トゲハテナブロック (hatena burokku) is Japanese for interrogation block ⍰.

The app is in /src. Assets called by the app needs to be compiled using some front-end tooling, detailed below.

1. Duplicate .env.example to .env and edit it.
2. Run npm install to install all the required packages and tools.
3. Run npm run dev and open the URL returned by the CLI.

Note: the Service Worker is broken while developing. Ignore it and check it using the production mode.

1. Duplicate .env.example to .env and edit it.
2. Run npm install to install all the required packages and tools.
3. Run npm run build to compile the app. The compiled app goes in /public.
4. Run npm run preview to preview it on your machine.

Tests use Playwright and can be run with npm test. They can all run locally or in a GitHub action.

Before running tests locally:

• npx playwright install pulls the headless browsers used by the test;
• (optional) set the test URL in the PW_BASE_URL entry of your .env.

You can also play with Playwright GUI by running npm run test:ui.

External URLs are tested on merge requests using Lychee in a GitHub Action. To test URLs locally install Lychee and run npm run test:absolute-links.

When running the tests locally, the results are in /tests/results:

• {timestamp}.json: JSON report of the test suites;
• html/index.html: HTML report of the latest test suites.

When running in a GitHub action, the “summary” view of the GitHub Action has an artifact section at the very bottom. The artifact archive contains the same HTML report as in html/index.html when you run tests locally.

• Files in /src/public are copied as is (respecting the directory structure in /src/public) in the build directory.
• Lint JavaScript: npm run lint to scan for errors or npm run lint-fix to try an automatic fix.

When adding a new block, sounds must be provided in FLAC and WAV (WAV is the fallback). Then, their volume needs to be adjusted.

ffmpeg -i my-sound.wav -c:a flac my-sound.flac

Normalizing sounds means (in our case, but there is more) aligning the maximum volume of several sounds, so that there’s no sound effect sounding too loud or too quiet compared to the others.

First, run ffmpeg -i my-sound.wav -filter:a volumedetect -f null /dev/null to detect the volume profile (official doc) on your audio files. The audio profile of a file could be:

[Parsed_volumedetect_0 @xxx] mean_volume: -19.3 dB
[Parsed_volumedetect_0 @xxx] max_volume: -2.7 dB
[Parsed_volumedetect_0 @xxx] histogram_2db: 2
[Parsed_volumedetect_0 @xxx] histogram_3db: 8

Among your files, the loudest will have its max_volume closer to 0 dB. You can normalize to whatever file you want by running a command on each audio file in order to increase or decrease their volume by a few dB.

If the audio file sounds too loud (max_volume: -2.7 dB), we can normalize it to another one having (for example) max_volume: -8 dB. The gap between the 2 files is 6 dB (5.3 dB rounded to the upper integer), so we’ll decrease the loudest file by 6 dB by running:

ffmpeg -i my-sound.wav -filter:a "volume=-6dB" my-sound-normalized.wav

Thanks to themushroomkingdom.net and all the people running Nintendo fans sites since the previous century!

Last deployment: