Allowing adding mfas to ignore for stacktrace logging

[Trond M]

I find the `stacktrace: true` option for my repo to be extremely helpful when working on large applications, but I was wondering if we could perhaps make it even more helpful by allowing the user to ignore certain calls in the stacktrace?

My problem is specifically with the `assign_new` from Phoenix.LiveView.Utils which wraps a large part of my database calls.

I see from the current implementation that it returns the last call before the repo call

```

defp last_non_ecto([{mod, _, _, _} | _stacktrace], repo, last)
       when mod == repo or mod in @repo_modules,
       do: last

  defp last_non_ecto([last | stacktrace], repo, _last),
    do: last_non_ecto(stacktrace, repo, last)

  defp last_non_ecto([], _repo, last),
    do: last

```

I tried adding this as a test:

```

defp last_non_ecto([{mod, fun, arity, _} | _stacktrace], repo, last)
       when mod == repo or mod in @repo_modules or {mod, fun, arity} == {Phoenix.LiveView.Utils, :assign_new, 3},
       do: last

```
and it immediately became more helpful for me. 

I'm not sure of the API here, but would it be possible to allow the user to add mfas to ignore here?

[Greg Rychlewski]

Maybe we can give an option to let the user specify an extraction function. Otherwise it defaults to last_non_ecto

[Trond M]

Yes, that sounds like a good way to solve this!

[Felipe Stival]

Hi,

Maybe it would be better if we supported more stacktrace "levels" instead of custom filtering? Then you could get something like:

```

└ Phoenix.LiveView.Utils.assign_new/3

    └ MyApp.SomeLive.handle_event/3

```

I can see it being more helpful than the current log, and it's probably easier to configure for users. Maybe we can even default to 2.

--
You received this message because you are subscribed to the Google Groups "elixir-ecto" group.
To unsubscribe from this group and stop receiving emails from it, send an email to elixir-ecto...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/elixir-ecto/a21c3a18-5d2c-416c-b2b9-e7f18e9647e2n%40googlegroups.com.

[Trond M]

That's a good point! I like that solution