On Thu, Jun 21, 2012 at 5:10 PM, Mark Fortner <phidia...@gmail.com> wrote:
> Gary, > There was some talk a while back about implementing File System-specific > Operations. I think what Mario had in mind was supporting version control > system functionality through VFS. It strikes me that this might be the > best way for implementing functionality that makes use of the JSch bells > and whistles without breaking encapsulation. > Hi Mark, I am not sure I understand what you are proposing. Gary > Regards, > > Mark > > > On Thu, Jun 21, 2012 at 2:00 PM, Gary Gregory <garydgreg...@gmail.com > >wrote: > > > Hi All: > > > > WRT VFS-422: [SFTP] Allows to create other channels in SftpFileSystem ( > > https://issues.apache.org/jira/browse/VFS-422) > > > > I'll like to solicit feedback from the list on this issue. I've cleaned > up > > the proposed patch in the ticket and attached it back to the issue as > > > > > https://issues.apache.org/jira/secure/attachment/12532940/vfs-422-v2-by-gg.diff > > > > First and foremost is whether this feature should be in VFS at all. > > > > Internally, VFS does not use the feature, so it is a convenience to the > > client application. > > > > A simpler solution from VFS' POV would be to make the Jsch Session object > > accessible via a public getSession() method. The client would then be > > responsible for the resources created and be free to use all of the Jsch > > bells and whistles. > > > > One can argue that this breaks encapsulation and surfaces VFS' private > > parts (pun intended). But because the point of the feature is get to > > another Jsch object (a Channel) through a Session, encapsulation would be > > broken anyway, so this point is moot IMO. > > > > A second issue has to do with implementation of the method > > com.jcraft.jsch.CommonsVFSChannelFactory.createChannel(AtomicLong, > > SftpChannelType) > > > > - Should finalization be used? > > - Should the counts be managed through the connect and disconnect methods > > instead? > > - There is a lot of repeated code in the method, should a delegate/proxy > to > > a channel be used instead such that there is no code duplication? > > > > Thank you, > > Gary > > > > -- > > E-Mail: garydgreg...@gmail.com | ggreg...@apache.org > > JUnit in Action, 2nd Ed: <http://goog_1249600977>http://bit.ly/ECvg0 > > Spring Batch in Action: <http://s.apache.org/HOq>http://bit.ly/bqpbCK > > Blog: http://garygregory.wordpress.com > > Home: http://garygregory.com/ > > Tweet! http://twitter.com/GaryGregory > > > -- E-Mail: garydgreg...@gmail.com | ggreg...@apache.org JUnit in Action, 2nd Ed: <http://goog_1249600977>http://bit.ly/ECvg0 Spring Batch in Action: <http://s.apache.org/HOq>http://bit.ly/bqpbCK Blog: http://garygregory.wordpress.com Home: http://garygregory.com/ Tweet! http://twitter.com/GaryGregory
