collegegeek.org

Deluge 1.0.0 was just released, and will soon be is now available here

http://launchpad.net/~deluge-team/+archive

Here’s what I’ve been hacking on for the last couple of days:

I’ve been developing a Search plugin for Deluge 1.0.  Unlike my previous search plugin from 0.4 and 0.5, this plugin shows the results inside the client, without having to open a web browser.  Attached is a screenshot, I’ll be releasing the source very soon.

The plugin works by parsing RSS feeds, and as such it is currently only compatible with Mininova and Isohunt.  It also requires python-feedparser to be installed.  I’m looking to get more search engines supported, but at the moment, it can only support search engines that offer search results in RSS form.

Deluge In-Client Search

Update: I now have experimental support for The Pirate Bay, but due to the nature of their RSS feed, it doesn’t display Seeders, Peers, or Size, and it only searches the torrents’ tags.

Sometime after I get back to Kentucky, I’m thinking of building a server for running virtual machines using VMware Server 2.0. The idea is to virtualize as many of my machines as possible.  Virtual servers, in particular, benefit from this by being able to be easily moved or cloned.  Rather than having to take a server down for upgrades, you have the  ability to clone the VM, apply the upgrade to the clone, then switch them out, resulting in nearly zero downtime.

In addition to running web and mail servers, I could also virtualize the backends to some other applications, namely Deluge and MythTV, so that the computers they ran on didn’t need to be on 24/7.  MythTV is a little tricky due to the lack of support for PCI passthrough in VMware, but I could of course use USB tuners.  I could also use a virtual machine as my primary desktop, and use a small, low powered, thin client to access it over VNC or SSH.

I recently looked into the hardware I’d want for such a server and came up with the following:

• Intel Core 2 Quad Q6600 2.4GHz CPU
• Intel BOXDG31PR Micro ATX Motherboard
• Kingston 4GB (2×2GB) DDR2-800 Memory
• Seagate 1TB SATA2 Hard Drive
• APEX DM-387 Slim Micro ATX Case

The build would be another slim micro ATX build, similar to Cervantes and have a quad-core CPU, 4GB of ram, and a terabyte of storage.  In total, it came to a little under $600 for the entire build (I love how cheap hardware is getting) but I think I’ve decided to hold out for Nehalem before I build it.  With native quad cores and hyper threading, Nehalem should work well for this kind of project.

The third release candidate is now available and has been uploaded to my repository.  The changelog can be viewed here.

Notable changes in this release include a new Blocklist plugin and upstream bugfixes from libtorrent.

I’m suprised I completely forgot to post anything about RC 1 here a few days ago, so I’d like to go ahead and introduce the second RC release of Deluge 1.0.  Previously known as Deluge 0.6, Deluge 1.0 is a complete rewrite from the 0.5.x branch (which was in turn a complete rewrite of 0.4, which was a nearly complete rewrite of 0.3 :P).
Read the rest of this entry »

So, the last couple days I’ve been playing with the latest release of my former project, Deluge.  With version 0.6 comes support for a client/server mode of operation, meaning the downloading can run on a different computer as the user interface.

For those of you running Ubuntu, you can get up to date SVN builds of Deluge 0.6 from my PPA: https://launchpad.net/~zachtib/+archive.  Otherwise, you’ll have to wait for an official RC release or build it yourself from source.

Several things have changed since the 0.5 days.  First off, 0.6 is a complete rewrite.  It uses XMLRPC to communicate between the client and server, which allows it to run on separate computers.  Today I tested this using VMware and ran the front and backend on separate virtual machines.

The user interface has also been completely overhauled.  Personally, I think it now looks a lot more like uTorrent with the Labels pane on the left hand side.  In addition, the preferences dialog has been redesigned as well.

For those that don’t have a need for a separate frontend and backend, a “Classic” mode is available in the preferences dialog that will emulate the old method by automatically starting and stopping the backend with the client.

I’m very happy with the direction the client is going, and plan on using it for the forseeable future.