Ruben Taelman

Some overhead due to the federated actor was to be expected. But 12% is a bit much indeed, we should be able to optimize this.

Sure, but may not be needed if we'd be able to reduce the overhead to ~1%.

Good idea! Would be great in combination with the new HTTP timeout functionality @Tpt is working on #70.

Yes, definitely! PR is certainly welcome. Can probably be done in a backward-compat manner I guess.

Yeah, I also spent some time on getting this to work a while ago, but gave up because it was taking up too much time (and there were more urgent...

Thanks @jaxoncreed! I will look into this soon!

> @pietercolpaert Yes, it would seem that this would only work in a NodeJS environment, so I guess you should have a separate configuration for an in-browser environment. Could you...

Note to self: before merging, we should do an in-depth manual check to see if everything still works correctly with this. For example, we should test and inspect TPF queries.

The reason we use `.transform` instead of direct `map` and `filter` calls, is because of the `autoStart` flag. If we would move to `map` and `filter`, then they'd need a...

It looks like your tsconfig settings are much stricter than comunica's, which causes your compiler to complain about things in the linked comunica project. One solution may be to (temporarily)...