|
// Licensed to the .NET Foundation under one or more agreements.
// The .NET Foundation licenses this file to you under the MIT license.
// See the LICENSE file in the project root for more information.
using System.Threading;
using System.Threading.Tasks;
using Microsoft.CodeAnalysis.Storage;
namespace Microsoft.CodeAnalysis.Classification;
internal static class SemanticClassificationCacheUtilities
{
public static async Task<(DocumentKey documentKey, Checksum checksum)> GetDocumentKeyAndChecksumAsync(
Document document, CancellationToken cancellationToken)
{
var project = document.Project;
// We very intentionally persist this information against using a null 'parseOptionsChecksum'. This way the
// results will be valid and something we can lookup regardless of the project configuration. In other
// words, if we've cached the information when in the DEBUG state of the project, but we lookup when in the
// RELEASE state, we'll still find the entry. The data may be inaccurate, but that's ok as this is just for
// temporary classifying until the real classifier takes over when the solution fully loads.
var projectKey = new ProjectKey(SolutionKey.ToSolutionKey(project.Solution), project.Id, project.FilePath, project.Name, Checksum.Null);
var documentKey = new DocumentKey(projectKey, document.Id, document.FilePath, document.Name);
// We only checksum off of the contents of the file. During load, we can't really compute any other
// information since we don't necessarily know about other files, metadata, or dependencies. So during
// load, we allow for the previous semantic classifications to be used as long as the file contents match.
var checksums = await document.State.GetStateChecksumsAsync(cancellationToken).ConfigureAwait(false);
return (documentKey, checksums.Text);
}
}
|