- 18 3月, 2015 13 次提交
-
-
由 Kevin Pilch-Bisson 提交于
Handle null AccessorList in expression bodied properties
-
由 Tomáš Matoušek 提交于
Mark SyntaxFacts.IsLambdaBody as obsolete
-
由 Charles Stoner 提交于
Re-enable EvaluateInitializerExpression test Fixes #1300
-
由 Jared Parsons 提交于
Convert FakeSign to a NuGet reference
-
由 Jared Parsons 提交于
Fixed an errant bad line ending
-
由 Charles Stoner 提交于
-
由 Tom Meschter 提交于
Delay-load Roslyn services
-
由 Tomas Matousek 提交于
-
由 Charles Stoner 提交于
-
由 Jared Parsons 提交于
-
由 Tom Meschter 提交于
Commit bcb2c8c7 added an `[Import]` of `IDiagnosticAnalyzerService` to our Solution Explorer code. Unfortunately, this causes us to compose and load effectively all of Roslyn whenever the Solution Explorer is opened, even if no C# or VB code is loaded. The fix is to delay pulling in the `IDiagnosticAnalyzerService` until we actually need to make use of it, which is guaranteed to be after we have loaded a C# or VB project (and thus we've already loaded Roslyn).
-
由 Pilchie 提交于
Fixes #1297 a crash in line separators. I also took a look around for other places where we access BasePropertyDeclarationSyntax and tried to tighten them up too.
-
由 David Poeschl 提交于
Qualify member accesses with type name in nameof
-
- 17 3月, 2015 27 次提交
-
-
由 David Poeschl 提交于
Fixes #1193 During expansion, directly accessed members need to be qualified with "this." or "TypeName.". Prior to this change, we always qualified instance members with "this." and static members with "TypeName.", but this technique fails in nameof contexts where an instance member can be referenced from a static method, resulting in an improper "this." qualification. We can only legally use "this." when both the member being accessed and the context from which we are accessing it are non-static. However, this change updates the expanders to always use "TypeName." qualification in nameof contexts regardless of the staticness of either member because it is always legal and feels more natural in nameof expressions (where no actual value is being accessed).
-
由 Manish Vasani 提交于
Handle unsupported diagnostics reported by analyzers. Fixes #252 : If an analyzer reports a diagnostic with an unsupported diagnostic ID, i.e. no descriptor returned by SupportedDiagnostics has that ID, then throw an ArgumentException in ReportDiagnostic method. This exception would be turned into an analyzer diagnostic by the driver and reported back to the analyzer host. Also fix a few tests that were reporting diagnostics with unsupported ID!
-
由 mavasani 提交于
-
由 mavasani 提交于
Fixes #252 : If an analyzer reports a diagnostic with an unsupported diagnostic ID, i.e. no descriptor returned by SupportedDiagnostics has that ID, then throw an ArgumentException in ReportDiagnostic method. This exception would be turned into an analyzer diagnostic by the driver and reported back to the analyzer host. Also fix a few tests that were reporting diagnostics with unsupported ID!
-
由 Shyam N 提交于
Sort nodes by location before determining contiguous spans encompassed b...
-
由 Shyam N 提交于
CommonSyntaxNodeExtensions.GetContiguousSpans() was incorrectly assuming that the supplied nodes will be in the order in which they appear in source. This helper is used in the analyzer that produces diagnostics to grey out unused usings (i.e. to produce one diagnostic for a set of contiguous unused using directives). Due to the above incorrect assumption, we would end up producing a single diagnostic spanning multiple unused usings sometimes and multiple diagnostics at other times. Often different invocations of the analyzer for the same source file would produce different set of unused usings diagnostics. While this does not cause observable problems in regular usage - this is a problem for tests and led to flaky behavior where test would pass sometimes and fail at other times.
-
由 Tomáš Matoušek 提交于
VB EnC: enable editing methods with lambdas and LINQ - merged
-
由 Ravi Chande 提交于
Use TryGetValue when looking up cached SyntaxContexts
-
由 Charles Stoner 提交于
Use Unicode escape sequences for control characters
-
由 CyrusNajmabadi 提交于
Make the IsLowPriority API take a cancellation token as well.
-
由 Charles Stoner 提交于
-
由 Tomas Matousek 提交于
Changes to the syntax associated with BoundQueryBody caused slight changes in error span of some rarely reported errors. Update tests.
-
由 Charles Stoner 提交于
-
由 Tomas Matousek 提交于
-
https://github.com/tmat/roslyn由 Tomas Matousek 提交于
Conflicts: src/Compilers/CSharp/Portable/Syntax/SyntaxUtilities.cs src/Compilers/VisualBasic/Portable/Lowering/LambdaRewriter/LambdaFrame.vb src/Compilers/VisualBasic/Portable/Symbols/AnonymousTypes/AnonymousTypeManager_Templates.vb src/Compilers/VisualBasic/Portable/Syntax/SyntaxUtilities.vb src/Compilers/VisualBasic/Test/Emit/CodeGen/CodeGenClosureLambdaTests.vb src/EditorFeatures/VisualBasicTest/EditAndContinue/ActiveStatementTests.vb
-
由 Tomas Matousek 提交于
-
由 Charles Stoner 提交于
Preserve Key modifiers from existing anonymous type properties Fixes #1292
-
由 Cyrus Najmabadi 提交于
-
由 CyrusNajmabadi 提交于
Make the IWorkCoordinatorPriorityService async and appropriately cascade out the asynchrony.
-
由 Charles Stoner 提交于
-
由 Charles Stoner 提交于
-
由 Tomas Matousek 提交于
-
由 Charles Stoner 提交于
-
由 Charles Stoner 提交于
-
由 Cyrus Najmabadi 提交于
Originally this was synchronous because it was used through codepaths that all were inherently synchronous. Over the last year the WorkCoordinator has become increasingly async, and we're now at the point where this API can be async too. This is a win for TypeScript because we have to currently implement this API in blocking fashion. This has shown up in profiles, and making it async should be a nice win for us.
-
由 Jonathon Marolf 提交于
Handle using static
-
由 Jonathon Marolf 提交于
Responding to PR feedback, fixing some typos and simplifying some of the boolean logic.
-