Not a day goes by when I stop dreaming of building a MythTV box. I think I have been a closet home entertainment freak for a few years now. I remember when I lived at home, I would spend hours trying to get optimal speaker placement with my shitty old speakers that I inherited from my brother. Then, when at University, I popped out to Currys to buy a CDR and came back with a surround sound system. I was in bliss.

The MythTV thing really intrigues me. Sure, the sensible side of my brain tells me I should go out and buy a Sky+ box, but my hacker instinct tells me to build it from scratch, and then I get all the cool things that MythTV affords me such as DVD ripping, MAME, audio jukebox, RSS reading, weather the permanent upgradability that you get with a free software project. The whole geek cred thing is important too. It would be nice to have something like MythTV to show off to friends who come over. 😛

The one thing that has stood in my way for quite a while has been a means of controlling the sky box from the MythTV box. I once thought the Red Remote could do it, but it seems maybe not. While browsing around the net last night however, I discovered a homebrew concoction that Joseph Heenan has built. It is a USB device that hangs out the back of the MythTV box and plugs into the RF2 output on the Sky box. The Sky box can then be controlled. Woo!

I think people like Joseph are inordinately cool. The guy went into a garage and knocked up a solution to a problem and selling his efforts to others. 🙂

Nice relaxing weekend. I started work on my dedicated music site. I have been wanting to separate out my music collection from jonobacon.com and move it to somewhere else to make it a little more than just a section on my existing website. Luckily, much of the functionality for adding and viewing songs was there, and I also hacked in the ability to add comments for songs. I tried to make the site look a little more interesting than my traditionally fairly predictable designs. I am quite pleased with the result. I have registered the domain (I will keep the name of the domain quiet for now :P) and I hope to have it uploaded reasonably soon.

A few people have asked why Sacraficial Life is spelled ‘Sacraficial’ as opposed to ‘Sacrificial’. The reason is in the general subject matter of the song. The song is basically about the sacrifice of someone to a cause. To make the title a little more interesting I incorporated Sacra into the title; sacra is plural for sacred bones. 🙂

Tonight I wrote up some thoughts about Linspire asking Freespire to change their name. When I first read about this I was pretty annoyed with their rationale, particularly given their history. Go and read The little pot Linspire calling the kettle black.

URGENT: By the way, Saturday September 17th is my birthday! Woo! Why not go and buy a Bacon something nice. Stuck for ideas? No problem, go check out my Amazon Wishlist.

For those of you with a penchant for distorted guitars, double bass drumming and growly vocals, you should stop what you are doing right now and go and listen to my new tune, Sacraficial Life.

Yesterday I had a meeting with Jake Stride from Senokian about Enterprise Groupware Server. EGS is going to be released soon and it is a PHP based CRM and project tracking tool. It competes with the likes of SugarCRM and I am really impressed with it. Jake and his team have worked really hard and I am intrigued to see how well it gets taken up by the Open Source community. This is not only a higher quality CRM than Sugar, but it is also fully GPLed. I have been chatting with the Jake for a few weeks about some plans for it, and they sound awesome. 🙂

Had a meeting with IBM yesterday and it was great to see Cliff and Colin again. We chewed the fat about various topics and projects that we are working on and IBM are going to help in bringing more people from the West Mids to OpenAdvantage for free training and consultancy. I am also giving a talk at a seminar IBM are running later in the month.

Went to the LUG last night, drunk some beer, talked some Linux and rolled into bed at 4.30am. Three hours later I got up for work and both Aq and I were unamused at the lack of sleep and the tenderness of stomach after such a heroic drinking session.

You know, computers often bemuse me. Today I tried to run OpenOffice.org Draw 1.1.3. It usually works fine, but today it didn’t want to come out and play. I clicked the icon, nothing happened. I clicked again, the taskbar lied to me that it was loading the application, and subsequently buggered off to wherever taskbars tend to hang out. I fired up a terminal and tried to run it manually. It then told me this:

jono@ubuntu:/opt/lampp/htdocs/sites/faq$ oodraw

OpenOffice.org for Debian - see /usr/share/doc/openoffice.org/README.Debian.gz
openoffice.org:  You have no entry for OpenOffice.org 1.1.3 in ~/.sversionrc,
yet the directory /home/jono/.openoffice/1.1.3 exists.
Please remove /home/jono/.openoffice/1.1.3 and try again.

It had problems loading a configuration file. Given that 99.99999999999998% of users run OpenOffice.org by pointing and clicking, why on earth couldn’t it just back it up, remove it and fix it or even simply inform me of the problem in a dialog box:

This small defect is catastrophic to users. The user would be sat there, the day before their report is due and not able to start OpenOffice.org. Not good at all…

Thanks to Seb I got the new Hula web interface up and running:

Its a tad slow in parts, but it really shows the potential for creating truly dynamic web interfaces. I am convinced that Hula is going to set a standard for many Open Source web applications, and its about time too. 🙂

Oh, and filed this.

Well, it seems like we may be on the way to finding out what the problem is with the server. It looks like there may not be a memory leak at all and the load is simply peak traffic. Matt Bloch from Bytemark kindly had a look at the box and threw some pretty rigorous tests at it. It seems we simply need to tune the bugger. This is another class example of Bacon and Aq Bodge It and Scarper SysAdmin Services. Any suggestions on tuning Apache, MySQL, Python?

Had a meeting this morning with a company who are interested in Open Sourcing a product. Its a pretty tough decision to make, and this case was fairly straightforward with no complex IP issues. It was a good chance for them to sit down and brainstorm through the issues and get an idea of what they want to achieve and how to do it. The product they want to work towards seems quite compelling and I am confident that the Open Source process can bring some real value to their business. In cases such as this it is important to step back and identify if you product can sail, what you aim to get out of the Open Source process and how you balance your corporate and community responsibilities. It is a fairly inexact science, but there are social and technical trends that can help in this area. Fun stuff. 🙂

Thanks for all the comments about which SVN viewer to use. It think it is safe to say that Trac is a highly recommended choice. If I can make it fit in with my current directory structure (this could even be done with mod_rewrite) I think I will run it. This is a shame in some ways because I have just spent two days hacking something similar in PHP. 😛

Last night I added GNOME iRiver to the repository (svn). I was gonna do some hacking but I felt like kicking back and watching 24. Watched two or three episodes and zonked out.

Now, I would like to solicit some advice from you folks. The server that jonobacon.com is running on has recently been upgraded to Debian Sarge, but something is eating all the memory and the machine is becoming unusable, usually requiring a reboot. We want to find out exactly what is doing this. Does anyone know of a way of finding out how much memory each unique process is using? top doesn’t help as it reports shared memory (e.g. Python may report 10MB but this is shared between all the Python processes). We basically want to find out how the memory is being sucked up. Ideally it would be great if we could have a list of processes and how they eat memory and the end of the list is when the machine goes down. This can tell us which software is buggering it up. Any suggestions?

Back to work tomorrow…

Went out for dinner last night with friends, and then when we got in I figured I should finally get around to sorting my subversion server out. The result is https://svn.jonobacon.com/. The only project that is in there at the moment is the XAMPP Control Panel (svn). I will probably get GNOME iRiver in there tonight. I also have some other code I plan on dropping in at some point soon.

Setting up the server was reasonably easy, but there were some gotchas:

• Make sure that the db and dav directories have the correct permissions for writing too.
• Make sure that when you have set up the server you set up the correct permissions, otherwise you server is wide open.
• Make sure you create the trunk, branches and tags directories inside your project module. This does not seem to be automated on the import.
• Make sure you use the fsfs filesystem as it seems to work faster, better and be more recoverable
• ViewCVS is a pain for setting up with Subversion
• WebSVN is installed on my server but I don’t think it is that good. Suggestions?

So, with the code now available, it will be easier to create patches. To check out the XAMPP Control Panel, just do:

svn co https://svn.jonobacon.com/svn/xamppcp/trunk/

Next, hack in your changes. Check which files have changed:

svn status

Then, generate the patch:

svn diff > patch

and mail me it. 🙂

Got into bed last night and we started watching Season 4 of 24. Sooz fell asleep through the first episode, but I carried on watching. I then proceeded to watch eight episodes back to back. Woke up today a bit knackered. Good job I was off work on holiday. 🙂

Tonight I started hacking a new PHP script that I plan on releasing. The basic idea is to provide a nice and easy way of providing software releases on your homepage. I have deliberately designed the script so that it can be dropped very easily into any homepage. I have written quite a lot of it and I am currently hacking the admin section, making it as easy as possible to manage software releases. When I have got it into a fit state I will release it. Not sure when this will be though. 😛

I was also pleased to receive an email from Kai Seidler (who is behind XAMPP) extolling many nice words about the XAMPP Control Panel. Kai also said that if I needed anything adjusting in XAMPP itself that would be no problem. Thanks a lot Kai. I have already started discussing a few things for error checking that could be useful.

Right, now back to 24…

You know, Banger is cool. In fact, sometimes he is so cool, he looks good enough to eat:

Today Sooz and I took it easy. I played some guitar, watched some TV and hacked on my new site. Don’t you just love time off? 🙂

Still having a few server troubles. We are trying to nail down the source of the problem. News forthcoming…

It seems my first release of the XAMPP Control Panel had a bug in it. Thanks to oris for reporting it. The bug assumed that a GConf key was set that stores the XAMPP path location. This prevented it from running. I have fixed the code and released a bugfix release, 0.5.1.

Headed over to Aqs tonight and we upgraded angel. The upgrade went well and everything seems in pretty good shape. I was pleased to see that the PHPWM mailing list also experienced a smooth transition. Fingers crossed all of this should result in a far more stable box. 🙂
—–