Uncluttered/README.org

#+title: Uncluttered
#+author: Xavier Brinon
#+date: [2023-11-24 Fri]
#+startup: indent
* Org shortcut
- Evaluate the line, apply the config in place :: ~C-c C-c~
* Webserver
** Guile
So we need a webserver to display the HTML page. Since I'm keen on trying fancy
pancy standard stuff in Webdev, let's do it to with GNU.
Let's try and choose a webserver in [[https://www.gnu.org/software/guile/manual/html_node/index.html][Guile]].

#+name: install guile
#+begin_src shell
  guix install guile
  guile --version
#+end_src

#+RESULTS: install guile
: guile (GNU Guile) 3.0.9
: Copyright (C) 2023 Free Software Foundation, Inc.
:
: License LGPLv3+: GNU LGPL 3 or later <http://gnu.org/licenses/lgpl.html>.
: This is free software: you are free to change and redistribute it.
: There is NO WARRANTY, to the extent permitted by law.
** or Rust ?
We're exploring Rust as well, so maybe there is some fancy stuff here as well. I
mean while we wait and learn about Guile anyway.
Well, there's [[https://rocket.rs/][Rocket]] for Rust but I guess we can wait.
** Nah, let's stick to Esbuild
Because this is what the book is doing and we don't want to waste too much time
(it's 9:29pm) and I'm too tired for that.
* Set up
** Repo
Setting up the project is done via the repo template on github.[fn:1]
Let's clone it from gitlab, or gitea ?
** Esbuild
Recommended way is via curl, see the [[https://esbuild.github.io/getting-started/#other-ways-to-install][install page]].
#+name: curl esbuild
#+begin_src shell
  curl -fsSL https://esbuild.github.io/dl/v0.19.7 | sh
#+end_src

Now that you have ~esbulid~, install it in your local path if you fancy.
We can start a server from here.
#+name: start server
#+begin_src shell
  esbuild --serve=8888 --servedir=.
#+end_src

#+RESULTS: start server
:  > Local:   http://127.0.0.1:8888/
:  > Network: http://192.168.1.140:8888/
:  > Network: http://192.168.122.1:8888/
:  > Network: http://10.0.3.1:8888/
:  > Network: http://172.17.0.1:8888/
** HTTPS
Local servers are considered secure by default. There is no need to
over-engineer the setup. At least at this point.
As for CORS and other security rules, we will fix them as they arise.
No need to fix a problem until it actually becomes a pbm. The solution will be
measured agains the reality of the situation. This is the sure way to avoid
scope creeps.
** Firefox
Because you don't want to share the development setup with other projects or
the usual browsing, e.g. same cookie origins, indexdb, and localstorage, ...
we can launch firefox with a dedicated profile.
Assuming we all know where our firefox is launched from and that the esbuild
server is running.
#+name: firefox launcher
#+begin_src shell
  #!/bin/bash
  set -eu

  DATA_DIR="$(mktemp -d -t 'firefox_unsafe_data_dir.XXXXXXXXXX')"
  /usr/local/bin/firefox/firefox-bin \
      -profile $DATA_DIR \
      -no-remote \
      -devtools \
      -url http://localhost:8888 \
      > /dev/null 2>&1 &!
#+end_src

#+RESULTS: firefox launcher

Any good script should be checked against [[https://www.shellcheck.net/][shellcheck]]
Looks like there is a [[https://how-to.dev/how-to-set-up-a-dev-server-with-esbuild][good doc]] about esbuild for dev server.
** HTML file
The html file contains the testing libraries and loads the tools.
It is using Mocha and Chai[fn:2].
*** And then some
Let's create 3 HTML files:
- chai+mocha.html
- tape.html
- qunit.html


And see which one works best.
*** Chai+Mocha
There is a [[https://mochajs.org/#running-mocha-in-the-browser][page]] that describe the starter HTML page. It's using *unpkg* for both
libraries.
I'm going to do just that. Let's download the *css* and *js* file from the URL
as used in the template html file.



[[https://www.chaijs.com/guide/installation/][Installing]] Chai is done by downloading the =chai.js= file from the repo.
Last realease is =4.3.10= as per the [[https://github.com/chaijs/chai/blob/v4.3.10/chai.js][URL]] but the exports says ~4.3.8~ :shrug:
Putting that in a folder =test-utils/=
#+name: download chai
#+begin_src shell
  cd test-utils
  curl https://github.com/chaijs/chai/blob/v4.3.10/chai.js -O
#+end_src

#+RESULTS: download chai

As for mocha, we would use the unpkg URL directly.

#+begin_src html
  <!DOCTYPE html>
  <html lang="en">
    <head>
      <meta charset="utf-8" />
      <title>Mocha Tests</title>
      <meta name="viewport" content="width=device-width, initial-scale=1.0" />
      <link rel="stylesheet" href="/test-utils/mocha.css" />
      <!--
          <script type="importmap">
           {
           "imports": {
           }
           }
           </script>
      -->
    </head>
    <body>
      <div id="mocha"></div>

      <script src="/test-utils/chai.js"></script>
      <script src="/test-utils/mocha.js"></script>

      <script class="mocha-init" type="module">
        mocha.setup({ ui: "tdd", timeout: "2000"})
        mocha.checkLeaks()
        window.addEventListener("load", function () {
          // Callback so that the test results get reported to CI
          mocha.run(function (failures) {
            window._testResults = {
              done: true,
              failures,
              succeeded: failures ? false : true,
            }
          })
        })
      </script>
      <!-- Include the test files as script elements. -->
      <!-- <script src="/path/to/file.js" type="module"></script> -->
    </body>
  </html>
#+end_src

Displaying the file via ~esbuild~ will show 0 tests run.
*** Tape
*** Qunit

* Footnotes
[fn:2]I don't know, possible alternatives are:
- Tape
- Qunit
They have really good scores in [[https://libraries.io/search?q=&platforms=NPM&sort=dependents_count][libraries.io]]

[fn:1]Which causes 2 problems:
- It's github
- I want to set up the project from scratch :shrug: