[
https://issues.apache.org/jira/browse/DIRMINA-874?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13163667#comment-13163667
]
Emmanuel Lecharny commented on DIRMINA-874:
-------------------------------------------
There is a bloody nasty bug in Java 6 :
http://bugs.sun.com/bugdatabase/view_bug.do?bug_id=6302954
Using the attributes the way we do in MINA 3 currently is doomed to fail in
some cases like :
boolean needClientAuth = session.getAttribute(NEED_CLIENT_AUTH);
It works well in Eclipse, but when compiled using the Sun/Oracle JVM, you get :
[ERROR] Failed to execute goal
org.apache.maven.plugins:maven-compiler-plugin:2.0.2:compile (default-compile)
on project mina-core: Compilation failure: Compilation failure:
[ERROR]
/Users/elecharny/mina/3.0/core/src/main/java/org/apache/mina/session/SslHelper.java:[130,57]
type parameters of <T>T cannot be determined; no unique maximal instance
exists for type variable T with upper bounds boolean,java.lang.Object
In order to fix the compilation error, the code should be :
boolean needClientAuth =
session.<Boolean>getAttribute(NEED_CLIENT_AUTH);
No need to say that I produce such code nearly on a daily basis, in the
smallest room of my appartment...
I may reconsider using your approach, Christian, seems more 'clean' :)
> Typesafe AttributeKeys
> ----------------------
>
> Key: DIRMINA-874
> URL: https://issues.apache.org/jira/browse/DIRMINA-874
> Project: MINA
> Issue Type: Improvement
> Components: Core
> Affects Versions: 3.0.0-M1
> Reporter: Christian Schwarz
> Attachments: mina3.zip
>
>
> In Mina 2 we can add attributes to a session and get them as well. Because
> IoSession#getAttribute(Object key) returns Object we always have to remember
> what attribute type is associated with the key and cast the result to the
> expected type.
> Mina 2 example:
> final static Object KEY = new AttributeKey(SourceClass.class,"myKey");
> ...
> session.set(KEY,"myAttribute");
> String attributeValue= (String)session.get(KEY);
> Instead of using plain Object-keys, the key type should contain information
> about its attributes. The aim is to get type-safe access to attributes.
> Assume we have the following new AttributeKey-Class:
> /**
> * @parmeter T Type of the referenced Attribute
> */
> class AttributeKey<T> {
> public TypesafeAttributeKey(Class<T> attributeType, String attributeKey){
> ...
> }
> }
> The IoSession should have Attribute related accessors like these:
>
> void setAttribute(AttributeKey<T> key, T value);
> T getAttribute(AttributeKey<T> key);
> So in Mina 3 the example could look like this:
> final static AttributeKey<String> KEY = new
> AttributeKey<String>(String.class,"myKey");
> ...
> session.set(KEY,"myAttribute");
> String attributeValue=session.get(KEY);
> These 2 cases won't compile:
> session.set(KEY,new Date());
> Integer attributeValue=session.get(KEY);
> This pattern would simplify the use of Attributes, because the programmer
> don't have to care about the types and can concentrate on more improtant
> things. In my humble opinion the Objekt keys should be removed at all, i
> don't see the purpose for such unspecific keys.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators:
https://issues.apache.org/jira/secure/ContactAdministrators!default.jspa
For more information on JIRA, see: http://www.atlassian.com/software/jira
