Add test watching and some unit tests (#7)

* Move lint after test and to separate workflow

It can be annoying to have the tests not run just because linting fails
so run it after tests instead (this is to make sure devs see it since
failing to lint before pushing is common).

Additionally make lint a separate step when in CI since otherwise a lint
failure would make it look like a test failure.

* Automatically fix linting issues after running tests

It turns out having to manually run this is just kind of annoying.

* Add watch command for tests from the CLI

* Add tests for utils

Also remove an unused function bubbleError.

* Add tests for workspaces

* Stub out extension tests

* Add a way to directly set the coder binary path

* Set command for coder on Windows

It is failing to find the binary in the PATH.

Author: code-asher

.github/workflows/lint.yaml

@@ -0,0 +1,24 @@
+on:
+  push:
+    branches:
+      - master
+  pull_request:
+    branches:
+      - master
+
+name: Lint Extension
+
+jobs:
+  lint:
+    runs-on: ubuntu-20.04
+
+    steps:
+      - uses: actions/checkout@v2
+
+      - uses: actions/setup-node@v2
+        with:
+          node-version: '16'
+
+      - run: npm ci
+
+      - run: npm run lint

.github/workflows/test.yaml

@@ -35,7 +35,7 @@ jobs:
 
       - run: npm ci
 
-      - run: npm run test
+      - run: npm run test:ci
         env:
           # This tells the application on which display it should run. Set to
           # the display we started on Linux via Xvfb. The other platforms do not

.vscode/launch.json

@@ -24,7 +24,7 @@
       "outFiles": [
         "${workspaceFolder}/out/**/*.test.js"
       ],
-      "preLaunchTask": "npm: test:watch"
+      "preLaunchTask": "npm: tsc:watch"
     }
   ]
 }

.vscode/tasks.json

@@ -19,7 +19,7 @@
     },
     {
       "type": "npm",
-      "script": "test:watch",
+      "script": "tsc:watch",
       "problemMatcher": "$tsc-watch",
       "isBackground": true,
       "presentation": {

README.md

@@ -13,18 +13,20 @@
 From here you have a few choices:
 
 - Run through VS Code:
-  - `Run Extension` task
-  - `Extension Tests` task
-- Run `npm run test` from the CLI
-  - On Linux to run headless install and run `xvfb`
-    - Refer to CI test workflow for more details
+  - Launch `Run Extension` through the debug panel
+  - Launch `Extension Tests` through the debug panel
+- Run `npm run test` or `npm run test:watch` from the CLI
+  - On Linux to run a headless install use `xvfb`
+    - Refer to CI test workflow for more details about `xvfb`
 
 To fix automatically fixable linting issues run:
 
 ```shell
 npm run lint -- --fix
 ```
 
+Linting will be automatically fixed when you use `npm run test` or `npm run test:watch`.
+
 ## Known issues
 
 - Context menu action `Show Logs` is blocking, and blocks until a rebuild completes

fixtures/coder

@@ -0,0 +1,30 @@
+#!/bin/sh
+# Coder CLI mock
+
+need_login() {
+  echo "fatal: failed to read session credentials"
+  echo "  | hint: did you run \"coder login [https://coder.domain.com]\"?"
+  exit 1
+}
+
+workspaces() {
+  cat ./workspaces.json
+  exit 0
+}
+
+main() {
+  cd "$(dirname "${0}")"
+
+  # Use CODER_MOCK_STATE to force certain states.
+  case "${CODER_MOCK_STATE-}" in
+    "need_login") need_login ;;
+  esac
+
+  # Mock the output based on the passed arguments.
+  case "$*" in
+    "envs ls --output json") workspaces ;;
+    *) echo "$* is not mocked" ; exit 1 ;;
+  esac
+}
+
+main "$@"

fixtures/workspaces.json

@@ -0,0 +1,17 @@
+[
+  {
+    "id": "workspace-id",
+    "name": "workspace-name",
+    "cpu_cores": 1,
+    "memory_gb": 2,
+    "updated": false,
+    "image_tag": "image-tag",
+    "image_id": "image-id",
+    "gpus": 4,
+    "updating": false,
+    "latest_stat": {
+      "container_status": "ON"
+    },
+    "disk_gb": 5
+  }
+]