Dennis Felsing

I think this might come from https://github.com/MaterializeInc/materialize/pull/24161

@petrosagg This is still failing commonly in CI, have you had a chance to take a look? https://buildkite.com/materialize/nightlies/builds/2446#018845cf-6c66-435f-8382-f90b62c38c8d

This is still happening often: https://buildkite.com/materialize/nightlies/builds/5665#018c8358-86f9-4580-9830-a6a55539120f

I don't think it should be part of DDNet. You can work on it independently.

This is probably a result of https://github.com/ddnet/ddnet/issues/7917#issuecomment-1987236079

Yeah, but that code is very experimental and doesn't even work correctly. I've heard that Araq is working on making this work in a proper way.

@bkirwi Just to make sure, is https://github.com/MaterializeInc/materialize/pull/23279/commits/776eb3936fd9ad6f2a8a31970db9d73f85b3d5c8 correct? Might the vectors be still used somewhere later, referencing the already overwritten content?

Seems to happen quite often, just saw it again in the same test: https://buildkite.com/materialize/tests/builds/69074#018bdfa0-a7a0-4bc2-9a27-b59ea03db595 We might want to revert so tests pipeline stays green.

I couldn't repro this locally, but it happened again in CI: https://buildkite.com/materialize/nightlies/builds/6803#018e18e5-e2ce-4bc0-b42c-b483e8180d59

Looks like walkDir changed its definition and you also need to pass the parameters relative & checkDir. Maybe like this: ```nim for f in concat(walkDir(".", true, false), walkDir(".", true, false)):...