Make `raise` statements show the correct function name

[angelcaru]

Closes #137

[Almas-Ali]

This should also accept just function name. Ex. raise somefile.TypeError without ().

[Almas-Ali]

It would be nice to get like this:

Radiation (most recent call last):
  File ./test_lib.rn, line 3
TypeError

raise rlib.TypeError
      ^^^^^^^^^^^^^^

if called rlib.TypeError this only. no arguments and no default will be executed while no ().

[angelcaru]

What, like this?

TypeError = "type error"
raise TypeError

instead of

fun TypeError() {
    return "type error"
}
raise TypeError

[Almas-Ali]

What, like this?

TypeError = "type error"
raise TypeError

instead of

fun TypeError() {
    return "type error"
}
raise TypeError

Radiation (most recent call last):
  File ./test_lib.rn, line 6, in <program>
RuntimeError: Illegal operation for ("type error", "type error")

raise TypeError
      ^^^^^^^^^

[angelcaru]

OK, now I'm even more confused

[angelcaru]

So if the function returns null don't print a message? So

fun TypeError() {
    return "type error"
}
raise TypeError

would give TypeError: type error, while

fun TypeError() {} # implicitly returns null
raise TypeError

would give TypeError?

[Almas-Ali]

When it is not defined yet, it raise errors 2 times. It could be for the super() call on error file.

Radiation (most recent call last):
  File ./test_lib.rn, line 3, in <program>
Radiation (most recent call last):
  File ./test_lib.rn, line 3
RuntimeError: Attribute 'rrError' does not exist

raise rlib.rrError()
      ^^^^^^^^^^^^

raise rlib.rrError()
      ^^^^^^^^^^^^

[Almas-Ali]

Some times it leaves closing parent on error.

[angelcaru]

? Can't reproduce

[Almas-Ali]

? Can't reproduce

What do you mean?

[angelcaru]

What is even the issue?

[Almas-Ali]

What is even the issue?

Check the last error message on the screenshot.

[angelcaru]

It looks fine to me...

[Almas-Ali]

It missed the closing parenthesis.

[angelcaru]

That's such a minor thing though. It doesn't really matter

[Almas-Ali]

That's such a minor thing though. It doesn't really matter

Well you are right. But this looks weird to me.

[angelcaru]

We have all sorts of similar issues with error reporting. I don't see a reason why this one specifically needs to block the PR. For instance:

[Almas-Ali]

We have all sorts of similar issues with error reporting. I don't see a reason why this one specifically needs to block the PR. For instance:

I am not blocking the PR here. Just wanted to share my opinion on this related topic.

In this case we need to identify this issue with more various perspective and create a new issue linking all found bugs on radiation traceback formatting to fix it.