feat: add Playwright E2E test infrastructure#533
feat: add Playwright E2E test infrastructure#533DSanich wants to merge 10 commits intogetAlby:masterfrom
Conversation
…d oauth error handling tests
📝 WalkthroughWalkthroughThis PR introduces comprehensive end-to-end testing infrastructure using Playwright. It adds test fixtures, test suites covering LNClient, NWC, OAuth, WebLN, and amount resolution modules, Playwright configuration, CI integration, and supporting documentation and dependencies. Changes
Estimated code review effort🎯 3 (Moderate) | ⏱️ ~25 minutes Poem
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. ✨ Finishing touches
🧪 Generate unit tests (beta)
Tip Issue Planner is now in beta. Read the docs and try it out! Share your feedback on Discord. Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
|
Review the following changes in direct dependencies. Learn more about Socket for GitHub.
|
|
Warning Review the following alerts detected in dependencies. According to your organization's Security Policy, it is recommended to resolve "Warn" alerts. Learn more about Socket for GitHub.
|
|
@rolznz @reneaaron Hello! Following this discussion getAlby/sandbox#15 (comment), I started working on e2e tests at https://github.com/getAlby/js-sdk. I'm happy to share this pull request with you. The library's core modules are now covered with real e2e tests, ensuring greater security for business logic. |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 1
🧹 Nitpick comments (13)
e2e/fixtures/lnclient.html (1)
10-10: UnusedLNimport.
LNis imported but never used in this fixture.Proposed fix
- import { LNClient, LN } from "/dist/esm/lnclient.js"; + import { LNClient } from "/dist/esm/lnclient.js";🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@e2e/fixtures/lnclient.html` at line 10, The import brings in LN but it is unused; remove the unused symbol by changing the import that currently lists both LNClient and LN to import only LNClient (i.e., update the import statement that references LNClient and LN so it no longer includes LN) to eliminate the unused import and any linter warnings.package.json (1)
91-92:devDependenciesalphabetical ordering.
serve(line 91) is listed beforeqrcode-terminal(line 92), breaking the alphabetical sort of devDependencies.Proposed fix
- "serve": "^14.2.4", "qrcode-terminal": "^0.12.0", + "serve": "^14.2.4",🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@package.json` around lines 91 - 92, The devDependencies list is out of alphabetical order: swap the positions of the "serve" and "qrcode-terminal" entries so that "qrcode-terminal" comes before "serve" (maintain existing version strings and trailing commas), ensuring the devDependencies array is sorted alphabetically by package name.e2e/fixtures/oauth.html (1)
10-10: UnusedOAuth2Bearerimport.
OAuth2Beareris imported but never used in this fixture. Remove it to keep the fixture clean.Proposed fix
- import { Client, OAuth2Bearer } from "/dist/esm/oauth.js"; + import { Client } from "/dist/esm/oauth.js";🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@e2e/fixtures/oauth.html` at line 10, The import statement currently brings in an unused symbol — remove OAuth2Bearer from the import so only the used Client is imported; update the import that reads something like import { Client, OAuth2Bearer } ... to import { Client } ... and ensure no other references to OAuth2Bearer remain in the file..github/workflows/ci.yml (1)
20-25: Remove redundantyarn buildin CI, as E2E tests trigger it via Playwright's webServer.Line 20 runs
yarn build, then line 25 runsyarn test:e2e, which executesplaywright test --project=chromium. Playwright's webServer configuration (line 20 ofplaywright.config.ts) runsyarn build && serve . -p 3000, so the build executes twice unnecessarily. Since the E2E tests depend on the built output being served, the explicit build at line 20 can be removed (if unit tests don't require it) or the webServer command can be optimized to skip the redundant build using an environment variable or a CI-specific script.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In @.github/workflows/ci.yml around lines 20 - 25, Remove the redundant top-level "yarn build" CI step because Playwright's webServer (configured in playwright.config.ts and invoked by the "yarn test:e2e" / "playwright test --project=chromium") already runs "yarn build && serve . -p 3000"; update the CI workflow by deleting the "yarn build" run step (or alternatively modify the webServer command to skip the build when running in CI via an environment variable or a CI-specific script) so the build is not executed twice.e2e/webln.spec.ts (1)
60-68: Extract repeatedmockAuthobject to reduce duplication.The identical
mockAuthobject is constructed in every test. Extract it to a shared constant or a helper inside thedescribeblock:♻️ Suggested refactor
test.describe("webln", () => { + const mockAuth = { + token: { access_token: "x" }, + getAuthHeader: () => ({ Authorization: "Bearer x" }), + }; + test.beforeEach(async ({ page }) => {Then pass
mockAuthinto eachpage.evaluatevia the second argument, or serialize it as needed.Note: since
getAuthHeaderis a function, it can't be serialized bypage.evaluate. You could instead inline it in a shared snippet string or keep the object literal insideevaluatebut extract just the repeated structure into a helper that returns the evaluate callback.Also applies to: 76-84, 90-98, 105-114, 120-132, 139-152
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@e2e/webln.spec.ts` around lines 60 - 68, Extract the repeated mockAuth construction used inside multiple page.evaluate calls into a shared helper or constant within the describe block to avoid duplication; for example, create a getMockAuth() helper (or a stringified snippet) and use it when constructing the provider in tests that reference OauthWeblnProvider, but remember functions like getAuthHeader can't be serialized into page.evaluate so either pass a serializable representation or use the helper to return the full evaluate callback (or inline the function inside the evaluate body) and update all places that construct mockAuth (lines creating mockAuth for OauthWeblnProvider) to use that helper.e2e/fiat-amount.spec.ts (2)
5-11: EUR test reuses USD mock data — not truly testing EUR-specific conversion.The
mockRatesResponsehascode: "USD", but it's also served for the EUR rate request (theRATES_APIregex matches all currency paths). The EUR test passes only because the mock rate is identical. Consider adding a separate EUR mock (e.g., with a differentrate_float) and routing based on the URL path to verify the currency-specific calculation path.Also applies to: 52-61
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@e2e/fiat-amount.spec.ts` around lines 5 - 11, The USD mock (mockRatesResponse) is reused for the EUR test so the EUR-specific conversion isn't actually validated; create a separate mock object (e.g., eurMockRatesResponse with code: "EUR" and a different rate_float/rate_cents) and update the request stub that matches RATES_API to inspect the request URL/pathname (or currency segment) and return mockRatesResponse for USD paths and eurMockRatesResponse for EUR paths; update any stubs/expectations referenced around the existing mock usage (including the handler logic used by the EUR test) so each currency test receives its corresponding mock response.
27-50: The first two tests (USD converts…andFiatAmount is interoperable…) are near-duplicates.Both create
USD(1)and callresolveAmount. The first asserts> 0and themillisatrelationship; the second asserts the exact value1000. These could be a single test asserting exact values + the millisat relationship, reducing redundancy.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@e2e/fiat-amount.spec.ts` around lines 27 - 50, Duplicate tests: both tests create USD(1) and call resolveAmount; collapse them into one. Replace the two tests ("USD converts to satoshis via resolveAmount" and "FiatAmount is interoperable with Amount") with a single test that imports USD and resolveAmount, calls const fiatAmount = USD(1); const resolved = await resolveAmount(fiatAmount); then assert resolved matches the shape (satoshi and millisat are Numbers), assert resolved.satoshi === 1000, and assert resolved.millisat === resolved.satoshi * 1000; remove the redundant test to avoid duplication.e2e/oauth.spec.ts (1)
175-199: Minor: Double page navigation in custombase_urltest.The
beforeEachalready navigates to the fixture page (line 82-83), then this test navigates again (line 187-188) after adding the custom route. The first navigation is wasted work. Consider moving this test to its owndescribeblock without the sharedbeforeEach, or skipping the redundant navigation.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@e2e/oauth.spec.ts` around lines 175 - 199, The test "Client uses custom base_url" contains a redundant page.goto because the shared beforeEach already navigates to the fixture; remove the extra navigation inside that test or move the test into its own describe block without the shared beforeEach. Specifically, in the test named Client uses custom base_url remove the await page.goto("/e2e/fixtures/oauth.html") (or relocate the whole test into a new describe that does not use the existing beforeEach) so the custom route is added then the already-navigated page is used to evaluate the Client and call accountBalance.playwright.config.ts (2)
19-24:serve .exposes the entire project root.The web server command serves the repository root, making all files (including source, configs,
node_modules, etc.) accessible on port 3000. This is fine for local dev and CI isolation, but worth noting. If you'd prefer a tighter surface, you could serve only thedistande2edirectories (e.g., by creating a small serve config or symlinking into a dedicated folder).🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@playwright.config.ts` around lines 19 - 24, The webServer configuration in playwright.config.ts currently uses the command string "yarn build && serve . -p 3000" which exposes the entire repo root; change the webServer.command to serve only the built output (e.g., "yarn build && serve ./dist -p 3000" or a dedicated test-serve folder containing dist and e2e assets) or create a small serve config/symlink to a dedicated folder and point webServer.command at that folder, keeping the existing webServer.url, reuseExistingServer, and timeout settings unchanged; update any scripts that rely on the old path accordingly.
14-18: Three browser projects may significantly increase CI time withworkers: 1.Running Chromium, Firefox, and WebKit sequentially in CI could triple test execution time. Consider running only Chromium in CI (matching the PR description's "uses Chromium by default") and reserving cross-browser testing for periodic/nightly runs, or increase the worker count.
♻️ Example: CI-only single-browser config
- projects: [ - { name: "chromium", use: { ...devices["Desktop Chrome"] } }, - { name: "firefox", use: { ...devices["Desktop Firefox"] } }, - { name: "webkit", use: { ...devices["Desktop Safari"] } }, - ], + projects: process.env.CI + ? [{ name: "chromium", use: { ...devices["Desktop Chrome"] } }] + : [ + { name: "chromium", use: { ...devices["Desktop Chrome"] } }, + { name: "firefox", use: { ...devices["Desktop Firefox"] } }, + { name: "webkit", use: { ...devices["Desktop Safari"] } }, + ],🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@playwright.config.ts` around lines 14 - 18, The Playwright config currently runs three browser projects (projects array with names "chromium", "firefox", "webkit") which will slow CI when workers: 1; update playwright.config.ts to only include the "chromium" project for CI or make the projects list conditional (e.g., check process.env.CI) so CI uses only Chromium while local/dev runs include all three, or alternatively increase the global "workers" value for CI; modify the projects array or add the CI-branch logic around it and ensure any references to "projects" and "workers" are adjusted accordingly.e2e/fixtures/nwc.html (1)
12-13: Test NWC URL is duplicated across multiple files.The same
exampleNwcUrlstring is hardcoded here, ine2e/nwc.spec.ts(line 3-4), and ine2e/lnclient.spec.ts(line 3-4). Consider extracting it to a shared test constants file (e.g.,e2e/fixtures/constants.ts) to keep it DRY and simplify future updates.🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@e2e/fixtures/nwc.html` around lines 12 - 13, The string assigned to exampleNwcUrl is duplicated across tests; extract it into a single exported constant (e.g., export const EXAMPLE_NWC_URL) in a new shared test constants module and replace the inline exampleNwcUrl declarations in the tests with imports from that module so all tests reference the same EXAMPLE_NWC_URL; update import statements in the files that currently declare exampleNwcUrl and remove their local definitions.e2e/README.md (2)
23-24: Consider clarifying which browsers are included.The comment mentions "All browsers" but doesn't specify which browsers Playwright will run against. Consider adding clarity for users unfamiliar with Playwright's defaults.
📝 Suggested clarification
-# All browsers (requires: yarn test:e2e:install first) +# All browsers - Chromium, Firefox, WebKit (requires: yarn test:e2e:install first) yarn playwright test🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@e2e/README.md` around lines 23 - 24, The "All browsers" note is ambiguous; update the README entry around the "All browsers (requires: yarn test:e2e:install first)" heading and the "yarn playwright test" command to explicitly list Playwright's default browser targets (Chromium, Firefox, and WebKit) and mention that running "yarn test:e2e:install" will install the browser binaries; e.g., change the heading text to name the three browsers and add a short sentence clarifying that Playwright runs tests against Chromium, Firefox and WebKit by default and that the install step downloads those browser binaries.
33-44: Consider adding context about the test execution model.The structure documentation is comprehensive, but could benefit from a brief explanation of how tests execute (e.g., that Playwright starts a webServer automatically to build and serve fixtures, tests load modules from /dist/esm, etc.). This would help users understand the test infrastructure better.
📝 Suggested addition
## Structure +Tests run against a local webServer (configured in `playwright.config.ts`) that automatically builds and serves fixtures. Tests load SDK modules from `/dist/esm` in a real browser environment. + - `e2e/` — e2e test directory - `e2e/fixtures/` — HTML fixtures for browser tests🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed. In `@e2e/README.md` around lines 33 - 44, Add a short paragraph to the e2e/README.md that explains the test execution model: state that Playwright (via playwright.config.ts) automatically starts a webServer to build and serve files from e2e/fixtures, that browser tests load built modules from /dist/esm, and that some specs use mocked APIs (e.g., oauth.spec.ts, webln.spec.ts) while others run without a relay (nwc.spec.ts); mention any required build step before running tests and how tests are started (e.g., via the Playwright test runner).
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@package.json`:
- Around line 54-55: The package.json npm script "format" has a glob mismatch
with "format:fix" for examples (format excludes .ts while format:fix includes
it); update the "format" script's examples glob to match "format:fix" (change
'examples/**/*.(js|jsx)' to 'examples/**/*.(js|jsx|ts)') so both "format" and
"format:fix" use the same file set, and verify both scripts' globs remain
consistent across other entries.
---
Nitpick comments:
In @.github/workflows/ci.yml:
- Around line 20-25: Remove the redundant top-level "yarn build" CI step because
Playwright's webServer (configured in playwright.config.ts and invoked by the
"yarn test:e2e" / "playwright test --project=chromium") already runs "yarn build
&& serve . -p 3000"; update the CI workflow by deleting the "yarn build" run
step (or alternatively modify the webServer command to skip the build when
running in CI via an environment variable or a CI-specific script) so the build
is not executed twice.
In `@e2e/fiat-amount.spec.ts`:
- Around line 5-11: The USD mock (mockRatesResponse) is reused for the EUR test
so the EUR-specific conversion isn't actually validated; create a separate mock
object (e.g., eurMockRatesResponse with code: "EUR" and a different
rate_float/rate_cents) and update the request stub that matches RATES_API to
inspect the request URL/pathname (or currency segment) and return
mockRatesResponse for USD paths and eurMockRatesResponse for EUR paths; update
any stubs/expectations referenced around the existing mock usage (including the
handler logic used by the EUR test) so each currency test receives its
corresponding mock response.
- Around line 27-50: Duplicate tests: both tests create USD(1) and call
resolveAmount; collapse them into one. Replace the two tests ("USD converts to
satoshis via resolveAmount" and "FiatAmount is interoperable with Amount") with
a single test that imports USD and resolveAmount, calls const fiatAmount =
USD(1); const resolved = await resolveAmount(fiatAmount); then assert resolved
matches the shape (satoshi and millisat are Numbers), assert resolved.satoshi
=== 1000, and assert resolved.millisat === resolved.satoshi * 1000; remove the
redundant test to avoid duplication.
In `@e2e/fixtures/lnclient.html`:
- Line 10: The import brings in LN but it is unused; remove the unused symbol by
changing the import that currently lists both LNClient and LN to import only
LNClient (i.e., update the import statement that references LNClient and LN so
it no longer includes LN) to eliminate the unused import and any linter
warnings.
In `@e2e/fixtures/nwc.html`:
- Around line 12-13: The string assigned to exampleNwcUrl is duplicated across
tests; extract it into a single exported constant (e.g., export const
EXAMPLE_NWC_URL) in a new shared test constants module and replace the inline
exampleNwcUrl declarations in the tests with imports from that module so all
tests reference the same EXAMPLE_NWC_URL; update import statements in the files
that currently declare exampleNwcUrl and remove their local definitions.
In `@e2e/fixtures/oauth.html`:
- Line 10: The import statement currently brings in an unused symbol — remove
OAuth2Bearer from the import so only the used Client is imported; update the
import that reads something like import { Client, OAuth2Bearer } ... to import {
Client } ... and ensure no other references to OAuth2Bearer remain in the file.
In `@e2e/oauth.spec.ts`:
- Around line 175-199: The test "Client uses custom base_url" contains a
redundant page.goto because the shared beforeEach already navigates to the
fixture; remove the extra navigation inside that test or move the test into its
own describe block without the shared beforeEach. Specifically, in the test
named Client uses custom base_url remove the await
page.goto("/e2e/fixtures/oauth.html") (or relocate the whole test into a new
describe that does not use the existing beforeEach) so the custom route is added
then the already-navigated page is used to evaluate the Client and call
accountBalance.
In `@e2e/README.md`:
- Around line 23-24: The "All browsers" note is ambiguous; update the README
entry around the "All browsers (requires: yarn test:e2e:install first)" heading
and the "yarn playwright test" command to explicitly list Playwright's default
browser targets (Chromium, Firefox, and WebKit) and mention that running "yarn
test:e2e:install" will install the browser binaries; e.g., change the heading
text to name the three browsers and add a short sentence clarifying that
Playwright runs tests against Chromium, Firefox and WebKit by default and that
the install step downloads those browser binaries.
- Around line 33-44: Add a short paragraph to the e2e/README.md that explains
the test execution model: state that Playwright (via playwright.config.ts)
automatically starts a webServer to build and serve files from e2e/fixtures,
that browser tests load built modules from /dist/esm, and that some specs use
mocked APIs (e.g., oauth.spec.ts, webln.spec.ts) while others run without a
relay (nwc.spec.ts); mention any required build step before running tests and
how tests are started (e.g., via the Playwright test runner).
In `@e2e/webln.spec.ts`:
- Around line 60-68: Extract the repeated mockAuth construction used inside
multiple page.evaluate calls into a shared helper or constant within the
describe block to avoid duplication; for example, create a getMockAuth() helper
(or a stringified snippet) and use it when constructing the provider in tests
that reference OauthWeblnProvider, but remember functions like getAuthHeader
can't be serialized into page.evaluate so either pass a serializable
representation or use the helper to return the full evaluate callback (or inline
the function inside the evaluate body) and update all places that construct
mockAuth (lines creating mockAuth for OauthWeblnProvider) to use that helper.
In `@package.json`:
- Around line 91-92: The devDependencies list is out of alphabetical order: swap
the positions of the "serve" and "qrcode-terminal" entries so that
"qrcode-terminal" comes before "serve" (maintain existing version strings and
trailing commas), ensuring the devDependencies array is sorted alphabetically by
package name.
In `@playwright.config.ts`:
- Around line 19-24: The webServer configuration in playwright.config.ts
currently uses the command string "yarn build && serve . -p 3000" which exposes
the entire repo root; change the webServer.command to serve only the built
output (e.g., "yarn build && serve ./dist -p 3000" or a dedicated test-serve
folder containing dist and e2e assets) or create a small serve config/symlink to
a dedicated folder and point webServer.command at that folder, keeping the
existing webServer.url, reuseExistingServer, and timeout settings unchanged;
update any scripts that rely on the old path accordingly.
- Around line 14-18: The Playwright config currently runs three browser projects
(projects array with names "chromium", "firefox", "webkit") which will slow CI
when workers: 1; update playwright.config.ts to only include the "chromium"
project for CI or make the projects list conditional (e.g., check
process.env.CI) so CI uses only Chromium while local/dev runs include all three,
or alternatively increase the global "workers" value for CI; modify the projects
array or add the CI-branch logic around it and ensure any references to
"projects" and "workers" are adjusted accordingly.
| "format": "prettier --check '**/*.(md|json)' 'src/**/*.(js|ts)' 'examples/**/*.(js|jsx)' 'e2e/**/*.(js|ts)' 'playwright.config.ts'", | ||
| "format:fix": "prettier --loglevel silent --write '**/*.(md|json)' 'src/**/*.(js|ts)' 'examples/**/*.(js|jsx|ts)' 'e2e/**/*.(js|ts)' 'playwright.config.ts'", |
There was a problem hiding this comment.
Glob mismatch between format and format:fix for examples.
format uses 'examples/**/*.(js|jsx)' while format:fix uses 'examples/**/*.(js|jsx|ts)'. The format check will miss .ts files in examples that format:fix would rewrite, leading to silent drift.
Proposed fix
- "format": "prettier --check '**/*.(md|json)' 'src/**/*.(js|ts)' 'examples/**/*.(js|jsx)' 'e2e/**/*.(js|ts)' 'playwright.config.ts'",
+ "format": "prettier --check '**/*.(md|json)' 'src/**/*.(js|ts)' 'examples/**/*.(js|jsx|ts)' 'e2e/**/*.(js|ts)' 'playwright.config.ts'",📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| "format": "prettier --check '**/*.(md|json)' 'src/**/*.(js|ts)' 'examples/**/*.(js|jsx)' 'e2e/**/*.(js|ts)' 'playwright.config.ts'", | |
| "format:fix": "prettier --loglevel silent --write '**/*.(md|json)' 'src/**/*.(js|ts)' 'examples/**/*.(js|jsx|ts)' 'e2e/**/*.(js|ts)' 'playwright.config.ts'", | |
| "format": "prettier --check '**/*.(md|json)' 'src/**/*.(js|ts)' 'examples/**/*.(js|jsx|ts)' 'e2e/**/*.(js|ts)' 'playwright.config.ts'", | |
| "format:fix": "prettier --loglevel silent --write '**/*.(md|json)' 'src/**/*.(js|ts)' 'examples/**/*.(js|jsx|ts)' 'e2e/**/*.(js|ts)' 'playwright.config.ts'", |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@package.json` around lines 54 - 55, The package.json npm script "format" has
a glob mismatch with "format:fix" for examples (format excludes .ts while
format:fix includes it); update the "format" script's examples glob to match
"format:fix" (change 'examples/**/*.(js|jsx)' to 'examples/**/*.(js|jsx|ts)') so
both "format" and "format:fix" use the same file set, and verify both scripts'
globs remain consistent across other entries.
|
Hi, thanks for your PR. I don't think Playwright makes sense here, the JS SDK has no frontend files. |
|
@rolznz Thanks for you answer 🙏
Playwright lets us verify that:
I created minimal HTML fixtures that load the SDK — they simulate how a real app would consume the library. This is a common pattern for testing JS libraries that target the browser. The webln providers (OauthWeblnProvider, NostrWeblnProvider) are explicitly browser-based; E2E tests are the right way to validate them. What do you think? |
|
I think we need to do a plan here. Some of your tests are mixing multiple things and is hard to follow. For Playwright, what are the minimum tests that we can do that actually are needed? I'd like most of the tests to be done for the JS SDK directly, not for the browser. As it is right now it's very difficult for me to review. |
|
@rolznz yes, sorry that PR has become so monstrous. As for minimizing Playwright, there is the following idea: One test — load the bundled SDK in a browser and run a simple operation (e.g. SATS(10) or Client instantiation). This validates that the build works in a browser environment. Jest runs in Node and can't catch browser-specific bundle issues. Everything else (Amount, FiatAmount, oauth Client, nwc parsing, LNClient constructor) can be moved to Jest with fetch/API mocks — they don't require browser APIs. I can refactor it. For example, cancel this PR and share a new one with minimal Playwright usage, and the next PR will be porting my current implementation to Jest. I'm open to any advice regarding this plan 🙏 |
|
@DSanich that would be great, thanks. Small PRs that do a single thing are best for us right now. For jest, please also check out the tests we already have and if you're ok to double check with me which ones are missing and are useful to add. We don't have e2e tests that use faucet.nwc.dev yet. |
|
@rolznz Yes, I would be happy if you could point out which tests are missing so that we could add them selectively (for example, several tests in one PR within one module) to make your review more comfortable. I am currently preparing a PR with the minimum required tests using Playwright. |
|
@DSanich for inspiration you could look at the CLI tests: https://github.com/getAlby/cli/tree/master/src/test Ideally if you could first make a PR with a single test so we can more easily review the test setup. E.g. A test that creates a wallet and then checks the balance, it should be 10_000 sats. Later we can add more tests. |
|
@rolznz Hi! Yes, that's a great idea. I'll create such a test using Jest and publish a PR today. |
2 participants
Summary
Adds Playwright-based E2E tests to verify SDK behavior in a real browser environment.
What's included
Infrastructure
Test coverage by module
Scripts
yarn test:e2e— run E2E tests (Chromium)yarn test:e2e:install— install Playwright browsersyarn test:e2e:ui— run with Playwright UIyarn test:e2e:headed— run in headed modeHow to run
yarn install
yarn test:e2e:install # one-time
yarn test:e2e
Summary by CodeRabbit
Tests
Documentation
Chores