Ubuntu Linux: followup

I’ve solved the problem with my ATi Radeon 7000 graphic card. I’ll leave this for future reference, you never know…

The Symptoms

ATi Radeon 7000, in combination with an Acer AC711 monitor (17″), just goes wild under XFree86. With the default configuration, X loads with an resolution of 640×480, and a virtual resolution of 1024×768. No matter what modeline I put inside my XF86Config-4 file, and no matter what option I specified for the radeon driver, I could get X to use an higher resolution.

The Solution

The only way of making a ATi Radeon 7000 work with my 17″ monitor was to enable the Radeon framebuffer device inside the kernel (you might want to load the module during the boot sequence, but I strongly suggest recompiling the kernel with the framebuffer device built-in), and change the Device section inside the /etc/X11/XF86Config-4 file like this:

Section "Device"
        Identifier      "ATi Radeon 7000 VE (RV100 QY)"
        Driver          "radeon"
        Option          "UseFBDev"      "true"

I’ve highlighted the line you should add

By telling X to use the framebuffer device, the card successfully reaches a usable resolution, and the virtual desktop is gone for good.

Ubuntu Linux

Yesterday, I’ve downloaded the first official release of Ubuntu Linux – codenamed Warty Warthog. It seems – as I’ve been reading on Planet GNOME and Planet Debian – that this distro is the Next Best Thing® for desktop usage, so I decided to give it a try (currently, on my box I’m running Debian Unstable with some packages from Experimental).

I’ve turned my Dummy Mode on, and launched the installation. Well, not exactely “turned off” – since I actually had to carefully choose the right partition inside the partition table, in order to keep my precious / intact.

The installation worked pretty much flawlessy: it was better than any installation procedure I’ve done in years; it seemed more like a LiveCD booting than a disk installation.

Unfortunately, here cometh trouble. I’ve a crappy Radeon 7000 graphic card and an Acer AC711 monitor; this combination yields major PITA, since X seems to work only when using the kernel’s framebuffer device, by setting the “UseFBDev” option to “true” inside the /etc/X11/XF86Config-4 file. Ubuntu’s kernel image comes with the Radeon framebuffer device compiled as a module – thus I’m unable to activate it before GDM spawns.

Hence, I’ll have to turn my Dummy Mode off, and recompile a kernel with the radeonfb module compiled statically and invoked inside the kernel command line.

So long for a dummy mode installation.

Late Night Hacking

Another night of late hour hacking session.

Trash applet

I’ve hooked up a working version of Michiel’s trash applet; it now has a stub menu entry that recalls the online help, and I also added the possibility to open up a Nautilus window with the trash: URI.

GNOME Trash Applet
Uh oh, time to empty the trash bin…


I’m still learning ORBit2 implementation of CORBA, in order to write the user daemon serving the GnomePimClient object. I thought about dumping the entire concept of the user daemon and using FAM to watch and update the XML files.


I’ve never learned proper JavaScript programming; I always dissed it, maybe beacuse of the bad habit, in the DotBomb era, of using it for everything (especially when not needed), making web pages practically unusable. Last night I had to write a simple countdown function which periodically updates the content of a XHTML entity; in just ten minutes I grasped the basics of the Document Object Model structure exposed inside the language interpreter, and had my function basically being written by itself. JavaScript (or, better, ECMAScript) surely has come a long way to finally being usable for serious client-side web programming.

Trash Applet

Skimming through the gnome-devel list, I found the announcement of a simple Trash applet for the GNOME panel, made by Michiel Sikkes (here‘s the site, with a screenshot).

I think it’s a great idea: sometimes the Trash icon is simply buried under screen clutter, especially with the new spatial paradigm that Nautilus uses. The panel, on the other hand, is always on top.

Since Michiel’s code is something short of a proof of concept, I fleshed it out, adding state recognition, and some menu shortcuts which will open a Nautilus window showing the trash contents, or empty the trash bin. It works pretty well, for a two hours hack. ;-)

This actually is the first time I touch some C code, besides a project for a class final exam, in a very long time – I had almost forgot how much fun is hacking just for fun.

Gnome2::Print (and future plans)


It’s been a long way, but in the end the Perl binding for libgnomeprint (and libgnomeprint-ui) has been finally completed. Hurray for me. :-)

I’ve just committed the code that binds Gnome2::Print::Paper and its methods, inside the 0.93 release of the Gnome2::Print module. I began maintaining this module the past august, soon after this email; soon after that, I began the Gnome2::GConf binding, that ended up inside the modules included inside the GNOME Platform Bindings. I’m proud of both, even though I rarely used the Gnome2::Print binding – I’m not really into printing data.


As soon as I return to own my life, after this semester’s finals, I’ll be able to begin working on a GStreamer Perl binding set. I already gave a preliminary look on this library, but didn’t really dig into it. I will also try and complete the GTK2-Perl tutorial, which is still missing a large chunk of chapters.