Re: [w3ctag/design-reviews] Multiple import maps (Issue #980)

OK, with respect to non-determinism, it would be good to have some text on that in the specification.  Or even some way to manage it (can a script that performs dynamic loading test whether a particular import map has been seen?).

The spec PR seems to be a little out of sync with what you just said (it's written in the usual impenetrable language, so I'm not 100% here) with what you say.  It says something about the map being fixed and then propagated throughout each top-level module load.

>> If a mapping is used, but does not resolve successfully (perhaps because no mapping exists) is that a state that needs to persist? Why is that safe?

I think that you indirectly answered this one with your non-determinism answer.  My question, expanded:

It appears as though there is a requirement that the map, once queried, cannot change.  Does that apply only in cases where the query results in a successful resolution, or does it apply in other cases:

1. An attempt is made to load "foo" and "foo" doesn't exist in the map.  Does that mean that a future addition to the map cannot add a mapping for "foo"?
2. An attempt is made to load "foo" and "foo" exists in the map, but the resolution fails (an HTTP 404 error on the indicated resource, say).  Does that make "foo" available for overriding in a future addition to the map?

These are not particularly obvious to me from the explanatory material, though I'll admit to not having gone through all of your updated examples.

-- 
Reply to this email directly or view it on GitHub:
https://github.com/w3ctag/design-reviews/issues/980#issuecomment-2415656822
You are receiving this because you are subscribed to this thread.

Message ID: <w3ctag/design-reviews/issues/980/2415656822@github.com>