v0.1.465

fix: Fix generator variable allocation bug

The code changes in `src/ast/node/control.rs` fix a bug related to variable allocation in generator functions. Previously, when a variable was used in a generator, it needed to be stored in the context. However, there was an issue where the variable was not properly allocated in the context, leading to incorrect behavior.

This commit addresses the bug by correctly allocating the variable in the context when inside a generator. It also adds a check to ensure that the allocation is only performed when necessary.

I also added some code in iter test, to digging deeper into possibility of implement effect system based on generator.

v0.1.464

fix: debug build

v0.1.463

fix: use global variable in generator fn

Previously, using global variables in generator would generate
wrong ir code to load it twice.

This commit also add a new cli flag `nocheck` to skip check stage.

A small change in iter test was made to investigate the possibility of
implementing effect system based on our generator.

v0.1.462

fix: salsa pointer compare issue

See https://github.com/salsa-rs/salsa/issues/536

There's a fix in the main stream, but that fix is using an unstable api
`std::ptr::addr_eq` which will cause compile failure in our ci.

Anyway, I decided to swicth salsa to my custom fork before it went
stable to avoid possible similar issues.

v0.1.461

chore: update salsa to the latest version

Well, previous salsa version has a [serious bug](https://github.com/salsa-rs/salsa/issues/473)
which makes lsp deadlock from time to time,
I've fixed it in April, but somehow forget to update salsa orz.

v0.1.460

fix: skip repl in mac test

v0.1.459

fix: deconstruct assignment doesn't work on tuple

this commit disabled most of implicit cast logic to ensure the type inference logic is
working as expected

v0.1.458

feat: more escapable types && fix docs

v0.1.457

fix: a panic case & add some docs

v0.1.456

fix: repl wasm