Dwight Guth

From what I can tell, this doesn't actually affect the coverage of the code at all; the only reason it shows the coverage going down is that there are some...

Well, if your main concern is eliminating overhead when the user doesn't care about this additional information, that's going to be difficult to completely eliminate unless everything to do with...

@arximboldi did you have any preference for or against either of the solutions I proposed here?

Looks like your CI is no longer working. We had the same problem with nix a couple days back. It seems to be an incompatibility between the latest version of...

I've tested this code now with both `make check` and against our own integration tests, and it seems to provide the necessary functionality. I intend to write some unit tests...

This is the code that makes use of this functionality: https://github.com/kframework/llvm-backend/blob/immer_update/runtime/collect/migrate_collection.cpp Essentially, it needs to be able to take each object on the heap containing a collection or iterator, and...

Looks like this is still failing CI due to the SSH key issue, but it is now ready for review again.

@arximboldi This is still failing due to the missing SSH key even though #193 is now merged in. Are you sure you correctly solved the issue?

I mean, there's room for maybe putting some of the traversal stuff in our code into the library, I agree. That doesn't really affect this particular change though because the...

Well, it's certainly the case that the relocation info struct is specifically designed for use by a relocating garbage collector, but any garbage collector which moves objects will need this...