeset, explaining the problem, solution implementation, design rationale, dependencies added, and test verification.
Linked Issues check	✅ Passed	The PR fully addresses issue #3006 requirements: parses production _redirects.yml to collect redirect targets [#3006], creates stub files for real targets that exist in docs/ but not in mock [#3006], and injects stubs before TOC generation so they appear in navigation [#3006].
Out of Scope Changes check	✅ Passed	All changes are directly scoped to issue #3006: Setup.fs redirect handling, LinkReferenceFile.fs golden test updates for new redirects/links, and necessary YamlDotNet dependency addition.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.

_{✏️ Tip: You can configure your own custom pre-merge checks in the settings.}

✨ Finishing Touches

✨ Simplify code

• Create PR with simplified code
• Commit simplified code in branch redirect-fix

---

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[coderabbitai]

🧹 Nitpick comments (2)

tests/authoring/Framework/Setup.fs (2)

75-79: Redundant pattern branches.

Both None cases return outerFromKey, making the when hasAnchors guard unnecessary:

 let effectiveTo =
     match toVal with
     | Some t -> t
-    | None when hasAnchors -> outerFromKey
     | None -> outerFromKey

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/authoring/Framework/Setup.fs` around lines 75 - 79, The match for
computing effectiveTo is redundant because both None branches return
outerFromKey; simplify the logic by removing the unnecessary guard and collapse
the branches: match toVal with Some t -> t | None -> outerFromKey, keeping
symbols effectiveTo, toVal, hasAnchors, and outerFromKey in mind when updating
the code.

---

95-100: Unsafe downcast could throw if YAML structure is unexpected.

The direct cast :?> YamlMappingNode will throw InvalidCastException if kv.Value is not a mapping node. Given this parses a controlled production file, the risk is low, but a defensive pattern match would be safer:

♻️ Optional defensive fix

 match
     rootMap.Children
     |> Seq.tryPick (fun kv ->
         match kv.Key with
-        | :? YamlScalarNode as k when k.Value = "redirects" -> Some(kv.Value :?> YamlMappingNode)
+        | :? YamlScalarNode as k when k.Value = "redirects" ->
+            match kv.Value with
+            | :? YamlMappingNode as m -> Some m
+            | _ -> None
         | _ -> None)

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@tests/authoring/Framework/Setup.fs` around lines 95 - 100, The match uses an
unsafe downcast (kv.Value :?> YamlMappingNode) which can throw if the YAML node
isn't a mapping; update the pattern to defensively match kv.Value as a
YamlMappingNode (e.g., match kv.Value with :? YamlMappingNode as m -> Some m | _
-> None) so the Seq.tryPick branch only returns a mapping node when it's safe;
adjust the expression around rootMap.Children and the kv match accordingly to
avoid InvalidCastException.

🤖 Prompt for all review comments with AI agents

Verify each finding against the current code and only fix it if needed.

Nitpick comments:
In `@tests/authoring/Framework/Setup.fs`:
- Around line 75-79: The match for computing effectiveTo is redundant because
both None branches return outerFromKey; simplify the logic by removing the
unnecessary guard and collapse the branches: match toVal with Some t -> t | None
-> outerFromKey, keeping symbols effectiveTo, toVal, hasAnchors, and
outerFromKey in mind when updating the code.
- Around line 95-100: The match uses an unsafe downcast (kv.Value :?>
YamlMappingNode) which can throw if the YAML node isn't a mapping; update the
pattern to defensively match kv.Value as a YamlMappingNode (e.g., match kv.Value
with :? YamlMappingNode as m -> Some m | _ -> None) so the Seq.tryPick branch
only returns a mapping node when it's safe; adjust the expression around
rootMap.Children and the kv match accordingly to avoid InvalidCastException.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: ce7570e4-629b-4206-b96c-c653201e01d2

📥 Commits

Reviewing files that changed from the base of the PR and between dc4a854 and 62be0bf.

📒 Files selected for processing (3)

• tests/authoring/Framework/Setup.fs
• tests/authoring/Generator/LinkReferenceFile.fs
• tests/authoring/authoring.fsproj