feat: add finally return support

[jameslahm]

Fix #7

[playXE]

Oh wow! That's nice work! But seems like there's some regression as now half of previously passing test262 tests do not pass:

Results:
Total tests: 78695
Passed tests: 6187
Ignored tests: 24233
Failed tests: 48275 (panics: 1497)
Conformance: 7.86%

dev branch:

Results:
Total tests: 78695
Passed tests: 12579
Ignored tests: 24233
Failed tests: 41883 (panics: 1244)
Conformance: 15.98%

EDIT: You can see how much tests pass in Actions

[jameslahm]

Hi, I encountered a segment fault error when I run sl examples/hello-world.js without .startupshot. I feel like there has some problem in add_conservative in gc.rs. Do you have this problem?

[jameslahm]

It seems that I found it. I will push and try to run test262.

[playXE]

Hi, I encountered a segment fault error when I run sl examples/hello-world.js without .startupshot. I feel like there has some problem in add_conservative in gc.rs. Do you have this problem?

Unfortunately, I will not have access to PC for one or two days so Idk where the problem is. I actually tested the new GC when I was merging it and hello-world was working fine. Can you please check that you run the latest dev branch GC?

[jameslahm]

The problem should be here. And I have added the index check. It works fine now.

starlight/crates/starlight/src/gc/space_bitmap.rs

Lines 84 to 93 in df79b32

	pub fn test(&self, object: *const u8) -> bool {
	let object = object as usize;
	let offset = object as isize - self.heap_begin as isize;
	let index = Self::offset_to_index(offset as _);
	let mask = Self::offset_to_mask(offset as _);
	if index >= self.bitmap_size / size_of::<usize>() {
	return false;
	}
	let atomic_entry = unsafe { *self.bitmap_begin.add(index) };
	(atomic_entry & mask) != 0

EDIT: I have checked that I am on the latest dev branch

[jameslahm]

And I continue to look at the problem of regression testing 🤣