[
https://issues.apache.org/jira/browse/GROOVY-8275?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16115509#comment-16115509
]
John Wagenleitner commented on GROOVY-8275:
-------------------------------------------
For #2 and #3, looks like adding local variable support was part of GROOVY-4876
as a way to [enable usage in
scripts|https://github.com/apache/groovy/commit/a0c8d2418286d46182ee647d5f3c6d48212a508f#diff-43aa4979f66a79c1de9464bf9059cf4eR97].
Based on GROOVY-4876 I'm not sure if this is by design; that the presence of
the annotation on a local variable turns it on for all others in the script.
For example:
{code}
@Newify
def x = Integer.new(1)
def y = Integer.new(2)
def z = Integer.new(3)
assert x + y + z == 6
{code}
It does lead to unexpected results when used in a Class.
{code}
class Foo {
int bar() {
def x = Integer.new(1)
}
int baz() {
@Newify
def y = Integer.new(2)
}
}
assert new Foo().bar() == 1
assert new Foo().baz() == 2
{code}
> Newify transform docs missing support for local variable declaration
> --------------------------------------------------------------------
>
> Key: GROOVY-8275
> URL: https://issues.apache.org/jira/browse/GROOVY-8275
> Project: Groovy
> Issue Type: Documentation
> Components: xforms
> Reporter: Eric Milles
> Priority: Minor
>
> Javadocs and online documentation for the Newify AST transform is not quite
> clear in a couple instances.
> 1) Online docs http://groovy-lang.org/metaprogramming.html#xform-Newify
> Ruby-style section could be more clear IMO if the {{([Tree,Leaf])}} were
> removed since it only enables the Python-style expressions.
> 2) Javadoc and online docs do not mention that {{@Newify}} can be applied to
> a local variable declaration, like: {{@Newify def o = Object.new()}}.
> 3) The scoping of the transform is pretty clear for the class, method and
> field/property cases. However for the local variable case, it appears that
> the transform applies throughout the containing scope of the local variable.
> That is, I can use the extra constructors anywhere in the block to
> re-initialize the variable. {{@Newify def o = Object.new(); ... o =
> Object.new() ...}}. Is this the intended behavior? My intuition was that
> the initializer expression was the only place the transform applied, like in
> the case of a field.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
