The Philoshpy of Gentoo

Ever since I loaded up Ubuntu temporarily while waiting for Paludis to release a non-alpha, then failing to install with the alpha, and then reloading Ubuntu I’ve found myself questioning the point. Why Gentoo?  In fact I had a coworker the other day ask me the same question. He said to me, “You know, in a couple of years you’ll get lazy and tired of all that crap. You’ll just want it to work.”

Which leads me to wonder. Why? Why do Gentooers do what they do? Are the benefits worth the trouble? Are the potential risks worth the gains?

For now my answer is still “yes.” Tomorrow… who knows.

Enjoy the Penguins!

The Paludis Plot

The Paludis Plot, as I have coined it, is the conspiracy theory that the paludis developers, lead by Ciaran McCreesh, are attempting to rework Gentoo in order to only work properly with their package manager, paludis, and rework portage (the tree behind the package manager) and the ebuild system solely in thier image.

Being a paludis “fan boy” and a long time Gentoo user I can honestly say that paludis does not dictate ebuild development and they are not trying to mold Gentoo into some kind of strange paludis cult that will disillusion a large portion of Gentoo users. Thats the kind of statement I’ll get flamed for, but let me explain before you flame. Then you can flame on my friends, just like that guy in the movie. The problem here stems from several places.

First many of the paludis developers are also the same developers who are working on the ebuild system. This, I know, looks extremely suspicious, but they’re not trying to mold ebuilds into paludis only scripts. Thats ridiculous, not to mention it goes against the open source philosophy and their own personal goals. They’re personal goals then come into question.

Which leads us to the second reason; they’re goals. They’re goals are to make Gentoo better. To make a ports like system that works, in their words, more sanely than portage. And to give users a choice. They don’t care if you portage, paludis, or pkgcore. They really don’t and they’ll be the first to tell you that. They enjoy their tight nit community of Gentoo savvy people. If you’ve never been there go sit and #paludis and listen but do not speak. Unless your a member of the click you’ll probably get lost in the technical conversations that tend to ensue when people aren’t asking questions about paludis itself.

Finally, though, in the end when it comes down to it, I can’t think of anyone else I’d rather let lead the way in reworking Gentoo innards. Don’t get me wrong, they’re are smart non-paludis developers that are probably just as Gentoo savvy and have plenty of good suggestions. And I advocate that they all provide them. They key words there though are “smart” and “good.” Ciaran, and a portion of his comrades, come off poorly a lot, but if you can read past that, then read their code… thats why I’m a “fanboy.” (Which seems to carry a bad connotation and sounds a tad homosexual. Not a big fan of that word really…)

Enjoy the Penguins!

Excel Macros

Have you written macros for Excel? They’re easy, most of the time their annoying, and the editor given to you to write them in is absolutely horrendous. Other than that I really enjoy writing them. Strange isn’t it? I think its because its like witting a whole program that works flawlessly (for what you want, probably not for other people) and you can do quickly. Great ego booster if you can write VB (which is easy).

Enjoy the Penguins!

Uht oh, should of used Linux!

So I’m at work and I’m trying to defragment my hard drive. According to Firefox “defragment” is not a word. My guess is thats a product of a great number of its developers not really being Windows kind of guys. Anyways, I analyze the drive and it says 25% fragmented. Yeah, you don’t need to inform me how bad that is. So it runs for some where around 15 – 20 minutes and then stops. Upon looking back up hearing the hard drive stop I see that its quit but a majority of the stupid graph is still red! Upon reading the report I’m only down to 22%. What on earth? What kind of lame program can only defragment 3% of my drive. So I proceeded to check the files that could not be moved list. It, amazingly, only had one file in it. It was the log file for the mysterious Dr. Watson. Someone who has been in Windows since at least the 3.11 days but disapeared somewhere around 98 or XP, but yet, has existed behind the scenes the whole time. Just waiting to pop out and smack you.

Why couldn’t it move this log file then?  Upon further inspection the file “drwtsn32.log” was, and I shit you not, 25,034,688 KB in size. Thats right, I had a 23.8G log file on my computer. There is no need to continue on…

Enjoy the Penguins!

Vim problems and tarballs

In lieu of my recent install problems with Gentoo I installed Arch again just so I could have an OS up and running. I plan on trying to install Gentoo again, probably sometime around this weekend. I saw D. Robbins put his own stages which I will probably use. They’ll at least be more up to date the official ones. His goals behind doing this though make me wonder? Predictions predictions!

Anyway, naturally one of the first things I install after getting the OS itself installed along with X and Gnome is Gvim. It comes with vim by default but Gvim is sexier and gives me an easy way out if I need it. So naturally I went online, downloaded all my color themes from vim.org along with the secure modelines script by McCreesh and went off to the races. Thats the when the proverbial turd hit the fan.

I’ve complained about this before. Yet no one seems to care. If I considered Arch “my distro” like I think of Gentoo then I would probably be a little more proactive about it, but instead I will lay back and complain till I get my way. The default vimrc and gvimrc that it comes with break everything!! The worst part about this is they claim that by my local vimrc will override the global one in /etc. Lie! And trust me on this, I have a long, fairly complete vimrc, there isn’t much left to put in there. Thus my vimrc would not work for the life of me till I remembered this and deleted the damn things. Now everything works as I remember it working. Most people don’t realize it but Gentoo actually comes with a global vimrc as well. But the folks at Gentoo were smart enough not include anything in there that screws with your visual settings! It drives me nuts. I need to go through pacman’s documentation and pray there is a “don’t update this file” option or something.

Enjoy the Penguins!

What Gentoo Needs

If Gentoo wants to remain one of the prominent distributions in this world I believe some things need to change. Gentoo is, in my opinion, starting to become crusty. I mean all distributions have their issues but I don’t use every distribution just Gentoo so I can’t comment on everyone.

1. A new website – Gentoo’s website is terribly outdated. While fairly functional it looks terrible. The layout could definitely use some updating. What should it look like? I’m not sure but there are a lot of websites out there that I feel have much more attractive, yet still functional, designs.
   Good Examples:
   • Gnome
   • Arch Linux
   • Ruby

   Bad Examples:

   • Debian
   • Django
   • Linux.org
2. More active development – I have no gripe with 90% of our current development team. In fact many of them do a good job. But the sad truth of it is many of them don’t appear to be doing such a wonderful job. The whole fiasco with Daniel Robbins a while back is only one point to look at. Another good example is the Release Team. Its December 2007 and we’re still rocking 2007.0. Where’s .1 or .2? If we want to stay on top of the game we’re going to need to release more often than once a year. Period.
3. Package Manger War – Gentoo needs to end the package manager war. Having three viable package managers has its ups but it has a lot of downs too. Mainly its confusing to new and current users. Not to mention it looks bad to the rest of the Linux community. While most people would claim we should not care what they think, I believe their attitude is going to affect the minds of those searching for a distribution. And not in a good way. Most importantly, this is the kind of war that causes distributions to fork. Forking is not generally a good thing and with inevitably lead to at least one dead distribution. Or an eventual re-merger. Beryl is a good example of this.Things can be done to fix this though. Or at least get the ball rolling in the right direction. The package manager specification needs to be completed and certified. In my personal opinion portage should be dumped all the way around. Even taking up pkgcore instead of paludis as the official PM would, in my opinion, be a step toward the right direction. Though I would prefer paludis personally; portage is starting to show its age either way and needs to go. The back end needs considerable updating as well. Its no secret the ebuild used by Gentoo has its flaws. Replacing it with something more robust while fixing up the tree would be beneficial to say the least.

   I want to make it clear, though, that I don’t advocate rushing the specification through the process just to have one. This is one thing we need to take our time with but on the same hand we can’t slack off with.

Those are the three biggest issues I see. Comments, good and bad, welcome. I do censor foul language though…

Enjoy the Penguins!

Just give me one that works…

Fed up with my ever failing Gentoo install I decided it was time I just start dual booting with another distro I know I can get installed. So I downloaded Arch Linux, figured out how to burn ISOs from the command line and fired it up. I only downloaded the net install because its only 33MB.

I rebooted, selected everything in the “core” and “dev” package category to keep it simple for me. Unlike a lot of Gentoo-ers keeping it as lean as possible is not an obsession of mine. So I went through install as usual, configured up my files real nice, and rebooted. Everything was going real smooth until I finished the install of xorg. When installing yelp it returned some funny error about “command not found.” Well thats totally ridiculous. How can the command not be found. What command was missing for the installation of that package I don’t remember right now but it was after that I noticed my nvidia driver was not working either. It kept returning a module not loaded in kernel error. Come to find out every command revolving around modules was missing; rmmod, modprobe, you name it I didn’t have it.

This morning on #archlinux they recommended I check my profile. I’ve never really investigated this profile business so I’ll have to do a little research. Its definitely something I should be more familiar with. This would probably also explain why after installing fluxbox and firefox, “firefox” was not a recognized command.

This all leads to my final Arch Linux gripe. Why on earth is the default fluxbox menu full of programs I don’t have installed?

Enjoy the Penguins!

I was right

I was right, Paludis does not display summary information about conf files needing to be updated if the install fails. I have this bad feeling that me not realizing that till it was to late is what is now causing all of my econf errors. I just find it awfully strange that every major set of packages I try to install (Gnome, Xfce) all fail with econf.

According to the ticket though this should be fixed by 0.26. Thank goodness!

Also, I was chatting on IRC today. It looks like custom output formats and custom color schemes are waiting on a good patch. I think you should submit those by the end of December at the latest. Thank you.

Enjoy the Penguins!

About to give up

As previously mentioned I’m in the process of reinstalling Gentoo. Things just never get any better. Apparently somewhere along the way I failed to to do a update-modules which meant I didn’t have a modules.conf file. And then somewhere else I also lost my udev.conf file. Where on earth it went I have no idea. Well I managed to fake the udev.conf with a empty file name udev.conf but I ended up having to use the modules.conf file from the liveCD just so I could boot. After doing that paludis finall told me I had something like 40+ conf files that needed updating. Why I failed to notice that before is beyond me. It makes me wonder if it even ever told me that. Does paludis even check for new conf files if it ends in error instead of success?? I have a feeling it does not and thats probably why none of those files were ever updated.

After making it through that ordeal I could finally boot and work. So I finished compiling Xorg and then went on to compile Gnome. Everything started going great until I got to libgnome. Which is where I proceeded to get an econf failed message. Discouraged I decieded to just screw Gnome I’ll go with Xfce. Sounds nice except for the fact that to eventually ended in an econf failed message. Doubly discouraged I went back just for the hell of it and tried Gnome again. This time, for who knows what reason, it made it past libgnome but failed once again with an econf error on another package. Hoping I could just do this over and over again till they both finished I tried to start the Xfce install again. I had no such luck though. Both my Gnome and Xfce installs are now both stuck with an econf error.

This all leads me to another point a tad off topic that really bothers me. The default location paludis wants to put my names cache is in one of portage’s folders, I think its /etc/portage. Well last night I hoping that maybe it was just paludis screwing up so I proceeded to re-sync portage and test it. At the end of the portage sync it proceeded to delete paludis’ cache! Whats up with that?? Something is not right there.

Enjoy the Penguins!