jackrabbit-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mark Slater (JIRA)" <j...@apache.org>
Subject [jira] Commented: (JCR-328) TransientRepository with LocalFileSystem eventually causes Repository data to be stored at path '/'
Date Mon, 20 Mar 2006 22:17:00 GMT
    [ http://issues.apache.org/jira/browse/JCR-328?page=comments#action_12371151 ] 

Mark Slater commented on JCR-328:

Sure. That would make my JUnit code somewhat cleaner.

> TransientRepository with LocalFileSystem eventually causes Repository data to be stored
at path '/'
> ---------------------------------------------------------------------------------------------------
>          Key: JCR-328
>          URL: http://issues.apache.org/jira/browse/JCR-328
>      Project: Jackrabbit
>         Type: Bug
>   Components: core
>     Versions: 0.9
>     Reporter: Mark Slater
>     Assignee: Jukka Zitting
>      Fix For: 1.0

> I'm using a TransitoryRepository for my unit testing, with the repository's file system
specified as:
>     <FileSystem class="org.apache.jackrabbit.core.fs.local.LocalFileSystem">
>         <param name="path" value="${rep.home}/repository"/>
>     </FileSystem>
> I noticed today that when I run my unit tests Jackrabbit is creating four directories
at the root of my hard drive: "meta", "namespaces", "nodetypes", and "data". I tracked the
problem the fact that when a LocalFileSystem is closed, it sets the "root" to null - an invalid
state. But when using a TransitoryRepository, the invalid state is never discovered because
the LocalFileSystem object itself is not released, or re-initialized. It is simply used to
create BasedFileSystem objects in RepositoryImpl. Calls to BasedFileSystem defer to the LocalFileSystem
object that now has a null root. Inside the LocalFileSystem, all the calls to Java's io.File
constructor have a "null" parent parameter, causing File to fall back to its single argument
constructor which sees the path "/meta" and happily creates files at the root of the disk.
> I'm not sure what the best solution is, but some thoughts I've had are:
> - don't set the "root" property to null when closing a LocalFileSystem
> - make RepositoryConfig re-init the FileSystem variable when it is accessed.
> - don't cache the RepositoryConfig in TransitoryRepository (this might also require a
new constructor that takes a class-path resource for the repository configuration file)

This message is automatically generated by JIRA.
If you think it was sent incorrectly contact one of the administrators:
For more information on JIRA, see:

View raw message