Extensions: address some follow-up comments #78847

jcouv · 2025-06-06T20:20:41Z

This brings down the number of references to the test plan down below 100.
Relates to test plan #76130

jcouv · 2025-06-06T20:25:14Z

src/Compilers/CSharp/Portable/Binder/ForEachLoopBinder.cs

@@ -1535,7 +1535,7 @@ private MethodArgumentInfo FindForEachPatternMethodViaExtension(SyntaxNode synta
                returnRefKind: default,
                returnType: null,
                ref extensionUseSiteInfo,
-                acceptOnlyMethods: true); // Tracked by https://github.com/dotnet/roslyn/issues/76130 : Test effect of acceptOnlyMethods value


📝 this is not observable. If we got an extension property returned in the result here, the way we look at the overload resolution result below would ignore it #Resolved

Could you add more details? Which line would ignore it? Why?

In that scenario, methodGroupResolutionResult.OverloadResolutionResult is null.
The found property is stored in methodGroupResolutionResult.OtherSymbol.
A relevant test is ExtensionMemberLookup_PatternBased_ForEach_DelegateTypeProperty

…nversion_*)

…ed by ExtensionMemberLookup_PatternBased_ForEach_DelegateTypeProperty

…ethod

AlekseyTs · 2025-06-10T12:21:49Z

src/Compilers/CSharp/Portable/Binder/Binder_Expressions.cs

@@ -8016,7 +8016,7 @@ private BoundExpression MakeMemberAccessValue(BoundExpression expr, BindingDiagn
                Debug.Assert(typeArgumentsOpt.IsDefault);
                if (!receiver.HasErrors)
                {
-                    diagnostics.AddRange(resolution.Diagnostics); // Tracked by https://github.com/dotnet/roslyn/issues/76130 : test dependencies/diagnostics
+                    diagnostics.AddRange(resolution.Diagnostics);


diagnostics.AddRange(resolution.Diagnostics);

It looks like no new test was added. Is there an existing test that observes diagnostics reported here? #Closed

The commit message for this says "covered by multiple tests including GetCompatibleExtension_Conversion_*"

AlekseyTs · 2025-06-10T12:27:57Z

src/Compilers/CSharp/Portable/Binder/ForEachLoopBinder.cs

@@ -249,7 +249,7 @@ private BoundForEachStatement BindForEachPartsWorker(BindingDiagnosticBag diagno
                            hasErrors = true;
                        }
                    }
-                    else if (getEnumeratorMethod.GetIsNewExtensionMember() && getEnumeratorMethod.ContainingType.ExtensionParameter.RefKind == RefKind.Ref) // Tracked by https://github.com/dotnet/roslyn/issues/76130: add test coverage for 'ref readonly' and 'in' 
+                    else if (getEnumeratorMethod.GetIsNewExtensionMember() && getEnumeratorMethod.ContainingType.ExtensionParameter.RefKind == RefKind.Ref)


else if (getEnumeratorMethod.GetIsNewExtensionMember() && getEnumeratorMethod.ContainingType.ExtensionParameter.RefKind == RefKind.Ref)

It doesn't look like any of the added tests verify that we report ErrorCode.ERR_RefLvalueExpected. #Closed

Does existing test TestGetEnumeratorPatternViaRefExtensionOnNonAssignableVariable cover this?

Yes, that's the existing test

AlekseyTs · 2025-06-10T12:38:08Z

src/Compilers/CSharp/Portable/Lowering/ExtensionMethodReferenceRewriter.cs

@@ -113,7 +113,7 @@ static BoundExpression updateCall(
                    }
                    else
                    {
-                        argumentRefKinds = argumentRefKinds.Insert(0, argumentRefKindFromReceiverRefKind(receiverRefKind)); // Tracked by https://github.com/dotnet/roslyn/issues/76130 : Test this code path
+                        argumentRefKinds = argumentRefKinds.Insert(0, argumentRefKindFromReceiverRefKind(receiverRefKind));


argumentRefKinds = argumentRefKinds.Insert(0, argumentRefKindFromReceiverRefKind(receiverRefKind));

It looks like no new test was added. Is there an existing test that covers this code path including "interesting" receiverRefKind values? #Closed

Will add (RefReadonlyExtensionParameterWithOneExplicitRefArgument)

AlekseyTs · 2025-06-10T12:41:41Z

src/Compilers/CSharp/Portable/Symbols/Source/SourceNamedTypeSymbol_Extension.cs

@@ -93,12 +93,12 @@ internal sealed override ParameterSymbol? ExtensionParameter

            if (containingType is null)
            {
-                return null; // Tracked by https://github.com/dotnet/roslyn/issues/76130 : Test this code path
+                return null;


return null;

It looks like no new test was added. Is there an existing test that covers this code path? #Closed

The corresponding commit says "Covered by BadContainer_TopLevel"

AlekseyTs · 2025-06-10T12:42:51Z

src/Compilers/CSharp/Portable/Symbols/Source/SourceNamedTypeSymbol_Extension.cs

            }

            if (_lazyExtensionInfo is null)
            {
-                Interlocked.CompareExchange(ref _lazyExtensionInfo, new ExtensionInfo(), null); // Tracked by https://github.com/dotnet/roslyn/issues/76130 : Test this code path
+                Interlocked.CompareExchange(ref _lazyExtensionInfo, new ExtensionInfo(), null);


Interlocked.CompareExchange(ref _lazyExtensionInfo, new ExtensionInfo(), null);

It looks like no new test was added. Is there an existing test that covers this code path? #Closed

TryGetOrCreateExtensionMarker is always called first, as part of initialization of members, so _lazyExtensionInfo is always initialized by the time we get here.

AlekseyTs · 2025-06-10T13:02:35Z

src/Compilers/CSharp/Portable/Symbols/Metadata/PE/PENamedTypeSymbol.cs

            }

            if (!method.IsStatic && ExtensionParameter is null)
            {
-                return null; // Tracked by https://github.com/dotnet/roslyn/issues/76130 : Test this code path
+                throw ExceptionUtilities.Unreachable();


throw ExceptionUtilities.Unreachable();

I would prefer to keep the original code even if our tests do not hit this code path at the moment. Returning null is a reasonable thing to do in the given situation, and, if we really really want to, we can test this behavior by calling the API directly. I am fine with throwing when this.ContainingType is null, because that can be treated as an invalid API usage on a vary very broken declaration. However, I wouldn't make consumer responsible for checking whether ExtensionParameter is null before calling this API. #Closed

AlekseyTs · 2025-06-10T13:06:01Z

src/Compilers/CSharp/Portable/Symbols/Metadata/PE/PENamedTypeSymbol.cs

@@ -2045,7 +2043,7 @@ private MethodDefinitionHandle TryGetExtensionMarkerMethod()
                    {
                        if (!foundMarkerMethod.IsNil)
                        {
-                            return default; // Tracked by https://github.com/dotnet/roslyn/issues/76130 : Test this code path
+                            return default;


return default;

If PENamedTypeSymbol_09 unit-test gets here, I would expect it to report an error of some sort. #Closed

AlekseyTs · 2025-06-10T13:09:33Z

src/Compilers/CSharp/Portable/Symbols/Metadata/PE/PENamedTypeSymbol.cs


                    if (candidate.Arity != @this.Arity + method.Arity)
                    {
-                        continue; // Tracked by https://github.com/dotnet/roslyn/issues/76130 : Test this code path
+                        continue;


continue;

If PENamedTypeSymbol_10 added to cover this code path, I would expect an error of some sort. #Closed

AlekseyTs · 2025-06-10T13:11:32Z

src/Compilers/CSharp/Portable/Symbols/Metadata/PE/PENamedTypeSymbol.cs

                foreach (var member in @this.ContainingType.GetMembers(method.Name))
                {
                    if (member is not MethodSymbol { IsStatic: true } candidate)
                    {
-                        continue; // Tracked by https://github.com/dotnet/roslyn/issues/76130 : Test this code path
+                        throw new System.Exception("JCOUV2");
+                        continue; // Tracked by https://github.com/dotnet/roslyn/issues/76130 : Test this code path TODO2
                    }



// Tracked by #76130 : Consider comparing accessibility as well TODO2

Are we checking somewhere that extension type itself is public?

AlekseyTs · 2025-06-10T13:14:49Z

src/Compilers/CSharp/Test/Emit3/Semantics/ExtensionTests2.cs

+    [Fact]
+    public void PENamedTypeSymbol_11()
+    {
+        // Accessibility mismatch between skeleton and implementation


// Accessibility mismatch between skeleton and implementation

I do not see the mismatch. Why does the test report an error? #Closed

AlekseyTs · 2025-06-10T13:20:11Z

src/Compilers/CSharp/Test/Emit3/Semantics/ExtensionTests2.cs

+    [Fact]
+    public void PENamedTypeSymbol_15()
+    {
+        // parameter type mismatch, instance method


// parameter type mismatch, instance method

Is there a similar test for a static method? #Closed

AlekseyTs · 2025-06-10T13:27:22Z

src/Compilers/CSharp/Portable/Symbols/Metadata/PE/PENamedTypeSymbol.cs

@@ -406,14 +406,14 @@ internal sealed override ParameterSymbol? ExtensionParameter
                {
                    var methodSymbol = getMarkerMethodSymbol(@this, uncommon);

-                    // Tracked by https://github.com/dotnet/roslyn/issues/76130 : do we want to tighten the flags check further? (require that type be sealed?)
-                    if (methodSymbol.DeclaredAccessibility != Accessibility.Private ||
+                    if (!methodSymbol.ContainingType.IsSealed ||


!methodSymbol.ContainingType.IsSealed ||

It feels like the check is misplaced. If we say that extension block must be sealed, then we shouldn't treat the enclosing type as IsExtension, and, therefore, should never get here. #Closed

I wasn't sure because we've kept the checks that make the nested type not count as an extension to the minimum (typekind check and existence of a method with right name). Most detailed checks are delayed and don't mean IsExtension is false. I'm good either way

AlekseyTs · 2025-06-10T13:34:57Z

Done with review pass (commit 26). I was reviewing commit by commit. #Closed

AlekseyTs · 2025-06-12T12:19:40Z

src/Compilers/VisualBasic/Test/Symbol/SymbolDisplay/SymbolDisplayTests.vb

@@ -6060,14 +6060,15 @@ static class E
            Dim comp As Compilation
            If useMetadata Then
                Dim libComp = CreateCSharpCompilation("c", text, parseOptions:=parseOptions)
-                comp = CreateCSharpCompilation("d", code:="", parseOptions:=parseOptions, referencedAssemblies:={libComp.EmitToImageReference()})
+                comp = CreateCSharpCompilation("d", code:="", parseOptions:=parseOptions, referencedAssemblies:=libComp.References.Concat(libComp.EmitToImageReference()))


libComp.References.Concat(

I am curious what prompted this change. #Closed

When references are given to this method, the default references are not passed. So the compilation would lack System.Object and the type would not be considered IsExtension

AlekseyTs · 2025-06-12T12:28:14Z

Done with review pass (commit 28)

AlekseyTs

LGTM (commit 28)

jjonescz

LGTM

* Simpli * Simplify * Simplify * docs * docs * Clear * Update src/Workspaces/Core/Portable/FindSymbols/Shared/AbstractSyntaxIndex_Persistence.cs * Remove method * Fix ILBuilder visualization (#78764) Turns out that we do have visualization of in-progress IL building, but it's been broken since 2020, when we renamed Roslyn.Test.Utilities to Microsoft.CodeAnalysis.Test.Utilities. It was then further broken by the change a few months ago to pass IL visualization formats along. This gets the debugger display working again. * Remove EditorFeaturesWpf TestComposition (#78769) After the EditorFeatures.WPF => EditorFeatures refactoring, this TestComposition became equivalent to the EditorFeatures one, so just removing the WPF one. * Fix renames across different extensions * Add docs * Add docs * Update src/VisualStudio/Core/Def/PackageRegistration.pkgdef * Explicitly reset * Reset state * Add docs * Simplify * Revert * Simplify * Simplify * Docs * Update src/VisualStudio/Core/Def/Commands.vsct Co-authored-by: David Barbet <[email protected]> * Update src/VisualStudio/Core/Impl/SolutionExplorer/SymbolTree/SymbolItemContextMenuController.cs Co-authored-by: David Barbet <[email protected]> * Simplify * Simplify * multi notify * only if it changed * Don't move backwards * Docs * Update src/VisualStudio/Core/Impl/SolutionExplorer/SymbolTree/SymbolTreeChildCollection.cs * Ensure sln load uses project absolute paths * lint * Rename field to make it clearer what it's used for It's only for dynamic files, so make that clear. * Don't refresh dynamic files under a lock When we were getting the notification of a dynamic file being changed, we were taking the project and workspace locks before fetching the actual content. This is really expensive now that the fetching itself might not be instant. Fundamentally this change is just switching the work to acquire the locks to get the setup info, then releasing the locks and calling the provider to get the info. I'm putting this around a batching queue so we can deduplicate lots of events at once, and also to ensure ordering so if we were to try refreshing the document more than once we don't end up in a case where we have two threads trying to put different versions of the document into the workspace and getting the ordering wrong. Fixes #78734 * Fix cast before we unsubscribe This was otherwise throwing an exception every shutdown. * Adjust implementation to respect diagnostic flag and update tests * Remove workarounds for bugs that have been fixed Closes #77995 * Remove a workaround we had for an old version of Copilot This bug was fixed a long time ago. * Switch return to continue * Fixed multi-variable declaration support in RemoveUnnecessarySuppressions. * Removed whitespace. Co-authored-by: DoctorKrolic <[email protected]> * Simplified TestRemoveDiagnosticSuppression_Attribute_MultiVariableDeclaration with inline data. * Expanded remove unnecessary suppression tests, removed unnecessary loop and fixed partial method/property support in AbstractRemoveUnnecessaryInlineSuppressionsDiagnosticAnalyzer. * Directly create virtual project when dotnet run-api is missing for now * Avoid dereferencing null CheckConstraintsArgs.CurrentCompilation (#78729) Fixes #78430. * Updates to unblock dartlab pipeline (#78793) * update timeout * wip * Add main to main-vs-deps flow (#78798) * Create branch-merge2.jsonc * Update main-merge.yml * Rename branch-merge2.jsonc to branch-merge-2.jsonc * Update PublishData.json * Update and rename branch-merge.jsonc to main-to-main-vs-deps-branch-merge.jsonc * feedbacl * make clear main is for validation * EnC: Simplify diagnostic reporting (#78708) * Refactoring of extension methods in source packages (#78620) * Change namespace of OneOrMany and TemporaryArray to MS.CA.Collections * Cleanup Enumerable and ImmutableArray extensions * Update imports/usings * Fix VB remove-imports not showing up * add the schedule back to main-merge (#78807) * Fix LSP references for using alias * Clean up HasDuplicates extension method (#78808) * Extensions: Do not consider extension block methods as entry points, consider their implementation methods instead (#78667) * Decouple EditorFeatures.Test.Utilities from VS services Microsoft.CodeAnalysis.EditorFeatures.Test.Utilities depended on Microsoft.VisualStudio.Editor which brought along various VS layer services. Further, we had some stub implementations of those VS services there. Digging into this this coupling wasn't really necessary: 1. There were EditorFeatures tests using the editor adapters stub, but those tests pass just fine without it, so that was easy to move to the VS specific tests. 2. StubVsServiceExporters didn't appear to be used either, so unsure why those were in this layer at all. Those were also moved. 3. We had implementations of some settings services, which it's easier to just delete and remove the options persister from the composition in the first place. This also better ensures that we don't have code that might try creating registry keys from ILocalRegistry which we absolutely don't want. * Extensions: mark skeleton type with 'specialname' (#78690) * Fix await completion in an async iterator * Extensions: fix lowering of implicit conversion on receiver in pattern-based scenarios (#78685) * Add test * Find metadata reference for alias * Add a fallback path when launching the BuildHost * refactor * Added notes to update GetBuildHostPath if packaging changes * Update comments * Add heuristic for the loaded from nuget package scenario * Fix corruption of sliding text window when trying to peek backwards. (#78774) Co-authored-by: Fred Silberberg <[email protected]> * Track changed text instead of clearing immediate window. * EnC: Partial solution updates (#78744) * Restore some parts of the progression api some legacy components (like codemap) use. * Remove unused internal APIs * Remove unused internal APIs * Remove unused internal APIs * Remove unused internal APIs * Make static * Simplify * Avoid ignored directive errors in disabled regions (#78158) * Extensions: use specific tracking issues for different areas (#78834) * Extensions: relax inferrability rule (#78758) * Remove blank line output from the node writer proportional to the number of non-abstract nodes (#78844) * Revert "Refactoring of extension methods in source packages (#78620)" (#78850) This reverts commit 917401d, broke typescript * Switch off of mutlidictionary (#78851) This is from ICSharpCompiler, and is therefore causing unnecessary dll loads. * Switch behavior of "Go to definition" and "Go to implementation" for partial members * Fix LSP tests * Fix peek test * Switch to existing helpers for multi-dictionary use * call TryGetValue to get dictionary entry (#78863) * Remove more progression code * Remove more progression code * Remove more progression code * Remove more progression code * Remove more progression code * Remove more progression code * Remove more progression code * Remove more progression code * Add VSTypeScriptAsynchronousTaggerProvider2 for TypeScript that avoids TemporaryArray (#78866) * Remove more progression code * Remove more progression code * Remove more progression code * Remove more progression code * Fix build break (#78870) * Add more ETW events to trace assembly loading (#78840) * Add more ETW events to trace assembly loading * Update src/Compilers/Core/Portable/CodeAnalysisEventSource.Common.cs Co-authored-by: Joey Robichaud <[email protected]> * Put correct alc reference * Use binary literals --------- Co-authored-by: Joey Robichaud <[email protected]> * Introduce EA layer for TS to integrate with classification * Simplify * Add CommandLineResource API (#78679) * Revert "Revert "Refactoring of extension methods in source packages (#78620)" (#78850)" This reverts commit e083be9. * Fix * Revert "Fix" This reverts commit 574935d. * Revert "Revert "Revert "Refactoring of extension methods in source packages (#78620)" (#78850)"" This reverts commit 27ae586. * Lint * Remove all progression code * Revert * remove code * Fix * Reduce allocations under ContextMutableIntervalTree (#78885) There are some closure allocations unnecessarilly happening during ContextMutableIntervalTree construction. Simply changed a couple methods to static and added a parameter for the extra bit of data they need to get rid of those allocations. This is a very small improvement, only about 0.1% (6 MB) of allocations during typing in the razor speedometer test. * Reduce allocations in the ImageElementConverter and ImageIdConverter Read methods (#78881) * Reduce allocations in the ImageElementConverter and ImageIdConverter Read methods These methods show up in the typing scenario in the razor speedometer test as about 0.9% (63 MB) of allocations. 1) Changed ImageIdConverter to be more like ImageElementConverter and not create a JsonDocument object to query 2) Changed several Utf8JsonReader.GetText calls to instead use Utf8JsonReader.CopyString 3) Changed JsonElement.GetString and new Guid(...) to instead use Utf8JsonReader.GetGuid() Note that if this PR is merged, I'll also try to make a change to vslanguageserverclient to also do the same as that code has the same issues. * Fix issue serializing null options in VB * Yield in task continuation to prevent stack overflow * Move to .NET 10 Preview 5 * Remove RemoteControl workaround * Lint response * Lint response * Lint response * Lint response * Lint response * Fix issue with not resetting valueLength (#78890) * Fix issue with not resetting valueLength #78881 was just merged with a small bug, in that valueLength wasn't reset before it was used a second time. If the typeName is > 64 chars, then this would have thrown a different exception than it should have down a couple lines. * remove unused method * Lint response * Lint response * fix compiler side * Simplify workspace hookup in syntactic tagger * Move * Switch to ITextBuffer2 * REvert * REvert * Update src/EditorFeatures/Core/InlineDiagnostics/AbstractDiagnosticsTaggerProvider.cs * Update src/EditorFeatures/Core/StringIndentation/StringIndentationTaggerProvider.cs * Update src/EditorFeatures/Core/Tagging/AsynchronousViewportTaggerProvider.cs * REvert * remove more workspace registration code * Move to .NET 10 Preview 5 (#78906) * Move to .NET 10 Preview 5 * Lint response * Lint response * Lint response * Lint response * Lint response * remove unused method * Lint response * Lint response * fix compiler side * more --------- Co-authored-by: Cyrus Najmabadi <[email protected]> * Fix * Inline Hints - do not allow double click for collection expression type hint (#78900) * do not allow double click for collection expression hint * Update src/EditorFeatures/Test2/InlineHints/CSharpInlineTypeHintsTests.vb Co-authored-by: Cyrus Najmabadi <[email protected]> * comment --------- Co-authored-by: Cyrus Najmabadi <[email protected]> * Add tests * Revert "Update code to use null propagation (#78733)" This reverts commit d785549, reversing changes made to a8808be. * Revert "Move to .NET 10 Preview 5 (#78906)" This reverts commit a7fa681. * fix warning * Revert "Update to using unbound nameof(X<>) expressions (#78731)" This reverts commit 6a09280, reversing changes made to a7fa681. * Revert "Update to using simple untyped lambdas (#78732)" This reverts commit a8808be, reversing changes made to 6a09280. * Fix OOP crash issue with copilot change analysis * Re-enable IDE0051 (#78919) This re-enables the unused member analyzer and removes all of the methods it flagged as unused. * Extensions: address some follow-up comments (#78847) * Pass missing Hot Reload result properties (#78929) * Fix deadlock if an MSBuild task is writing to stdout When we switched over to communicating over a named pipe rather than stdin/stdout, we were still redirecting stdin and stdout. This had the side effect that if a build task was directly writing to standard out, the build would eventually deadlock since we weren't reading from the other side. I thought about simply not redirecting stdin/stdout, but I could imagine other problems might come up if we were to have multiple build hosts trying to share stdin/stdout. So now we'll log stdout the same way we log stderr, and explicitly close stdin so readers won't deadlock waiting for input. Fixes #78766 * Revert * Collect stats * Extensions: allow cref references to extension members (#78735) * Fix scoped variance checks involving ref struct interface implementation (#78883) * Remove now unused Progression references and related hacks * Expose a couple of things to Razor * Remove more unneeded references * Obsolete api * [main] Source code updates from dotnet/dotnet (#78805) * [VMR] Codeflow a528f84-a528f84 [[ commit created by automation ]] * Update dependencies from https://github.com/dotnet/dotnet build 270315 No dependency updates to commit * [VMR] Codeflow f5faea9-f5faea9 [[ commit created by automation ]] * Update dependencies from https://github.com/dotnet/dotnet build 270450 No dependency updates to commit * Update dependencies from https://github.com/dotnet/dotnet build 270603 No dependency updates to commit * Update dependencies from https://github.com/dotnet/dotnet build 270662 No dependency updates to commit * [VMR] Codeflow 86826d3-86826d3 [[ commit created by automation ]] * Update dependencies from https://github.com/dotnet/dotnet build 271018 No dependency updates to commit * [VMR] Codeflow 32a2620-32a2620 [[ commit created by automation ]] * Update dependencies from https://github.com/dotnet/dotnet build 271181 No dependency updates to commit * [VMR] Codeflow 25357a9-25357a9 [[ commit created by automation ]] * Update dependencies from https://github.com/dotnet/dotnet build 271343 No dependency updates to commit * Update dependencies from https://github.com/dotnet/dotnet build 271417 No dependency updates to commit * Revert incorrect Roslyn.sln changes --------- Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com> Co-authored-by: Viktor Hofer <[email protected]> * Code simplification * Specify single reducer * Simplify * Add test * Only add annotations to potential references * :Extract type * Docs * move up * Expose `IsIterator` as a public API (#78813) * Expose `IsIterator` as a public API * Fix workspace generated method symbol implementation * Implement new property in yet another non-compiler-owned method symbol implementation * SemanticSearch.ReferenceAssemblies * Verify public code path in C# * Add negative tests * Add C# local function tests * Test VB lambda iterators * Revert accessibility change * Disable warnings * Use explicit type * Simplify lambda tests * Test interface property * Add additional VB case * Update enum values * PR Feedback --------- Co-authored-by: Cyrus Najmabadi <[email protected]> Co-authored-by: Cyrus Najmabadi <[email protected]> Co-authored-by: Todd Grunke <[email protected]> Co-authored-by: David Barbet <[email protected]> Co-authored-by: David Barbet <[email protected]> Co-authored-by: Jason Malinowski <[email protected]> Co-authored-by: John Douglas Leitch <[email protected]> Co-authored-by: DoctorKrolic <[email protected]> Co-authored-by: Rikki Gibson <[email protected]> Co-authored-by: AlekseyTs <[email protected]> Co-authored-by: Ankita Khera <[email protected]> Co-authored-by: Tomáš Matoušek <[email protected]> Co-authored-by: Julien Couvreur <[email protected]> Co-authored-by: Joey Robichaud <[email protected]> Co-authored-by: Jan Jones <[email protected]> Co-authored-by: DoctorKrolic <[email protected]> Co-authored-by: Chris Sienkiewicz <[email protected]> Co-authored-by: Joey Robichaud <[email protected]> Co-authored-by: tmat <[email protected]> Co-authored-by: Jared Parsons <[email protected]> Co-authored-by: David Wengier <[email protected]> Co-authored-by: dotnet-maestro[bot] <42748379+dotnet-maestro[bot]@users.noreply.github.com> Co-authored-by: dotnet-maestro[bot] <dotnet-maestro[bot]@users.noreply.github.com> Co-authored-by: Viktor Hofer <[email protected]>

jcouv self-assigned this Jun 6, 2025

jcouv added Area-Compilers Feature - Extension Everything The extension everything feature labels Jun 6, 2025

jcouv commented Jun 6, 2025

View reviewed changes

jcouv added 26 commits June 6, 2025 15:10

TODO tests

555104f

Remove TODO2

949909f

Remove (covered by multiple tests including GetCompatibleExtension_Co…

345f794

…nversion_*)

Cover ReportQueryInferenceFailedSelectMany

1406899

Cover BindForEachPartsWorker

b901a1a

Cover foreach GetEnumeratorInfoCore

76dfcbc

Remove comment related to FindForEachPatternMethodViaExtension. Cover…

180e701

…ed by ExtensionMemberLookup_PatternBased_ForEach_DelegateTypeProperty

Cover AbstractFlowPass.VisitDelegateCreationExpression

c43b3bd

Remove follow-up with argumentRefKinds (covered)

2df86b3

Cover HasAsyncMethodBuilderAttribute

61ff12c

Cover PEMethod.GetUseSiteInfo

3e81bb2

Covered by BadContainer_TopLevel

bb6bb0f

Unable to cover a case in TryGetCorrespondingExtensionImplementationM…

2055cbc

…ethod

WIP on retargetnig

fe0fc8f

Cover retargeting

472999d

WIP on PENamedTypeSymbol

95d9efb

Cover TryGetExtensionMarkerMethod

4dbc235

Arity mismatch

b4b0545

Check accessibilities match

3101758

More tests

16f406b

More tests

ad81fe4

More tests

072e496

Skeleton type must be sealed

f0db3ee

Remove baseline test

eb9678a

All-in-one

9a92502

Mark issue as MQ

7c82c68

AlekseyTs reviewed Jun 10, 2025

View reviewed changes

jcouv added 2 commits June 11, 2025 19:38

Address feedback

97c4eaa

Fix VB test

bc3ae54

jcouv requested review from jjonescz and AlekseyTs June 12, 2025 05:58

AlekseyTs reviewed Jun 12, 2025

View reviewed changes

AlekseyTs approved these changes Jun 12, 2025

View reviewed changes

jcouv merged commit 225e7f2 into dotnet:main Jun 12, 2025
24 checks passed

jcouv deleted the extensions-cover branch June 12, 2025 16:31

dotnet-policy-service bot added this to the Next milestone Jun 12, 2025

jjonescz reviewed Jun 12, 2025

View reviewed changes

This was referenced Jun 14, 2025

[Automated] PRs inserted in VS build main-10713.56 #78974

Closed

[Automated] PRs inserted in VS build feature.debugger.main-10716.51 #78998

Closed

dotnet-bot mentioned this pull request Jun 17, 2025

[Automated] PRs inserted in VS build feature.Wix5-10716.138 #79003

Closed

RikkiGibson modified the milestones: Next, 18.0 P1 Aug 19, 2025

Extensions: address some follow-up comments #78847

Extensions: address some follow-up comments #78847

Uh oh!

Conversation

jcouv commented Jun 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jcouv Jun 6, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

AlekseyTs commented Jun 10, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

jcouv commented Jun 6, 2025 •

edited

Loading

jcouv Jun 6, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 10, 2025 •

edited

Loading

AlekseyTs commented Jun 10, 2025 •

edited

Loading

AlekseyTs Jun 12, 2025 •

edited

Loading