Use loose template literals to avoid bug in nft

[superdav42]

vercel/nft#314

This PR will fix proskomma from breaking next js builds. The option is documented in https://babeljs.io/docs/en/babel-plugin-transform-template-literals#loose . I'm really not familiar with the downsides of using loose but the tests pass just fine. I'm guessing it has potential to break old browser or node versions but I couldn't find details on which ones may be affected.

[DanielC-N]

Hello,
For this, please see these ressources :

• Babel "loose" already exists in presets env so you don't have to use @babel/plugin-transform-template-literals
• Also there is a problem with the usage of "loose" : in case it's used in other technologies, "loose" option may result in other non-wanted behavior and cause conflicts in the future. Please see this : https://github.com/babel/rfcs/blob/main/rfcs/0003-top-level-assumptions.md#motivation

Try this babel.config.js and see if this solves the problem:

module.exports = {
  "assumptions": {
    "noDocumentAll": true,
    "pureGetters": true,
    "iterableIsArray": true,
    "ignoreToPrimitiveHint": true
  },
  "presets": [
    [
      "@babel/preset-env",
      {
        "targets": {
          "esmodules": true,
        },
        useBuiltIns: "entry",
        corejs: "3.25.1"
      }
    ],
  ],
};

I think what you're looking for is "ignoreToPrimitiveHint"

[mandolyte]

It did seem to get further, but after some minutes, it has not completed. Looks to be hung again. Here is the babel.config.js I have ended up with.

module.exports = {
  "assumptions": {
    "noDocumentAll": true,
    "pureGetters": true,
    "iterableIsArray": true,
    "ignoreToPrimitiveHint": true
  },
  "presets": [
    [
      "@babel/preset-env",
      {
        "targets": {
          "esmodules": true,
        },
        useBuiltIns: "entry",
        corejs: "3.25.1"
      }
    ],
    [ "@babel/preset-react"]
  ],
};

[superdav42]

I think what you're looking for is "ignoreToPrimitiveHint"

Yes, that's exactly what we need. I just updated this PR to the config you proposed and it solved the problems we were having with the concat() chains.