Why don't we just store a lookup for each import and then traverse the lookups also for the imports. When doing that I already think about caching but maybe that will also be combinable with the cache idea in #2 ?
This idea originally came to my mind when I thought about whether the current lookup tables are just taking too much memory when stuffing all the imports from the other modules in. But I think it actually doesn't..
The question mostly is what saves us more nerves and performance
Open questions:
How would we handle/discover loops/cycles in imports?
Alternative to #5
Why don't we just store a lookup for each import and then traverse the lookups also for the imports. When doing that I already think about caching but maybe that will also be combinable with the cache idea in #2 ?
This idea originally came to my mind when I thought about whether the current lookup tables are just taking too much memory when stuffing all the imports from the other modules in. But I think it actually doesn't..
The question mostly is what saves us more nerves and performance
Open questions: