forgejo/tests/e2e
Mai-Lapyst 51735c415b Add support for workflow_dispatch (#3334)
Closes #2797

I'm aware of https://github.com/go-gitea/gitea/pull/28163 exists, but since I had it laying around on my drive and collecting dust, I might as well open a PR for it if anyone wants the feature a bit sooner than waiting for upstream to release it or to be a forgejo "native" implementation.

This PR Contains:
- Support for the `workflow_dispatch` trigger
- Inputs: boolean, string, number, choice

Things still to be done:
- [x] API Endpoint `/api/v1/<org>/<repo>/actions/workflows/<workflow id>/dispatches`
- ~~Fixing some UI bugs I had no time figuring out, like why dropdown/choice inputs's menu's behave weirdly~~ Unrelated visual bug with dropdowns inside dropdowns
- [x] Fix bug where opening the branch selection submits the form
- [x] Limit on inputs to render/process

Things not in this PR:
- Inputs: environment (First need support for environments in forgejo)

Things needed to test this:
- A patch for https://code.forgejo.org/forgejo/runner to actually consider the inputs inside the workflow.
  ~~One possible patch can be seen here: https://code.forgejo.org/Mai-Lapyst/runner/src/branch/support-workflow-inputs~~
  [PR](https://code.forgejo.org/forgejo/runner/pulls/199)

![image](/attachments/2db50c9e-898f-41cb-b698-43edeefd2573)

## Testing

- Checkout PR
- Setup new development runner with [this PR](https://code.forgejo.org/forgejo/runner/pulls/199)
- Create a repo with a workflow (see below)
- Go to the actions tab, select the workflow and see the notice as in the screenshot above
- Use the button + dropdown to run the workflow
  - Try also running it via the api using the `` endpoint
- ...
- Profit!

<details>
<summary>Example workflow</summary>

```yaml
on:
  workflow_dispatch:
    inputs:
      logLevel:
        description: 'Log Level'
        required: true
        default: 'warning'
        type: choice
        options:
        - info
        - warning
        - debug
      tags:
        description: 'Test scenario tags'
        required: false
        type: boolean
      boolean_default_true:
        description: 'Test scenario tags'
        required: true
        type: boolean
        default: true
      boolean_default_false:
        description: 'Test scenario tags'
        required: false
        type: boolean
        default: false
      number1_default:
        description: 'Number w. default'
        default: '100'
        type: number
      number2:
        description: 'Number w/o. default'
        type: number
      string1_default:
        description: 'String w. default'
        default: 'Hello world'
        type: string
      string2:
        description: 'String w/o. default'
        required: true
        type: string

jobs:
  test:
    runs-on: docker
    steps:
      - uses: actions/checkout@v3
      - run: whoami
      - run: cat /etc/issue
      - run: uname -a
      - run: date
      - run: echo ${{ inputs.logLevel }}
      - run: echo ${{ inputs.tags }}
      - env:
          GITHUB_CONTEXT: ${{ toJson(github) }}
        run: echo "$GITHUB_CONTEXT"
      - run: echo "abc"
```
</details>

Reviewed-on: https://codeberg.org/forgejo/forgejo/pulls/3334
Reviewed-by: Earl Warren <earl-warren@noreply.codeberg.org>
Co-authored-by: Mai-Lapyst <mai-lapyst@noreply.codeberg.org>
Co-committed-by: Mai-Lapyst <mai-lapyst@noreply.codeberg.org>
2024-06-28 05:17:11 +00:00
..
actions.test.e2e.js
debugserver_test.go
e2e_test.go
edit-comment.test.e2e.js
example.test.e2e.js
explore.test.e2e.js
markdown-editor.test.e2e.js Implement tab indentation and list continuation in the new editor (#3848, #3849) (#4072) 2024-06-15 13:40:31 +00:00
markup.test.e2e.js
README.md
right-settings-button.test.e2e.js
utils_e2e.js
utils_e2e_test.go

End to end tests

E2e tests largely follow the same syntax as integration tests. Whereas integration tests are intended to mock and stress the back-end, server-side code, e2e tests the interface between front-end and back-end, as well as visual regressions with both assertions and visual comparisons. They can be run with make commands for the appropriate backends, namely:

make test-sqlite
make test-pgsql
make test-mysql

Make sure to perform a clean front-end build before running tests:

make clean frontend

Install playwright system dependencies

npx playwright install-deps

Run all tests via local act_runner

act_runner exec -W ./.github/workflows/pull-e2e-tests.yml --event=pull_request --default-actions-url="https://github.com" -i catthehacker/ubuntu:runner-latest

Run sqlite e2e tests

Start tests

make test-e2e-sqlite

Run MySQL e2e tests

Setup a MySQL database inside docker

docker run -e "MYSQL_DATABASE=test" -e "MYSQL_ALLOW_EMPTY_PASSWORD=yes" -p 3306:3306 --rm --name mysql mysql:latest #(just ctrl-c to stop db and clean the container)
docker run -p 9200:9200 -p 9300:9300 -e "discovery.type=single-node" --rm --name elasticsearch elasticsearch:7.6.0 #(in a second terminal, just ctrl-c to stop db and clean the container)

Start tests based on the database container

TEST_MYSQL_HOST=localhost:3306 TEST_MYSQL_DBNAME=test TEST_MYSQL_USERNAME=root TEST_MYSQL_PASSWORD='' make test-e2e-mysql

Run pgsql e2e tests

Setup a pgsql database inside docker

docker run -e "POSTGRES_DB=test" -p 5432:5432 --rm --name pgsql postgres:latest #(just ctrl-c to stop db and clean the container)

Start tests based on the database container

TEST_PGSQL_HOST=localhost:5432 TEST_PGSQL_DBNAME=test TEST_PGSQL_USERNAME=postgres TEST_PGSQL_PASSWORD=postgres make test-e2e-pgsql

Running individual tests

Example command to run example.test.e2e.js test file:

Note: unlike integration tests, this filtering is at the file level, not function

For SQLite:

make test-e2e-sqlite#example

For PostgreSQL databases(replace mysql to pgsql):

TEST_MYSQL_HOST=localhost:1433 TEST_MYSQL_DBNAME=test TEST_MYSQL_USERNAME=sa TEST_MYSQL_PASSWORD=MwantsaSecurePassword1 make test-e2e-mysql#example

Visual testing

Although the main goal of e2e is assertion testing, we have added a framework for visual regress testing. If you are working on front-end features, please use the following:

  • Check out main, make clean frontend, and run e2e tests with VISUAL_TEST=1 to generate outputs. This will initially fail, as no screenshots exist. You can run the e2e tests again to assert it passes.
  • Check out your branch, make clean frontend, and run e2e tests with VISUAL_TEST=1. You should be able to assert you front-end changes don't break any other tests unintentionally.

VISUAL_TEST=1 will create screenshots in tests/e2e/test-snapshots. The test will fail the first time this is enabled (until we get visual test image persistence figured out), because it will be testing against an empty screenshot folder.

ACCEPT_VISUAL=1 will overwrite the snapshot images with new images.