[
https://issues.apache.org/jira/browse/GROOVY-11159?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17756010#comment-17756010
]
Eric Milles commented on GROOVY-11159:
--------------------------------------
Yes I think in a vacuum there would be separate types for TypeDefinition and
TypeReference. I'm trying to work out how to separate out the typedef
information when isPrimaryNode is modified by ClassHelper and subclasses.
> ClassNode: separate type definition information
> -----------------------------------------------
>
> Key: GROOVY-11159
> URL: https://issues.apache.org/jira/browse/GROOVY-11159
> Project: Groovy
> Issue Type: Improvement
> Reporter: Eric Milles
> Priority: Minor
>
> ClassNode is used to represent type definitions as well as type references.
> There are a lot of data members that apply only to type definitions.
> However, every reference pays the cost. ClassNode could have a single
> reference to another class (essentially a struct) that holds all the
> definition information. With this, type references would only have one
> reference (initialized to null) instead of all those fields.
> Fields:
> * modifiers
> * syntheticPublic
> * superClass
> * interfaces
> * mixins
> * objectInitializers
> * constructors
> * methods
> * methodsList
> * fields
> * fieldIndex
> * properties
> * module
> * compileUnit
> * staticClass
> * scriptBody
> * script
> * innerClasses
> * lazyInitDone
> * lazyInitLock
> * clazz
> * enclosingMethod
> * recordComponents
> * transformInstances?
> * permittedSubclasses
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
