[ 
https://issues.apache.org/jira/browse/MNG-7999?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17805109#comment-17805109
 ] 

ASF GitHub Bot commented on MNG-7999:
-------------------------------------

rmannibucau commented on code in PR #1367:
URL: https://github.com/apache/maven/pull/1367#discussion_r1447320254


##########
maven-core/src/main/java/org/apache/maven/internal/impl/DefaultLookup.java:
##########
@@ -24,6 +24,8 @@
 

Review Comment:
   thinking a bit out loud there but I'm starting to wonder if we should expose 
guice Injector instead of this container wrapper which basically propagate the 
status-quo in terms of number of containers in the API and have a single 
lookup(Type) method.
   Probably something to think about before releasing this since ultimately 
this will need to move from internal to public cause it is a common need to be 
able to lookup public components programmatically (scripting case being the 
extreme example).





> Confine Plexus as much as possible, switch to Lookup
> ----------------------------------------------------
>
>                 Key: MNG-7999
>                 URL: https://issues.apache.org/jira/browse/MNG-7999
>             Project: Maven
>          Issue Type: Task
>          Components: Core
>            Reporter: Tamas Cservenak
>            Assignee: Tamas Cservenak
>            Priority: Major
>             Fix For: 4.0.0, 4.0.0-alpha-11
>
>
> Confine plexus use in core as much as possible. Probably almost all uses may 
> be replaced with API Lookup.
> Cleanup unused members as well, I found at least two components:
> * org/apache/maven/lifecycle/DefaultLifecycleExecutor
> * org/apache/maven/lifecycle/internal/LifecycleStarter
> but I bet there are more...
> Given clients should _inject_ and not manually instantiate components, ctor 
> changes should be hidden from them (managed by DI). Hence, changes like these 
> are "safe" as long as client code "plays by the rules".



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to