Others

I think you may need to test this in an actually no-std environment. I think (off the top of my head) that some of our dependencies (in particular parking-lot) may...

Thought about this more, here are my concerns: a) I consider this blocked on having a CI step that runs the tests in a no-std environment so we continuously validate...

There are two background threads--the collector thread and the dropper thread. Both exist as optimizations to prevent stopping compute--they shouldn't be necessary. I think that replacing the dropper thread should...

Your test is not doing what you think it is. Even if this crate is no-std, our dependencies may not be. You should add an additional step to build the...

`rustup target install thumbv6m-none-eabi; cargo build --no-default-features --target thumbv6m-none-eabi` I think that's the command you want to ensure that we build on `no-std` targets :)

@setupminimal I do like the work you have done here so far, but there is a ways to go. Do you want to keep iterating on this, or should I...

prevent collection should optionally block if there is a collection happening

Cool! I think we want `prevent_collection` to block until there is no collection going on. And when the handle drops, we want to make sure the collector has a chance...

@alekratz Was thinking about this a bit more. Realized a few things: 1) If `Collector::gc_lock` becomes a RWLock, that might be a start of the implementation 2) If we implement...

Has some tests now, still needs audited for races