Do not schedule re-indexing of files if they haven’t been modified since an in-progress index was requested

ahoppen · ahoppen · commit d090fb070d14 · 2025-02-27T12:00:38.000-08:00
This fixes the following problem, which caused files to get indexed twice: You open a workspace, which schedules initial indexing. While we are waiting for the build graph generation from the build server, the build servers sends `buildTarget/didChange` notification to SourceKit-LSP (eg. because it has finished loading the build graph). This causes all files in the changed targets to be scheduled for re-indexing since new files might be present in the updated targets. And since we have already started indexing of those files, we assumed that we need to index them again because the contents might have changed.

To fix this, keep track of the file’s modification time at the time at which we scheduled indexing for it. If that time hasn’t changed and we have an in-progress indexing operation for it, there is no need to schedule another one.
diff --git a/Sources/SemanticIndex/CheckedIndex.swift b/Sources/SemanticIndex/CheckedIndex.swift
@@ -11,7 +11,7 @@
 //===----------------------------------------------------------------------===//
 
 #if compiler(>=6)
-import Foundation
+package import Foundation
 @preconcurrency package import IndexStoreDB
 package import LanguageServerProtocol
 import SKLogging
@@ -188,6 +188,20 @@ package final class CheckedIndex {
     return checker.fileHasInMemoryModifications(uri)
   }
 
+  /// Determine the modification date of the file at the given location or, if it is a symlink, the maximum modification
+  /// time of any hop in the symlink change until the real file.
+  ///
+  /// This uses the `CheckedIndex`'s  mod date cache, so it doesn't require disk access if the modification date of the
+  /// file has already been computed.
+  ///
+  /// Returns `nil` if the modification date of the file could not be determined.
+  package func modificationDate(of uri: DocumentURI) -> Date? {
+    switch try? checker.modificationDate(of: uri) {
+    case nil, .fileDoesNotExist: return nil
+    case .date(let date): return date
+    }
+  }
+
   /// If there are any definition occurrences of the given USR, return these.
   /// Otherwise return declaration occurrences.
   package func definitionOrDeclarationOccurrences(ofUSR usr: String) -> [SymbolOccurrence] {
@@ -327,7 +341,7 @@ private struct IndexOutOfDateChecker {
   private let checkLevel: IndexCheckLevel
 
   /// The last modification time of a file. Can also represent the fact that the file does not exist.
-  private enum ModificationTime {
+  enum ModificationTime {
     case fileDoesNotExist
     case date(Date)
   }
@@ -495,7 +509,7 @@ private struct IndexOutOfDateChecker {
     }
   }
 
-  private mutating func modificationDate(of uri: DocumentURI) throws -> ModificationTime {
+  mutating func modificationDate(of uri: DocumentURI) throws -> ModificationTime {
     if let cached = modTimeCache[uri] {
       return cached
     }
diff --git a/Sources/SemanticIndex/SemanticIndexManager.swift b/Sources/SemanticIndex/SemanticIndexManager.swift
@@ -52,42 +52,53 @@ private struct OpaqueQueuedIndexTask: Equatable {
   }
 }
 
-private enum InProgressIndexStore {
-  /// We know that we need to index the file but and are currently gathering all information to create the `indexTask`
-  /// that will index it.
-  ///
-  /// This is needed to avoid the following race: We request indexing of file A. Getting the canonical target for A
-  /// takes a bit and before that finishes, we request another index of A. In this case, we don't want to kick off
-  /// two tasks to update the index store.
-  case creatingIndexTask
+private struct InProgressIndexStore {
+  enum State {
+    /// We know that we need to index the file but and are currently gathering all information to create the `indexTask`
+    /// that will index it.
+    ///
+    /// This is needed to avoid the following race: We request indexing of file A. Getting the canonical target for A
+    /// takes a bit and before that finishes, we request another index of A. In this case, we don't want to kick off
+    /// two tasks to update the index store.
+    case creatingIndexTask
+
+    /// We are waiting for preparation of the file's target to be scheduled. The next step is that we wait for
+    /// preparation to finish before we can update the index store for this file.
+    ///
+    /// `preparationTaskID` identifies the preparation task so that we can transition a file's index state to
+    /// `updatingIndexStore` when its preparation task has finished.
+    ///
+    /// `indexTask` is a task that finishes after both preparation and index store update are done. Whoever owns the index
+    /// task is still the sole owner of it and responsible for its cancellation.
+    case waitingForPreparation(preparationTaskID: UUID, indexTask: Task<Void, Never>)
+
+    /// We have started preparing this file and are waiting for preparation to finish before we can update the index
+    /// store for this file.
+    ///
+    /// `preparationTaskID` identifies the preparation task so that we can transition a file's index state to
+    /// `updatingIndexStore` when its preparation task has finished.
+    ///
+    /// `indexTask` is a task that finishes after both preparation and index store update are done. Whoever owns the index
+    /// task is still the sole owner of it and responsible for its cancellation.
+    case preparing(preparationTaskID: UUID, indexTask: Task<Void, Never>)
+
+    /// The file's target has been prepared and we are updating the file's index store.
+    ///
+    /// `updateIndexStoreTask` is the task that updates the index store itself.
+    ///
+    /// `indexTask` is a task that finishes after both preparation and index store update are done. Whoever owns the index
+    /// task is still the sole owner of it and responsible for its cancellation.
+    case updatingIndexStore(updateIndexStoreTask: OpaqueQueuedIndexTask, indexTask: Task<Void, Never>)
+  }
 
-  /// We are waiting for preparation of the file's target to be scheduled. The next step is that we wait for
-  /// preparation to finish before we can update the index store for this file.
-  ///
-  /// `preparationTaskID` identifies the preparation task so that we can transition a file's index state to
-  /// `updatingIndexStore` when its preparation task has finished.
-  ///
-  /// `indexTask` is a task that finishes after both preparation and index store update are done. Whoever owns the index
-  /// task is still the sole owner of it and responsible for its cancellation.
-  case waitingForPreparation(preparationTaskID: UUID, indexTask: Task<Void, Never>)
+  var state: State
 
-  /// We have started preparing this file and are waiting for preparation to finish before we can update the index
-  /// store for this file.
-  ///
-  /// `preparationTaskID` identifies the preparation task so that we can transition a file's index state to
-  /// `updatingIndexStore` when its preparation task has finished.
+  /// The modification time of the time of `FileToIndex.sourceFile` at the time that indexing was scheduled. This allows
+  /// us to avoid scheduling another indexing operation for the file if the file hasn't been modified since an
+  /// in-progress indexing operation was scheduled.
   ///
-  /// `indexTask` is a task that finishes after both preparation and index store update are done. Whoever owns the index
-  /// task is still the sole owner of it and responsible for its cancellation.
-  case preparing(preparationTaskID: UUID, indexTask: Task<Void, Never>)
-
-  /// The file's target has been prepared and we are updating the file's index store.
-  ///
-  /// `updateIndexStoreTask` is the task that updates the index store itself.
-  ///
-  /// `indexTask` is a task that finishes after both preparation and index store update are done. Whoever owns the index
-  /// task is still the sole owner of it and responsible for its cancellation.
-  case updatingIndexStore(updateIndexStoreTask: OpaqueQueuedIndexTask, indexTask: Task<Void, Never>)
+  /// `nil` if the modification date of the file could not be determined.
+  var fileModificationDate: Date?
 }
 
 /// Status of document indexing / target preparation in `inProgressIndexAndPreparationTasks`.
@@ -229,8 +240,8 @@ package final actor SemanticIndexManager {
     let preparationTasks = inProgressPreparationTasks.mapValues { inProgressTask in
       return inProgressTask.task.isExecuting ? IndexTaskStatus.executing : IndexTaskStatus.scheduled
     }
-    let indexTasks = inProgressIndexTasks.mapValues { status in
-      switch status {
+    let indexTasks = inProgressIndexTasks.mapValues { inProgress in
+      switch inProgress.state {
       case .creatingIndexTask, .waitingForPreparation, .preparing:
         return IndexTaskStatus.scheduled
       case .updatingIndexStore(updateIndexStoreTask: let updateIndexStoreTask, indexTask: _):
@@ -340,8 +351,8 @@ package final actor SemanticIndexManager {
     await waitForBuildGraphGenerationTasks()
 
     await withTaskGroup(of: Void.self) { taskGroup in
-      for (_, status) in inProgressIndexTasks {
-        switch status {
+      for inProgress in inProgressIndexTasks.values {
+        switch inProgress.state {
         case .creatingIndexTask:
           break
         case .waitingForPreparation(preparationTaskID: _, indexTask: let indexTask),
@@ -424,7 +435,7 @@ package final actor SemanticIndexManager {
   private func filesToIndex(
     toCover files: some Collection<DocumentURI> & Sendable,
     indexFilesWithUpToDateUnits: Bool
-  ) async -> [FileToIndex] {
+  ) async -> [(file: FileToIndex, fileModificationDate: Date?)] {
     let sourceFiles = await orLog("Getting source files in project") {
       Set(try await buildSystemManager.sourceFiles(includeNonBuildableFiles: false).keys)
     }
@@ -433,6 +444,7 @@ package final actor SemanticIndexManager {
     }
     let deletedFilesIndex = index.checked(for: .deletedFiles)
     let modifiedFilesIndex = index.checked(for: .modifiedFiles)
+
     let filesToReIndex =
       await files
       .asyncFilter {
@@ -442,13 +454,13 @@ package final actor SemanticIndexManager {
           return false
         }
         return true
-      }.compactMap { (uri) -> FileToIndex? in
+      }.compactMap { uri -> (FileToIndex, Date?)? in
         if sourceFiles.contains(uri) {
           if !indexFilesWithUpToDateUnits, modifiedFilesIndex.hasUpToDateUnit(for: uri) {
             return nil
           }
           // If this is a source file, just index it.
-          return .indexableFile(uri)
+          return (.indexableFile(uri), modifiedFilesIndex.modificationDate(of: uri))
         }
         // Otherwise, see if it is a header file. If so, index a main file that that imports it to update header file's
         // index.
@@ -465,15 +477,7 @@ package final actor SemanticIndexManager {
         if !indexFilesWithUpToDateUnits, modifiedFilesIndex.hasUpToDateUnit(for: uri, mainFile: mainFile) {
           return nil
         }
-        return .headerFile(header: uri, mainFile: mainFile)
-      }
-      .filter {
-        switch inProgressIndexTasks[$0] {
-        case .waitingForPreparation, .creatingIndexTask:
-          return false
-        default:
-          return true
-        }
+        return (.headerFile(header: uri, mainFile: mainFile), modifiedFilesIndex.modificationDate(of: uri))
       }
     return filesToReIndex
   }
@@ -651,7 +655,7 @@ package final actor SemanticIndexManager {
         return
       }
       for fileAndTarget in filesAndTargets {
-        switch self.inProgressIndexTasks[fileAndTarget.file] {
+        switch self.inProgressIndexTasks[fileAndTarget.file]?.state {
         case .updatingIndexStore(let registeredTask, _):
           if registeredTask == OpaqueQueuedIndexTask(task) {
             self.inProgressIndexTasks[fileAndTarget.file] = nil
@@ -667,9 +671,9 @@ package final actor SemanticIndexManager {
       self.indexProgressStatusDidChange()
     }
     for fileAndTarget in filesAndTargets {
-      switch inProgressIndexTasks[fileAndTarget.file] {
+      switch inProgressIndexTasks[fileAndTarget.file]?.state {
       case .waitingForPreparation(preparationTaskID, let indexTask), .preparing(preparationTaskID, let indexTask):
-        inProgressIndexTasks[fileAndTarget.file] = .updatingIndexStore(
+        inProgressIndexTasks[fileAndTarget.file]?.state = .updatingIndexStore(
           updateIndexStoreTask: OpaqueQueuedIndexTask(updateIndexTask),
           indexTask: indexTask
         )
@@ -694,34 +698,62 @@ package final actor SemanticIndexManager {
     // We will check the up-to-date status again in `IndexTaskDescription.execute`. This ensures that if we schedule
     // schedule two indexing jobs for the same file in quick succession, only the first one actually updates the index
     // store and the second one will be a no-op once it runs.
-    let outOfDateFiles = await filesToIndex(toCover: files, indexFilesWithUpToDateUnits: indexFilesWithUpToDateUnit)
+    var filesToIndex = await filesToIndex(toCover: files, indexFilesWithUpToDateUnits: indexFilesWithUpToDateUnit)
       // sort files to get deterministic indexing order
-      .sorted(by: { $0.sourceFile.stringValue < $1.sourceFile.stringValue })
-
-    if outOfDateFiles.isEmpty {
-      // Early exit if there are no files to index.
-      return Task {}
-    }
-
-    logger.debug("Scheduling indexing of \(outOfDateFiles.map(\.sourceFile.stringValue).joined(separator: ", "))")
-
-    // Sort the targets in topological order so that low-level targets get built before high-level targets, allowing us
-    // to index the low-level targets ASAP.
-    var filesByTarget: [BuildTargetIdentifier: [FileToIndex]] = [:]
+      .sorted(by: { $0.file.sourceFile.stringValue < $1.file.sourceFile.stringValue })
 
     // The number of index tasks that don't currently have an in-progress task associated with it.
     // The denominator in the index progress should get incremented by this amount.
     // We don't want to increment the denominator for tasks that already have an index in progress.
     var newIndexTasks = 0
-    for file in outOfDateFiles {
-      if inProgressIndexTasks[file] == nil {
+    var alreadyScheduledTasks: Set<FileToIndex> = []
+    for file in filesToIndex {
+      let inProgress = inProgressIndexTasks[file.file]
+
+      let shouldScheduleIndexing: Bool
+      switch inProgress?.state {
+      case nil:
         newIndexTasks += 1
+        shouldScheduleIndexing = true
+      case .creatingIndexTask, .waitingForPreparation:
+        // We already have a task that indexes the file but hasn't started preparation yet. Indexing the file again
+        // won't produce any new results.
+        alreadyScheduledTasks.insert(file.file)
+        shouldScheduleIndexing = false
+      case .preparing(_, _), .updatingIndexStore(_, _):
+        // We have started indexing of the file and are now requesting to index it again. Unless we know that the file
+        // hasn't been modified since the last request for indexing, we need to schedule it to get re-indexed again.
+        if let modDate = file.fileModificationDate, inProgress?.fileModificationDate == modDate {
+          shouldScheduleIndexing = false
+        } else {
+          shouldScheduleIndexing = true
+        }
+      }
+      if shouldScheduleIndexing {
+        inProgressIndexTasks[file.file] = InProgressIndexStore(
+          state: .creatingIndexTask,
+          fileModificationDate: file.fileModificationDate
+        )
+      } else {
+        alreadyScheduledTasks.insert(file.file)
+
       }
-      inProgressIndexTasks[file] = .creatingIndexTask
     }
     indexTasksWereScheduled(newIndexTasks)
+    filesToIndex = filesToIndex.filter { !alreadyScheduledTasks.contains($0.file) }
+
+    if filesToIndex.isEmpty {
+      // Early exit if there are no files to index.
+      return Task {}
+    }
+
+    logger.debug("Scheduling indexing of \(filesToIndex.map(\.file.sourceFile.stringValue).joined(separator: ", "))")
+
+    // Sort the targets in topological order so that low-level targets get built before high-level targets, allowing us
+    // to index the low-level targets ASAP.
+    var filesByTarget: [BuildTargetIdentifier: [(FileToIndex)]] = [:]
 
-    for fileToIndex in outOfDateFiles {
+    for fileToIndex in filesToIndex.map(\.file) {
       guard let target = await buildSystemManager.canonicalTarget(for: fileToIndex.mainFile) else {
         logger.error(
           "Not indexing \(fileToIndex.forLogging) because the target could not be determined"
@@ -768,9 +800,9 @@ package final actor SemanticIndexManager {
           if case .executing = newState {
             for file in filesToIndex {
               if case .waitingForPreparation(preparationTaskID: preparationTaskID, indexTask: let indexTask) =
-                self.inProgressIndexTasks[file]
+                self.inProgressIndexTasks[file]?.state
               {
-                self.inProgressIndexTasks[file] = .preparing(
+                self.inProgressIndexTasks[file]?.state = .preparing(
                   preparationTaskID: preparationTaskID,
                   indexTask: indexTask
                 )
@@ -807,7 +839,7 @@ package final actor SemanticIndexManager {
         // `.waitingForPreparation` here  because we don't have an `await` call between the creation of `indexTask` and
         // this loop, so we still have exclusive access to the `SemanticIndexManager` actor and hence `updateIndexStore`
         // can't execute until we have set all index statuses to `.waitingForPreparation`.
-        inProgressIndexTasks[file] = .waitingForPreparation(
+        inProgressIndexTasks[file]?.state = .waitingForPreparation(
           preparationTaskID: preparationTaskID,
           indexTask: indexTask
         )
