[ 
http://jira.codehaus.org/browse/MECLIPSE-294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_112566
 ] 

Joerg Schaible commented on MECLIPSE-294:
-----------------------------------------

IMHO sorting is not a good solution either, although bad practice - sometimes 
two jars contain same classes and the order of the classpath defines which are 
used. It should simply keep the "declaration" order. Therefore internally any 
instance of HashMap should be replaced by one of a LinkedHashMap.

> Sort items in the generated .classpath
> --------------------------------------
>
>                 Key: MECLIPSE-294
>                 URL: http://jira.codehaus.org/browse/MECLIPSE-294
>             Project: Maven 2.x Eclipse Plugin
>          Issue Type: Improvement
>    Affects Versions: 2.3
>            Reporter: Max Bowsher
>         Attachments: eclipse-sort.patch
>
>
> Currently, the ordering of items on the generated .classpath is controlled by 
> the JVM's HashSet ordering.
> One effect of this is that the project-07 and project-33 tests are currently 
> commented out, because it's impossible to write an expected/.classpath file 
> that works everywhere.
> More generally, it's bad that the classpath ordering can vary between 
> platforms and JVM orderings.
> Therefore, I think it would be a good idea to sort the classpath. This would 
> be done by invoking Collections.sort in 
> AbstractIdeSupportMojo.doDependencyResolution. IdeDependency already 
> implements Comparable, on the basis of groupId+artifactId+type. Classifier 
> should be added to the properties considered in compareTo.
> Patch attached as described above.

-- 
This message is automatically generated by JIRA.
-
If you think it was sent incorrectly contact one of the administrators: 
http://jira.codehaus.org/secure/Administrators.jspa
-
For more information on JIRA, see: http://www.atlassian.com/software/jira

        

Reply via email to