I already made some progress in #64 You can take some inspiration from it & close #64

while fixing the failing tests, I found that running cargo test -p metassr-bundler --lib results in a segfault.

The weird thing is that loading the bundle script in the actual code doesn't bring any problem like that.
This behavior requires some investigation. However I opened #73 to track that.

CC: @viferga @hulxv

We need to fix that also
it's related to the watcher

https://github.com/metacall/metassr/actions/runs/21854867036/job/64415054320?pr=68#step:7:292

error[E0308]: mismatched types
  --> crates/metassr-watcher/src/lib.rs:40:26
   |
40 |         Ok(FileWatcher { watcher, sender })
   |                          ^^^^^^^ expected `Debouncer<FsEventWatcher, NoCache>`, found `Debouncer<FsEventWatcher, FileIdMap>`
   |
   = note: expected struct `Debouncer<_, NoCache>`
              found struct `Debouncer<_, FileIdMap>`
              

Regarding the segfaulted test #73 .. after some debugging with Codex, I found that running with
NODE_PATH=/home/fahd/metassr/crates/metassr-bundler/tests/node_modules/ cargo test -p metassr-bundler
and insuring we have the dependencies installed solved the problem.

@viferga @hulxv I can add the env variable to dockerfile but are you sure that this is the intended behavior from MetaCall?
Should we ignore this test bundling_works like we do here crates/metassr-build/src/client/mod.rs, until we have full node enviroment to run it in.

I removed the second metacall init because it's not needed there.

Thanks @hulxv for your effort. we can merge this now. other macos & windows problems we face now is on MetaCall distributables side and will be fixed soon, i should open an issue with this. cc: @viferga