fix: prevent stale downloads from reviving removed repositories

public/i18n/locale-en.json

@@ -1,5 +1,6 @@
 {
   "ERRORS": {
+    "repository_job_cancelled": "The repository changed or was removed while processing.",
     "unknown_error": "Unknown error, contact the admin.",
     "unreachable": "Anonymous GitHub is unreachable, contact the admin.",
     "request_error": "Unable to download the file, check your connection or contact the admin.",

src/core/Repository.ts

@@ -470,22 +470,30 @@ export default class Repository {
    * @param status the new status
    * @param errorMessage a potential error message to display
    */
+  public protectLifecycle = false;
+
   async updateStatus(status: RepositoryStatus, statusMessage?: string) {
     if (!status) return this.model;
-    this._model.status = status;
-    this._model.statusDate = new Date();
-    this._model.statusMessage = statusMessage;
-    if (!isConnected) return this.model;
-    await AnonymizedRepositoryModel.updateOne(
-      { _id: this._model._id },
-      {
-        $set: {
-          status,
-          statusDate: this._model.statusDate,
-          statusMessage,
+    const statusDate = new Date();
+    if (isConnected) {
+      const result = await AnonymizedRepositoryModel.updateOne(
+        {
+          _id: this._model._id,
+          ...(this.protectLifecycle ? {
+            status: { $nin: [RepositoryStatus.REMOVING, RepositoryStatus.REMOVED,
+              RepositoryStatus.EXPIRING, RepositoryStatus.EXPIRED] },
+            anonymizeDate: this._model.anonymizeDate,
+          } : {}),
         },
+        { $set: { status, statusDate, statusMessage } }
+      ).exec();
+      if (this.protectLifecycle && result.matchedCount === 0) {
+        throw new AnonymousError("repository_job_cancelled", { httpStatus: 410 });
       }
-    ).exec();
+    }
+    this._model.status = status;
+    this._model.statusDate = statusDate;
+    this._model.statusMessage = statusMessage;
   }
 
   /**

src/queue/processes/downloadRepository.ts

@@ -24,6 +24,9 @@ export default async function (job: SandboxedJob<RepoJobData, void>) {
   await connect();
 
   const repo = await getRepository(job.data.repoId);
+  if ([RepositoryStatus.REMOVING, RepositoryStatus.REMOVED,
+    RepositoryStatus.EXPIRING, RepositoryStatus.EXPIRED].some((status) => status === repo.status)) return;
+  repo.protectLifecycle = true;
   const token = await getToken(repo);
   const tokenKey = token.slice(-8);
 
@@ -88,6 +91,7 @@ export default async function (job: SandboxedJob<RepoJobData, void>) {
       logger.info("downloaded", { repoId: job.data.repoId });
     } catch (error) {
       clearInterval(statusInterval);
+      if (error instanceof Error && error.message === "repository_job_cancelled") return;
       if (tickPromise) await tickPromise;
 
       // Rate-limited: delay the job and free the worker slot
@@ -123,6 +127,7 @@ export default async function (job: SandboxedJob<RepoJobData, void>) {
     }
   } catch (error: unknown) {
     clearInterval(statusInterval);
+    if (error instanceof Error && error.message === "repository_job_cancelled") return;
     if (tickPromise) {
       try {
         await tickPromise;

test/production-regressions.test.js

@@ -55,6 +55,25 @@ describe("production regressions", function () {
     expect(emitted).to.equal(false);
   });
 
+  for (const status of ["removing", "removed", "expiring", "expired"]) {
+    it(`ignores delayed downloads for ${status} repositories`, async function () {
+      stub(db, "connect", async () => {});
+      stub(db, "getRepository", async () => ({ status }));
+      stub(gh, "getToken", async () => { throw new Error("must not download"); });
+      await require("../src/queue/processes/downloadRepository").default({ data: { repoId: "repo" } });
+    });
+    it(`does not let an active worker overwrite concurrent ${status}`, async function () {
+      stub(db, "isConnected", true);
+      const model = new RepoModel({ repoId: "repo", status: "download", anonymizeDate: new Date(1000) });
+      const repo = new Repository(model); repo.protectLifecycle = true;
+      stub(RepoModel, "updateOne", filter => ({ exec: async () => ({ matchedCount:
+        require("sift").default(filter)({ _id: model._id, anonymizeDate: model.anonymizeDate, status }) ? 1 : 0 }) }));
+      try { await repo.updateStatus("ready"); throw new Error("expected rejection"); }
+      catch (error) { expect(error.message).to.equal("repository_job_cancelled"); }
+      expect(repo.status).to.equal("download");
+    });
+  }
+
   it("denies another GitHub identity access through a reused coauthor username", async function () {
     const User = require("../src/core/User").default;
     const UserModel = require("../src/core/model/users/users.model").default;