JSON config file

[KristjanESPERANTO]

I haven't dealt with inlang that deeply yet, but I find it very exciting and am considering using it 🙂 Thank you in advance!

The first question I asked myself was why the configuration happens in a JavaScript file. Wouldn't it be a better approach to separate the data from logic in a separate JSON file?

[samuelstroschein]

Off topic

I find it very exciting and am considering using it 🙂 Thank you in advance!

Always exciting to read messages like this one!

How can I help you set up inlang?

[KristjanESPERANTO]

First I need to find the time to try an implementation on a small project. You can't help me before then. But thanks for the offer! I might come back to that later 🙂

[samuelstroschein]

The config file is purposefully JavaScript to enable programmatic control and plugins.

I observed that every team localizes differently. Not giving teams programmatic control and an easy way to leverage plugins to configure inlang would severely limit integration possibilities. A few examples:

• the file format of resources vastly differs e.g. JSON, YAML, ARB, PO.
• detecting messages in various source codes (React, Flutter, iOS, HTML, JS, etc) is a complex task that we can't cover alone but by providing a JS config, developers can write their own solutions.

[samuelstroschein]

Here's an illustration of how complex the config can become and why the full power of code is beneficial in #376. (We assume the following will be hidden in plugins).

The code detects the referencing of messages in source code:

ideExtension: {
      messageFinders: [
        async (/** @type {{ "documentText": string; }} */ args) => {
          const regex = /{?t\(['"](?<messageId>\S+)['"]\)}?/gm;
          const str = args.documentText;
          let match;
          const result = [];

          while ((match = regex.exec(str)) !== null) {
            const startLine = (str.slice(0, Math.max(0, match.index)).match(/\n/g) || []).length + 1;
            const startPos = match.index - str.lastIndexOf('\n', match.index - 1);
            const endPos = match.index + match[0].length - str.lastIndexOf('\n', match.index + match[0].length - 1);
            const endLine = (str.slice(0, Math.max(0, match.index + match[0].length)).match(/\n/g) || []).length + 1;

            if (match.groups && 'messageId' in match.groups) {
              result.push({
                messageId: match.groups['messageId'],
                position: {
                  start: {
                    line: startLine,
                    character: startPos
                  },
                  end: {
                    line: endLine,
                    character: endPos
                  }
                }
              })
            }
          }
          return result;
        }
      ],

[KristjanESPERANTO]

Okay, that's an interesting approach. Thanks for the explanation! 💡

[samuelstroschein]

Surprise #1140