chore(tests): switch to vitest MCP-65 #363

gagik · 2025-07-14T17:59:55Z

This is a proposal to switch to vitest in light of continued issues with ESM modules and jest. Jest support for ESM modules is still experimental and it will likely require constant patching to keep it working as we end up adopting more ESM-based modules. This also applies to extra TypeScript processing.

vitest is tried and tested now and is essentially a drop-in replacement for jest (minus our minor usage of jest-extended, which is replaceable).

Pros of vitest

Native TypeScript, ESM module support means little to no configuration and extra work required compared to configurating jest or chai/mocha/sinon.
All in one package compared to chai/mocha/sinon.
It has better typed helper methods such as vi.mocked.
It has an official VSCode extension that just works without any extra setup 💘
There's potential to add type testing which is quite cool.
Its performance is similar, maybe even a bit faster? vitest vs jest.

Cons of vitest

As it is a more recent project, there's not much usage of it in the mongodb-js organization outside of small projects. I do believe that this is largely because of older age of most other codebases. That said, it is being adopted in other places: vitest is used by the https://github.com/mongodb/chatbot for example.

Possible alternatives:

Configuring jest: Probably would take even more effort than this switching to vitest right now and we may run into future issues as its support is experimental.
Using chai/mocha/sinon: We may have better time with this and would align with other mongodb-js codebases if that matters. Would likely again require a lot more extra setup than the built-in vitest work but be less experimental than jest.
Give up on ESM: this may be a good argument in light of VSCode only supporting CommonJS right now. However, I think it's odd to downgrade to an older standard when it's clear from perspective of future of ECMAScript that ESM is what new libraries should use. I found this thread useful in learning more about ESM pros.

In my opinion minimal configuration is preferable and vitest is mature enough and well worth being adopted.

Copilot

Pull Request Overview

This PR replaces Jest with Vitest as the test runner to simplify ESM/TypeScript support and reduce configuration overhead.

Adds a vitest.config.ts for Vitest setup
Updates TS config and package dependencies to reference Vitest types and remove Jest
Rewrites all tests to use Vitest APIs (vi, expect, describe, etc.)

Reviewed Changes

Copilot reviewed 43 out of 46 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
vitest.config.ts	New Vitest configuration
tsconfig.jest.json	Switched `"types"` from Jest to Vitest
package.json	Removed Jest deps, added `vitest`
jest.config.cjs	Deleted (old Jest config)
eslint.config.js	Updated ESLint to include `vitest.config.ts`
tests/unit/*/.test.ts, tests/integration/**	Replaced Jest imports, mocks, and assertions with Vitest

tests/integration/helpers.ts

tests/integration/tools/mongodb/metadata/listDatabases.test.ts

package.json

tsconfig.jest.json

Copilot · 2025-07-14T18:01:40Z

tests/unit/telemetry.test.ts

 import logger, { LogId } from "../../src/common/logger.js";
 import { createHmac } from "crypto";
+import type { MockedFunction } from "vitest";


[nitpick] The manual as unknown as typeof mockApiClient casting is complex. You can simplify typing by using const mockApiClient = vi.mocked(ApiClient, { shallow: true }); to get strongly typed mocks without workarounds.

coveralls · 2025-07-14T18:50:46Z

Pull Request Test Coverage Report for Build 16275099623

Details

0 of 26 (0.0%) changed or added relevant lines in 2 files are covered.
291 unchanged lines in 26 files lost coverage.
Overall coverage increased (+2.3%) to 77.524%

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
eslint.config.js	0	13	0.0%
vitest.config.ts	0	13	0.0%

Files with Coverage Reduction	New Missed Lines	%
src/helpers/EJsonTransport.ts	1	86.11%
src/helpers/indexCheck.ts	1	96.0%
src/tools/atlas/create/createAccessList.ts	1	90.54%
src/tools/atlas/read/inspectAccessList.ts	1	90.0%
src/tools/atlas/read/listOrgs.ts	1	88.89%
src/tools/mongodb/metadata/collectionStorageSize.ts	1	95.65%
src/common/atlas/apiClientError.ts	2	85.9%
src/helpers/container.ts	2	73.17%
src/telemetry/telemetry.ts	2	91.85%
src/telemetry/eventCache.ts	3	80.56%

Totals
Change from base Build 16220557964:	2.3%
Covered Lines:	2833
Relevant Lines:	3610

💛 - Coveralls

gagik · 2025-07-14T19:01:46Z

tests/integration/tools/mongodb/create/createIndex.test.ts

@@ -8,6 +8,7 @@ import {
    expectDefined,
 } from "../../../helpers.js";
 import { IndexDirection } from "mongodb";
+import { expect, it } from "vitest";


these imports can actually be avoided by opting in to use globals which would make this PR look less scary. However I'm not a fan of globals in general as they're a bit too magical and hide real dependencies.

I think using globals in this particular situation is kind of fine, but not a strong opinion.

yeah another thing with globals was that every time it required eslint/typescript/etc. configuration to make it apply to some files but not others so this ended up being quicker

nirinchev

Looks reasonable - no strong feelings on vi vs the rest. Looks like the coverage reporter is slightly different though and will now report 2 lines for each missed line (now seems to be counting the closing curly brace). Probably not a blocking issue.

tests/integration/helpers.ts

nirinchev · 2025-07-14T21:19:45Z

tests/integration/tools/mongodb/create/createIndex.test.ts

@@ -8,6 +8,7 @@ import {
    expectDefined,
 } from "../../../helpers.js";
 import { IndexDirection } from "mongodb";
+import { expect, it } from "vitest";


I think using globals in this particular situation is kind of fine, but not a strong opinion.

tests/integration/tools/mongodb/create/insertMany.test.ts

nirinchev · 2025-07-14T21:22:41Z

tests/integration/tools/mongodb/metadata/listCollections.test.ts

            expect(items2).toHaveLength(2);
-            expect(items2.map((item) => item.text)).toIncludeSameMembers([
-                'Name: "collection-1"',
-                'Name: "collection-2"',
-            ]);
+            expect(items2.map((item) => item.text)).toEqual(
+                expect.arrayContaining(['Name: "collection-1"', 'Name: "collection-2"'])
+            );


Can this be .strictEqual(...)?

a lot of these toIncludeSameMembers tests are not order-sensitive but toStrictEqual is so generally no

Ah, true. Kind of annoying there's no built-in matcher for arrays ignoring the order.

we can add custom matcher; honestly will be great to use that kind of stuff more, will add it now

…gagik/vitest

gagik · 2025-07-15T11:07:58Z

tests/integration/tools/mongodb/read/aggregate.test.ts

@@ -20,7 +21,7 @@ describeWithMongoDB("aggregate tool", (integration) => {
    validateThrowsForInvalidArguments(integration, "aggregate", [
        {},
        { database: "test", collection: "foo" },
-        { database: test, pipeline: [] },


this is actually best argument against globals, we had an accidental usage of jest's test global here.... I'm actually surprised the test was passing...

nirinchev

LGTM once eslint is happy. Also, if there's an easy fix for the coverage reports being akward, we should apply it, otherwise we'll have to live with it

chore(tests): switch to vitest

6eed483

Copilot AI review requested due to automatic review settings July 14, 2025 17:59

gagik requested a review from a team as a code owner July 14, 2025 17:59

Copilot AI reviewed Jul 14, 2025

View reviewed changes

fix: remove traces of jest

463b6cc

gagik force-pushed the gagik/vitest branch from 8f67434 to 463b6cc Compare July 14, 2025 18:15

gagik added 2 commits July 14, 2025 20:25

fix: add and fix vitest linting

3b03327

fix: coverage

2d63956

gagik force-pushed the gagik/vitest branch from 777997e to 2d63956 Compare July 14, 2025 18:33

gagik added 2 commits July 14, 2025 20:36

fix: use --exclude

6e24d64

fix: increase hook timeout

3f10cbd

gagik commented Jul 14, 2025

View reviewed changes

nirinchev reviewed Jul 14, 2025

View reviewed changes

fix: from feedback

52ccf24

gagik force-pushed the gagik/vitest branch from f950f7a to aea600a Compare July 15, 2025 09:45

feat: use custom toIncludeSameMembers matcher

cbebf3e

gagik force-pushed the gagik/vitest branch from aea600a to cbebf3e Compare July 15, 2025 09:46

gagik added 2 commits July 15, 2025 11:47

Merge branch 'main' of github.com:mongodb-js/mongodb-mcp-server into …

8fc12bc

…gagik/vitest

fix: remove unneeded globals

3bf00e6

gagik commented Jul 15, 2025

View reviewed changes

fix: remove unneeded syntax

5c4fe85

nirinchev approved these changes Jul 15, 2025

View reviewed changes

fix: re-add message

b254655

gagik enabled auto-merge (squash) July 15, 2025 12:09

gagik merged commit b12db06 into main Jul 15, 2025
17 checks passed

gagik deleted the gagik/vitest branch July 15, 2025 12:11

gagik changed the title ~~chore(tests): switch to vitest~~ chore(tests): switch to vitest MCP-65 Jul 15, 2025

fmenezes mentioned this pull request Jul 15, 2025

chore: fix local run and vulnerabilities #360

Open

1 task

chore(tests): switch to vitest MCP-65 #363

chore(tests): switch to vitest MCP-65 #363

Uh oh!

Conversation

gagik commented Jul 14, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Pros of vitest

Cons of vitest

Possible alternatives:

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull Request Overview

Reviewed Changes

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Copilot AI Jul 14, 2025

Choose a reason for hiding this comment

Uh oh!

coveralls commented Jul 14, 2025

Pull Request Test Coverage Report for Build 16275099623

Details

💛 - Coveralls

Uh oh!

gagik Jul 14, 2025

Choose a reason for hiding this comment

Uh oh!

nirinchev Jul 14, 2025

Choose a reason for hiding this comment

Uh oh!

gagik Jul 15, 2025

Choose a reason for hiding this comment

Uh oh!

nirinchev left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

nirinchev Jul 14, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

nirinchev Jul 14, 2025

Choose a reason for hiding this comment

Uh oh!

gagik Jul 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nirinchev Jul 15, 2025

Choose a reason for hiding this comment

Uh oh!

gagik Jul 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

gagik Jul 15, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

nirinchev left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

gagik commented Jul 14, 2025 •

edited

Loading

gagik Jul 15, 2025 •

edited

Loading

gagik Jul 15, 2025 •

edited

Loading

gagik Jul 15, 2025 •

edited

Loading