:Linux for the rest of us

I found a weird glitch in amarok last night when I was messing around with the context browser. Basically, anything I clicked on that called an external URL or a scripting language didn’t work. External URLs would “download” a file to a local cache directory (I put that in quotes because even thought there as a download dialog and a progress bar, nothing was actually ever downloaded) and I’d be given an error message about either the file not existing, or it being an invalid format. For scripts, I’d get the actual code in a browser window - nothing at all was executed. I looked in my amarok preferences, and my browser was set to “KDE Default Browser” - I use konqueror, mostly, so that was fine (or so I thought). So I started browsing the amarok bugs, and found a few similar bugs filed and closed with snippets of SVN commits or similar in the closing comments, some citing the browser as the issue, not amarok. I decided to change my browser from “Default” to Konqueror, and BINGO - things worked fine again. Weird, but I’ll take it.

On another note, lyrc.com seems to have been domain-jacked. I can’t pull any lyrics from the site, and actually browsing there brings up a generic bullshit search-engine-whore placeholder site rather than a real site. Not sure if they just forgot to renew/registrar lock their domain, or if they closed their doors. Either way, it’s a bummer…

No Comments

Well, it’s been quite a while since I’ve had the time or motivation to write my thoughts down regularly. This isn’t what I want to apologize for, though. I’d like to apologize for many of the negative things I’ve said about several linux projects in the past, Gnome especially. I’ve been very hard on gnome on occasion in the past, and honestly it stemmed from frustration with the current state of a lot of the “Linux Desktop.” I suppose a lot of the rest of my frustration stemmed from the current state of linux distros - I still can’t wrap my head around 6 month releases, and I still have a few major bones to pick with most linux distros. For instance, I really appreciate all of the things kubuntu does to simplify installation on a laptop, but I really dislike the amount of control they take away from me. This leads to some very conflicted feelings while using it.

Still, talking shit in a blog is more or less useless, and singularly unproductive. After reading Thom Holwerda’s recent op-ed article over on osnews.com I realized that I’ve not only been unproductive, but have been a complete asshole. People working on these projects obviously believe in them, and I certainly believe in linux and the entire FOSS development process. I should be decent enough to be constructive in my criticisms no matter how frustrated I may be.

I’m planning on removing a number of posts I have made here in the past - certainly they will have been archived by google or archive.org or wherever, but I feel like it would be the right thing to do in any case. Not because I’m trying to hide what I’ve said in the past - I said those things, and I certainly stand by them - but because I feel the folks working on the projects I’ve bashed in the past deserve more and better. Criticism doesn’t have to be insulting, and mine will hopefull not be in the future.

So, sorry to anyone who I’ve offended in the past during one of my rants - I promise to try to do better and be more constructive in the future.

1 Comment

The last major upgrade in Arch brought a number of changes with it, one of them being kernel 2.6.19. I use a mobo with an ata_piix sata controller, and it’s suggested that you do a number of things to keep that upgrade from breaking your install, especially name your disk devices in a new way. If you look in /dev/disk, you’ll probably see an number of by-* directories in there if you’re running an even semi-recent kernel (for instance, edgy now uses /by-uuid to name all of its disk partitions — take a look in your /etc/fstab if you’re running it). I decided that using by-label is a lot easier for my own needs since “/dev/disk/by-label/boot” is a hell of a lot easier for me to remember than “/dev/desk/by-uuid/e692e3dc-94c1-4a3e-9de2-77dc7301b393.” The thing is that by default, on Arch at least, there is no /dev/disk/by-label directory. This confused me for a while since running the commands to create labels didn’t add this directory - until next boot, that is. This confused me for the better part of an evening. Here’s a quick and dirty how-to:

The first step is to make your labels — make a label for every partition you mount at boot time as well as swap. Since I use ext3 the command is “e2label /dev/whatever labelname for example “e2label /dev/sda1 boot.” If you use a different filesystem, google around for your label creation command - they are all very similar. For your swap partition the command is going to be “mkswap -L labelname /dev/whatever” - for example “mkswap -L swap /dev/sda4.” Once this is done, reboot and verify that the things in /dev/disk/by-label are pointed at the correct partitions. The next step is to edit your /etc/fstab and /boot/grub/menu.lst to reflect these changes. Mine ended up looking like this:


/dev/disk/by-label/boot /boot ext3 defaults 0 1
/dev/disk/by-label/slash / ext3 defaults 0 1
/dev/disk/by-label/home /home ext3 defaults 0 1
/dev/disk/by-label/swap swap swap defaults 0 0
/dev/disk/by-label/multimedia /multimedia ext3 defaults 0 1


and


# (0) Arch Linux
title Arch Linux [vmlinuz26]
root (hd0,0)
kernel /vmlinuz26 root=/dev/disk/by-label/slash vga=791 ro
initrd /kernel26.img


Apparently you can use the form LABEL=labelname rather than /dev/disk/by-label/labelname, at least in fstab, but I always prefer absolute paths. Anyhow, I upgraded to 2.6.19 completely painlessly after following this set of steps and didn’t have any of the problems many other people reported. I didn’t have to add the kernel option “earlymodules=piix” - my setup worked fine without it - but you may need to if you have an intel mobo and use the onboard controller.

Have fun!

No Comments

Man - it’s been a loooooong time. Life has been keeping me busy, what can I say. I have a ton of stuff I want to blog about, but I think the first and easiest will be frugalware. Recently, I started messing around with a new distro, frugalware (can I say frugaware again w/o linking or getting to the point?). Frugalware is a distro that is part Arch, and part Slackware, I guess. It uses the Slack’s init scripts and installer, and Arch’s pacman to deal with packages - that’s pretty much what I mean. In any event, it’s *that* variety of distro, not at all *buntu-esque, and not a whiff of debian. Be prepared for some “hands dirty” time is what I’m trying to say.

So, I installed it on a whim, but also on the Recommendation of “Wally Balljacker” over at the Linux Link Tech Show forum. A word about the installer - it took me the better part of two days to get frugal installed because the installer kept bombing out - I have no idea why, but it did. I sha’d the isos I’d downloaded as well as the cd’s I burned, and everything came back correct, but I re-downed the isos just to be sure, same deal. It kept barfing on the checksums of random packages - never the same one(s) as far as I could tell. In the end, I got it installed by using the netboot image and installing base and networking only. This got me a base system I could build the rest of my desktop on, which worked out fine. However, I was unimpressed with that experience, to put it mildly.

Frugalware has a stable and a current release. I think this is fantastic. For some applications, I absolutely do not want the latest and greatest, I want stability. That said, I switched to -current pretty quickly. I’m not sure what gets backported to -stable, or whether there are security updates for it. I imagine there are, but you’ll have to look around their site to find out for sure. Also, I don’t know how difficult it has been historically to upgrade from one stable release to the next. Still, this is a good thing in many ways for many people.

First impressions - nice polish, and large repos. Seriously. Once I got KDE up and running, and had installed the basic stuff I always want (all of it was there, more or less), I was very pleased. There is a lot of polish here for what appears to be a fairly small distro. Nice fonts, the win codecs, dvdcss, etc. It’s all there, OOTB. For me, nice fonts is a big one - I’m not very good with font problem, and they are part of the guts of a system I don’t want to become overly familiar with. Installing fonts? Fine. *Tweaking* fonts? No thanks. In any event, I was very pleased with what I saw. I was also pleased with the large number of packages available - again, it’s amazing how many packages there are, and how up to date they are - I seriously thing there are only a handfull of devs. That must take s lot of work.

There was also a lot of attention paid to the boot-up and login process - if you don’t like to look at your machine booting, you don’t have to. There is a grub splash, a boot splash, and a KDM login screen that all follow what I think is a very elegant theme. I was simply amazed it wasn’t the standard framebuffer deal in the upper lefthand corner with tons of “ugly text” scrolling by. Now, I don’t generally want a splash screen myself, but that’s me - I guess most folks do, and they get it here. In spades. Nice, pretty spades. Good stuff.

If you’re familiar with ABS in the Arch world, there is a very similar tool for Frugalware called repoman (”Repoman’s always intense!” - sorry, couldn’t help myself) that does much the same thing. They are no more difficult to put together than PKGBUILDS - they basically follow the exact same format, with a few extra mandatory fields. In fact, you can take PKGBUILDS from Arch, and with a little doctoring have them install packages in frugalware. Sometimes it’s as simple as adding an Arch=() line, sometimes it’s a bit more of a PITA. Still, it can be done, and that’s awesome, especially considering ABS is IMO one of the best features Arch offers. If you like getting your hands dirty, you can get them as dirty as you like here.

Now, on to the stuff I’m not crazy about. I already covered the install, but I have to assume that’s far more the exception than the rule. Here’s some other stuff:

I’m not crazy about the use of the slack inits - I just don’t like them. I’m sure it’s my ignorance speaking here, but I don’t see how they are in any way better than sysv inits - sure, they’re easier to manage in that you don’t end up having to write complex start/stop/restart scripts, but IMO they are just as obtuse and just as ugly. Again, that’s me. I’m sure a diehard slacker can tell me that I’m an asshat and list 10k ways they are obviously different. Good. I’ll still prefer the arch inits. I’m sure someone else will point out there aren’t really “inits” - fine, I’ll take /etc/rc.conf over /etc/rc.d/rc*.d/* - I haven’t seen a use for runlevels in years. At the same time, if you’re going to use them, don’t have X start in every MFing one. I’m looking at you, *buntu…

The documentation is kind of lacking. The install docs are great, as are the repoman docs (though these could use a lot more examples of working scripts) - the wiki, the mailing list, and the fora are, well, deserted. I’ve been on the mailing list for an entire week at this point, and I’ve seen a total of 5 messages, including one I sent. The first page of every forum has posts as old as a year old. The wiki, well, as far as I can tell, it isn’t actually maintained. I’m left wondering whether most of the know-how is exchanged on irc, or whether the community is so small that this is all the traffic it generates. On the one hand it’s great to be able to speak one-on-one with a dev quickly and easily, but on the other, I’m worried Frugalware may simply implode like Rubix did. It also makes it very hard to do research w/o having to post in the fora and wait for answers - searching the fora generally turns up very little, or very old information.

Finally, there doesn’t appear to be anyplace like linuxpackages or the AUR where users can share their FrugalBuilds. This kind of sucks, but as I mentioned, the Frugalware repo is amazingly large for such a small/new project. I haev to assume that lots of the builds were borrowed from Arch, but still, someone has to maintain them, and the stuff in -current is all very up to date. Hell, Frugalware tends to run newer versions of pacman than Arch does - not sure that’s a Good Thing, but it’s a fact.

Still, that’s not much in the bad column, at all, and 99% of it isn’t really “bad” in any event. I’d definitely recommend Frugal to anyone who likes either Slack or Arch. Hell, I’d recommend it to most folks, actually, assuming that the installer problems I had weren’t commonplace. Of course, it’s still not going to replace Arch as my main distro, but it might find a home on my lappy. Time will tell. In the meantime, I give Frugalware two big thumbs up!

No Comments

Bummer. I was really hoping Rubix would take off - it had a lot of potential. I’m kind of surprised that he left Arch out of his parting words and instead gives props to slack and debian. Oh, well, Arch has been very good to me for quite a while now - hardly a hiccup for months, and I’ve already switched over to initcpio which looks exciting. I wish Joshua luck - he was always very courteous on his forums.

Read all about it here.

No Comments

Obviously everyone knows about google earth being released for linux by now. I installed it on my arch box (via the AUR package) and it is sweeeeeeeet. I have a decent, but not insane video card (nvidia 6600gt) and it is very fast and very fluid. Big-Ups to google for treating the linux community like the count - we do!

1 Comment

Well, I’ve got dapper running on my lappy now - I had it on one of my testing drives as well. Upgrading was asstastic - a dist-upgrade completely fucked up my laptop install. I’m sure that someone can point out that I did something wrong, and didn’t read something important somewhere, but in the end, I should be able to do a dist upgrade w/o rebooting to a useless system. In general, if Ubuntu really think they can give either windows or macintosh a run for their money, this kind of shit will have to be addressed. I do not mind having to go to third parties to get drivers, fonts, and codecs - I fully understand why this is the way it is. I DO NOT want to have to reinstall my system every 6 months to remain current though, and I cannot imagine anyone outside of the linux zealot community wanting to do so either.

Once I had freshly reinstalled kubuntu, I quickly discovered that a) hibernate appears to be mostly broken at this point, and b) it also take a LOT longer to do than it used to once I got it sorted out. The basic problem is that fn+f12 no longer causes my thinkpad to hibernate - I did a lot of fooling around, and eventually set my lid button up to hibernate the lappy through kcontrol. Actually hacking the various scripts didn’t work, and unless I was root, I couldn’t hibernate the lappy via the cmd line. Now that it’s sorted out, it still sucks - it takes (very litterally) ~1 to 1:30 to bring it out of a suspended state - that’s something like a minute longer than it used to take. I wish I understood why ubuntu decided swsusp is a better solution than suspend2 considering they patch the living hell out of the kernel anyway. Oh well.

It’s always fun to remember why I love arch so much - occasionally I pine for a more “professional” distro, or for a distro that has more polish, or is less difficult to maintain. This fabled beast simply doesn’t exist as far as I can tell. In the end, there hasn’t ever been a distro on any of my boxen I like better than arch, and in general it works better than any other distro as well. While dealing with breakage from upgrades sucks ass, it’s still better than any of the alternatives. I’m thinking about giving it another run on my lappy once I have some free time.

Honestly, I understand why so many people say that linux on the desktop is doomed.

No Comments

I’m really liking amarok 1.4 - it’s truly amazing. The iPod interface has really, really improved, and so has its handling of stuff you’d want to stick on your iPod - you can even choose to have your potcasts automatically sync. Awesome! However, I really dislike the way they’ve chosen to start naming podcasts - it makes playing them (or even using them) in any other program a total PITA. Essentially, the “new way” is to include the entire url of the feed in the name of the actual mp3 - the tags still reflect the correct info, but if you need to find the actual mp3 you’ve downloaded outside of amarok - good luck. There have been multiple bug reports filed which have been closed with “We MEANT to do it that way” as the reason. I think its fairly obvious I’m not the only person who thinks this new standard bites the bone. It also crashes mysteriously - so far it’s never crashed while I’m using it, but it’s very common for me to come back to my computer after a few hours away to find that I no longer have amarok running. This could be an arch thing, and isn’t a big deal, but it’s something I’ve noticed.

On a personal level, and this is going to be very cunty on my part, I’m a little disappointed with the devs. I recently donated $50 to the amarok team - not a big sum, but then I didn’t have to donate it either. Usually when I donate to projects I get an email with “Thanks!” or something in it. Not even a whisper from the amarok folks - that’s a little lame. Most folks who use open source use it because it is free as in $$$ - seriously, I stopped posting on several boards because of the rampant “WHY THE HELL WOULD I PAY FOR SOFTWARE?!?!?!?!” bullshit I see on most of them. I think that if devs want more donations they might want to at least have an auto-generated reply saying “Thanks for donating to our project!” Anyhow, I’ll probably donate to them again next major point release, but like I said, I am/was a little disappointed (and am being a little cunty).

3 Comments

Yeah - it’s been quite a while. I have a 1 month old baby boy (my second) and that’s kind of put the brakes on most spare-time activities of mine. I figured I’d better pop in and update this sucker though, or risk becoming the next Allan Metzler.

So, here’s a tip for those of you using wine on Arch (and perhaps other distros as well) - I launched wine yesterday, for the first time in a while only to be greeted with mega-bizarro glyphs in the place of my regular fonts. These could have been a real language, or they could have been crap - I couldn’t tell, though, since they were also teensy (6 point font?). My verbose console logging showed hundreds of these:

err:font:ReadFontDir Can’t open directory “/usr/bin/../lib/../share/wine/fonts/”

which is just as well since that directory doesn’t exist. I searched all over the web for info on what config file to munge, and found nothing outside some posts about suse, and a non-existant file in /etc. So, then I started looking for what wine expected, and tried to download the ms-ttf files into the directory they go into “naturally” when installed - this did nothing. Finally, I did a recursive ls of my ~/.wine directory, and found a few different directories that looked promising. After a bit of tinkering, I did the following:

cd .wine/drive_c/windows/
rmdir fonts !!! ### NOTE: if you have fonts in this dir, you may want to “mv fonts fonts.old” instead ###!!!!
ln -s /usr/share/fonts/TTF fonts

This worked perfectly, restoring my fonts to their original glory (?). This will probably work on any distro, but you’ll need to find where your distro stores its ttf fonts and link to that directory instead.

A side note (yeah, it’s another rant):

Wine is a great project - I’m firmly of the opinion that it’s a good thing, and piss on the folks who say it stifles creativity in the linux community, or whatever. The wine documentation, however, sucks (and the wine devs don’t want you making it easier on yourself by using the winetools package, either, for some reason). But if the wine documentation is bad, the wine “how to fix this problem you’re seeing” documentation is non-existent. As I mentioned, I couldn’t find a fix for my problem, or even info on how fonts are handled anywhere. That was easy enough to fix after some digging - my next problem is not even fixable as far as I can tell, and there are tons of reports of it all over the net, and yet not a single explanation of wtf the issue is. I cannot connect to the internet (or even network) with certain apps, and get lots of these in the logs:

“fixme:ras:RasEnumConnectionsW RAS support is not implemented! Configure program to use LAN connection/winsock instead!”

I found many reports of this issue in various forums, and mailing lists, but again, not a single fix - generally the suggestions were “Don’t use that app, use the linux one instead!” This isn’t a fix. Hell, there is even a bug report, and a bug fix issued with this issue referenced - the bug and fix? Make the error message more readable. As far as I can tell, all that was done was a reformatting of the error message. There is simply no way for me to change the settings for an app if I have no information (or even a flipping hint) available to me - running IE and going through its connection settings was useless, as was running inetwiz (which crashed constantly). It’s too bad, since IME, this tends to be the rule rather than the exception when speaking of linux projects and documentation.

No Comments

Hmmm. Well, here’s a rundown of what I discovered this week:

- the kubuntu flight6 cd is useless. I’ve downloaded the iso three times on different boxen, and I can’t boot from it on a single machine I own. That’s more or less all I have to say about that.

- If the number of packages available for fc4 is slim, the number for fc5 is non-existant. Looking at my favorite 3rd party repos, I see none of them have started packaging stuff for fc5, and any of them that have packages for fc4 have a very small selection. I’m not sure what this says about redhat/fedora on the whole, but it makes running it kind of unappealing.

- I’m still not at all comfortable with slackware or vector. It seems absurd to be to have to search around on multiple fora just to figure out which version of which third party repository’s package will do what you want/need. After fucking around for two days trying to get hal and dbus working a while ago, I gave up. Someone since then has pointed me to the packages I’d want to use, but I honestly doubt that I’ll be revisiting those grounds any time soon. Can someone please explain to me what’s “simple” about slack? As far as I can tell, the sysinit procedure they use is at least as bad as sysV, and certainly no better. I far, far prefer either redhat’s or debian’s /etc layout. I guess I just don’t get what all the fuzz is about.

- Rubix is a promising distro - very promising IMO. However, nearly everything I tried to add from the community repo was either built against packages in current, or depended on packages in current. So essentially that means to run rubix, and have much at all in the way of packages, you have to run the unstable branch. I’d really like to see creapkg ported from arch (maybe I’ll make a stab at it) - at least then installing software wouldn’t require writing multiple pkgbuilds from scratch. Still, it has a lot of good points - the rc.conf enhancements are really nifty, and te less-bleeding aspect combined with the more-security-conscious aspect are a recipe for a real winner IMO. I really, really hope it gains some traction.

So - even if arch farts in my face once every few weeks, it’s still the best option right now, I guess. Like I said - the only thing I dislike about it is the bleeding edgeness - otherwise it’s what linux should be, for me. Well, ok, it would be nice if it had debian’s absurdly huge repos too, but let’s keep this jalopy on Real Street…

No Comments