baseURL

[guybedford]

I know it's no longer strictly necessary, but consider running the browser loader in NodeJS for example - it can be useful to be able to set what the base path should be for normalizing unnormalized URLs.

It can also be nice to be able to do the following in the browser:

System.baseURL = '/js/';
System.import('blah');

And to be able to control that resolution. I think it's an important feature to keep here.

[guybedford]

Also perhaps it's about time to call it baseURI rather?

[domenic]

I think this feature should be provided by custom loaders, not the built-in ones. Also, it's URL; URI is an obsolete term that doesn't correspond to anything in real software (i.e. all software just uses a single parser, known as the URL parser).

[guybedford]

So we make it stick to document.baseURI / process.cwd(), and then if users want more than that they can add a custom resolve step?

[domenic]

Yeah. I don't think there should be a loader-specific way of customizing the base URL of a document. It would be weird to have that apply only to resources loaded through the loader (JS files, maybe HTML imports, what else?) but not others (CSS files, images, ...?)

[guybedford]

Ok sure I can go with that - I guess it is specifically the NodeJS case that makes it necessary, and a custom resolve can make it work out there fine, or even chdir.

[justinbmeyer]

but not others (CSS files, images, ...?)

There is a way https://developer.mozilla.org/en-US/docs/Web/HTML/Element/base

[domenic]

Yes, that was precisely my point, base element for some resources and loader baseURL for others is bananas.

[justinbmeyer]

Thinking about it some more ... is a design goal of this spec to work independent of environment? Outside of the browser, I could see how some "base" would be extremely important.

The Node case does make it necessary. How do you balance influencing factors of node vs browser vs Rhino vs Nashhorn etc?

[domenic]

My point of view is that base is an environment-specific thing, just like the actual mechanics of translating a URL or filename into a stream of bytes is an environment-specific thing. In node it is process.cwd() modifiable with process.chdir. In the browser it's document.baseURI modifiable with the <base> tag. In Rhino it's something else.

[caridy]

@justinbmeyer if you need some sort of custom base, you can rely on sites() to define the mapping for some of your files.

If we introduce baseURL, you have this big problem:

<script>
System.baseURL = '/static/whatever/';
System.import('path/to/foo.js');
</script>
<script src="path/to/foo.js" type="module"></script>

That's very confusing, because you break the reflection between the script type module and the Imperative api to load the exact same module since you have the ability to modify that value. Even more important, this breaks the mental model that developers have today when they think about fetching things over the network.

Essentially, we want to stick to the principle described by @domenic.