Start with chapter 8's `tasks.tea`, `task_list.tea`, and saved `tasks.txt`. There are no source changes in this chapter. We'll check the program, run its tests, and build the executable you've been working toward. ## Check and format ```bash tea check tasks.tea task_list.tea tea fmt tasks.tea task_list.tea tea fmt --check tasks.tea task_list.tea tea test "$PWD/task_list.tea" ``` The checker should report zero errors and warnings. The formatter may report files it changed; its `--check` invocation should then succeed without requesting changes. The four main walkthrough tests should pass. If you kept the optional test from chapter 8, you'll see five instead. `tea check` validates source without executing it. Tests exercise behavior, while building also verifies that your program can be linked into a native executable. ## Build and run ```bash tea build tasks.tea ./bin/tasks list --all ``` The build reports the output at `bin/tasks`; it may mention a cached build. The executable prints: ```text 1. [ ] Learn Tea 2. [x] Build a CLI 3. [x] Take a break 4. [ ] Write a test ``` It reads the same `tasks.txt` because you ran it from the same directory. The entry file imports `task_list.tea`, so Tea compiles that module as part of the program. You don't build the two files separately. ## Use it as a command-line tool ```bash ./bin/tasks add "Build something useful" ./bin/tasks done 1 ./bin/tasks list ``` ```text Added task 5. Completed task 1. 4. [ ] Write a test 5. [ ] Build something useful ``` You can copy the executable to another compatible machine without its Tea source files. Keep platform and architecture in mind: a native executable built for your machine isn't automatically a binary for every operating system. Your data file is separate from the executable, and its location remains relative to the directory where you run the command. ## Check the command's exit status ```bash ./bin/tasks list --unknown echo $? ``` The tool writes `tasks: invalid arguments; run tasks --help for usage` to standard error, and `echo` prints `2`. Help and successful operations return `0`; malformed saved data or file failures return `1`. Unlike the current script launcher, the compiled executable preserves these exact codes. ## Final source The complete two-file source is in [Organize and test](/docs/tutorial-test-and-organize). The repository also contains [runnable checkpoints](https://github.com/bentruyman/tea-lang/tree/main/examples/tutorial/tasks) for every source-changing chapter. You can revisit any checkpoint without recreating earlier exercises. ## Try it yourself Add an `open` command that marks a completed task as open again. Reuse the number validation from `done`, add a test, and update the help text.

Design hint

Extract the “text to valid task index” portion of `complete_task` into a private helper. Have both operations call it, then set the `done` field to the appropriate value. Keep returning whether a task changed so repeating either command doesn't rewrite the file unnecessarily.

## Where to go next You started with two lines of Tea and ended with a native tool that owns its data, validates input, handles failures, and has tests. For your next program, reuse the same progression: get a useful result, extract functions when they help, introduce structured data, and test the logic separately from the command interface. Try a task search command, a different storage location supplied by an argument, or a new CLI entirely. [Browse the examples](/examples) for ideas, use the [language guides](/docs) to deepen your understanding, and keep the [standard-library reference](/reference) nearby.