Pietro Albini

We currently rely on Docker to detect OOMs. If the implementation is not huge I'd love a PR to fix this.

Unfortunately that's not a OOM Crater can currently detect. The current detection code checks whether the kernel had to OOMKill a process, while that error is the allocator refusing to...

My worry with using that flag by default is that some crates could output *a lot* of output, making logs either too big or too unreadable to be useful.

I'd be OK with merging such PR.

A new rustwide will need to be released and Crater will have to be updated to use it.

Yeah, totally forgot about this. I'll think it over and write the mentoring instructions sometime in the future.

Yeah, I thought about this and I think the only viable option is to move the check to the agents. A flag `filter-reverse-dependencies=lazy_static` (bikesheddable) can be added to the experiment...

So, at the moment the OOM detection is done by asking the Docker daemon whether that container OOMed. If in some cases it doesn't detect that we'll need to resort...

@hdevalence the next one needed is https://github.com/rust-lang/crater/issues/443

> Building the docker container locally takes forever, so it'd be great if we could just publish it ourselves (as a tarball or something). Wouldn't this work? ``` docker pull...