Add migrate-xunit-to-mstest skill from dotnet/skills by Evangelink · Pull Request #54727 · dotnet/sdk

Evangelink · 2026-06-11T19:55:59Z

Imports the migrate-xunit-to-mstest skill from dotnet/skills into this repo so the xUnit → MSTest migration currently in progress (see #54722, #54723, #54724, #54725, #54726) has a documented, repeatable procedure that contributors and agents can follow consistently.

What's added

.github/skills/migrate-xunit-to-mstest/SKILL.md — verbatim from upstream
.github/skills/migrate-xunit-to-mstest/references/mapping-cheatsheet.md — verbatim from upstream

Why land this first

These small migration PRs all follow the same mechanical mapping table ([Fact]→[TestMethod], Assert.Equal→Assert.AreEqual, the Assert.Contains arg-reversal, StringValues ambiguity, Assert.Collection expansion, etc.). Centralizing the procedure as a discoverable skill means reviewers and future contributors don't need to rediscover the gotchas from each PR's description.

Validator note

.github/skills/ValidateSkill.cs flags SKILL.md as 546 lines (advisory limit is 500). The validator is not CI-enforced (no workflow consumes it; it's mentioned only as a checklist item in .github/skills/AGENTS.md). I'm intentionally mirroring upstream verbatim so future re-syncs stay trivial. If the team prefers a trimmed copy, happy to fold the longer sections (steps 8, 11, 12) into references/.

Source

Upstream commit (main at time of import): https://github.com/dotnet/skills/tree/main/plugins/dotnet-test/skills/migrate-xunit-to-mstest
License: MIT (same as this repo)

Imports the migrate-xunit-to-mstest skill from dotnet/skills@main (plugins/dotnet-test/skills/migrate-xunit-to-mstest) into this repo so the xUnit -> MSTest migration that's currently in progress (dotnet#54722, dotnet#54723, dotnet#54724, dotnet#54725, dotnet#54726) has a documented, repeatable procedure that contributors and agents can follow consistently. Files added (verbatim from upstream): - .github/skills/migrate-xunit-to-mstest/SKILL.md - .github/skills/migrate-xunit-to-mstest/references/mapping-cheatsheet.md Note: SKILL.md is 546 lines, slightly above the 500-line "Progressive Disclosure" guideline checked by .github/skills/ValidateSkill.cs. The validator is advisory (not CI-enforced) and we are intentionally mirroring the upstream content verbatim to keep future syncs trivial. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot

Pull request overview

Adds a new agent skill to document and standardize the ongoing xUnit → MSTest migration workflow in this repo, imported from dotnet/skills, along with a mapping cheatsheet reference.

Changes:

Add .github/skills/migrate-xunit-to-mstest/SKILL.md describing an end-to-end migration workflow and decision points.
Add .github/skills/migrate-xunit-to-mstest/references/mapping-cheatsheet.md with detailed attribute/assert/fixture/lifecycle mapping tables.

Show a summary per file

File	Description
.github/skills/migrate-xunit-to-mstest/SKILL.md	Adds the primary migration skill workflow and guidance.
.github/skills/migrate-xunit-to-mstest/references/mapping-cheatsheet.md	Adds the detailed xUnit→MSTest mapping reference used by the skill.

Copilot's findings

Files reviewed: 2/2 changed files
Comments generated: 4

- Bumps MSTest.Sdk to the latest internal preview to pick up the newest 4.3 assertion APIs (Assert.ContainsSingle, Assert.Contains for strings with the more natural (needle, haystack) signature, etc.). - Applies the assertion mapping flagged by the migrate-xunit-to-mstest skill in this repo (.github/skills/migrate-xunit-to-mstest, PR dotnet#54727): Assert.HasCount(1, x) -> Assert.ContainsSingle(x) (one occurrence in AspireResourceLauncherCliTests.cs) Verified: 58/58 tests still pass. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…le(x) Applies the assertion mapping flagged by the migrate-xunit-to-mstest skill (.github/skills/migrate-xunit-to-mstest, PR dotnet#54727). Two occurrences in ParserTests.cs converted from `Assert.HasCount(1, parseResult.Errors)` to the more idiomatic `Assert.ContainsSingle(parseResult.Errors)`, which matches the xUnit original (`Assert.Single`) more directly. Verified: 9/9 tests still pass. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…tain for strings The migrate-xunit-to-mstest skill (.github/skills/migrate-xunit-to-mstest, PR dotnet#54727) notes that MSTest 3.8+ added Assert.Contains(needle, str) / Assert.DoesNotContain(needle, str) overloads for strings -- with the natural (needle, haystack) argument order matching xUnit's Assert.Contains. With the 4.3.0-preview MSTest.Sdk bump in dotnet#54722, these are now available throughout the migration. Replaced: - StringAssert.Contains(haystack, needle) -> Assert.Contains(needle, haystack) - Assert.IsFalse(haystack.Contains(needle)) -> Assert.DoesNotContain(needle, haystack) Touched files: - BrowserRefreshMiddlewareTest.cs - BrowserScriptMiddlewareTest.cs - ResponseStreamWrapperCompressionTest.cs Note: Assert.IsFalse(dict.ContainsKey(key)) call sites are unchanged (no Assert.DoesNotContainKey API exists). Verified: 110/110 tests still pass. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…le(x) Applies the assertion mapping flagged by the migrate-xunit-to-mstest skill (.github/skills/migrate-xunit-to-mstest, PR dotnet#54727). Two occurrences in ParserTests.cs converted from `Assert.HasCount(1, parseResult.Errors)` to the more idiomatic `Assert.ContainsSingle(parseResult.Errors)`, which matches the xUnit original (`Assert.Single`) more directly. Verified: 9/9 tests still pass. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…tain for strings The migrate-xunit-to-mstest skill (.github/skills/migrate-xunit-to-mstest, PR dotnet#54727) notes that MSTest 3.8+ added Assert.Contains(needle, str) / Assert.DoesNotContain(needle, str) overloads for strings -- with the natural (needle, haystack) argument order matching xUnit's Assert.Contains. With the 4.3.0-preview MSTest.Sdk bump in dotnet#54722, these are now available throughout the migration. Replaced: - StringAssert.Contains(haystack, needle) -> Assert.Contains(needle, haystack) - Assert.IsFalse(haystack.Contains(needle)) -> Assert.DoesNotContain(needle, haystack) Touched files: - BrowserRefreshMiddlewareTest.cs - BrowserScriptMiddlewareTest.cs - ResponseStreamWrapperCompressionTest.cs Note: Assert.IsFalse(dict.ContainsKey(key)) call sites are unchanged (no Assert.DoesNotContainKey API exists). Verified: 110/110 tests still pass. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Folds learnings discovered while migrating dotnet/sdk to MSTest.Sdk: - MSTest.Sdk implicit global using note - Assert.IsExactInstanceOfType<T> (MSTest 4.1+) for xUnit Assert.IsType<T> - Assert.AreSequenceEqual (MSTest 4.3+) over CollectionAssert.AreEqual - AwesomeAssertions namespace clarification Keeps the sdk-local copy in sync with the upstream PR. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Evangelink · 2026-06-11T20:42:19Z

Pushed an update folding learnings from the actual migration (PRs #54722-#54726) into the skill, mirroring dotnet/skills#756:

MSTest.Sdk provides Microsoft.VisualStudio.TestTools.UnitTesting as an implicit global using -- don't add <Using> items or per-file usings.
Assert.IsExactInstanceOfType<T> (MSTest 4.1+) is the proper xUnit Assert.IsType<T> equivalent.
Assert.AreSequenceEqual (MSTest 4.3+) over CollectionAssert.AreEqual for element-wise compare.
AwesomeAssertions ships in the FluentAssertions namespace (no source changes needed).

Once dotnet/skills#756 merges, this branch will be in sync with upstream.

… equivalence Addresses review comment from copilot-pull-request-reviewer: the Step 6 summary table conflated xUnit's exact-type Assert.IsType<T> and assignable Assert.IsAssignableFrom<T> into a single row, which implies both map interchangeably even though their semantics differ. Split into three rows -- IsType<T> -> IsExactInstanceOfType<T>, IsNotType<T> -> IsNotExactInstanceOfType<T>, IsAssignableFrom<T> -> IsInstanceOfType<T> -- and explicitly call out the silent-weakening trap. The cheatsheet already mapped these correctly; this aligns the SKILL.md summary so the at-a-glance table doesn't suggest a wrong mapping. Mirrored in dotnet/skills#756. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

… equivalence Address dotnet/sdk#54727 review comment: the Step 6 summary table conflated xUnit's exact-type Assert.IsType<T> and assignable Assert.IsAssignableFrom<T> into a single row, which implies both map interchangeably even though their semantics differ. Split into three rows -- IsType<T> -> IsExactInstanceOfType<T>, IsNotType<T> -> IsNotExactInstanceOfType<T>, IsAssignableFrom<T> -> IsInstanceOfType<T> -- and explicitly call out the silent-weakening trap. The cheatsheet already maps these correctly; this aligns the SKILL.md summary with the cheatsheet so the at-a-glance table doesn't suggest a wrong mapping. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…le(x) Applies the assertion mapping flagged by the migrate-xunit-to-mstest skill (.github/skills/migrate-xunit-to-mstest, PR dotnet#54727). Two occurrences in ParserTests.cs converted from `Assert.HasCount(1, parseResult.Errors)` to the more idiomatic `Assert.ContainsSingle(parseResult.Errors)`, which matches the xUnit original (`Assert.Single`) more directly. Verified: 9/9 tests still pass. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

…tain for strings The migrate-xunit-to-mstest skill (.github/skills/migrate-xunit-to-mstest, PR dotnet#54727) notes that MSTest 3.8+ added Assert.Contains(needle, str) / Assert.DoesNotContain(needle, str) overloads for strings -- with the natural (needle, haystack) argument order matching xUnit's Assert.Contains. With the 4.3.0-preview MSTest.Sdk bump in dotnet#54722, these are now available throughout the migration. Replaced: - StringAssert.Contains(haystack, needle) -> Assert.Contains(needle, haystack) - Assert.IsFalse(haystack.Contains(needle)) -> Assert.DoesNotContain(needle, haystack) Touched files: - BrowserRefreshMiddlewareTest.cs - BrowserScriptMiddlewareTest.cs - ResponseStreamWrapperCompressionTest.cs Note: Assert.IsFalse(dict.ContainsKey(key)) call sites are unchanged (no Assert.DoesNotContainKey API exists). Verified: 110/110 tests still pass. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

* migrate-xunit-to-mstest: add learnings from dotnet/sdk migration Refines the skill based on a real migration of 5 test projects in dotnet/sdk: - MSTest.Sdk implicit global using: do not add `<Using Include='Microsoft.VisualStudio.TestTools.UnitTesting' />` or per-file `using` in MSTest.Sdk projects (it's already in scope). Note Option A (`MSTest` metapackage) still needs the per-file using. - `Assert.IsExactInstanceOfType<T>` (MSTest 4.1+) is the proper single-call equivalent of xUnit's exact-type `Assert.IsType<T>`; previous guidance silently degraded it to assignable semantics (closes #755). - `Assert.AreSequenceEqual` (MSTest 4.3+) is the modern element-wise equivalent of xUnit's `Assert.Equal` on `IEnumerable<T>`, avoiding the `CollectionAssert` + `.ToList()` dance and the MSTEST0065 trap on plain `AreEqual`. - Clarify `AwesomeAssertions` ships in the `FluentAssertions` namespace, so it's a no-source-change swap. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * SKILL.md: split IsType/IsAssignableFrom row to avoid implying drop-in equivalence Address dotnet/sdk#54727 review comment: the Step 6 summary table conflated xUnit's exact-type Assert.IsType<T> and assignable Assert.IsAssignableFrom<T> into a single row, which implies both map interchangeably even though their semantics differ. Split into three rows -- IsType<T> -> IsExactInstanceOfType<T>, IsNotType<T> -> IsNotExactInstanceOfType<T>, IsAssignableFrom<T> -> IsInstanceOfType<T> -- and explicitly call out the silent-weakening trap. The cheatsheet already maps these correctly; this aligns the SKILL.md summary with the cheatsheet so the at-a-glance table doesn't suggest a wrong mapping. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> * SKILL.md: resolve Step 3/Step 4 conflict on the MSTest using for Option B Address review feedback on #756: Step 3 said to skip the per-file 'using Microsoft.VisualStudio.TestTools.UnitTesting;' on Option B (MSTest.Sdk provides it as an implicit global using), but Step 4 unconditionally instructed replacing xUnit usings with that MSTest using -- so an agent following Step 4 literally would re-add the redundant using on Option B projects. Also drop the dangling 'Step 4 rewriter' phrasing in Step 3 (there is no rewriter in Step 4, just a list of mechanical rewrites the agent applies). Step 4 now explicitly says to skip the MSTest using on Option B and only remove the 'using Xunit;' lines, matching Step 2's note and Step 3's branch instructions. Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com> --------- Co-authored-by: Evangelink <amaury@microsoft.com> Co-authored-by: Copilot <223556219+Copilot@users.noreply.github.com>

Copilot AI review requested due to automatic review settings June 11, 2026 19:56

github-actions Bot added the Area-Infrastructure label Jun 11, 2026

Copilot started reviewing on behalf of Evangelink June 11, 2026 19:56 View session

JeremyKuhne approved these changes Jun 11, 2026

View reviewed changes

Copilot AI reviewed Jun 11, 2026

View reviewed changes

Comment thread .github/skills/migrate-xunit-to-mstest/SKILL.md

Comment thread .github/skills/migrate-xunit-to-mstest/SKILL.md

Comment thread .github/skills/migrate-xunit-to-mstest/SKILL.md

Comment thread .github/skills/migrate-xunit-to-mstest/SKILL.md

Evangelink merged commit bf132dc into dotnet:main Jun 12, 2026
11 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add migrate-xunit-to-mstest skill from dotnet/skills#54727

Add migrate-xunit-to-mstest skill from dotnet/skills#54727
Evangelink merged 3 commits into
dotnet:mainfrom
Evangelink:evangelink/add-migrate-xunit-to-mstest-skill

Evangelink commented Jun 11, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Evangelink commented Jun 11, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

Evangelink commented Jun 11, 2026

What's added

Why land this first

Validator note

Source

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Copilot's findings

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Evangelink commented Jun 11, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants