* Make Java nodes text-available
* Introduce shared JavaccToken in pmd-core
* Use factory to produce char streams
Tests are still on java-grammar,
since they use the DSL & newer
AST structure.
This is to prepare for other changes
that concern all javacc languages and
should not be done on java-grammar
* Remove `Dimensionable`, remove its methods from the former implementations (except from ASTArrayDimsAndInits, which is itself deprecated)
* The varargs ellipsis is now represented as an ArrayTypeDim.
* This affects FormalParameter and LambdaParameter
Closes#997. All forms of type annotations are now supported.
Among the different possible categorisations
of unary expressions, this is probably the
most logical and easiest to document.
Here's a comparison of different possible
categorisations. Note: `_++` is the postfix
increment operator, while `++_` is the prefix
one - idem for decrement. The last one is the
one implemented by this commit.
\## 6.0.x
```
Unary = { -, + }
UnaryNotPlusMinus { !, ~ }
PreIncrement = { ++_ }
PreDecrement { --_ }
Postfix { _++, _++ }
```
* Very assymmetric, splits operators based on parsing
concerns
\## Before #1890:
```
Unary = { -, + , !, ~ }
PreIncrement = { ++_ }
PreDecrement { --_ }
Postfix { _++, _++ }
```
* Minor simplification
\## #1890:
```
Unary = Prefix \ { ++_, --_ })
Increment ( { ++ , -- } x (postfix, prefix) )
```
* Names are weird (IncrementExpr may be decrement, Unary != Increment
even though semantically, Increment \subset Unary)
* No possibility to introduce a supertype (what would it be?)
* But easy to match all increment/decrement expressions
\## JLS (also, Eclipse):
```
Prefix = { !, ~, -, +, ++_, --_ }
Postfix = { _++, _-- }
```
* Both can have super interface UnaryExpr
* This allows matching all increment/decrement expressions easily too
* Easiest to document, JLS like, AST like
* Fits well with `InfixExpr`
