refactor: use Idx<T> instead of macro in arena

[baszalmstra]

This PR refactors the Arena Id's a little bit so that we need fewer macros and get a clearer coupling between Arenas, their data and their ids. Similar to how rust analyzer refactored this.

I also renamed a few Ids to indicate that they are "Local" to their container (TypeRefId is local to a specific TypeRefMap and is now called LocalTypRefId). I found that to be clearer.

[codecov]

Codecov Report

Merging #266 into master will increase coverage by 0.33%.
The diff coverage is 76.38%.

@@            Coverage Diff             @@
##           master     #266      +/-   ##
==========================================
+ Coverage   78.75%   79.08%   +0.33%     
==========================================
  Files         215      214       -1     
  Lines       12857    12803      -54     
==========================================
  Hits        10125    10125              
+ Misses       2732     2678      -54     

Impacted Files	Coverage Δ
crates/mun_codegen/src/ir/body.rs	84.23% <ø> (-0.14%)	⬇️
crates/mun_hir/src/adt.rs	90.00% <ø> (ø)
crates/mun_hir/src/resolve.rs	86.11% <ø> (ø)
crates/mun_hir/src/source_id.rs	80.00% <ø> (ø)
crates/mun_hir/src/arena/map.rs	64.28% <61.90%> (-9.05%)	⬇️
crates/mun_hir/src/arena.rs	59.57% <72.72%> (-4.32%)	⬇️
crates/mun_hir/src/expr.rs	87.65% <75.00%> (ø)
crates/mun_hir/src/expr/scope.rs	93.75% <88.88%> (ø)
crates/mun_hir/src/type_ref.rs	61.36% <88.88%> (ø)
crates/mun_hir/src/code_model.rs	79.73% <100.00%> (ø)
... and 13 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update d9ebc18...9c93702. Read the comment docs.

[Wodann]

My only concern is the reduced codecov, otherwise it looks good.

It seems that codecov already has some good recommendations for improving the coverage.

[baszalmstra]

My only concern is the reduced codecov, otherwise it looks good.

It seems that codecov already has some good recommendations for improving the coverage.

Oh I only just noticed you also commented. The code coverage of the overall project actually increases because the MR removes a lot of code. Other than that (besides the Arena code, which is tested) I only really renamed stuff so I didn't feel like adding more tests would be necessary.

I do see that there are a few functions in there that dont seem to be used at all.