lak

I just moved the standard setting thing to the top of the script, it's still C++14. one of the dependencies didn't like that the standard wasn't being set before it...

I'm struggling to get the environment set up for running these tests on my local machine. if it wasn't the fasm2frames changes then I haven't the slightest clue what could...

> it possibly has different semantics than the original code Technically true. Previously if that address had already been written to it would overwrite it, whereas with `.insert` it will...

Also it looks like with that example data, it's the -good output that has the extra data, not the -bad output? Were these outputs made with or without compression enabled?

`.insert_or_assign` would have been a better way to do this but unfortunately that's a C++17 feature

It should have, the uncompressed codepath should be practically untouched now

The bitread output appears to be the same as the known good version with my test core project. I still haven't managed to run any of the other tests myself.

Is there a reason the tests were cancelled? Did they just time out?

All green now, anything else I need to do?

Any update on this?