At first I thought that Ars Technica being acquired by Condé Nast was the weirdest-timed April Fools joke I’d ever seen; that’s mostly because Scary-go-Round is wrapping up a storyline about Shelly writing for Condé Nast and racking up silly expenses and being lost on a tropical island. It’s not a name I’ve encountered in serious prose before.
Packages for the dependencies of KDE4 for SPARC (32 and 64 bit) are on BionicMutton now; not all of them, but there’s a Qt package (4.4.0) for those with a pressing desire to run designer on SPARC. These are prepared with my current quality standards for Solaris packages: it compiles, so ship it. The same packages for x86 are also available but a good deal older; I’ll be updating them on BionicMutton when the build on our x86 cruncher is done. Those are the packages I use every day for my KDE4 desktop, so they should be pretty good for workstation use.
“Pretty good” sums up Shane’s continuing experience with KDE4. I see a lack of documentation on what Plasma / the desktop can do is annoying him this fine tuesday morning.
This morning I decided that having two complete checkouts of CVSDude and KDE was a bit much within my home network. In the interest of saving space and network resources, I want to have one checkout and use NFS to dish it out. Sun provides a 10-minute guide which doesn’t work out of the box on either of my Solaris machines, but Radu has a longer how-to that illustrates the setup better and how to trouble-shoot it. In particular, there are services that need to be set up and configured before you can start the NFS client or server.
Breakage overnight was minimal. A brief chat with Jos about Strigi has pushed some more patches upstream. PIM seems to be doing new and peculiar things, like taking the address of a bitfield. It’s Marc’s code, which means I’m more likely to assume it’s a compiler bug than bad code, especially since the error printed by the compiler is singularly useless: just ‘Error: Cannot take the address of a bit field.’ with no line number, no filename, nothing. Changing bool multipleCertificatesAllowed : 1 to drop the bitfield specifier : 1 gets things to compile, but that doesn’t seem right. I need to fiddle around to reproduce the problem in a smaller context and my initial stab at it didn’t work. So, chalk up one more un-mergeable patch.
Lastly, some kind words on the Eee. We now have two, a pink one and a white one (his & hers, you know), which means I get to watch Mignon use one of them as an inexperienced KDE / Eee user (although she remembers twm and vi from way back at university). It’s surprising how much is discoverable and just works. Printing was a brief moment of confusion because Postscript/default comes up in the printer dialog, but adding an Epson RX420 was straightforward. One more happy user.
The Wayback Machine ⏲ does not archive everything. Broken links are marked with a 💔.