groovy-notifications mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Paul King (JIRA)" <j...@apache.org>
Subject [jira] [Resolved] (GROOVY-8300) Groovy should have an @AutoFinal annotation
Date Tue, 12 Sep 2017 06:17:03 GMT

     [ https://issues.apache.org/jira/browse/GROOVY-8300?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Paul King resolved GROOVY-8300.
-------------------------------
       Resolution: Fixed
    Fix Version/s: 2.5.0-beta-2

Proposed PR merged

> Groovy should have an @AutoFinal annotation
> -------------------------------------------
>
>                 Key: GROOVY-8300
>                 URL: https://issues.apache.org/jira/browse/GROOVY-8300
>             Project: Groovy
>          Issue Type: New Feature
>            Reporter: Paul King
>            Assignee: Paul King
>             Fix For: 2.5.0-beta-2
>
>
> When defined on a method or constructor, all parameters of that method/constructor will
be final. When applied on a class, the annotation applies to all methods/constructors.
> Discussion points:
> # Should visibility matter, e.g. would it apply to all public, protected, private, etc.
members, or does there need to be some ability to customise? *Current thinking*: it should
apply to all methods. We can always introduce an additional annotation attribute at a later
point if there is demand.
> # Should there be support for 'includes' or 'excludes' functionality? *Current thinking*:
don't support this for now. The annotation can be applied on multiple methods/constructors
if finer granularity than the class level is required.
> # Should the annotation apply to fields/properties? Applying 'final' to a field/property
seems to be more efficient than applying @AutoFinal but would there ever be a desire to apply
such an annotation to all fields/properties? *Current thinking*: don't support this for now.
An annotation attribute could be added later if we wanted to optionally turn such extra cases
on.
> # Should it apply to synthetic methods/constructors added by the compiler/other transforms?
*Current thinking*: Don't apply to synthetic methods unless we find special cases were it
is needed.
> # What phase should this run at? It needs to be before the (incubating) final variable
analyser is run. *Current thinking*: SEMANTIC_ANALYSIS
> # We should have tests to confirm that there are no adverse interactions with other language
features, e.g. default parameter values. *Note*: we should have a test case that covers this.



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

Mime
View raw message