Skip to content

fix issues

fix issues #11693

Workflow file for this run

name: Local development
# **What it does**: Can you start the local server like a writer would do?
# **Why we have it**: Our CI is often heavily geared on testing in "production"
# that historically we've been known to break local
# development sometimes.
# **Who does it impact**: Engineers, Contributors.
on:
merge_group:
pull_request:
permissions:
contents: read
jobs:
local-dev:
if: github.repository == 'github/docs-internal' || github.repository == 'github/docs'
runs-on: ${{ fromJSON('["ubuntu-latest", "ubuntu-20.04-xl"]')[github.repository == 'github/docs-internal'] }}
steps:
- name: Check out repo
uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 # v4.1.1
- uses: ./.github/actions/node-npm-setup
- uses: ./.github/actions/get-docs-early-access
if: ${{ github.repository == 'github/docs-internal' }}
with:
token: ${{ secrets.DOCS_BOT_PAT_READPUBLICKEY }}
# Note that we don't check out docs-early-access, Elasticsearch,
# or any remote translations. Nothing fancy here!
- name: Disable Next.js telemetry
run: npx next telemetry disable
# The Playwright test, with the env vars we set here, takes care of
# starting a server and shutting it down when it's done.
# That's why it's important this step comes before the `npm start &`
# step below.
- name: Run Playwright tests
env:
# This is what local dev contributors are expected to do.
PLAYWRIGHT_START_SERVER_COMMAND: 'npm start'
# This is so that timeouts aren't retried, which can lead to
# tests not exiting at the end with a non-zero. Otherwise,
# by default failures are marked as "flaky" instead of "failed".
PLAYWRIGHT_RETRIES: 0
TEST_EARLY_ACCESS: ${{ github.repository == 'github/docs-internal' }}
run: npm run playwright-test -- playwright-local-dev
- name: Start server in the background
run: npm start > /tmp/stdout.log 2> /tmp/stderr.log &
- name: View the home page
run: |
echo "Going to sleep a little to wait for the server to start"
sleep 10
curl --fail --retry-connrefused --retry 5 http://localhost:4000/
- name: Run basic tests
run: npm run test-local-dev
- if: ${{ failure() }}
name: Debug server outputs on errors
run: |
echo "____STDOUT____"
cat /tmp/stdout.log
echo "____STDERR____"
cat /tmp/stderr.log
- name: Pre-commit hooks should prevent bad Markdown edits
run: |
set -e
# This test assumes this one file always exists
ls content/get-started/start-your-journey/hello-world.md
# Not sure if it matters but we're in a detached HEAD state
# after the actions/checkout action.
git checkout -b my-new-branch
# Also, do this so you don't get errors from git about this
# not being set up before your first commit attempt
git config user.name github-actions
git config user.email [email protected]
# To know what will fail the markdown lint, see src/content-linter/style/github-docs.js
# Add some NOT valid Markdown to it
# In this case an internal link with a hardcode /en/ prefix.
echo "This *is** not valid [Markdown](/en/foo)" >> content/get-started/start-your-journey/hello-world.md
git commit -a -m "this should fail"
exit_code=$?
if [ $exit_code != 0 ]; then
echo "That SHOULD have failed, but it DIDN'T"
exit 1
else
echo "As expected, it failed :)"
fi