This has been haunting me for months now, but I have finally implemented working library code, and an application around the AppServer/AppClient system. The application is the couple DaMon/Probe (I'm looking for more generic names for these applications):

• Probe is an AppServer accepting AppClients to connect. It can currently supply the uptime and load average values;
• DaMon is an AppClient, asking the Session on which port to find the "Probe" service, and then connecting to it (both remotely if specified via the environment).

I have a lot of other stuff to do now unfortunately, so further work on these may be put on hold for at least a week.

Would you like to propose news, or submit bugs a little less anonymously, it is now possible to register on the web site. A valid e-mail address is enough to proceed.

I have finally added dynamic javascript actions support to the PHP Explorer, allowing files deletion, and also improved the CVS browser at the same time. Generally speaking DaPortal has been improved both in CSS and content, mostly in the administration part though.

The libSystem is hopefully going on, with the first messages exchanged a few minuts ago. It's not functional yet, but I hope I'll be able to actually use it in some cases some time soon. I'll definitly have to update the DeforaOS documentation if this turns well, as it would likely be the base for the whole system.

The code rush from August has been slowed down by real life projects. Among other things, the shell has been improved a bit, the initial AppServer/AppClient design is almost finished, and there's a bit of code for it.
On the web site, anonymous bug reports should work now.

Following yesterday's effort to allow a wider access to the web site, a german translation has also been started. Of course I'll need extra feedback on this one.
Moreover, if you feel like helping the translation project, please contact me and I'll write down a small tutorial to allow submissions.
Please also note that the contents translations will be directly possible from the website when the code is done.

It's been some time without news, but these days I am more active than I've been for quite some time.
I am particularly working on the shell, which is improving quite fast, even though I expect some design problems when it will come to tricky parts (quoting, expansions, etc). Hopefully it is already handling enough to satisfy primal needs, so I'll try to boot it some time, and maybe even release the first Defora/Linux system. This would also require a bit of work on the libc.
I am also planning to start the code for the AppServer/AppClient system, for which I have designed some parts already.
Eventually made also a few cosmetic improves to DaPortal, like a minimal multi-language support. French is the first one to be supported.
Overall, I can say I'm starting to feel more excitement than doubt on this project for the first time ever, and it feels good. Time will tell.

A quick message to apologize for the couple days last week, when my servers have been down. I have disabled the water boiler and since then it doesn't black out anymore. As if I needed cold showers...
Anyway, motivation is still going up and down, I am quite a bit bored with PHP/SQL at work but it's still interesting enough with DaPortal to go on. You are welcome to try the first basic functionalities of projects' CVS browsing and bug reporting (you'll need an account at the moment to actually report one).
I expect a lot from this project tracking interface, but I can't wait to get this usable enough so that I can get back to C code. But first, go to bed.

It's finally done, the current orientation of my Defora project runs DaPortal again. It's maybe its 5th or 6th rewrite, but my intuition is it should last for a while. Its code is available on the project CVS repository, I hope it is going to be secure enough.

A word on DeforaOS now, I feel guilty everyday that I don't work and commit something somewhere, while as and the libc are so close to become usable. I hope I'll find motivation soon enough to code something not in PHP: I am thinking about the message passing system, maybe with "init" as the proof application.

I have also started to use some of my POSIX utils implementations instead of Debian ones on my current desktop system. So far I am quite happy with them, and I'll certainly test the ongoing bugtracking system with them.

Probably nobody but me noticed, but I have shut down the anonymous CVS access for a few days. This is because of the latest security vulnerability discovered, which should be fixed here within a few other days.

About the projects, I have written a basic GUI for GEDI and will start to write actual production soon hopefully, PHP Explorer supports basic permissions listing in the CVS, and I have added (way too) few lines to as and the libc. About the libc in particular, I think it could be actually "working" in the next times I'll work on it.

I have improved a bit PHP Explorer, it can now display the images on the server as thumbnails. I thought this was a time for a complete release, so I did and I have even included the icons in the package this time. As usual it will be on front of freshmeat during one day hopefully, but I have also submitted it to the PHP Québec contest. You never know...

I have found the courage to look further into the problems with the libc, and I am ending with something new this time: maybe _start() should actually come from a separate object. If this really is the answer, I feel stupid. After all crt.o certainly has a purpose, if you see what I mean.

To conclude, I could announce a new incoming project. It is called "GEDI", and it will certainly be developped using this project's resources. I feel frustrated because that's one more of the too many projects I am currently trying to work on (and can't handle), but it is also the first serious project I may start and end with a friend. So, CaseK will certainly gain write access to the CVS tree. GEDI is the french abreviation to "Gtk+ based Integrated Development Environment", as GIDE already exists, and interestingly enough it sounds like "jedi".
To say a word about the project itself, I could say that I hate IDEs. That's why we are going to try to write one the UNIX way, using existing elements of the system (Make, GCC, GDB, VIM, CVS, ...), because that's how we feel it should be done. We don't want the bloat of integrated editors, splash screens, tens of windows, dozens of tabs of preferences dialogs. We want it to ease the development of our projects, including tracking time of the various tasks.