"Unbounded type variable" definition may need clarification

[ljnelson]

I'm not trying to be pedantic, but the term "unbounded type variable" shows up in the spec. Strictly speaking it is nonsensical since type variables have one or more bounds.

I assume it means something like, simply, T. Does that mean that T extends Object is also supposed to be considered an "unbounded type variable" (I assume yes)? There are probably other edge cases I'm not thinking about right now.

[ljnelson]

(Weld says that an "unbounded type variable" is one whose reflective representation has null bounds, an empty bounds array, or a bounds array that has exactly one element whose type is Object.)

[ljnelson]

Related: https://redhat.atlassian.net/browse/CDI-391