Vinicius Lourenço

Nothing more sad than a green CI and a conflict, @RedYetiDev can you rebase?

@GeoffreyBooth I changed to `TIMING_MODULE_CJS`, we can also add support for `esm` under `TEST_MODULE_ESM`, and then we can measure both cases. I didn't found any good alternative for this name,...

@richardlau Maybe I'm not familiar with trace_events but from what I see in the source code, I think it won't be as useful as having it as `NODE_DEBUG` since to...

@joyeecheung I changed the implementation to match what we discussed, I liked this new way, it will print in the `stderr` if the `NODE_DEBUG=module_cjs` or it will trace if the...

 If we use the `chrome://tracing`, we can see what module is loading which module (I think we can improving this visualization).

I fixed the tests and I also changed a little bit the visualization for `trace`, I updated the PR to include the new graph and the new logging example.

I moved the new code entirely to `debuglog` as suggested by Joyee, I added more tests to ensure both `trace-events` and `NODE_DEBUG` are working correctly, and I also checked only...

@joyeecheung Should we support ending the trace even if the `_load` throws an unexpected error? I was creating a trace for `pnpm` and for some reason, the trace was not...

From what I debug, the error is thrown when the module is not found. Also, I read the source code of the functions that are called and there are other...

I added the `wrapModuleLoad` to implement the logic of tracing & recursion tracking. But I also noticed a lot of places where `_load` is called directly, should we rewrite those...