...

this would require it to

1. copy the archive locally 2.
2. apply the change
3. 3. repack the archive
4. 4. copy the archive back to its place
   

...

• pop, imap (useable for emails)
• mime (for the email itself - pretty much like an archive)
  

Consider this url:

zip:mime:imap://user@mailserver/INBOX/folders/newMail.mime!/attachment1.zip!/zipentry.txt

...

• generic XML
• ant (only a variant of XML? Cant we just set a dtd or scheme using some sort of configurationOption ?)
• java structure (by class, by source?)
  

...

Todos

The following is a list of items that need to be completed. Contributions are welcome!

Release 1.1

• moveTo should handle a directory move if it moves from a different filesystem.
  

Open

• More documentation (status, file naming etc).

• Fix the TODO items
• Add more providers:
  • rsync
  • subversion
  • nfs
  • cvs
  • jdbc filesystem
  • xml filesystem
  • jndi
  • imap
  • local mirror
  • spidering http
  • ...
• JNDI integration.
• Formalise the provider API.
• WebDAV Provider:
  • Add plain http support, and auto-detect dav resources.
  • Add set last-modified.
  • HTTPS
• Zip/Jar Provider:
  • Extract an AbstractLayerFileSystem out of ZipFileSystem.
  • Track changes to the parent layer. Eg when the parent layer is deleted, mark all the files in the fs as 'does-not-exist'.
  • Add support for writing to zip/jar files.
• URL Provider:
  • Support attributes.
• HTTP Provider:
  • Support attributes.
  • HTTPS support.
• The local disk caching mechanism also needs more work. Needs to check last-modified time. Replicator needs to be more configurable.
• Add persistent replicator.
• Finish support for junctions: Make ancestors of a junction point visible, fire events when junction is added or removed, tests.
• Add support for federation (ie transparently crossing file system boundaries, such as drilling down into the contents of a Jar file).
• Add an equivalent of the fileScanner Jelly tag.
• Add an equivalent of Ant path, fileset, dirset, filelist, etc. Ideally, these can be abstracted into a single data type.
• Allow selectors, name mappers, and filters to be specified for the Ant tasks.
• Add capabilities to FileObject.
• Attributes and attribute schema.
• Handle file canonicalisation better (for cases like case-insensitive file systems, symbolic links, name mangling, etc).
• Add more selectors: XPath, Ant style, regular expression.
• Add adaptor (NodePointerFactory?) for use with JXpath.
• Add content-changed, attribute-changed, and move events to FileListener. Maybe split into structure and content listeners.
• Get/set the file permissions.
• Automatically checksum and/or verify remote files.
• Look at adding native code for fine-grained control over permissions, file monitoring, faster moves, etc. Must be optional - the thing should still build and run without the native code.
  

Library upgrades

The following describes changes in upgrades dependencies which need to be considered:

webdavlib 2.1

• resource.listWebdavResources() do no longer list directories?
  Figure out what to do
  

wishes

The best chance to get those added is by contribute them

...