Distinct labels flagged as duplicate

[lhstrh]

This bug has been found by @lsk567. The following code yields the error Duplicate AttrParm 'value':

reactor Source {
  output out: int
  timer t(1 nsec, 10 msec)
  state s: int = 0

  @label(value="Foo")
  reaction(startup) {= lf_print("Starting Source"); =}

  @label(value="Bar")
  reaction(t) -> out {=
    lf_set(out, self->s++);
    lf_print("Inside source reaction_0");
  =}
}

AFAIK, this should be equivalent to the following code, which does not yield errors:

reactor Source {
  output out: int
  timer t(1 nsec, 10 msec)
  state s: int = 0

  @label("Foo")
  reaction(startup) {= lf_print("Starting Source"); =}

  @label("Bar")
  reaction(t) -> out {=
    lf_set(out, self->s++);
    lf_print("Inside source reaction_0");
  =}
}

The problem appears to me that uniqueness checks are done at the level of the eContainer, and while the attributes are associated with the syntactic element they precede, which are different, they in the same container. There must be a way to override this behavior but I'm not sure how. Perhaps @a-sr knows?

[a-sr]

We already use an adjusted uniqueness validator for LF, so this would be the point of entry to fix this. The error is created by checkDescriptionForDuplicatedName (from NamesAreUniqueValidationHelper), so we need to override this method, but I have to take a closer look at how it works.

[lhstrh]

Ah, I was looking at that class but didn't understand which method we needed the override. Thanks!

[lhstrh]

I noticed that checkDescriptionForDuplicatedName is deprecated. Are you sure this is the method to override?

[a-sr]

Yes, I checked it with the debugger. Apparently, the fact that we use a costomized implementation leads us into the deprecated API but don't ask me why they implemented it like this.

[lhstrh]

The implementation looks a bit opaque to me. Would you know how to fix it?

[a-sr]

Not yet, I would have to take a closer look, but I can do that.

[lhstrh]

@a-sr, have you been able to find out anything regarding this issue?

[a-sr]

It turns out to be a bit more tricky than I thought, but I am on it.

[a-sr]

It turned out easier than I thought, see #2147