Aidan Lee

Yes, I remember thinking it might make sense to move it into another file when I was originally working on it. Hopefully once I get around to replacing my custom...

This is one of those undocumented pitfalls when using `cpp.Struct` and similar, there's no real way around it. I do have new extern techniques on the way which aims to...

Hi Hugh, Thank for the info. What I've been doing for large byte arrays is to root it and then copy chunks of the arrays data into a `std::vector` and...

yeah, modifying the array after sending to write is a tricky one which can only really be solved by copying the entire array when the user makes the write call....

Thanks for the tips, I see what you mean about the large list locking. Perhaps it makes sense to come at this from a different angle then, if arrays had...

OK, I've completely re-jigged this merge. MemType changes have been reverted and it adds pinning spport. Pinning and unpinning checks the header to see if it lives on the LOH...

I thought of the atomic operations just as I was going to bed, I've changed things around now though so that `mPinnedAllocs` isn't needed any more. There is now an...

Yeah, thats fair. The only way to solve it without the extra copy would probably be to introduce an extra BytesData type instead of that being a typedef to an...

Probably should have opened this as a draft to begin with. I'm not planning on adding anything else so if this looks good its probably good to merge.

~~These failing GC tests are very odd. Take the `testFunc` test, the object passed into the do not kill is correctly identified as a closure object, however after a GC...