During this lab, you will build a basic client-side application using Angular, which will connect to a server-side todo API similar to the one you created in the previous lab. Your Angular application should enable you to handle user input, and display data returned from the server in nice ways. As always, you'll be expected to make good use of the version control and project management tools available to you:
- Write good commit messages
- Use branching and pull requests for features
- Test things
- Document issues and plans
- Etc., etc.
Your specific tasks for this lab can be found in the LABTASKS.md file in this repository.
⚠️ One thing to keep in mind is that the Angular developers provide two major updates to Angular each year. This lab is built using Angular 15. Pay attention to the version of Angular being used in examples and on-line documentation that you find. Most of the time, it won't matter very much, but there are times when something you find doesn't match what we're doing. If things seem odd, look at the versions for the example or documentation you're looking at just in case there's a mismatch that matters.
As in the previous lab, you'll be using VS Code and GitKraken. Once you've all joined your group using GitHub classroom, you can clone your repository using the command line or GitKraken:
- From the file menu, choose Clone Repo
- Choose GitHub.com in the middle column (as the source location of your repo)
- Browse to the location you'd like to put the local copy of this project repo
- Select the correct repo from the list of repositories
- Select Clone the repo!
Launch Visual Studio Code, and then choose File -> Open Folder…
. Navigate to your clone
of the repo and choose Open
.
You may see a dialog that looks like this if you don't already have the recommended extensions:
Don't worry if you don't get the dialog, it is probably because you already have them all installed.
Like in previous labs, click "Install All" to automatically install them.
Before you start working you will need to install the dependencies for the client.
- Move into the
client
directory (cd client
) - Run
npm install
⚠️ I (Nic) am not 100% sure that this will actually be an issue for you. Please let me know if if is or it isn't (e.g., whether you see the red "ESLINT" on the bottom bar).
Since this is the first time we will be using ESLint there is an additional step to make sure the VS Code extension is working in the project. When you first open a TypeScript file you will see at the bottom right that ESLint is disabled.
Click the red "ESLINT" to open this dialog:
Click "Allow Everywhere" to enable ESLint.
You can also open this dialog with the following steps:
- Hit
CTRL + SHIFT + P
(⌘ + ⇧ + P
on Macs) to open the Command Palette. You can also find this by going to the "View" menu and clicking "Command Palette..." - Start typing and select "ESlint: Manage Library Execution". That should open a dialog seen above.
To run your project you'll have to start both the server and the client.
- The run Gradle task (
./gradlew run
in theserver
directory) will still run your Javalin server, which is available atlocalhost:4567
. - The build task will still build the server, but not run it.
- The test task will run all the JUnit tests.
- The check task will run Checkstyle along with the JUnit tests.
The major difference here is that the client side of your project is, effectively, an entirely separate project from your Javalin server. We've included a full API for the todos, which you implemented in Lab 2, so there is no need to copy your old project over.
We don't expect you to have any reason to actually make any changes on the server side of the project, although you're welcome to look it over and ask questions about it. We don't really "object" if you make changes to the server code, but you should probably think twice about doing it. If you think you're "fixing" something by changing the server code, there's a good chance the problem isn't where you think it is.
Once you have successfully run npm install
, in order to serve up the client side of your project, you will run
ng serve
(from the client
directory as well). This will trigger the various tools in the
client side portion of the project to build and host your client side
application on their own little web-server, available by default at localhost:4200
. If your server is running, you will be able to see data for users if you navigate to the right place in the project.
To recap, here are the steps needed to run the project:
- Go into the
server
directory and enter./gradlew run
. - In a different terminal (or terminal tab), go into the
client
directory and enterng serve
. - You can then go to
localhost:4200
in your favorite web browser and see your nifty Angular app.
There are now more testing options! You can test the client, or the server or both.
In this lab there's no particular need to test the server since you're probably not changing anything there. All the tools are there, however, if you want to run the tests and see what that looks like.
From the client
directory:
ng test
runs the client unit (Karma) tests- This will pop up a Chrome window with the results of the tests.
- This will run "forever", re-running the tests whenever you make a
change. Both in your terminal and the Chrome
window will display updated results whenever the tests are re-run. Typing CTRL-C in the terminal window will end
the
ng test
process and close the generated Chrome window.
- You can add
ng test --no-watch
if you just want to run the tests once instead of going into the "run forever" mode. ng test --code-coverage
runs the client tests and generates a coverage report- It generates a coverage report in your client directory:
client/coverage/client/index.html
. - Right click on
index.html
and selectCopy path
and paste it into your browser of choice. You can also drag and dropindex.html
onto the tab area of your browser and it will open it.
- It generates a coverage report in your client directory:
- You can combine these with things like
ng test --no-watch --code-coverage
to run the tests once and compute the test coverage stats.
We have included a tool called ESLint which helps analyze the code and catch various errors. You will most likely see it directly in VS Code as yellow and red underlines. You can also directly run the linter on the entire client by running ng lint
. This will check the whole client project and tell you if there are any issues.
From the server
directory:
./gradlew test
runs the server tests once.- It generates a report you can find in
server/build/reports/tests/test/index.html
.
- It generates a report you can find in
./gradlew test jacocoTestReport
runs the server tests once and creates a coverage report- It generates a coverage report you can find in
server/build/jacocoHtml/index.html
in addition to the regular report generated by thetest
task.
- It generates a coverage report you can find in
End to end (E2E) testing involves the whole software stack rather than one part of it. Our E2E tests look at the behavior of both the client and server and how they interact by simulating what a real user would do with it.
We use Cypress for our end-to-end tests. There are a few ways to run the E2E tests. They are all started from the client
directory and require the server be running at the same time (./gradlew run
in the server
directory).
ng e2e
both builds and serves the client and runs through all the Cypress end-to-end tests once.ng e2e --watch
builds and serves the client but just opens Cypress for you to be able to run the tests you want without closing automatically.- This is the same as running
ng serve
followed bynpm run cy:open
(ornpx cypress open
). If you are already runningng serve
it will be easier to do this rather than closing it and runningng e2e
.
- This is the same as running
If you're running this on a Mac, macOS will run a validation check the first time you run a given version of Cypress. Just wait for that to finish, and then everything should move forward fine.
When you first run Cypress --watch
it'll ask you to choose your
preferred browser for testing. You can use any browser installed on
your computer, but just sticking with "Electron" is probably the
simplest.
After you've chosen a brower, you should get a window with the main Cypress view, which looks something like this:
You can click on any of the integration test files to run their tests or run them all. When you run a set of tests like the tests in user-list.cy.ts
, you will see the results of the tests:
There are a lot of neat things you can do here like inspect each test and find which selectors to use in the tests you are writing. If click
on the test for filtering by company, for example, we can see the steps
of that test. We can use those to see exactly what's being selected and asserted. If we hover over step six, for example (find .user-card-company
), we see that the company fields are highlighted
for the two users that work for "OHMNET":
This can be extremely useful for debugging failing E2E tests, and can also be a nice way to better understand the structure of the HTML generated by Angular.
We encourage you to look through some of the Cypress documentation linked in the "Resources" section below.
There are three GitHub Actions workflows set up in your repo:
- Server Java - JUnit tests for the server (
gradle-build
) - Client Angular - Karma tests (
ng-test
) and ESLint linting (ng-lint
) for the client - End to End - Cypress tests for end-to-end testing
There are badges above that show the status of these checks on the master branch.