Merge pull request #8 from skx/7-allow-name-to-be-specified

Allow the script-name to be specified in the workflow file.

README.md

@@ -1,38 +1,51 @@
+
 # GitHub Action for Running tests
 
 This repository contains a simple GitHub Action which allows you to run a shell-script every time an event occurs within your repository.
 
-If your shell-script terminates with an exit-code of 0 that is regarded as a pass, otherwise the action will be marked as a failure.
+* [GitHub Action for Running tests](#github-action-for-running-tests)
+  * [Overview](#overview)
+  * [Enabling the action](#enabling-the-action)
+  * [Sample Configuration](#sample-configuration)
+  * [Advanced Configuration](#advanced-configuration)
+
+
+## Overview
+
+This action allows you to run a shell-script when your workflow action is triggered.  If your script terminates with an exit-code of 0 that is regarded as a pass, otherwise the action will be marked as a failure.
 
 The expectation is that you'll use this action to launch your project-specific test-cases, ensuring that all pull-requests, commits, or both, are tested automatically.
 
-Because the action ultimately just executes a shell-script contained in your repository, you can be as simple or complex as you can imagine.  For example a [golang](https://golang.org/) project might contain a simple script such as this:
+Because the action ultimately executes a shell-script contained in your repository you can be as simple or complex as you can like, for example a [golang](https://golang.org/) project might contain a script such as this:
 
     #!/bin/sh
-    # Run the go-vet tool
+    # Run the go-vet tool.
     go vet ./..           || exit 1
-    # Run the test-cases
+    # Run the test-cases, with race-detection.
     go test -race ./...   || exit 1
-    # Everything passed.
+    # Everything passed, exit cleanly.
     exit 0
 
-A C-based project might contain something like this:
+A C-based project could contain something like this instead:
 
     #!/bin/sh
     make && make test
 
 But as you can install/invoke arbitrary commands, and update them as your project grows, you can do almost anything you wish.
 
 
+
 ## Enabling the action
 
 There are two steps required to use this action:
 
 * Enable the action inside your repository.
   * You'll probably want to enable it upon pull-requests, to ensure their quality.
   * You might also want to enable it to run each time a push is made to your repository, for completeness.
-* Add your project-specific tests to the script `.github/run-tests.sh`.
-  * The exit-code of this script will determine the result.
+* Add your project-specific test-steps to a script in your repository.
+  * By default this action will execute `.github/run-tests.sh`, but you can specify a different name if you prefer.
+  * The exit-code of your script will determine the result.
+
 
 
 ## Sample Configuration
@@ -46,7 +59,7 @@ For example:
 * `.github/workflows/push.yml`
   * This is used when a commit is pushed to your repository.
 
-You can use content like this to invoke the action:
+The simplest example of using this action would be to create the file `.github/workflows/pull_request.yml` with the following contents:
 
 ```
 on: pull_request
@@ -61,4 +74,41 @@ jobs:
       uses: skx/github-action-tester@master
 ```
 
-You can also limit the tests to only executing when specific branches are updated, and chain events.  For more details please consult the [Github Actions documentation](https://developer.github.com/actions/).
+This example will run the default test-script, `.github/run-tests.sh`, every time a pull-request is created, edited, or updated.
+
+
+
+## Advanced Configuration
+
+As noted github actions can be launched on multiple events, for example pushes to branches, new releases, pull-request related events, and similar.
+
+Because you probably wish to run different tests/scripts on these different events it is possible to override the name/path of the shell-script which is executed on a per-event basis.
+
+For example you might wish to run more thorough tests upon pull-requests, and a smaller subset when a push is made to your `master` branch (on the assumption that commits there are rare, and the usual workflow will have ensured the full-tests will have been executed via pull-requests).
+
+As an example you might create a workflow for use solely with pushes to master, in the file `.github/workflows/push.yml`:
+
+```
+on:
+  push:
+    branches:
+    - master
+name: Push Event
+jobs:
+  test:
+    name: Run tests
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@master
+    - name: Test
+      uses: skx/github-action-tester@master
+      with:
+        script: .github/fast-tests.sh
+```
+
+Here we've done two things:
+
+* We've limited the action to only apply to pushes made to the `master` branch.
+* We've explicitly set the name of the testing-script to `.github/fast-tests.sh`
+  * With the expectation this script contains only "quick" tests.
+  * With slower tests being applied to pull-requests.

action.yml

@@ -4,6 +4,12 @@ author: 'Steve Kemp'
 branding:
   icon: eye
   color: black
+inputs:
+  script:
+    description: 'The path to the test-script to run, within the repository.'
+    default: '.github/run-tests.sh'
 runs:
   using: 'docker'
   image: 'Dockerfile'
+  args:
+      - ${{ inputs.script }}

entrypoint.sh

@@ -1,47 +1,74 @@
 #!/bin/bash
+#
+# The core of our action, it will be invoked with the name of the script
+# to execute - If the script doesn't exist then that is a fatal error.
+#
+# The script used to default to `.github/run-tests.sh`, but now the
+# name is passed as the first argument.
+#
 
 #
 # Terminate upon errors
 #
 set -e
 
+
+#
+# Get the name of the script to run.
+#
+script=$1
+
+
 #
-# If the script is present ..
-#
-if [ -e .github/run-tests.sh ]; then
-
-    #
-    # Ensure it is executable.
-    #
-    chmod 755 .github/run-tests.sh
-
-    #
-    # Run it.
-    #
-    .github/run-tests.sh
-
-    #
-    # If the script exits with a non-zero status-code
-    # then it will terminate this script due to the use
-    # of `set -e`.
-    #
-    # So when we reach this point we know:
-    #
-    #  1.  We've executed the test-script.
-    #
-    #  2.  The test-script passed.
-    #
-    # Exit cleanly because we're done.
-    #
-    exit 0
+# If the argument is empty that is a fatal error.
+#
+if [ -z "${script}" ]; then
+    echo "You must specify the script to execute as an argument."
+    exit 1
+fi
+
+
+#
+# If the script does not exist that is a fatal-error.
+#
+if [ ! -e "${script}" ]; then
+    echo "The supplied testing-script does not exist: ${script}"
+    exit 1
 fi
 
 
+#
+# Ensure the script is executable.
+#
+chmod 755 "${script}"
+
+
+#
+# Run it.
+#
+${script}
+
+
+#
+# If the script exits with a non-zero status-code
+# then it will terminate this script due to the use
+# of `set -e`.
+#
+# So when we reach this point we know:
+#
+#  1.  We've executed the test-script.
+#
+#  2.  The test-script passed.
+#
+# Exit cleanly because we're done.
+#
+exit 0
+
+
 #
 # There is no test-script within the repository.
 #
 # Show that, and exit with a failure-code.
 #
 echo "The repository does not contain a test-script '.github/run-tests.sh'"
 exit 1
-