10/15/07 11:11AM EST -- Socket to me! [Updated!] The sockets API used in rDDE is causing a LOT of problems. You'll notice that the current source copy in the SVN repo is broken. I mean *very* broken, as in not functioning at all. This is due in full to the fact that I can't seem to make the socket system I use with the TruThread Interprocess Communication system work at all. Clients get disconnected randomly, and while I know this is an alpha, I don't believe in releasing something that has zero usable functionality. UPDATE: robert_, a long-time friend of mine is now going to take over the inter-process communication system, leaving me to continue working on modifying the remainder of the core modules to be compatible with TruThread. (The IRC module is practically completed!) 10/1/07 1:51PM EST -- In Due Time, My Friend. I'm aware that this has been going painfully slow as of late in the public eye -- however, much has been going on outside of the Subversion repo or even this website -- the DJ web interface has been started (and boy is it going to look FANTASTIC) and it is taking a significant chunk of time to get anything done on it so far. Matt Sharpe will hopefully join me later this month in the development of that particular module, thereby easing off some of the tasks I have set for me. Lee Holder has also been busy and has barely scratched the surface on his request module and SAM integration. I'm going to try to finalize more of the core today to get it into Subversion this evening. Stick with us! :) 9/14/07 2:40PM EST -- TruThread I just want to take a quick moment to respond to some criticism about how I've been running the RadioDog loop. Yes, I know it's not the most efficient way to do things and that I'm really just making a poor emulation of threads. So let's move forward off of that. Today I've written a new class using a technology which I call TruThreading. This uses sockets to communicate between threads (I was going to just use signals and stacks but they're not cross-compatible with Windows.) And yes! These are actual, honest to God threads. Which officially makes RadioDog into a SMP (Symmetric Multi-Processing) program. Woo! 9/13/07 5:12PM EST -- Not dead yet! It's been over a month since I've written in here, whoops. Long story short, it's been a busy month for me. What between moving and my new job, I've been incredibly strapped for time. However, RadioDog's development hasn't been stopped quite yet. falconX (Lee Holder), a long time friend of mine, is now joining me in this project as it grows larger and larger. He will be working on integrating RadioDog with SpacialAudio's SAM3 and SAM4, and also working on the request system. In the meantime, Emess (Matt Sharpe) has been keeping in touch with the iDJc developers about putting support into that program so that RadioDog can communicate requests there as well. (Perhaps we can convince them to adopt the new .RPL format, which is currently the way RadioDog's internal playlist is stored). And don't think I've been slacking off either... I've been scouring the 'net for libraries I can use in the web interface. So far I've come up with two that I am definitely going to use (unless someone throws something better my way) -- Yahoo's YUI and PEAR's Image_Graph. These will provide a nicer interface for the dashboard along with graphical representations of a load of statistics. Also, it's looking more and more like MySQL will be a requirement instead of an option. APC has gone down the tubes, to be replaced by TCP sockets for interprocess communication, and I will actually be modularizing the entire RadioDog base code soon (to work better with the interprocess signals), so expect basically a full overhaul to be committed to SVN in the next few weeks. 8/12/07 9:56PM EST -- First field test We had the first field test of the RadioDog binary today... and it bombed. I'm not sure exactly why but judging from our tests it has something to do with the compilation process. In other news, the Windows binary actually doesn't work *at all* but the Windows script version works fine. I may or may not release the alpha under GPL in a couple days, uncompiled. Sorry guys, I was hoping for binary but until Roadsend (the PHP compiler I use) gets my bug reports, I'm quite stuck. 8/11/07 6:16PM EST -- C-c-c-razy! Apage43 pointed out today that using APC might be good for some local application but using UNIX/TCP sockets would be a better idea. I'm not sure yet, because cross-compatibility issues could well come into play, but at the very least the rDDE function will stay in as a TCP socket on both platforms... it just might become the DDE communication as well. I redeveloped the IceCast2 Library today, adding some better and less sketchy multi-source support... that does mean, though, that I'll be rebuilding all the previously written function calls from the IRC bot and web inclusion libraries. Ugh. Here's hoping it's an overall improvement. 8/9/07 8:45PM EST -- From here on in, I code without a script. RadioDog's programming was successful right up until yesterday, when box43 (my development system) bit the dust... it would have meant that development was suspended until almost September (which is lousy because I wanted 4.0Beta out before 9/1) but I managed to pick an old copy off the server... and updated it. At this point I'm switching my ideas for DDE from using APC memory access to using UNIX sockets. The problem is still cross-compatibility, though, so I'll probably just stick with APC, because hey, it's a huge step above using a temporary flatfile as DDE swap... or SQL tables.