Convert UTF-8 string literal analyzer/codefix to refactoring #61593
Conversation
| var node = (SyntaxNode?)await context.TryGetRelevantNodeAsync<ArrayCreationExpressionSyntax>().ConfigureAwait(false) ?? | ||
| await context.TryGetRelevantNodeAsync<ImplicitArrayCreationExpressionSyntax>().ConfigureAwait(false); |
There was a problem hiding this comment.
This is new. Is it better to not use the helpers? Or use the helpers to get ExpressionSyntax and then manually compare to these two array expression types?
| return; | ||
|
|
||
| var model = await document.GetRequiredSemanticModelAsync(cancellationToken).ConfigureAwait(false); | ||
| if (model.Compilation.GetTypeByMetadataName(typeof(ReadOnlySpan<>).FullName!) is null) |
There was a problem hiding this comment.
This is new because it wasn't relevant before the language change
| return SyntaxFactory.InvocationExpression( | ||
| SyntaxFactory.MemberAccessExpression( | ||
| SyntaxKind.SimpleMemberAccessExpression, | ||
| SyntaxFactory.LiteralExpression(SyntaxKind.UTF8StringLiteralExpression, literal), | ||
| SyntaxFactory.IdentifierName(nameof(ReadOnlySpan<byte>.ToArray)))); |
There was a problem hiding this comment.
This is new, to add the ToArray call
| var root = await document.GetRequiredSyntaxRootAsync(cancellationToken).ConfigureAwait(false); | ||
| var editor = new SyntaxEditor(root, document.Project.Solution.Workspace.Services); | ||
|
|
||
| editor.ReplaceNode(node, CreateUTF8StringLiteralToArrayInvocation(stringValue).WithTriviaFrom(node)); |
There was a problem hiding this comment.
This isn't really new, but removing the parameter array stuff simplified this bit, and removed 2 or 3 helper methods, and made trivia handling much simpler
|
Will look later today. Did you implement 'fix all' support here? Thanks! |
No, I assumed that was being done case-by-case basis for selected refactorings only. |
You are correct :) But it seems here to be useful. I could see a file with dozens of byte[]s that a user wants to bulk convert over :) Fix-All should be fairly simple to add. Would you be willing to do it here? |
|
Superseded by #61650 |
Due to a language spec change, where UTF-8 strings are
ReadOnlySpan<byte>rather thenbyte[], it no longer makes sense for there to be a code style analyzer to offer UTF-8 string literals, so this PR removes the old analyzer and code fix (added in #60647) and instead creates a refactoring that convertsnew byte[] { 65, 66, 67 }to"ABC"u8.ToArray(). The analyzer used to offer to convert parameter arrays too, but this is removed as it seemed a bit weird. Can always add it back later if we get feedback.Fixes #61517
The PR looks a lot bigger than it is, most of the work was just moving code around, and simplifying.