Our next release of Cherokee, 0.8.0, will once again have a native Windows package.
A few moments ago it was officially announced at the mailing list.

We had plans to finally fix it in the very near future, but Alvaro decided to speed things up a bit. This guy is amazing! ;)

I’ve been rather busy these days and haven’t been paying all the attention to the project I would have wanted, but everything started moving on Windows’ side of things just a couple of days ago. I really didn’t expect it to be ready this soon, but here it is: Cherokee Windows Build. The development branch already compiles and works on Microsoft’s OS. Just check out the latest SVN version and give it a try.

Alternatively if you can live without Windows and want something more stable, you can just download the latest official release, Cherokee 0.7.1 Cherokee 0.7.2 (as of June 12th).

A couple of months after the previous Cherokee stable release, and once a lot of work and testing has been done, here it is: Ladies and gentlemen, Cherokee 0.7.0 is out!. You could take a look at the official announcement, but what you should really do is try it out. Seriously. Every single one of us is working hard on the project -make that extra-extra-hard working in Alvaro‘s case-, and the results are clearly visible.

We have a lot of fresh ideas, and as always feedback and feature requests are more than welcome at the mailing list. The Summer is almost here, and this one will prove to be extremely productive. Just trust me for now. It will ;-)

Here is the download link.

A couple of days ago I mentioned this in the post about extending libStargate for a class assignment we had. The team was formed by Miguel, Gonzalo and myself, and we received a pretty good grade, by the way. I wanted to write again after having some patches sent to the main author of libStargate by the end of this week, but it turns out I won’t be able to properly test them until after that. Since a lot of friends have been having problems with their own projects, ours is going to be temporarily cannibalized until their time is up. Wish them luck! ;)

Anyway, I’ll post again once  I have the things properly tested and I’ve spoken to the people of IEArobotics. For now, here is the link to download Wiilson if you want to try it out.

As mentioned earlier, it is just a PWM Wiimote implementation to control a Skybot. The Wiimote is interfaced thanks to the excellent pywii library, Orienting the Wiimote makes the bot move and it goes faster or slower depending on the inclination of the controller. Changing from mode ‘A’ to ‘B’ (or ’1′ to ’2′) overrides this behavior and controls the bot with the direction keys of the Wiimote, being able to alter speeds with the ‘Plus’ and ‘Minus’ buttons. It’s not pretty, but it is a quick hack and works fine. If you want something better you’ll have to wait for the release of pybot, a Python module to control (extended) libStargate managed robots with PWM support. The only class present right now is to control the Skybot as it is out of the box, so PWM support is software based -you’d have to rewire it if you wanted a hardware version of this, much better but totally out of the scope-.

Coming up soon, at least as soon as our bot gets de-cannibalized if that is even a word!

By the way… I think I’m getting to love this stuff. I’ll have to buy a Skypic or Arduino for myself. You can’t ever play around enough with these things. And to think it was always there and I never knew!

Couldn’t let the day pass by without mentioning that Ubuntu 8.04 LTS (Hardy Heron) has finally been released today. You’ll have lots of information elsewhere, so I won’t lose our times telling you about it.

It has been working flawlessly with my new Thinkpad since February, so I’m pretty sure it is a polished product by now.

I just wanted to note that every time a release is made, some servers seem to go down due to the overwhelming demand. It’s sort of a Download Frenzy I suppose. If you can’t find a working direct link maybe you should try downloading via BitTorrent. It’s the only way I use, anyway.

Here you have some cached links for the torrents in case you need them.

• Kubuntu Desktop CD for 64-bit PC (AMD64)
• Kubuntu Desktop CD for PC (Intel x86)
• Ubuntu Desktop CD for 64-bit PC (AMD64)
• Ubuntu Desktop CD for PC (Intel x86)
• Ubuntu Server install CD for 64-bit PC (AMD64)
• Ubuntu Server install CD for PC (Intel x86)

Precisely one week after releasing the long awaited 0.6 version of this web server, Cherokee 0.6.1 has been released minutes ago. As I’ve said once and again, the project has been gaining momentum and here you have the download link to prove it.

It is mostly a maintenance release since most of the development efforts are being put towards the future release of version 0.7 with a lot of goodies. I’ll be writing about that when the times comes. For now the list of changes in 0.6.1 includes support for “Personal Web” -a missing feature in this release that was requested for the administrative interface-, a fix for a bug introduced in handler dirlist right before the previous release, and some other minor bugs in documentation, compilation and cherokee-admin.

As before, you’ll have more information at Alvaro’s site and the official Cherokee Project’s website. We’ll be waiting for your input. Happy hacking!

As I anticipated some weeks ago, Cherokee 0.6.0 has finally been released. It was a couple of days ago but it caught me away and with no Internet access until now.

Anyway, the trunk was branched on March 6th and after some exhaustive testing and a little polishing here and there … here it is!!!

You should take a look at Alvaro’s site and at the official Cherokee Project’s website. Suffice it to say a lot of work has been put into the new release, and the effort has been well invested. Work will resume shortly to improve an already impressive web server, and as advancements are made towards 0.7 a lot of new features will be added. A fast look at the TODO file will give you an idea of what is yet to come: a bunch of new handlers and modules (WSGI, AJPv13, WebDAV, mod_evasive, upload progress module), generic caching, new header entry, AIO based fdpoll, memcached support, chunked encoding, Dtrace hooks, better language support…

As I said previously, stay tuned. You’ll be able to expect a lot from this project in the near future. That is a given ;)

UPDATE: I forgot to mention the most important thing. Here is the original announcement of the release at the official mailing list.

I’ve just read some interesting news of what will be coming with Linux kernel 2.6.23. In short, as you can read in Gmane:

This interface allows the ability to write the majority of a driver in userspace with only a very small shell of a driver in the kernel itself. It uses a char device and sysfs to interact with a userspace process to process interrupts and control memory accesses.

You can have the full details of what I read at liquidat’s site.

I think this is awesome, and it was about time. It turns out this was announced a year ago by Greg Kroah-Hartman, who obviously knows his stuff since he’s been developing kernel drivers like forever.

As I was saying, these are wonderful news. It would be great if every hardware vendor developed open source drivers, but this hasn’t been going on, and being pragmatic it’s a lot better having a closed-source driver that not having a driver at all. I think this API will be a good thing for several reasons:

• Firstly, Linux kernel internals aren’t really that stable as they frequently suffer interface changes -for better or worse-. There aren’t many companies that develop drivers for Linux, and far too many of those spend close to nothing in maintenance. By providing a stable API there’s a chance that closed-source drivers won’t become obsolete so fast, and these could even be marginally more portable.
• A second reason would be the advantage of being able to use this as a testbed platform while developing the drivers. Tinkering in kernel space is a tricky question. Developing the same code in userspace with an API to link over eliminates many of the possible ways of trashing the system. By allowing most of the code to be run in userspace, poorly designed code is less likely to screw everything up. Which also brings us to the closed-source drivers’ subject. As long as these remain closed-source there’s no telling how bad the code can be. There’s no peer reviewing, no community bug fixes and so on. So hey, what the Hell! If it has to be run, at least let us be somehow protected.

To anyone interested in developing kernel modules there’s an excellent book by Jonathan Corbet, Alessandro Rubini, and Greg Kroah-Hartman that is licensed under Creative Commons: Linux Device Drivers, Third Edition. I’m currently going through it and it’s making wonders to what little knowledge I had on the subject.