Techcable

I don't believe the "applying zero offset to null pointer" bug is very serious. The one that does look interesting (besides integer overflow) is: ``` src/core/specials.c:207:88: runtime error: left shift...

In #255, @bakpakin mentioned an "original peg debugging script", but I can't find that anywhere :( I learned about [Margaret](https://gitlab.com/sogaiu/margaret) after filing this issue, and the debugging tools there look...

Hmm, so maybe this is a Mac specific bug then? Probably something to do with transition of `errno` There's a fair chance Mac diverges from the POSIX spec here (although...

I have removed `contains-value?`, removed the type error (it now returns false) and changed to support all iterable types. Note that the builtin `next` function throws a type error if...

> Remove the try. In dynamic languages, the usual idea is garbage in, garbage out. You misunderstood my point about the type error. “Test” functions are not special in that...

Unit tests added, documentation cleaned up & corrected, `try` wrapper removed. Should be very close to final submission (or dismissal). Remaining tasks: - Final check over documentation - Final rebase?

Agreed. I definitely need to clean up (and reduce the size of) the docstrings. The reference to `null` was just a sloppy mistake on my part. > names like `has-key?`...

Yes. Keep in mind that the original DynASM project for C (used as backend in LuaJIT 1.x, used as frontend in ) does *not* use a hash map for labels....

This should absolutely work with a JIT :smile: In fact, that's my intended use case. Right now multi-threading support doesn't really work correctly, though I'm working hard on it. The...

@playXe That is correct. A basic_safepount implicitly blocks the thread if it needs to do a collection. The safepoint does a quick comparison, comparing the current size of the heap...