Yep, another bad case of *Oof*… I blame those Thursdays… Never could get the hang of Thursdays… and sometimes they do seem to spill right out into the Friday afternoon, too!
Anyway, today's *oof* — just how deeply asleep must I have been so that, instead of a simple chmod 644 blah.cfg I tried to do a chmod u+w blah.cfg — and managed to actually type out chmod o=w blah.cfg, thereby bringing the poor file from 444 to 442 instead of the desired 644…
And in case you have no idea what I'm talking about at all — no worries, it seems that I myself have no idea whatsoever!
First *oof* for today: if your nVidia drivers (from the x11/nvidia-driver port) are misbehaving and mysteriously rebooting your machine, consider removing “options WITNESS” and “options DIAGNOSTIC” from your kernel configuration.
Which leads us directly to…
Second *oof* for today: if you remove “options WITNESS” from your kernel configuration, “-DNOCLEAN” is emphatically not your friend!
Still, we live and learn, and now I've finally got a working X.org installation on my three-weeks-new laptop.
It will be instructive. It will be interesting. It will be fun. Just two days left for the early-bird registration - as some'd say, be quick or be… okay, not necessarily dead, but at least poor - or at the very least, poorer by about 30 EUR :)
This time, to paraphrase a good old starship, nothing's gonna stop me now! Or rather… there's really nothing wrong with the original - nothing's gonna stop us now - 'cause I just don't believe that I'm going to be the only Bulgarian in attendance! :)
Well, it seems that the recent rumours about the demise of BlueTooth have been greatly exaggerated. Or, at least, concerning my Nokia 6310i phone, Iva's Cambridge Silicon Radio USB BlueTooth adapter, and FreeBSD 5.3-BETA2. In fact, thanks to the awsesome work by Maksim Yevmenkin for developing the BlueTooth stack support for FreeBSD, Pav Lucistnik for documenting it in the FreeBSD handbook to a point where all I had to do was follow the instructions and pretty much copy/paste the command lines after reading a couple of manpages for good measure, and undoubtedly countless other testers and contributors, I am writing this very blog entry over a Dial-Up over BlueTooth connection - and it seems to work quite well, doesn't it now? :)
And now it's off to the Faculty of Mathematics and Informatics for me, to torture the poor students who have come for a retake of the Network Security exam... Wish them luck! (them, not me, obviously ;)
Gee, take a week off and people come up with the strangest things :) First, there are three new FreeBSD committers in a single 24-hour period (congratulations again, Cheng-Lung, Dejan and Alexey!), and then this - do the daily CVSup, cvs -q up src/, build/installworld/kernel, reboot, and...
[roam@straylight ~]> uname -sr FreeBSD 6.0-CURRENT [roam@straylight ~]> uname -a FreeBSD straylight.m.ringlet.net 6.0-CURRENT FreeBSD 6.0-CURRENT #4: Mon Aug 23 15:33:28 EEST 2004 email@example.com:/fs/5/usr.obj/fs/5/usr.src/sys/WAYLIGHT i386 [roam@straylight ~]>
Welcome to the brand new world of 6.0-CURRENT, have a nice day! :)
However, at present I (and I suppose a couple of other developers) am more interested in the stability of the 5.x-STABLE branch (RELENG_5 in CVS), which is currently in code freeze pending the release of FreeBSD 5.3, so right now straylight is busy doing a "cvs -q up -rRELENG_5". Guess I'll be back with -CURRENT a couple of months after 5.3 is out, but then you never know :)
EuroBSDCon 2004 is scheduled for October 29-31 in Karlsruhe, Germany. I hope I'll be able to attend this year, although it's still too early to tell...
Since my laptop's out of commission for a while, and since I'd thoughtfully backed up a couple of its filesystems just a couple of days before this happened, I thought I'd just restore a couple of pieces of my home dir and go along with e-mail, ICQ, and all that jazz. Unfortunately, it turned out that things were not quite so simple.
My laptop currently (well, when it's working, I mean) quadruple-boots between FreeBSD 4.x-STABLE, FreeBSD 5.x-CURRENT, Debian Linux and Windows XP. Unfortunately, when I decided to do the backup, I did it via booting FreeBSD 5.x into single-user mode, then, for some stupid reason or other, had to mount the filesystem containing my home dir as read-write instead of read-only... and dump(8) decided that since it was dumping a live filesystem, it would just create a snapshot and make a backup of that. That's all fine and dandy, except that filesystem snapshots seem to be dumped as UFS2, *not* UFS1 - and now I have a big UFS2 dump which I cannot restore on a FreeBSD 4.x machine, since 4.x's dump(8) does not grok UFS2 :(
Guess it's really time to sacrifice another machine and find a way to connect the laptop's HDD to a "real" IDE controller, so I can use it again... and *then* make a real UFS1 dump for future reference.
And before anyone takes this entry the wrong way: no, I'm not blaming either FreeBSD 5.x for coming up with a better UFS format, or 5.x's dump(8) for making UFS2 dumps of snapshots. These are both great and logical decisions; the problem was yours truly, dumping a live filesystem for no good reason at all.
The night before, over a beer or five, Vasil Kolev, Georgi Chorbadzhiyski, Boyan Krosnov, and I had an interesting discussion about the difference in the development model, distribution model, and/or design philosophies between the various Linux and *BSD OS's (or should I say distributions?). When Vasil asked if there were any *BSD distributions besides Free/Net/OpenBSD, I couldn't come up with a lot of examples, although I knew there were some. Well, as a first approximation, there are TrustedBSD, EkkoBSD, The FreeBSD LiveCD Project, and just today I heard about a new one, LittleBSD, which even sports some anime artwork (and there's more around the site).
I know there are more, it's just that I cannot think of them at the moment.
I finally got around to porting my patch for the FreeBSD driver for Intel ICH chipsets to -CURRENT, and now I've got sound on my laptop again! For some reason, my 82801DB ICH4 chip does not want to play nice with FreeBSD's attempts to address its memory directly - could this be related to its being labeled as revision 2?
Anyway, if anyone else should stumble on the same problem, the patch I just sent to the sound@FreeBSD.org mailing list may be found in my highly-unofficial patchsets in the src5 collection: here or here.
Next step: figure out why the USB drivers won't have anything to do with the USB serial port that I'm trying to plug in.