Gentoo Announces 'Seeds'

rvale writes "Gentoo has announced a new project called Seeds. Aiming to provide out of the box images for various common tasks, it could be the answer to the common complaint that installing and customizing Gentoo takes too long. However, with other developers and Council members complaining that the project was improperly set up and those backing the project refusing to back off, lending weight to recent claims that Gentoo is suffering from management problems, will what could be a massive step forward degenerate into a repeat of the Sunrise disaster?"

Comments

They're in for a world of hurt

[GroeFaZ]

The Neo-Vickys will put all their resources into stopping them.

Re:They're in for a world of hurt

[cloricus]

Only the first thirty posts and there are disgusting posts from pissed off users... I hate to be the one to say it - though having to read through several sets of Debian issues and their comments on /. and seeing a lot of Gentoo users flaming I think I should - maybe Gentoo should get its own house in order before it attacks other distros?

Now for the constructive part of this post. Why is this even a problem? Seriously, sit down and talk through the issue, it's not that hard. They don't want to do what you want and you don't want them to do what they want...I've heard of this one before I think it's called life? Gentoo while a bit off some times is a damn good project and one of the shining stars for the Linux communities...and you keep the windows newb gamers off the rest of our backs so for that you get extra brownie points. So for the sake of every one else just talk it through instead of fighting...

Re:They're in for a world of hurt

[networkBoy]

Oh how I wish I still had points ...

Re:They're in for a world of hurt

[Quantenmechaniker]

Yay, Neal Stephenson, baby!

Ill informed post

[Toba82]

What exactly is this summary asking? Whether or not making gentoo installable without spending endless painful hours of your time on it is a good thing?

I think we all know the answer to that.

Re:Ill informed post

[ArcherB]

Whether or not making gentoo installable without spending endless painful hours of your time on it is a good thing?

That's already been done. It's called STAGE3

Re:Ill informed post

[KhanReaper]

I hate to admit it, but I think that many people have misinterpreted what Gentoo really is and for whom it is geared. Let's be candid: it's really not about excessive CFLAGS.

Take a good read of this article; it outlines some of the fundamental differences in philosophy between BSD and Linux. In some respects, Gentoo's portage system attempts to reconcile the differences between BSD's ports tree and the absence thereof in Linux as well as the concept of perpetual updates through make buildworld. (I know a guy who's maintained the same install of Gentoo on his laptop for over four years who has kept it up-to-date by using portage without a re-install. Talk about impressive for a Linux distribution.) Yes, it is true that Gentoo does not have a native pkg_add that FreeBSD does to install ports, but what Gentoo offers is as close to that as one can get in Linux; and it is one hell of an improvement on the base concept, might I say. In many respects, if you want to criticize Gentoo over having to compile things to keep it up-to-date, then BSD ought to be brought up for discussion.

Still, it is nice that Gentoo can be updated without having to perform a complete re-installation of the operating system. I hate to say it, but performing "s/old release/new release/g" on /etc/apt/sources.list, apt-get update, and apt-get dist-upgrade is not always as clear as one might expect. When the average user who lacks strong familiarity with dpkg's options is in this situation, I have seen the results: They are very depressing. And while it is true that emerge updates can break, they will at least teach the user in time how to deal with them and learn quite a bit. The same can be said about other distributions, too, so the exclusivity of this issue to Gentoo is really a moot point.

.

What about customization? Sure, some BSD packages may have makefile-based booleans, but in no way are the centrally documented or are they centrally documented. FreeBSD KNOBS comes close, but it still is not exhaustive. There is no real comparison with USE flags. If BSD had it so well, I wonder why people are trying to port portage to BSD. (I love BSD, mind you, so I am not being unreasonably harsh on it.)

What about fundamental design? It is meant to be flexible and dynamic. Ever notice how many directories are suffixed with ".d" in /etc on Gentoo? A lot are. Yes, some other distributions do use the enumerated ".d" directory paradigm, but none seem to do it as much as Gentoo. Gentoo seems to use ".d" directories whenever it can. So if a new package wants to add something to the path, it merely adds another entry to /etc/env.d which specifies this path. I find this system so great, that I've re-implemented it in Debian/Ubuntu across 100+ computers at my work for the special in-house, non FHS-friendly applications. Talk about a compelling innovation.

And when it comes to configuration changes, Debian has debconf, which allows some packages to preserve changes across updates through configuration file regeneration. While this is nice for preseeding, this is not helpful when there are major updates or when you've made hand-made modifications. Yes, dpkg will bring about a diff of the two files, but does dpkg's integrated configuration diff mechanism really hold its own against Gentoo's dispatch-conf? If you've used dispatch-conf, the answer is no.

Yes, it is true that there are some quality assurance failings with packages in Portage, but let's put that aside for a moment. When it comes to making packages for Gentoo, it certainly beats making them for Debian. Yes, Debian has its nice policy manual, but it is not always up to date or the easiest thing to read. Gentoo's documentation let's a first-time package builder build a package in very little time; whereas Debian or Redhat's syste

Re:Ill informed post

[Anonymous Coward]

Huzzah!

Re:Ill informed post

[Enderandrew]

And Sayonora and Kororaa and Emission, etc.

Re:Ill informed post

[ozbird]

It's also the default installation method as of 2006.0 (or thereabouts.)

Personally I prefer STAGE1, but you can use STAGE3 to get up and running quickly and then replace the binary packages with built-from-source later.

Re:Ill informed post

[ArcherB]

And Sayonora and Kororaa and Emission, etc.
VidaLinux, lxnay, Sabayon....

Excellent point! These distro's are even easier than a stage3 and still give you the ablility to update everything as you wish by setting up your use flags and running
emerge -uDav rebuild
(or something like that).

Re:Ill informed post

[Anonymous Coward]

Wait... there's Gentoo documentation that actually is useful, and expands upon important common topics and not the specialized one in 30000 people use topics?

I disbelieve.

Re:Ill informed post

[DAharon]

I completely agree with you that people who are unfamiliar with Gentoo, are largely ignorant of what it's like to actually run a Gentoo system. CFLAGS? I don't really run any fancy CFLAGS. USE flags? If a program can benefit from a certain use flag, I add it. It's Portage and the Gentoo toolset that really makes Gentoo a joy to work with. I am always hearing people complain about how hard it is to install, and how it always breaks. I always ask myself, "What the hell are they doing"? I have been running the same Gentoo install for at least 3.5 years now on my desktop. I play games, surf the internet, write papers, download stuff on it. It also acts as my wireless router, MythTV machine, and file/ftp server. I emerge -uD world at least once a week and encountering a problem is a rarity. I just don't get it. How can everyone be having such a devil of a time maintaining just a basic Gentoo "surf the internet" computer, when I've had such an easy time for so long?

Re:Ill informed post

[Pausanias]

Does emerge break things frequently? There seem to be too many posts for comfort on this and other Gentoo-related forums with users complaining about emerge breaking their system.

I'm an Ubuntu user here, and I've never had apt-get update break anything (without having been responsible for the breakage myself through fiddling with the configs). I think Ubuntu users would be furious if a security update wound up breaking their system.

As people who rely on a Linux computer, we really do put our fates in the hands of apt-get / yum / emerge, and it's a real feeling of betrayal when these tasks let us down. I remember in the old days of Red Hat 9 how horrible up2date was in this regard.

Perhaps I'm wrong, and it's only a few loud people doing this... but as a potential Gentoo user reading slashdot, I wouldn't get the most comfortable feeling about relying on emerge. It's not that I don't have the knowledge or desire to fix things if they're broken by the package manager. It's that I don't have the time. My work is too demanding for me to spend a lot of time cleaning up after broken packages.

And while it is true that emerge updates can break, they will at least teach the user in time how to deal with them and learn quite a bit.

Yes, it is true that there are some quality assurance failings with packages in Portage, but let's put that aside for a moment.

Re:Ill informed post

[Anonymous Coward]

If BSD had it so well, I wonder why people are trying to port portage to BSD. (I love BSD, mind you, so I am not being unreasonably harsh on it.)

To draw an analogy: if the Linux kernel was so well, why would Debian try to make Debian GNU/KNetBSD and Debian GNU/KFreeBSD (BSD kernel with Debian (APT, glibc, plus all normal other software)).

The answer to both questions is simple: because they can. I've not seen anyone in the *BSD community being particulary happy about Portage. Or adopting it. I've not seen Free|Open|Net|DragonflyBSD adopting it. They're happy with Pkgsrc mostly last time i checked. They're not against it either; why would they? What purpose does it serve? Its F/OSS after all.

The equiv to USE flags on my BSD systems is FLAVOR flag. It does more than enough for me on my BSD systems. If I don't like to compile I simply type pkg_add -r package_name and it just works by downloading that package from RELEASE. If I need a certain flavor, I add that. For example squid-2.5.STABLE12p1-transparent instead of squid-2.5.STABLE12p1. Ofcourse, that may have been updated in STABLE, and packages for non-x86 are not really updated, so then I need to a CVS update for my ports tree... and do it your way. I don't like to compile myself, would trust a MD5/SHA1/RMD160/GPGed repository more, but so be it. You guys automated that, made it a bit simpler, but if you know what you're doing such is not necessary. Problematic is it when you just took a lot of time to compile AND TEST your application, and it works (or doesn't). Who the hell would use Gentoo in a productive environment performing such tasks with dangerous USE flags?

Also, Gentoo is Linux, internally suffers from fragmentation, and its users frequently piss the hell out of F/OSS developers because of their clueless and bogus bug reports -- due to things such as USE flags. The only useful purpose I can think of is.. catching bugs in GCC ;) now I'm off eating rice crackers, laters...

Re:Ill informed post

[Anonymous Coward]

Security updates never break the system (although they may require config changes that are always very well-documented in the update process itself, and on the gentoo website/forums/wiki). Most people run at least some packages from the ~arch branch (unstable), though, and breakage does sometimes happen there (which is usually well-documented as well).

Java especially, has given trouble lately, but there is now a system in place that allows several 1.4 and 1.5 JDKs and JREs to coexist, with each user able to choose their own preferred JDK/JRE, and the system defaults set separately from that as well. Also, packages that require a certain version of the JDK to build properly can now specify that, so you can use JDK 1.5 for most things, and something that is not yet 1.5-compatible will still build and run just fine.

Re:Ill informed post

[rtyall]

I was under the impression that stage 1 installations are now redundant, you may as well just start form a stage 2 as the updates done afterwards negate anything done in a stage 1 install.

Re:Ill informed post

[cheater512]

Thats odd. 'emerge world' works just fine for me. :P

Re:Ill informed post

[cheater512]

The only time 'emerge world' will break something is if a new binary incompatible library comes out or something drastic changes.

A recent example is libexpat which is solved by running 'revdep-rebuild' overnight.

Re:Ill informed post

[chrismcdirty]

I concur. I've been running various distros for some time now, but I always come back to Gentoo. My MythTV/web/file server has been running it for somewhere between 3 and 4 years now, and never once have I had a major complication with the OS (MythTV hosing all of my recordings and me deleting the database like an idiot are not related to Gentoo). I've even switched the actual hardware it's running on, booted into a generic kernel, and everything worked fine.

Sure, it may take a little bit of time to get it up and running, but it's a learning process. And I usually don't have to deal with the dependency hell that some distros bring to the table when I try to upgrade a package that has a dependency version that is no longer in the tree.

Re:Ill informed post

[wolf31o2]

It is actually quite simple.

They are not the intended audience.

Gentoo is not something for the person who does not want to learn and does not want to tinker. As someone else in these comments posted, Gentoo asks you what you want to do, then proceeds to do *exactly* what you asked it. If you tell it to do something stupid, it does something stupid. The sheer number of people that see the whole pam-login/shadow blocker as an error just goes to show that the people do not understand Gentoo are the ones that complain the most. I'll be honest. I'm glad to see those people leaving in droves. Gentoo is not for them. It never was. People seem to forget that the Gentoo developers call Gentoo a metadistribution. We allow you to build what you want. This doesn't mean that everything is done for you. Over the past few weeks, I've been encouraging certain people to start using Ubuntu. These are the ones that don't read any documentation, and either ask a question for every prompt/response scenario, or bash everything they don't understand. Seriously, go to Ubuntu. You'll like it much better there. What Gentoo is really about is customization. That is its biggest strength. If you're not willing to put in the time, then you're definitely using the Wrong Tool(tm) for your job. Find a better tool.

Re:Ill informed post

[astralbat]

Yes, Gentoo breaks things all the time if your running on the unstable (~x86 for me). I love to keep up with the latest versions of software and unfortunately whenever I do a world compile every month or so it not only takes forever, but a handful of packages will fail to compile with various errors that I don't understand and then it's time to hit the forums and hope that someone else has the solution.

I've been using Gentoo for over 2 years and been using it exclusively for my Desktop for almost that length of time. It's taken me ages and ages to customize to my liking, but once it's there I don't have to reinstall my system ever again unless I upgrade to a different machine arch. Part of the pain should be accepted as upgrading GCC and glibc isn't exactly a simple thing to do and I don't know any other disto that allows you to upgrade your entire distribution without reinstall over the net.

Re:Ill informed post

[Anonymous Coward]

Too bad it doesn't work and you still have to compile stuff. Last time I tried Stage 3 it compiled for ages and then finally failed at some point. I'm no newbie to Linux (using it 12 years now) but I couldn't get the stupid thing to work right. After wasting a bunch of time searching for answers (not to mention the wasted time compiling) I gave up.

Re:Ill informed post

[Yggdrasil42]

I'm an Ubuntu user here, and I've never had apt-get update break anything (without having been responsible for the breakage myself through fiddling with the configs). I think Ubuntu users would be furious if a security update wound up breaking their system.

I don't like to point fingers, but Ubuntu recently had it's share of update problems too, even on the stable Dapper release. Two updates a few weeks apart that both broke Ubuntu's stable release. First the xorg-core issue (http://www.ubuntu.com/UpgradeIssue) and later another one, even after measures were taken after the first issue. My desktop was affected by this. Gentoo users have had similar problems in the past. I've personally had issues both when updating really fresh updates, and when updating after months of not updating.

In my opinion this kind of problem is worse for Ubuntu users, since Gentoo users are usually more familiar with the command-line and how to access forum websites from the CLI. Still it shows that no distribution is risk-free. In my experience Debian stable is the safest distro with the least downtime during updates. Ubuntu is a bit riskier and has to prove itself after the last two problems. Gentoo is even riskier, and hard to QA because of the many possible configurations.

For me, Gentoo is worth the occasional problem, but if you want to be safe go for Debian stable at the moment.

Re:Ill informed post

[Anonymous Coward]

Stage 3 plus GRP (binary packages) would be a better choice. Seriously, stage 3 gets you a baseline system with little in the way of applications. Most folks will want X, Gnome and/or KDE, OpenOffice or equivalent, media software, CD/DVD burning software, etc., and the first three on that list take a very long time to compile. Granted, when you update you'll still be rebuilding everything, but you'll at least have a usable machine while that's going on in the background.

Addressing the main article's concerns, though: package maintainers must be the most important workers in the system. When a distro stops updating, it dies. It's lifeblood. Sure, new projects are sexy, but they must be seen for what they are: risky experiments with little chance of success. Many, many projects fail in the business world -- go check out Hollywood sometime, and see how many failed projects there are compared to the successful ones. Sturgeon's Law applies in full force to everyone; it's just that with open source, all the dirty laundry is out there for everyone to see. Proprietary companies usually sweep failures under the rug and the public is none the wiser.

Re:Ill informed post

[richardablitt]

Or 'emerge --newuse world' if you just want to recompile the programs affected by the change in use flags.

Re:Ill informed post

[say]

I hate to say it, but performing "s/old release/new release/g" on /etc/apt/sources.list, apt-get update, and apt-get dist-upgrade is not always as clear as one might expect.

I agree. If you use Ubuntu, it is also quite unnecessary. The Ubuntu Update Manager GUIfies this - it pops up to ask you if you would like to upgrade. If you don't want to, it doesn't nag, but presents the option if you would like to update at a later time.

You have no GUI? Well, CLI users are SOL: They need to replace keywords in sources.list.

Re:Ill informed post

[fistfullast33l]

I agree, ~x86 breaks my build all the time. However, it's rarely Portage that is breaking. Nine times out of ten there is a compiler error that everyone else has picked up that breaks the build and forces the Portage team to roll back the version. I think it's hard to fault the distro when it's actually the developers who are breaking the code.

There have been many posts over what Gentoo's stated purpose is. When you read the link above regarding the comments on the mailing list (on LWN), you'll see that even the developers fight over its purpose. Personally, I think that is okay. I use Gentoo because it's highly customizable and I can get all the bells and whistles. The documentation is excellent and the chatroom and forums are really helpful. I say they stop worrying about what Gentoo's purpose and goals are and just keep doing what they're doing. I'm happy just the way it is (although I kind of preferred the old stage 1 build myself).

And another thing - if someone adds a feature to portage, it's not going to break or change the way I do things. I've never used the Graphical LiveCD and I probably never will. I think the freedom to add that should be included, just as I have the freedom to ignore it if I like. Very few things are forced on you by Gentoo, even Desktops/WM's like Gnome and KDE. It's all about choice, and Gentoo definitely personifies that more than other distros.

Re:Ill informed post

[OverlordQ]

I know a guy who's maintained the same install of Gentoo on his laptop for over four years who has kept it up-to-date by using portage without a re-install. Talk about impressive for a Linux distribution.)

Steps to recreate this 'amazing' feat:

1) Install Debian
2) apt-get update && apt-get upgrade
3) Profit!

Re:Ill informed post

[zootm]

Gentoo is not something for the person who does not want to learn and does not want to tinker.

Right on the money. The problem is that a lot of people get very religious over the issues, determined that the tinkering is "necessary" for a useful, efficient system (it's not, let's face it), and will tend to advocate Gentoo for every task, which it's neither designed nor particularly useful for.

The biggest problem with almost any given Linux distribution is its fans. The people who don't realise that just because they like something that it's not universally useful. I don't like recompiling things, or tinkering with USE flags and the like. The performance increases are, to me, negligable. The customisability of the system is not useful to me since I just want a working system in a short time, that works reasonably well. Gentoo is not the distribution for me — but that doesn't mean it can't be good for someone else. People get too absolutist over these things, arguing until they're blue in the face about irrelevant benefits.

Re:Ill informed post

[Enderandrew]

If you change your flags, and you're rebuilding for a different architecture, then you really want to do an emerge -uDn world at the VERY least. I throw in a v and p for good measure to see what I'm doing.

Can't help but think...

[spiritraveller]

that the headline would have read better if it said "Gentoo spills seed."

Just sayin.

No, bad

[tomstdenis]

NO MORE!

Spend more time on fucking Q & A. I'm tired of trying to talk people into Gentoo only to find out that the tree is half-fucked all the time [like packages marked stable requiring other libs NOT IN THE FUCKING TREE YET].

No more extras, fix the base!!!

This is the problem with OSS. Everyone wants to get famous for the next big breakthrough and nobody wants to maintain the shit.

Tom

Re:No, bad

[slapys]

"This is the problem with OSS. Everyone wants to get famous for the next big breakthrough and nobody wants to maintain the shit."

Seriously. I submitted several UI bugs to the Xfce bugzilla site recently and none of them were addressed. People want to develop fun new features, but unfortunately that's not all that software is.

Re:No, bad

[Rix]

Really, I find that most people would rather complain about what they don't like than actually do anything to help fix it.

Re:No, bad

[Anonymous Coward]

I find that most project owners would rather complain that other people don't fix things for them than actually do anything to maintain their own code.

Re:No, bad

[kfg]

Touche! Three minute limit and match.

KFG

Re:No, bad

[tomstdenis]

Um, that's such a inappropriately ignorant comment it's hard to really put this into context.

Let me see I can take a stab at it anyways...

My workstation is the product of about 800 Gentoo ports being installed. Roughly speaking probably around 200M lines of C, C++, Python and Perl source code. I don't have the time, nor the energy to go through all of them to fix them up because the developers are too lazy to maintain them. Frankly, this is why OSS sucks. In the non-free world you don't see Microsoft telling it's customers "You don't like explorer? Fix it yourself!"

I report bugs when I can but sometimes you don't know what to say. Like why did my UHCI controller stop working after I moved from 2.6.16 to 2.6.17? Fucked if I know. Why does my gige not work in 2.6.18 but it worked in 2.6.17? [actually this is a ICH8 fix I bet...]

Sure I could report it to the lkml, but chances are they'll need more than "Tyan 2882 and UHCI no work."

Tom

Re:No, bad

[penguinbrat]

NO MORE!

Here, here!!!!

I completely 2nd this - it was only a couple of weeks ago that I was aiming for the gentoo distribution for a server. Only to find this new fangled installer (that I did NOT like), and then after spending a couple of hours getting the base system installed, tried to update everything - and what do you know? PAM and shadow libraries conflicting whith each other. I had remove both (making logging in impossible obviously) and the update the whole system. Completely breaking the OS, just to upgrade is *NOT* a good idea - even for M$...

Re:No, bad

[Anonymous Coward]

Oh I wish like hell MS would give me the source and say that to me! As do a lot of others here I suspect.

Re:No, bad

[Anonymous Coward]

Ebuilds and/or packages, please, not ports.

And if you hate Gentoo so much, and if it can't get the job done for you, then isn't it your responsibility to use some other operating system that meets your needs and/or has the enterprise-level of reliability you want?

Seeing as Gentoo primarily just compiles *upstream* packages, then if it's broke for us, it's got to be broke for everyone, in all likelihood.

Signed,

A Gentoo Developer

Re:No, bad

[tomstdenis]

Hint: tarball your system drive. Burn to a DVD.

That way when Gentoo decides to do something like "lets move to glibc 2.4 and break all your coreutils, nobody needs ls anyways!" you can just untar and wait.

Tom

Re:No, bad

[fucktheriaa]

Yes - they will need more than "Tyan 2882 and UHCI no work." - in the case of kernel bugs stopping previously working functionality - you can use the excellent git bisect functionality to quickly (within a few reboots) narrow down exactly which patch broke the functionality. For instructions - see http://www.kernel.org/pub/software/scm/git/docs/ho wto/isolate-bugs-with-bisect.txt malc

Re:No, bad

[SuperBanana]

No more extras, fix the base!!!

Somewhere in the bug system is a YEARS OLD ticket, begging the tree maintainers to add one of the (otherwise very popular) alternatives to the horrendously bad etc-update tool. Numerous times people had "bumped" the ticket by replying to it, and asking what the hold-up was.

Maintainers replied with "we're BUSY", and the response was "FOR A WHOLE YEAR!?". Then the excuse changed to "we don't want to confuse the users" (emphasis added.) Apparently, gentoo users are considered too stupid or simple-minded to handle multiple choices for getting the job done. Yet gentoo has both kde and gnome, both vi and emacs.

Anyway, on a different note...from one of the emails in the article:

I'm just sick of new "projects" spawning off without being thought out in the least, and making us all look like jackasses. Is it honestly going to be the new "tradition" that every single new project that starts out is going to be completely undiscussed, poorly thought-out, poorly implemented, and cause us all to look like a bunch of fools for weeks on end before it *finally* gets into a half-way workable state? What ever happened to *talking* about something before going off and announcing it to the world as if it's some kind of completed project and ready for public consumption?

Is this guy new or something? I remember back in 1999 or so, there were about FIFTY different mp3 players in freshmeat that all did nearly the same thing. Everyone wants their own sandbox to play in. I joke that now it's web browsers; seems to be a tutti-frutti flavor popping out every month.

Re:No, bad

[tomstdenis]

fuck you, ebuilds/package/ports it's all the same idea. The *BSDs invented [modern] ports anyways. So don't think Gentoo was the first with scriptable installs.

That said, I question how much testing goes on, specially on the core platforms [e.g. x86 and x86_64]. Often I find packages will install and run fine first try on the x86 but take a -r1 or -r2 before it works on x86_64.

I use Gentoo because I like the idea of Linux and I like USE flags. Generally though my attachment isn't that deep. If Gentoo died tomorrow I'd just go to FreeBSD.

Tom

Re:No, bad

[tomstdenis]

That, or ... they could TEST THE FUCKING UPDATES.

I don't have the time to rebuild kernels and reboot all the time. Especially since this is a MULTI-USER box with a file share on it. I need it to be up and working. I really don't see how you can screw up a working USB driver. I mean EHCI still works but UHCI just gives "bad address" errors when probing the USB device [it detects that I plugged it in, just can't get past the probe stage].

You know, real companies who make proprietary software have boxes for testing. This is what Linux will need if they want to grow beyond a neato hobby. I mean hell, if I can buy spare boxes for processor testing surely there has to be enough people in the Linux project to buy boxes just to test the mobo/hardware.

Tom

Re:No, bad

[Stalyn]

Well maybe you should stick with 2.6.16? You don't necessarily have to upgrade everytime a new kernel comes out. Also I believe you report kernel bugs to http://bugzilla.kernel.org/. To report a bug all you need to do is list your hardware config, kernel config, dmesg and basic outline of your problem.

Re:No, bad

[penguinbrat]

Heh, I would agree with that if the system was already to go and completely functional with all the services I wanted. But I was just trying to install it in the first place, the *upgrade* I'm refering to is the first one you do after you build the kernel and reboot into your new system.

Re:No, bad

[cloricus]

Wait...the non-free world? How the hell is the non-free world any better at all? We've had a major roll over in the last ten days of our mission critical control system that practically runs the place. I've filed over 40 critical bugs and god knows how many others around the place have filed without telling me. Now the pilot and dev teams (third level support) are working closely with us to resolve as many as they can as quickly as possible though that is taking awhile as they are moving more and more of their customers over to the new system. This is all well and good and we've only lost a small percentage of productivity and lost only a few sales, which can be expected in a major upgrade, though the real killer is that there are several bugs we have filed which I know I could fix in five minutes that they have out right told us will never be fixed... This is a situation where OSS would win out and hell we have one of the better closed source proprietors that I know of. So try living in the real world first before you diss free software.

Re:No, bad

[whoever57]

Somewhere in the bug system is a YEARS OLD ticket, begging the tree maintainers to add one of the (otherwise very popular) alternatives to the horrendously bad etc-update tool. Numerous times people had "bumped" the ticket by replying to it, and asking what the hold-up was.

And a replacement has been there for years -- it's called "dispatch-conf".

Re:No, bad

[naasking]

you don't see Microsoft telling it's customers "You don't like explorer? Fix it yourself!"

No, Microsoft says, "You don't like explorer? Tough shit!" On rare occasions it's "You don't like explorer? Well I'm afraid it's far too integrated into Windows your Honour."

Re:No, bad

[fucktheriaa]

Really - if you "can buy spare boxes for processor testing" - why not use one to debug the problem? If you can't - just roll back to the previous kernel where it works. Man - you get this all for free - but you rant and rave like a 5yr old... even when given simple tools to narrow down the problem. Jeez, no pleasing some ppl...

Re:No, bad

[Kavli]

Couldn't possibly agree more!

Basic components are so broken that they doesn't even compile with the most common USE-flags. kdelibs-3.5.4-* wouldn't link unless the 'noutempter'-flag was present. FontConfig-2.4.1 compiles and links, but all applications that link against it segv's. That was just yesterday's experience. I really have to bite my teeth together before doing a world update. Most likely I have to spend 2-3 hours tricking my way through an upgrade, reading the source to see how I can tweak it.

I really like the concept that Gentoo offers, but not at any price. I'm very close to ditching Gentoo, but I don't have any real options where to go, except rolling my own.

People are complaining about the difficulty of installing Gentoo, but I think if it was easier, less experienced people would have to fight with the broken ebuilds and how that should work with plain users without programming experience, I don't know. Probably it would give the distribution so much negative PR that it would be totally abandoned in the industry.

Re:No, bad

[fucktheriaa]

And a new curses based option was developed with the help of Google's SoC project - conf-update http://sources.gentoo.org/viewcvs.py/gentoo-x86/ap p-portage/conf-update/

Re:No, bad

[Anonymous Coward]

This thread explains why there will always be a place for proprietary software you must pay for. That's so someone can be paid to do uninteresting but necessary stuff.

Re:No, bad

[tomstdenis]

my "spare boxes" are DIFFERENT PLATFORMS, one is a P4 32-bit, another is a Core2 64-bit and the other is a $7000 Opteron 2-way box.

I'm not buying spare IDENTICAL boxes just because Kernel developers are too immature to address the problem properly. I mean how do you get the right to brag "Oh, I'm a Kernel developer, see my code-fu muscles" when you can't invest the money to actually develop seriously?

You think MSFT has only one type of workstation in their QA labs? Very doubtful. Everyone is so quick to mock MSFT but then when the real punches come out they get all super defensive...

I bought my boxes because I implement crypto software and I take performance studies seriously. I even own various ARM and PPC kits to test on. These things aren't free. They cost money to buy and time to setup and use. But I do it because I take my projects seriously.

Tom

Re:No, bad

[TCM]

This is the problem with OSS. Everyone wants to get famous for the next big breakthrough and nobody wants to maintain the shit.

s/OSS/Linux/

There, fixed it for you.

Re:No, bad

[orasio]

I don't think that actually happens.
A project owner with an itch to scratch, will scratch it, they _do_ maintain their own code.
Maintaining their own code, but for other people, and for free, is something that they usually do, but some times they just don't want to. There is where you see some project owners complain, about users that want tailored free maintaining right now, without contributing.
There's no implied contract that says that the original coder will fix _your_ problems for free. That's one of the best things of the GPL, you have the power to make stuff happen yourself. Even if you don't know how to program, there are a lot of programmers that will do it for you, given the right compensation. That is the actual meaning of "free" as in "freedom", and not as in "beer".

Re: No, bad

[Black Parrot]

> Spend more time on fucking Q & A. I'm tired of trying to talk people into Gentoo only to find out that the tree is half-fucked all the time

Yes, for the past year or so I've been noticing more broken dependencies. Really annoying was a couple of *mm packages that got into an upgrade-downgrade cycle. Every time you did -uD world they'd want to switch to whatever they were before the last time.

That was a glaring annoyance, but hardly the only one. I've been working on "clean" install this week, and it hasn't been very clean at all. I've found both broken dependencies and documentation that tells you to do the wrong thing. I'm starting to sympathize with Joe Barr after all...

Re: No, bad

[tomstdenis]

I cried when they removed "--inject". I used that all the time to overstep the up-down cycles.

Or how about this,

I DON'T WANT GENTOO FUCKING SOURCES IN THERE. I'll use a vanilla kernel thank you.

What the hell is dependent on gentoo-sources anyways?

Tom

Re:No, bad

[Anonymous Coward]

Why don't YOU buy the devs spare boxes to work with then?!

Re:No, bad

[Dhalka226]

I mean how do you get the right to brag "Oh, I'm a Kernel developer, see my code-fu muscles" when you can't invest the money to actually develop seriously?

It's very easy: You look at the whole picture.

Your workstation is the result of 800 Gentoo ports being installed. 200M lines of source code. Your own words. Well. that sounds pretty fucking awesome to me. And you got all of that for free.

You're having a problem? Sorry. But even one problem, or five, or twenty--out of 800 installed packages--is still a damn good mark. And those developers you bitch and moan about being immature for not fixing your specific problem (without you even taking the time to report that problem) are clearly providing you a service that you like. If they weren't, the system wouldn't be acceptable for your use andyou would switch to something else.

Don't like the way that your problems aren't getting fixed in the OSS world? Well, you're free to go install Windows or OS X or some other, non-free-world operating system. There are NO problems with any of those OSs, I'm sure. And if there were, I'm sure Microsoft would be happy to release a patch just for you when you... well, don't report it. But I'm sure they'll get your report telepathically.

Seriously, you have one of the absolute worst attitudes I have ever seen. It's not just your posts here. I'm sad to say that I recognize your name because your attitude is constantly shitty. Maybe if you weren't such a high and mighty ass all the time, people would be more inclined to care what your problems are. Hell, if those "occasional" bug reports you file read anything like the drivel you spew here, no wonder your problems don't get fixed.

Re:No, bad

[gripped]

LFS - hard work but worth it at the time. Now I just don't have the time.

Gentoo - seemed like a good idea but project seems to have lost direction. Stable is too far behind and unstable just plain don't work in too many cases. Minor package updates can break things in a major way and any complaints are met with 'what do you expect in unstable ?'
I suspect many Gentoo developers wish to return to being a niche distro.

Ubuntu - I've settled with for now. I can compile anything I need a newer version of, and updates don't tend to bork the box.

Please feel free to piont out any erors in my speling grammer
I care a lot.

Re:No, bad

[tomstdenis]

Hahahaha, that's a funny retort, you think about it yourself or did your mother help you out with the big words?

I'm sorry, but why should we trust the future systems to people who can't afford a box to test their drivers on? This is where becoming a platform expert comes in handy. Instead of hacking any random bit just to get your name in the changelog you take ownership of something and then work towards that. Like if you maintain the ICH drivers ... you should own a few Intel boxes. Not just some 12 yr old box and a few databooks.

Tom

Re:No, bad

[tomstdenis]

That's such a bullshit unprofessional attitude.

"oh it doesn't work? fix it yourself."

How about take ownership of your projects?

It's being a bad parent. Nobody cares if your sperm shoot straight. If you can't invest in raising the child you're a bad parent.

Unfortunately, for some reason in the hacker/oss/etc world nobody cares about the longterm. They only care about the recent coolness. oooh, GLX rendered xterm!! AWESOME!!!

Similarly, if you can't maintain a project [and actually fix things before adding NEW content] you're a bad developer. People shouldn't praise and adore you [like most OSS press whores] but shun and shame you. As a user I do my part, I file bug reports. In my software that I release I don't blame the users [even the ones on Windows...]. I see if I can fix the problem and work on it. Why is that so hard?

Tom

Re:No, bad

[Anonymous Coward]

Seriously dude... You need to chill the fuck out.

*anon for a good reason*

Re:No, bad

[tomstdenis]

Yeah, but my USB ports work in Windows don't they. :-)

Though I half blame crappy mobo producers too... PICK A STANDARD AND STICK TO IT!!!

In this day and age, ethernet and USB controllers should be a dime a dozen. Not obscure and hard to clone...

Tom

Re:No, bad

[Anonymous Coward]

>This is the problem with OSS. Everyone wants to get famous for the next big breakthrough and nobody wants to maintain the shit.

And you have contributed?

Re:No, bad

[tomstdenis]

Yeah, well next time people are like "Linux market share is only 3%, boo fucking hoo" don't look this way for sympathy.

As much as I like Linux and the potential of GNU/Linux/OSS/etc somedays it drives me fucking mad. Two steps forwards, one step backwards.

As for chilling out, that comes next week at Toorcon. Horton plaza == Movie Theater + nice bars == fuck you americans I'll get drunk and insult strangers all I want. :-)

Tom

Re:No, bad

[Zach978]

Dude, get a grip. You're using a source based distro, it's not going to hold your hand through kernel module selection, hotplugging, etc. If you want that, go try one of these easier to use distros, because Gentoo, Source Mage and the like are pretty much "fix if your self" type of deals. And maybe if you weren't such and asshole they'd help you out a little more.

Re:No, bad

[tomstdenis]

Actually I have patches in Allegro and bug reports in GCC [that have been fixed or are being fixed]. I've just recently filed a lkml notice about the ICH8 in 2.6.18. That's how I contribute to these projects.

On my own, I have http://libtomcrypt.com/ which has three major projects that I support, even commercially [I don't get paid but I have companies asking for help for products that they sell... damn I need money...]. I certainly don't say "fuck you user, fix the fucking problem your damn god fucking self you lazy cunt!" I'm sure I'd be out of users fairly quickly.

Next question?

Re:No, bad

[Rix]

Brace yourself, your sense of entitlement may make this difficult for you to accept.

You are not Gentoo's customer. You did not give them any money. They do not owe you a damned thing. Calling the developers "lazy" is absurd, given that you seem to have contributed absolutely nothing. If Gentoo isn't working out for you, you are perfectly free to not use it, fix it, or hire someone to fix it. You have absolutely no basis to complain that the gift horse has ugly teeth.

Re:No, bad

[Anonymous Coward]

Yup, parent is %100 right on.

Everybody wants to go to the Party, Nobody wants to stay and clean up.

Re:No, bad

[tomstdenis]

You don't get to brag that you *donated* your time and effort to a project if you expect to be compenated for your troubles. Last I checked Gentoo was a free product, donated to the public [under various licenses] for all to use. If this means I can't gripe about shit getting broken without taking out my cheque book ... then it's not really a free project now is it?

I didn't pay for that toll highway in Toronto, but I certainly can't use it for free. Does that mean the highway was "donated" to the public?

I certainly don't demand money for my projects [though I'm not above begging at times]. I fix/upgrade/add new shit all the time for free. Even if the customers are people like IBM, Intel, whatever, making bank off my work. I would rather help others than demand they pay up.

Tom

Re: No, bad

[Wonko the Sane]

I DON'T WANT GENTOO FUCKING SOURCES IN THERE. I'll use a vanilla kernel thank you.

I solved that by adding gentoo-sources-2.6.99 to package.provided. I also masked vanilla-sources and gentoo-sources because I do all that myself, I don't need portage messing with my kernel sources.

Re: No, bad

[tomstdenis]

What I hate is the downloads of a 40MB package which then turns into 32000 files taking up 400MB...I could see if the kernel was 10MB and a dozen files or something but this is a bit excessive to just randomly force install...

Tom

Re:No, bad

[Nimey]

How about you shut the fuck up and just use a different distro? Nobody's holding a gun to your head to force you to use Gentoo or anything else.

I'm torn between thinking you're a whiner or that you're a troll.

Re:No, bad

[Rix]

Sure it is. The fact that people gave you a bunch of free shit doesn't give you the right to demand they give you more free shit, or different free shit. Do I have the right to demand that you add feature x to your projects, or that you fix bug y right now and before bug z?

Re:No, bad

[tomstdenis]

I know what you are saying.

But there is a diff between "Your shit won't build with GCC 4.1.1" and "I really want a port to J2ME this weekend if possible" [*].

I think overall I fulfill probably more than 90% of all user requests. I reject the ones that have nothing to do with my goals [e.g. porting to J2ME] and others which are just wrong [e.g. insecure or just inefficient].

Tom

[*] Don't laugh, I have people requesting stupid shit all the time. Micro-optimizers are the worst though. Telling me that if I use the lib in some obscure way [e.g. converting 10 million bit numbers to decimal] that it's inefficient. But the use cases for those projects are so small and obscure that it doesn't merit polluting the code base for 99.999% of the users....

Re:No, bad

[tomstdenis]

Um other distros suffer similar problems. Specially since many packages are universal.

And besides, bitching is how things get done in this world.

If I walk up to a LTC user and ask "what do you hate about it" and they just politely say "oh I love it, nothing to complain about", they're not doing me a favour. LTC isn't perfect and there are shortcommings in it that I don't know about [or haven't realized yet].

If we all just sit around holding hands and singing campfire songs then nothing gets improved.

Tom

Re:No, bad

[Anonymous Coward]

you don't see Microsoft telling it's customers "You don't like explorer? Fix it yourself!"

No, Microsoft says, "You don't like explorer? Tough shit!"

To an end user, how is this any worse than Jakub Moc saying "we don't care" and "don't waste our time"?

Re:No, bad

[the_bard17]

"Stable is too far behind" !?

Mind elaborating on this one a bit? I've been tempted to start playing around with a distro other than Gentoo, probably (K)Ubuntu... but I do hear this same complaint quite often. On Gentoo, I stick with the stable packages for the most part, and if something's too far behind, I unmask it and work with the "unstable" version. This usually works perfectly fine for me (with the one notable exception of Scorched3d & OpenAl... I'm getting horrible static).

Not that I'm trying to attack you, keep in mind... I'm genuinely curious. And I wouldn't mind if you detailed the update process for Ubuntu... I've been spoiled by that "emerge -uDv world", I suppose ;o)

Re:No, bad

[the_bard17]

Not to karma whore or anything, but the Gentoo wiki has a nice tip on dispatch-conf here.

And thanks for pointing this out... I've always used etc-update, and never known about a possible replacement for it.

Re: No, bad

[Wonko the Sane]

Having portage update my kernel tree doesn't work for me because I always make a few changes to the source, which gets eraticated if portage extracts a new kernel.
That automatic console blanking always annoys me. I best solution I found to disable it is to change "blank_inverval" to 0 in drivers/char/vt.c

Re: No, bad

[tomstdenis]

or just hit the shift key. It's what I do when I don't knwo what's going on in the shell.

Of course who looks at tty directly? Hello SSH....

Tom

Re: No, bad

[Wonko the Sane]

Of course who looks at tty directly? Hello SSH..

It happens sometimes. Not all my computers have X installed. Sometimes I use those computers and SSH into the other ones. Depends what room i'm in.

Re:No, bad

[frogstar_robot]

And besides, bitching is how things get done in this world.

For the most part, that certainly isn't true in the FOSS word. It is most especially true of the developers who are only doing it because they want to. If I'm a developer with limited time and I get three pleasant requests and a self-entitled bitchy one then guess which one is going straight to /dev/null?

One: You indicated that you had the needed functionality on 2.6.16. Roll back to it fer crying out loud. While we're on the subject, you don't have to upgrade to the latest kernel just because.

Two: You didn't pay anything for Gentoo. See my first paragraph. In the FOSS world bitching that is heeded is something you pay for in either professional grade bug reports, patches, or money. You want someone to be obligated to hear yer bitchin'? Pay for a RHEL or SuSE license.

Re:No, bad

[Anonymous Coward]

Can't afford A box? LOL, as if all of the ICH implementations are exactly the same. Let me guess, Tyan just takes an intel motherboard, scrapes off their logo and paints theirs on and marks it up $300 right?

Donate a motherboard, donate time on your motherboard, or take the day to figure out why your motherboard is different from the other ICH motherboards, since apparently either you're the only owner of an ICH motherboard that runs linux, or else everyone with an ICH motherboard has your problem and has your attitude and believes that the developers they're not paying should address their bug at the top of their priority list. Psychically, since they don't deserve to be told of the regression.

Or do what all the windows fanboys do, revert to the last working version when their spiffy new nvidia driver causes their SATA driver to BSOD because its not like anyone can test every possible hardware configuration in the PC world. (Buy a Mac, join the herd of people who think different, all alike!)

I googled and found the changelog where ICH8 gigabit ethernet was moved into e1000 to live in the "intel gigabit" driver in .18-rc2. Can't find anything for your ICH8 USB/UHCI problem though.

Re:No, bad

[Anonymous Coward]

Who pissed in your cereal this morning?

Re:No, bad

[Anonymous Coward]

well, at least you didn't pay for gentoo.

Re:No, bad

[tomstdenis]

Rolling back isn't always an option when there are SECURITY FIXES in between. Between 2.6.16 and 2.6.18 there are a few exploits/DoS.

Second, I agree that politely worded requests deserve more attention. The problem I was trying to point out is Gentoo needs more time spent on quality than quantity.

Tom

Re:No, bad

[frogstar_robot]

And I wouldn't mind if you detailed the update process for Ubuntu... I've been spoiled by that "emerge -uDv world", I suppose ;o)

#apt-get update
#apt-get dist-upgrade

The first command updates the database of available packages and their versions. The second downloads and installs anything that is newer than what you already have.

If I want something a bit more bleeding edge than usual but don't want to risk a wholesale upgrade, I'll add "deb-src" lines from the development distro and "apt-get source packagename" then build the package from source against my otherwise default environment. If it's just a game or something trivial there really isn't much risk. Backporting something like glibc I just won't do. Anymore I just want to use the damn thing and I'm no longer willing to hose large tracts of my system to play a silly little game that just got released yesterday.

Re:No, bad

[Fyre2012]

If Gentoo died tomorrow I'd just go to FreeBSD
Not trolling, but what's keeping you from going there now?

Re:No, bad

[ChaosDiscord]

Frankly, this is why OSS sucks.

You want stability? If Gentoo is as problematic as you claim, maybe you should move to a more mainstream, boring distribution. Debian's stable release isn't very exciting, but it has a good reputation for stability. Same goes for the various Red Hat repackages like CentOS. If you don't like the free support, check out what other distros offer; I hear good stuff about Ubuntu's community. If you want more certain support, go pay for a distribution like Red Hat and cough up for the support contract.

This is not an problem with Open Source, this is a problem with Gentoo. The other option is going to demand that you pay them, why not pay for support for OSS?

In the non-free world you don't see Microsoft telling it's customers "You don't like explorer? Fix it yourself!"

Indeed. Microsoft's answer is "You don't like explorer? Sucks to be you!"

Re:No, bad

[idonthack]

Or even worse: packages in stable that don't have the right checksums. You can't even install them.

Once I was emerging world and one of the files it downloaded for bash was "invalid". The file wasn't corrupt, but the portage tree had the wrong checksum. And since the install scripts were MD5'd also, I couldn't easily edit it to skip the check. I had to downgrade the version (which subsequently downgraded some other packages). Seriously, how hard is it to test the installer on one of the most vital packages in the system before you put it in the stable tree?

That was the day I switched to Ubuntu.

Re:No, bad

[frogstar_robot]

Rolling back isn't always an option when there are SECURITY FIXES in between. Between 2.6.16 and 2.6.18 there are a few exploits/DoS.

I don't know about Gentoo but other distros backport those security fixes and otherwise disturb as little as possible. Debian and Ubuntu have both been good in this regard.

Does Gentoo truly mandate full kernel version upgrades to get security/bugfixes? I'm not being snarky. I understand they have a "stable" branch as well. I'd think similar practices at least go on there.

Re:No, bad

[iabervon]

They don't need to spend more time on QA (Q & A?). What they really need is some automated testing which keeps packages from going to the site. They do a pretty good job of fixing things; the fundamental issue is that problems aren't obvious until packages have been released and people complain. (E.g., people testing a package as to whether it should be marked as stable are using unstable packages, obviously, so they're unlikely to realize that it requires unstable packages, and they're likely to have other stuff that's under development available, too). The infrastructure should take care of consistancy-testing, and delay releases as appropriate, with the package maintainer able to leave the release queued to happen as soon as the dependancies are suitable.

Re:No, bad

[stuuf]

Why does it have to be proprietary for someone to get paid to develop it?

Re:No, bad

[Muramasa]

Seriously. I submitted several UI bugs to the Xfce bugzilla site recently and none of them were addressed. People want to develop fun new features, but unfortunately that's not all that software is.

This is a really bad example. Xfce probably has one of the most pitiful developer to project-size ratios of any open source project. Pretty much all the modules are maintained by one person who also happens to have a fulltimejob.

I've been running the svn version of Xfce for quite some time and there haven't been any new features added in ages because they're currently in feature-freeze leading up to the 4.4 release.

Re:No, bad

[Anonymous Coward]

Talking about maintaining, how is libtomcrypt doing today?

And if Gentoo is so broken, why do people use it...? Debian works fine for me.

Re:No, bad

[Rix]

Not necessarily, no. "You're shit won't build with GCC 4.1.1" is certainly a bug, but it isn't necessarily a show stopping bug for the developer. "It works with gcc 4.1.0" is certainly a valid response.

In Gentoo's case, most of these are issues in portage, either that that particular version should be hard masked, or that gcc 4.1.1 should still be in ~arch (or hard masked). That particular package may be low on the priority list, or have no one, or an overworked someone maintaining it. If you aren't rolling up your sleeves to help out, you really don't have any basis to complain.

I've used Gentoo for years, and it meets my needs far better than any other distribution, but I'm also well aware of it's drawbacks. It really isn't the distro for someone who isn't willing to know what's going on under the hood, and get a little oil splattered in their face from time to time. It will likely never pass the "Mom test".

Re:No, bad

[davie]

"In the non-free world you don't see Microsoft telling it's customers "You don't like explorer? Fix it yourself!"

Exactly. Microsoft says, "You don't like Explorer? Go fuck yourself!"

Re:No, bad

[cp.tar]

I do have to ask... if you have a server which shouldn't go down, why would you perform every single kernel upgrade on it?

If everything was working, why play with the server?

Re:No, bad

[Anonymous Coward]

I rarely see that happen, though, so I assume it gets fixed reasonably quickly. FYI -- not that you should ever need to -- for a temporary fix you can run "ebuild whatever.ebuild digest" to fix the digest locally, although the it will get replaced on the next sync.

Re:No, bad

[Sr. Zezinho]

the gift horse has ugly teeth

Extreme nitpicking follows...

The thing about looking a horse in mouth is not about beauty, it's about age. When buying a horse people checked the wear of the back teeth to estimate the horse's age, just to be sure that the horse dealer was not lying about the age of the beast. That would be a very impolite thing to do when receiving a gift, because it means that you were trying to guess the value of the horse.

Re:No, bad

[Wylfing]

As a sibling post points out, this is totally false. I have been involved in several free projects, and at one time was the lead maintainer of one. It is true that sometimes you don't feel like bug-fixing. But sometimes you do feel like it, and it is pretty rewarding to send out a new release with no (or few) new features but a hundred bug fixes. It's just that, since it's my time, gratis, I feel like I should get to spend it how I want to.

And, frankly, who gives a rip about "beating" proprietary software? Not me.

Re:No, bad

[gripped]

As I remember there were occasions when emerging something from the stable tree had dependancies which required 'unstable'.

Far to often an update would result in time spent searching the gentoo forums for an answer to what had just messed up the system. An answer was normally easy to find but still time wasted. It wasn't clear which updates were security related and which weren't (though I'm sure that is a failing on my part).

Ubuntu has a notification on the desktop when I need to update. Click, Click done. It has not messed up anything as of yet. (apart from the fact that I have to manually edit my Grub settings after a kernel update because grub gets confused about my root drive. Grub did the same on gentoo though so I can't blame Ubuntu.

The fact is I do not have the spare time I used to have to tinker, and want something that more or less just works.

The next time I upgrade my hardware (x64) I may give OSX on intel a try. But that will be some time away yet.

Re:No, bad

[Anonymous Coward]

yeah there are -rcX updates to the vanilla kernel with the gentoo patchset to fix bugs and whatnot. Remote vulns are going to be in there.

"genpatches is a patchset applied to many of Gentoo's Linux kernel packages. It aims to support the entire range of Gentoo-supported architectures.

Each patchset is based on the initial stable release of the kernel's released at kernel.org (e.g. 2.6.10, 2.6.11, 2.6.12).

The patchset is split up into two parts: base and extras. base contains bug and security fixes, while extras includes some extra hardware support and some feature patches.

genpatches is mainly produced for gentoo-sources, which is the only kernel to include both base and extras."

gentoo-sources has this applied.

Re:No, bad

[Anonymous Coward]

Do me a favor Tom; Stop using OSS.

We honestly don't need you around.

Re:No, bad

[Rix]

Yes, thank you Pedantic Boy. Captain Obvious went that way.

Re:No, bad

[cheater512]

Really? Last time I checked MS required third parties to fix their flaws. There has been several slashdot articles about it.

Thanks to Gentoo I have learnt how to quickly and accurately diagnose errors and most of the time find the solution too.

Maybe you should learn how to give the devs the information they need instead of being a whiney baby complaining 'uhci no work'.

Re:No, bad

[anpe]

Sure I could report it to the lkml, but chances are they'll need more than "Tyan 2882 and UHCI no work."
That's right, but the procedure is fully documented here: http://www.kernel.org/pub/linux/docs/lkml/reportin g-bugs.html

Re:No, bad

[Anonymous Coward]

Where, where?

You mean "Hear, hear!".

I hope that helps. Have a nice day.

Re:No, bad

[emilper]

why update for the sake of updates, especially if you run such a mission critical setup ? You don't have to rebuild kernels and reboot all the time. Real companies ... yes ... like those that don't test or don't care if they have relevant SATA drivers on their install disks ?

Re:No, bad

[chrismcdirty]

As you've been quick to point out, you don't have the time to troubleshoot and bugfix. Which likely is a direct result of your high-paying, high-demand job. Maybe the kernel developers aren't as lucky as you to be able to buy every motherboard known to man and test it on every single one.

If you're going to complain so much, maybe you should try to get a refund.

Re:No, bad

[HikingStick]

If you are talking to people in the developer community, you have a valid point. For the rest of the user base, however, all they can do is inform the community of bugs and problems, along with providing a list of desired features. In the business world we have many names for it: validation testing, requirements mapping, and quality management to name a few...

Re:No, bad

[ebyrob]

Because no one seems to want to write an OSS license that lets you *gasp* charge for copies of your software.

(Actually, the GPL forbids it...)

Why is it a necessary rule that 1 user pays for what millions use for free?

Re:No, bad

[OldBus]

Wrong! As long as you provide a copy of the source code (or make the source code available at minimal cost), you can charge whatever you like for GPL'd software. The problem is that very few people pay for something if you can get it for free, therefore it is harder to make a profit (or break even). I'm not familiar with other open source licenses, but I believe Apple ship software with BSD licenses and they charge plenty of money (of course, they are also charging for proprietary stuff as well)

Re:No, bad

[Anonymous Coward]

Why do you need to charge for copies to get paid? What's wrong with getting paid to write the code?

Re:No, bad

[archen]

Um other distros suffer similar problems. Specially since many packages are universal.

Care to cite a distro where this happens? Redhat, SuSE, Debian... the all have testing before their respective releases to ensure that everything works the way it should.

I'm not defending Gentoo as I have the same problems as yourself. Hell I think they should split the damn tree per arch so I don't have to look at packages that don't even work on my machine. I'm also wondering if they snapshot the tree or not. Sometimes crap is broken and the answer is to "resync the tree". That's not reassuring as far a stability goes; where I can end up with broken crap just because of the random chance I synced at the wrong time.

If you want my opinion, don't even recommend Gentoo to people. I use it myself and am more or less okay with it, but I would never recommend it as a server, or even to other people. If someone wants to try out Linux, there are many more graceful distros for them to choose from. If someone asks me about Gentoo, or what I use I'll tell them about it and why I use it; aside from that I'd rather not referr someone to a distro that has a lot of problems I've had to work though.

Re:No, bad

[ebyrob]

Wrong! As long as you provide a copy of the source code (or make the source code available at minimal cost)

To paraphrase: You don't have to give away your software for free. (You only have to give away the source for media cost under a freely redistributable license)

Letting others give your source for free (as you even mention yourself) is tantamount to giving away your software.

Note: Apple mixes proprietary BSD licensed software with their proprietary OS features. (What's cool is they've even released a lot of their own stuff under an OSS license...) This is a lot closer to the "give a copy of the source with the license model" but it still holds back the source to Aqua. This wouldn't be possible using the GPL.

Re:No, bad

[Dan Ost]

Such mismatches are generally caused when you sync against a tree while it's
being updated. You can simply sync again after a reasonable wait or use ebuild
to create your own digests (wish I'd learned to use ebuild sooner...).

Re:No, bad

[ebyrob]

Who's going to pay for software anyone can use for free? (including them)

Re:No, bad

[Anonymous Coward]

[quote]Who's going to pay for software anyone can use for free? (including them)[/quote]

You presume that the software will get written in any case. That's not true, in fact that is premise of the this discussion. If it's worth X dollars to you to have the software written then you will pay up to X dollars. Of course you will try to minimize how much you pay and externalize the cost to someone else, but given the choice between no software or paying X dollars you will pay X dollars. So if the software has any value at all someone *will* pay. If you think someone else values the software at Y dollars then you will collaborate and each end up paying less.

It doesn't make sense to compensate software development any other way. You should be paying for labour and skills (which is what is truly valuable) and not copies (which have little intrinsic value). This would solve many problems with copyright as well.

Why would you be happier paying X dollars for software that only you can use as opposed to X dollars for software that other people will also benefit from? What do you lose?

Re:No, bad

[Tim C]

(Actually, the GPL forbids it...)

No it does not. The GPL forbids you from distributing binaries without making an offer of source availability, and making that source available more or less "at cost" - eg cost of media and reasonable handling charges.

You are free to charge whatever you want for the software itself - I could sell you gcc for $1,000,000 if I wanted to (and you were willing to pay it). What I can't do is prevent you from selling it for $100,000, and similarly you can't prevent your customers from selling it for $10,000, and so on.

Making a living selling GPLed software is hard, as you can't prevent your customers from destroying your market. There is nothing in the GPL that forbids you from trying, however, as long as your customers receive the full rights afforded them by the GPL. (It's precisely that clause that makes it so difficult, of course)

Re:No, bad

[Blakey Rat]

I know enough about computers to know:

1) When something usability-wise completely and utterly pisses me off. (To use a Microsoft example, entering OEM numbers after you install Windows... if you mis-type the first character in one of the 5-character groups, and hit delete to correct it, the entire previous group will become selected. Then when you type the correct character, you've erased part of the already-correct number already typed!)

2) When something is a bug. Usually, in Linux, this takes the form of "X doesn't work, but there's no error message telling me WHY X doesn't work." (For instance, install VLC on PPC Ubuntu and then use "Open Disk" to try to play a DVD. NO error message. Hit Play. No error message; also no DVD playing.)

What I don't know is enough programming to actually fix any of these bugs. Most of the stuff I used to put into bug databases for open source projects ended up either marked as "not fix" (even though they were obvious errors, like the aforementioned lack of dialogs), or they go unread until the next version comes out then get marked "invalid" because they apply to the old version. So I don't bother to submit bugs anymore. (Oh, and one or two times I've gotten the runaround. "We don't put up an error because it's program X's job to put up the error!")

Re:No, bad

[commanderfoxtrot]

I use Gentoo on a few boxes but haven't followed the mailing lists for a while now- too much volume.

Do they use vmware at all now? Having installed VMWare Server (free) last week, I am astounded at how useful it is- one could have multiple Gentoos (Seeds or not) constantly building and doing recursive QA. Very impressive stuff.

The VMWare site has downloads of difference virtual images: I've just installed a couple of Ubuntu images and ReactOS which I would never have time to do if they weren't neatly packaged as VMs.

Gentoo- release the Seeds as VMs!! Easier to end users to test and try out.

Re:No, bad

[Braino420]

His comment definately was not an ignorant one, and it holds true from my experience.

I really don't understand the logic in demanding (very rudely I might add) that someone work on your problems for free. You should try to show a little more respect towards people that offer you their hard work for free. No one is forcing you to use OSS. The developers don't owe you anything, I don't know how you could possibly see it that way.

However, I'm sure they would love to hear your constructive criticism, or better yet, you fixing some of the problems that bother you. This is how OSS works. Sure, you may think that no one wants to work on the bugs, even if it were true (which has not been my experience either), that's their prerogative.

In my mind, the only person who owes someone anything here is you. You owe them for letting you use their software for free.

Re:No, bad

[turbidostato]

"Who's going to pay for software anyone can use for free?"

The one that can't get it for free?

I.E. The one that wants a feature/bugfix that is not already there?

Re:No, bad

[Anonymous Coward]

"I don't have the time, nor the energy to go through all of them to fix them up because the developers are too lazy to maintain them. Frankly, this is why OSS sucks."

How can be somebody *so* wrong? (well, unless he's only a troll, of course).

What you are describing is why *buggy* code sucks, nothing to do with OSS.

"In the non-free world you don't see Microsoft telling it's customers "You don't like explorer? Fix it yourself!""

No. You won't see this. You'll only see "I am the only one with access to the source code, so I am the only one that can fix that bug. And I *won't* fix it".

That's the answer I recieved *thrice* from Microsoft (a bug on a hierarchical ActiveX grid component; a bug on the Visual Basic interpreter and a bug on the poledit parser). And that's exactly what I mean: *recognized* bugs tagged "won't fix".

So in both cases you're exactly stopped at the same spot.

Uh... No! With OSS you have all the options you have with proprietary software and then some more (access to the source code, just in case you want/can resolve the bug yourself).

Re:No, bad

[orasio]

Because no one seems to want to write an OSS license that lets you *gasp* charge for copies of your software.

(Actually, the GPL forbids it...)

Why is it a necessary rule that 1 user pays for what millions use for free?


You are trolling, or just wrong, as others stated.

The GPL doesn't forbid you from selling copies, you don't sound like someone who actually read it.

For example, Mysql will happily sell you a proprietary copy of their software, because the GPL _lets_ them *gasp* charge for their software, and some people are willing to pay.
What you can't usually do in practice (because of no value added) is sell more than a few copies of some other guys GPLed software.

Of course, the GPL doesn't help your copy-selling bussiness, because it is a copyleft license, its whole idea is to revert the effects of copyright, and disencourage the sale of copies by restricting users.

There are other licenses that let you include some source and keep distribution priviledges, like microsofts shared source and stuff. Of course, most people wouldn't call _that_ OSS (I don't have a clear definition of what exactly _you_ mean by "OSS"), and it wouldn't be free software.

The idea of 1 user paying for what other millions use for free is not that bad to some people. People who use the GPL, for example. Then, there are people like you. You just need to keep away from GPLed software, and you will be fine. Nobody is forcing it down your throat, there are some other alternatives, and you are free to use them.

Re:No, bad

[orasio]

Could you please elaborate how the hell are you going to make a proprietary software provider fix the bug that is bugging you?

Problem: Exchange doesn't have a good IMP implementation.
What is your solution? Who do you pay to do it?

I think that the proprietary developer is much less sensitive to your needs, because they have a monopoly on support, whereas a free software provider will have to comply to your needs, or you could just pay someone else to get your feature implemented.

I think that the problem in the free software - proprietary software debate is that most people who like proprietary don't understand that "free" doesn't mean "you don't pay for it", it means "you do whatever you want as a user with it". Free software doesn't mean you shouldn't pay for support, or fixes.

Re:No, bad

[Lodragandraoidh]

This is the problem with OSS. Everyone wants to get famous for the next big breakthrough and nobody wants to maintain the shit.

Whoa partner! You're using an awful broad brush to paint your picture.

Debian, Slackware, Ubuntu and other distros are rock solid and continually maintained.

Emacs, Python, Firefox, Open Office are some very big applications that are well maintained.

So - while some OSS projects may have serious maintenance problems, you can't say that for all of them. Additionally, I wouldn't be so quick, as some have, to believe proprietary software is any better maintained; I would argue just the opposite.

Re:No, bad

[Anonymous Coward]

"I don't have the time to rebuild kernels and reboot all the time. Especially since this is a MULTI-USER box with a file share on it. I need it to be up and working"

What the heck are you doing with vanilla kernels? It's stupid going against Linus Torvalds -no less, advice about vanilla 2.6 kernels being *UNSTABLE AND UNTESTED*. Just follow Torvalds' advise and use properly tested/stabilized kernels on a properly tested/stabilized Linux distribution (like Debian or Red Hat) and forget about problems *YOU* are the only guilty to have created.

"You know, real companies who make proprietary software have boxes for testing"

No. Real companies have proper structures. "Who make proprietary software" is an uneeded addendum unless you are just trolling.

Re:No, bad

[turbidostato]

"In Gentoo's case, most of these are issues in portage, either that that particular version should be hard masked, or that gcc 4.1.1 should still be in ~arch (or hard masked)."

While you have spotted the immediate cause, I'd say there's a deeper one too. Most problems on Gentoo come from the very basis of the distribution itself, that is a "source based rolling distribution".

How come that a package on "stable" depend on a library still masked? Because the packager (not a professional on a professional environment, for the most part) built it on his own system (a nasty mismatch of versions -due to Gentoo's "rolling" nature, and irreproducibility -due to Gentoo's "compile it on your own box") and gave the mythical microsoft answer: if it compiles (in my box), deliver!

But... That's Gentoo!

Of course we could ask for "blessed" build procedures, stabilized package versions, security backports... it wouldn't be impossible to get them on Gentoo, but it's highly improbable for this to happen: the people that aproaches Gentoo and makes it come to live just don't tend to think/work that way.

All in all, Gentoo is an awesome effort with a well desserved fame. It's just that (currently at least) its "phylosophy" doesn't fit well on dependable environments.

"I've used Gentoo for years, and it meets my needs far better than any other distribution"

Yes! And that's exactly the point for a community-based effort: it scratches your itch, serves well your needs and it's fun to tweak, so you use it. As other people have said, it's not as if Gentoo's project leader came pointing a gun to your head forcing you to use it.

Re:No, bad

[crystalattice]

What if you charged for pre-compiled binaries and offered the source for free? Then the lazy or non-tech people could pay for ease of use while the tech folks could build from scratch for free. Everyone's happy.

Re:No, bad

[Anonymous Coward]

Now why did you have to go and get the Gentoo fanboys all riled up? The foam from their mouths is running down my monitor.

Re:No, bad

[Rix]

True enough. People in the non developer community contribute with $. They're probably best going with a Redhat or Novell solution that will have those things.

Re:No, bad

[Rix]

Yep, there are certainly lots of projects out there that couldn't care less about usability. That's fine, and it's their perogative. All their code is available, though, so you can certainly build a more mouse friendly trap (or pay someone to do so.).

Re:No, bad

[Blakey Rat]

It's not that they don't care about usability, it's that they don't care about fixing bugs. Not showing an error dialog when an error occurs is a bug no matter how you look at it.

Re:No, bad

[Rix]

No, it isn't. Dialogs should be used very sparingly, and only on critical errors. Often stderr or a log file is a better choice. Regardless, it's entirely the perogative of the developer to decide what is intended functionality and what isn't.

Re:No, bad

[idonthack]

I synced once before the update, and again several hours later (the next day) once the error became apparent. The problem was still there.

Re:No, bad

[ebyrob]

You are trolling, or just wrong, as others stated.

The GPL doesn't forbid you from selling copies, you don't sound like someone who actually read it.

Actually I've read it several times, and quite frankly I'm not even sure it says what RMS thinks it says... (I know I don't fully understand it in a legal context) Of course, all that really matters is what the FSF decides to use it for in court. Judging by their track record so far, you have to be pretty far out of line before they get the big stick out. If you're at least trying to follow what you think it says, they'll probably just ask you nicely to behave at first.

For example, Mysql will happily sell you a proprietary copy of their software, because the GPL _lets_ them *gasp* charge for their software,

The "for pay" MySQL is under a seperate license. The only people paying for it are those who don't want to be encumbered by the normal rules of the GPL. (And also a couple other things under the seperate license) Saying that the GPL allows copies to be sold because it can be part of a dual license plan is like saying that water isn't wet because it can be mixed with sand (which is dry).

The definition of "open source" is just that. The source code is open and available to the users of the software. Perhaps the term open here is like an open tournament... All comers need access. But reading the typical definitions of open source I don't see anything about free copies to non-customers. Basically, I was just idly wondering why it is that no one sells software more like books, where you pay for a copy, but get the whole thing instead of a restricted source-less version.

Obviously this would be neither copyleft nor "free software". That certainly doesn't stop me wondering why it never seems to happen.

The idea of 1 user paying for what other millions use for free is not that bad to some people.

Ya, it's probably great for everyone except that first user.

You just need to keep away from GPLed software, and you will be fine. Nobody is forcing it down your throat, there are some other alternatives, and you are free to use them.

Actually I use and do my best to be part of the community for several pieces of GPL software. The GPL is great for what it was intended for, creating an intellectual commons that isn't easily usurped by greedy corporations. The problem is, not everyone who writes software is doing it for the same reasons, or agrees with the ideal of getting "all software" to eventually be under a GPL like license. Some of us think it's good to have some of both because we think there's some "crap work" in software development that's a lot less likely to get done under a free software model... Or, maybe we figure copyright is actually a somewhat useful model if properly limited and kept within reason.

As far as having things shoved down my throat, it's really more that the entire community appears to consider things only "open source" if they're 100% free (as in beer). I just don't quite understand why that is... You would think there could be a middle ground where source is available to paying customers (and their agents) only.

Note: Microsoft "shared source" is a pretty poor example of this kind of thing last I checked. Sun's at least been trying with the Java source (of course they give most of that software away to begin with, but the JDK is kind of a loss leader).

Re:No, bad

[ebyrob]

No it does not. The GPL forbids you from distributing binaries without making an offer of source availability, and making that source available more or less "at cost" - eg cost of media and reasonable handling charges.

That works for precisely one copy, since both source and binary must be made available under a redistributable license.

You are free to charge whatever you want for the software itself - I could sell you gcc for $1,000,000 if I wanted to (and you were willing to pay it). What I can't do is prevent you from selling it for $100,000, and similarly you can't prevent your customers from selling it for $10,000, and so on.

Is that supposed to mean something? Because this comes strait off the FSF site and it has never meant a damn to me no matter how many times I see RMS or someone else spout it off. I mean c'mon! Obviously you can't charge for something that people can download for free.

Oh wait, that was my whole point. You get to sell exactly one copy (or perhaps several all at once to various people) but after that the "selling software" part of the equation is pretty much all over. (Aside from a couple tricks like alternative licensing or "early access" etc which only work in certain markets anyways)

Note: This isn't to say Free/open source software isn't interesting and successful, quite the contrary, it's amazing to me that the economies of re-use (or something) allow this model to actually thrive in quite a few situations. I'm just curious why no one charges for copies and yet still provides source for free. I guess maybe it's because that would put one party in too much control over add-on work by others and it's actually the free sharing of code portions that makes Free/OSS a success in the first place.

Re:No, bad

[orasio]

The GPL doesn't forbid you from selling copies, you don't sound like someone who actually read it.
--
Actually I've read it several times, and quite frankly I'm not even sure it says what RMS thinks it says... (I know I don't fully understand it in a legal context) Of course, all that really matters is what the FSF decides to use it for in court. Judging by their track record so far, you have to be pretty far out of line before they get the big stick out. If you're at least trying to follow what you think it says, they'll probably just ask you nicely to behave at first.

The issue was not the enforceability of the GPL, it was that you said that the GPL forbids selling GPLed software , and it just does not, period.
In the context you used it first, it seemed like you were thinking about the author making money from his software, and that was why I mentioned the Mysql model.

Of course, the GPL is not good for making money as a middle man in a completely free distribution market, because it is intended to improve distribution, not to artificially restrict it. Of course, you could even make money as a distributor, like the time when RMS sold tapes with Emacs in it, everybody else could do it, too. Now that software distribution costs approach zero, it's a bad market for middle men, if there are no artificial restrictions.

The idea of 1 user paying for what other millions use for free is not that bad to some people.
--
Ya, it's probably great for everyone except that first user.

If you are talking about sharing development costs, the first user could be a group of users, like when the blender foundation raised 100000 euros to free its code.
There are lots of times where paying for something that will be free afterwards is fair for you. Maybe you are just interested in the software, and you don't consider it a competitive advantage, but only a way to keep costs down. A good example is government organizations. It would be more sensible for them to pay to develop stuff that is available to everyone than to pay several times for licenses of the same stuff.

As far as having things shoved down my throat, it's really more that the entire community appears to consider things only "open source" if they're 100% free (as in beer). I just don't quite understand why that is... You would think there could be a middle ground where source is available to paying customers (and their agents) only.


About the "free" as in "beer", I don't think so, blender for example was bought with money, and it is free software, and open source, anybody will tell you that.

Free software, like the FSF definition, "free" as in "freedom", is very specific, of course. It is about protecting the freedom of the user, at the expense of everything else when needed.

The thing with the "open source" term is that it doesn't mean much. Everybody can use that term to fit what they want. The thing you talk about does exist. Lots of companies buy software with the sources, but just don't get a license to redistribute. Unix was sold that way, back in the day. Custom software is often sold that way, too. The problem is that usually the actual use of the source is restricted by the same arrangement that restricts distribution, so it couldn't be considered "open".

You shouldn't worry , though. You can call whatever model you want "open source", most people don't know what it means, and those who do, don't agree, so it's a safe bet.

Re:No, bad

[ebyrob]

Of course, the GPL is not good for making money as a middle man in a completely free distribution market

It's arguably not as good at making money for the initial developer either (Otherwise we'd probably all be using it/Bill Gates is a lot richer than RMS, or even Linus Torvalds last I checked). Of course, so far it appears it can be more efficient at creating high quality software in several markets. It also makes the software more valuable to the end user once it's finished.

If you are talking about sharing development costs, the first user could be a group of users, like when the blender foundation raised 100000 euros to free its code.
There are lots of times where paying for something that will be free afterwards is fair for you. Maybe you are just interested in the software, and you don't consider it a competitive advantage, but only a way to keep costs down. A good example is government organizations. It would be more sensible for them to pay to develop stuff that is available to everyone than to pay several times for licenses of the same stuff.

Which is a good model when you can get it to work, but it requirees a lot of up-front organization on that "first sale".

About the "free" as in "beer", I don't think so, blender for example was bought with money, and it is free software, and open source, anybody will tell you that.

It's free as in freedom on purpose yes, but it is also "free as in beer" for everyone who wasn't part of that initial effort (whether as a developer or funder). This is a free rider issue. (Though whether it is a problem or a benefit depends on who you talk to...)

The thing you talk about does exist...Unix was sold that way, back in the day.

Maybe it is the case that the model of selling software, and giving source to customers only for free, isn't all that efficient at monetizing *or* sharing ideas which is why the practice isn't all that popular. (Or maybe, it just isn't all that noticable when it happens.)

You shouldn't worry , though. You can call whatever model you want "open source", most people don't know what it means, and those who do, don't agree, so it's a safe bet.

I'm not trying to mislabel anything here. Just understand what open source is supposed to mean, as well as something more about the market.

Does open refer just to the fact you can read the source? Or is it more like an open tournament where anyone can join up? Or, perhaps it becomes useless to categorize at some point and everything has to boil down to individual licenses in the end *shudders* (waaay too many of those to get an easy handle on).

--

Don't get me wrong, I'm not going to go on about creators inherent rights to extract money and/or control from every individual who touches their creation in some way. However, the flip side of this on a day job where my goal is making money, at least enough to live on, is that I'd like some kind of distribution model to customers that seems "fair". My company currently sells proprietary software to a rather small market. We don't currently offer source to our customers for free, though we have in the past. One thing we do provide is all major upgrades (which aren't "new products", and yes it's a gray area) for free to any customers under our maintenance contract.

What I'm trying to figure out is where open source fits into that business model outside of tools projects and spare-time endeavors off hours.

What's the difference?

[PhrostyMcByte]

I havn't used Gentoo since its early days, when there where no big binary downloads for it. My question is, if you aren't going to compile from source to get that extra level of customizability, what's the difference between Gentoo and say, Debian testing/unstable?

Re:What's the difference?

[really?]

Err ... completely different colour high horse to look down on other from?

Re:What's the difference?

[Rix]

There's value in having a system that works now, which can then recompile itself to your specifics in the background.

Re:What's the difference?

[Anonymous Coward]

Ok, i'll say what no one else wants to say: portage *looks* good when it's running. It's aesthtically pleasing. It makes you feel like you're in complete control of your system (whether or not that truly is the case). I currently no longer use Gentoo (i tinker too much, tinkering makes things break, you break things in gentoo and you have a 3-day reinstall ahead of you - i know, i used the bastard for about 2 years), and I have no stake in it's survival. But man...i've been looking at screenshots of people on gentoo, and i see that blue and green prompt, the output of emerge -up world, and i get nostalgic. It makes me want to install it again on my computer. Hell, it's not even my favorite package manager in the way it *Works*, but in terms of look...and feel...yeah, portage has it.

And, i've used other package managers. Most of them. Ports (open and freebsd), pkgsrc (net and dragonfly bsd), apt, pacman (my functional favorite), slackpacks, rpms, pkg_add in freebsd, solaris packages. None of them give you the same feeling that portage does.

So, making a seed might *functionally* put gentoo at the same level as debian and all other package-based distros...but portage will always appeal to a certain segment of users, and will make them feel like they're truly in control of their computer, even if they're not.

Plus, after you install the seed, you can change your USE and CFLAGs, recompile the whole system, and will have the same essential system as if you had built it all from source to begin with. Something that would be quite a bit more difficult in a debian-based distro that uses apt.

Re:What's the difference?

[ArcherB]

if you aren't going to compile from source to get that extra level of customizability, what's the difference between Gentoo and say, Debian testing/unstable?

apt-get is spelled different that emerge.

Seriously, apt-get replaced my kernel... probably because I told it to, but I was expecting at least to have the option to go back. I was hosed (before live CD's). Gentoo is a bit harder to replace the kernel, but you have the option of installing multiple kernels. I know I could have done it with Debian too, but I didn't think it was going to replace my only good kernel with a bad one. Also, I really prefer the Gentoo forums over any help I tried to receive with Debian.

Re:What's the difference?

[hobbesmaster]

In Ubuntu when you do "sudo apt-get install kernelnamehere" your new kernel is added as the default in grub, but every other kernel you've had on that system remains there as an option. It pushes windows down pretty far down the boot menu. :)

I'm not sure why debian doesn't do the same thing by default...

Re:What's the difference?

[batkiwi]

Honest question: What value is that?

Re:What's the difference?

[batkiwi]

What release of debian, what kernel? Did you file a bug?

I've NEVER had apt-get replace a kernel. It installs them along-side and sets the new one to the default.

Re:What's the difference?

[Rix]

I'm not exactly sure what you're asking me. The value of having most of x now and all of x later rather than none of x now and all of x later seems obvious.

Re:What's the difference?

[ArcherB]

What release of debian, what kernel? Did you file a bug?

It was quite a while ago and the first time I tried debian (yes, real debian), back when Mandrake was still Mandrake and Fedora was called RedHat. I was even more of a newbie then than I am today. I'm sure I could have booted off of a floppy and simply edited grub or lilo.conf, but I wasn't that advanced yet. I'm also sure that I had said yes to many "Are you sure you wanna screw your system up? (y/n)" messages, so any bug report I would have filed would have been printed out hung in a developers cubicle to remind him of the idiots are using his stuff.

Gentoo (or genkernel) doesn't seem to have the ability to edit grub.conf, or at least I haven't been able to get it to do so. However, since that debian thing, I've made it a point to edit my boot menus by hand anyway, so I actually prefer it that way.

Re:What's the difference?

[the_bard17]

I need a copy of OpenOffice now, on my P3-600 laptop with 384 MB of RAM. I can't afford to wait the two/three days it's gonna take OO to compile. So I install the openoffice-bin package (binary, mind you), do whatever it is that I've got to do, then set the laptop up to install the "from source" openoffice package when I get my next three day weekend.

Mind you, this doesn't bother me, mainly 'cause the last thing I'm going to want to do on my three day weekend is touch anything computer related. Lock me in a room with a set of books and some light classical music in the background. Or rent a cabin for me up in the middle of the Adirondacks, preferably without any electricity or other "modern amenities". Will I care that OO is taking 3 days to compile? Heck no. Just don't make me wait for it.

Re:What's the difference?

[RSevrinsky]

Full disclosure: I am a Gentoo user for the last 4 years, having fled
RedHat at home, and more recently Fedora Core 3 at work.

Compiling from source has very little to do with why I use Gentoo.
The only real impact on me from source compilation is the ability to
specify USE flags, and thus configure which parts of a package are
included. Binary packaging systems like RPM or apt simply can't handle
that degree of control.

No, for me, it's much more the dependency handling that keeps me in
the Gentoo world. Portage makes it far easier to install packages and
their dependencies than RPMs. I played with yum when using Fedora, but
never got it working quite right (maybe that's my fault).

Re:What's the difference?

[osu-neko]

if you aren't going to compile from source to get that extra level of customizability, what's the difference between Gentoo and say, Debian testing/unstable?

Last time I checked, Gentoo kept its packages more up to date. At one point I was using Gentoo because I needed a particular version or newer of a package and the version in Gentoo stable was newer than the version in Debian unstable (there'd been a bug filed against the Debian package a year ago noting that the package was no longer usable, but nothing had been done about the problem for a year, so I gave up and switched to Gentoo).

Had Gentoo had binary packages at the time, I'd probably have kept using it -- compiling from source is a big pain in the ass that, as far as I can tell, doesn't really buy you anything, but aside from the compile-from-source nastiness, Gentoo was a much better distribution.

Interestingly enough, I saw responses like this one or even worse on the Gentoo boards. It's like a lot of people were convinced Gentoo was a one-feature distro -- if you didn't want to compile from source, it had nothing to offer. Nothing could be further from the truth. It was better maintained and offered more options than Debian at the time. Don't know if that's still true today, or the current state of Gentoo's binary packages, but if they finally implemented some sort of decent binary package distribution system (for each package, not just a few large popular ones), I might take a look at them again.

Re:What's the difference?

[Anonymous Coward]

Seeds is entirely different from your typical binary installer. The entire point of Seeds is that they're downloadable system images, i.e. LiveCDs, Xen or VMWare systems. Ready to boot and run for whatever task they happen to be targeted at. So, I don't have to install my OS, I just have to download the system image and boot it in my virtual machine. This is extremely handy for server farms, for instance, where you want versioned and QA'd OS releases that you can push out to all of your nodes at will.

Yes, but look who's back......

[mickwd]

Daniel Robbins becomes a Gentoo Developer again.

Welcome back.

Re:Yes, but look who's back......

[Anonymous Coward]

Good, let's make him honorable dictator and kick gentoo back into shape.

Re:Yes, but look who's back......

[On Lawn]

Way cool.

He's always been a person who does something first, and then talks about it later. I remember way back in the Stampede days, it was all just an idea that code could be compiled to run faster. And the Stampede team just did it.

But what really made Gentoo nice was when DRobbins took from BSD the ability to turn on and off features for the whole system from the bottom up. Having watched Turbo Fredriksson chaff on Debian's package scheme (although I don't think it felt as painful to him as it looked to me) Gentoo was a welcome suprise.

But I have to say, I finally understand what the poster might be getting at. The problem is not Gentoo as much as it is a general problem that Gentoo has not solved for itself. And that is managers vs doers. The person linked as a complainer wants more management, and the people in the LWN article want more doing. Management is overhead if all you do with it is make decisions. Management streamlines and makes the doers' job easier when done right.

I remember the golden days of slashdot ended for me when I realized that it was filled with managers not doers. Everyone was giving opinions on what everyone else should do, from software development to technical development. I suppose it is only natural, slashdot is a place to talk and think. So that is who it attracts.

So it is funny to me that this debate that seems to have come full circle, and probably will continue to hound projects like Gentoo. The person who posted the controversy seems to have completely missed the topic, and his summary shows it. But I hope all the best for Gentoo. I'm sure they'll figure it out with a few hard knocks here and there. I don't think Gentoo will go anywhere, unless a better source distro comes along. And if so, I have no problem with that.

Yes, Gentoo is a mess

[Anonymous Coward]

Posting anonymously, because I'm a Gentoo developer and I don't feel like getting fired for speaking out against a certain clique.

Gentoo is, at this point, royally fucked, and this is a perfect illustration of why. The project no longer encourages technical discussion, debate or getting things done. Anyone trying to have technical discussion is called out and accused of flaming by the once great Seemant (who has not done any development himself for years) and his horde of fanboy minions (most noticably, Jakub) who skipped the usual recruitment process (Seemant throws a hissy fit any time any of his recruits are rejected for failing the quiz), who would rather that people did things without planning and jumped ahead with the kind of fuckups that OS X and Sunrise were than that anyone had a disagreement. Instead, it favors fancy announcements and poorly thought out publicity under the guise of 'making things easier for the users'.

If you look closely, you'll see that Gentoo has not actually done anything for about two years now. Even an attempt to change the color of the website failed after over a year of work. And this is a shame, because it has so much potential. Honestly, I don't know how to fix things. I don't have enough time or enough of a reputation to persuade people to learn from past mistakes (yes, this is Sunrise all over again).

Re:Yes, Gentoo is a mess

[GooglyWoogly]

Interesting to read all this. I've been using Debian for a long time, and about 6 months ago I had a crack at Gentoo on my laptop to see what all the interest was in Gentoo. I liked what I saw over the last couple of years before then, but I definately feel that there's been a lack of cohesion at Gentoo from an outsiders POV in recent times - although I can't put my finger as to why I feel like that. OT - whats all the fuss & hype about Ubunto/Kubunto ? I've install it a couple of times and I can't see any compelling reason for those distros compared to anything else like debian (apart from an easy install) or SuSE ? They just feel like a cut-down debian for some weird reason. Debian is still my disto of choice despite recent perceived debian admin problems, including lots of X stuff going wrong after upgrades (but that's unstable branch to be honest!)

Re:Yes, Gentoo is a mess

[Stalyn]

If you look closely, you'll see that Gentoo has not actually done anything for about two years now. Even an attempt to change the color of the website failed after over a year of work.

I have been using Gentoo for quite awhile and I don't have any gripes. Someone has to be doing something (maintaining packages) or I wouldn't have to emerge --sync and emerge -uD world every few days. But I guess if you consider changing the website color as doing something you could be right.

Re:Yes, Gentoo is a mess

[On Lawn]

I remember one early day on #gentoo a developer I recognized from the mindless cabal that was #devian-devel came a'trollin. The conversation went something like this:

debian-stud> Nice little distro you have here
generic1> Thanks
generic2> Welcome Thanks!
debian-stud> Its not very good now, but it could be
debian-stud> it has many violations of the FSB...

There were many replies, but this one seemed to typify the attitude of Gentoo at the time:

superdev>FSB compliant does not make a great distro

A great distro is one that people love. Why people love it is up to them, but chances are if you love it then someone else will too. They were doing what they love, and that was great. BTW, I am for FSB compliance but that is another matter.

I think that attitude still exists, and will prevail. That is just a grand philosophy about life. Either that happens or someone takes their experience with Gentoo and fixes it by making their own. In the LWN article we note how people hate forks, but its worked pretty well for X-org. But I digress, I'm not calling for or predicting a fork. Just that people who want to do what they love will prevail over people who want to tell people to do things for them.

Re:Yes, Gentoo is a mess

[Penguin Follower]

I have been using Gentoo for quite awhile and I don't have any gripes. Someone has to be doing something (maintaining packages) or I wouldn't have to emerge --sync and emerge -uD world every few days.

Except that as another user noted above, there's a problem with the base install where PAM and shadow libraries conflict with each other (obviously there are packates from the stage tarball that depend on each?) and it was a problem in the 2006.0 release. Something like that should have been fixed for the 2006.1 release. I had a lot of "fun" getting around that problem (and a lot of wasted time!).

I mean basically "out of the box", you've just finished the install and you reboot into your new gentoo system only to find a PITA of a problem the minute you go to install a package or set of packages that depends on Pam or shadow. I still have the 2005.whatever release still running on my main linux box because of this (I was smart and tried 2006.x on a different system first.)

Re:Yes, Gentoo is a mess

[MacJedi]

OT - whats all the fuss & hype about Ubunto/Kubunto ? ... They just feel like a cut-down debian for some weird reason.

I'm not sure if you are being serious or not but Ubuntu basically *is* a "cut-down" Debian. ;)

Re:Yes, Gentoo is a mess

[Stalyn]

Apparently the fix is emerge -C pam-login && emerge -u shadow. You should probably do a revdep-rebuild to be safe.

I think I encountered this bug awhile ago and did the same thing which I figured out myself. Also having built my system from ~arch I think I removed further problems. But you are right it should have been fixed.

Re:Yes, Gentoo is a mess

[agaffney]

Except that as another user noted above, there's a problem with the base install where PAM and shadow libraries conflict with each other (obviously there are packates from the stage tarball that depend on each?) and it was a problem in the 2006.0 release. Something like that should have been fixed for the 2006.1 release. I had a lot of "fun" getting around that problem (and a lot of wasted time!).

There was no "conflict". At some point in the past, pam-login had been broken out from shadow, and then it was decided it should be put back in. There was a blocker put in the newer version of shadow (with pam-login rolled back in) so you wouldn't have pam-login and shadow providing the same files at the same time. Blockers are a normal part of using Portage. This "fun" you speak of was simple to get around:

emerge -B shadow && emerge -C pam-login && emerge -1k shadow

This was covered on the mailing lists, GWN, and forums. Don't blame Gentoo because you can't use the tools or read.

I mean basically "out of the box", you've just finished the install and you reboot into your new gentoo system only to find a PITA of a problem the minute you go to install a package or set of packages that depends on Pam or shadow. I still have the 2005.whatever release still running on my main linux box because of this (I was smart and tried 2006.x on a different system first.)

You mean you've never updated the box since you installed over a year ago? If you've updated world recently, congrats, you're running a 2006.x release on that box.

Re:Yes, Gentoo is a mess

[Anonymous Coward]

I don't think things are quite that bad, but I do agree that Jakub Moc should stick to developing (if that's his forte) and stay the hell away from the website, Bugzilla and other user interaction roles.

Re:Yes, Gentoo is a mess

[Otter]

This was covered on the mailing lists, GWN, and forums. Don't blame Gentoo because you can't use the tools or read.

That pretty much sums up the problem with Gentoo, nowadays -- searching the forums is now a routine tool for getting your updates to work. It used to be that Portage was the biggest advantage of Gentoo; now it's one headache after another.

And we're not talking about an xmms extension or Firefox theme here. We're talking about fundamental login and security functionality, and the "simple" solution is searching the forums for "emerge -B shadow && emerge -C pam-login && emerge -1k shadow"!

Re:Yes, Gentoo is a mess

[wolf31o2]

Well, I'm not going to post anonymously. In fact, you're more than welcome to see exactly what I think. I know for a fact that this is going to make me a few enemies and probably piss off quite a few developers. Quite frankly, I agree with the poster here on many things, but definitely not on all of them. Gentoo really needs a few things to remain a top distribution. For one, we absolutely must stop doing this experimental crap and start focusing on improving and fixing what we already have in the tree. We need to focus on improving the quality of the distribution more than adding new "features" that do nothing more than make things easier on the lazy and the incompetent. There are simply too many Gentoo developers moving in too many directions. We have no focus. We have no direction. Worse yet, if we had one, we have no way of enforcing that we actually move towards it. Gentoo needs a good house-cleaning to remove some of the "problem children" and get us back to, oh, I don't know, maintaining packages and fixing bugs. Sure, that's not very glamorous, but we *are* a community-based Linux distribution. Perhaps we should get back to actually working on that, instead of trying to come up with new projects which sap resources from other places.

It is my personal opinion that the Gentoo developer community is too large and too diverse to properly work towards any real common goals. We have also diverged too far into essentially two camps, those that want the new whiz-bang features and want them now, and those that want a good, stable, reliable, and flexible system that is capable of meeting the demands put upon it. I definitely fall into the second category. Don't get me wrong, I'm all for experimental things, but they should be done outside of the scope of the main project and brought into the project once they've been proven.

I tend to believe that Gentoo needs more internal structure and needs more *well-designed* process to get things done. I don't think that red tape is the answer, but there has to be something done to solve this current anarchy. General development in the business world follows many stages, from initial design, through development, testing, QA, then deployment. In too many places, Gentoo developers are completely skipping the design and jumping straight into development. What this gives the world is a poorly designed product that is extremely hard to maintain and keep the quality up on over time. Beyond that, general testing an QA is being skipped in far too many places, or being done "after the fact" once something is in the wild.

I hope that the election of a new Gentoo council will bring about change to make Gentoo for the better, but truly fear that unless we start taking a hardline position on many of these new projects that we will fade into oblivion under the weight of our own garbage.

Re:Yes, Gentoo is a mess

[Penguin Follower]

That pretty much sums up the problem with Gentoo, nowadays -- searching the forums is now a routine tool for getting your updates to work. It used to be that Portage was the biggest advantage of Gentoo; now it's one headache after another.

Which is exactly why it took me a while to figure out what I needed to do to fix the problem -- I *DON'T* read the forums. Generally, I prefer to fix it myself if I can. If I can figure it out myself, then I have learned something deeper than being told what to do. That said, my solution wasn't nearly as concise as that one-liner he quoted. :/

Re:Yes, Gentoo is a mess

[agaffney]

Generally, I prefer to fix it myself if I can.

I figured this out on my own, but that's because I know how to use portage and what a blocker is.

Re:Yes, Gentoo is a mess

[wolf31o2]

Except that as another user noted above, there's a problem with the base install where PAM and shadow libraries conflict with each other (obviously there are packates from the stage tarball that depend on each?) and it was a problem in the 2006.0 release. Something like that should have been fixed for the 2006.1 release. I had a lot of "fun" getting around that problem (and a lot of wasted time!).

Umm... that's called a blocker and it is done intentionally to keep you from screwing up your own system. Perhaps you should read section 2 of the Handbook, that clearly explains how portage works, what a blocker is, and what you should do when you get one. No offense, but we can't fix stupid.

I mean basically "out of the box", you've just finished the install and you reboot into your new gentoo system only to find a PITA of a problem the minute you go to install a package or set of packages that depends on Pam or shadow. I still have the 2005.whatever release still running on my main linux box because of this (I was smart and tried 2006.x on a different system first.)

The state of the tree is consistent with itself. It had nothing to do with packages depending on PAM or shadow. The entire situation was that at one time, the "login" program was split out into a separate package when you merged "shadow" with USE="pam" and with newer versions of shadow, it was re-integrated. Because of this, one needed to first unmerge pam-login, which is why a blocker was added. This is normal and expected behavior on a Gentoo system. Now, the whole 2005.whatever comment just shows that you really don't know what you're talking about, since the release information is only important when you're booting from the CD and unpacking the tarball. Everything after that point would be "Gentoo Linux as of my last emerge --sync" not "Gentoo Linux 2005.whatever" as Gentoo is in constant motion. But you knew that, didn't you?

Re:Yes, Gentoo is a mess

[agaffney]

I only have one thing to say to this....damn straight!

Re:Yes, Gentoo is a mess

[arth1]

This was covered on the mailing lists, GWN, and forums. Don't blame Gentoo because you can't use the tools or read.

And here we have the Gentoo attitude in a nutshell. Instead of problems being fixed, workarounds are posted somewhere, and you're derided for not having checked all of the Gentoo mailing list, gmane, the .ebuild file, comments to unrelated bugs, and patches in a CVS repository overlay. A user can spend hours browsing gentoo.org documentation without finding what he needs, because someone not intimately familiary with how Gentoo is organised just won't know all the right places to look. And when he asks, he's being told off in a rude manner. Welcome to Gentoo.

Re:Yes, Gentoo is a mess

[arth1]

You mean you've never updated the box since you installed over a year ago? If you've updated world recently, congrats, you're running a 2006.x release on that box.

Wrong. You need to change the /etc/make.profile symlink to get to a new release version.
For most x86 users, that would be something like:

# file /etc/make.profile /etc/make.profile: symbolic link to `../usr/portage/profiles/default-linux/x86/2005.1'
# ls /usr/portage/profiles/default-linux/x86
2005.1 2006.1 dev make.defaults packages use.mask vserver
2006.0 ChangeLog gcc2 no-nptl parent virtuals xbox
# ln -fs ../usr/portage/profiles/default-linux/x86/2006.1 /etc/make.profile
# env-update && emerge -e system && emerge -e system && emerge -e world

Regards,
--
*Art

Re:Yes, Gentoo is a mess

[agaffney]

You people just don't get it. The pam-login/shadow blocker was not a bug. It was a fix for a bug. Dealing with blockers is a normal part of administrating a Gentoo system. If you can't handle it, go use Ubuntu.

Re:Yes, Gentoo is a mess

[wolf31o2]

No. That just gets you to the release profile, which is generally nothing more than the default USE the release was built using. Changing your profile link doesn't change your release, it simply changes the defaults the system will try to use. Otherwise, how could their possibly be non-release profiles? (Hint: because profiles are not solely tied to releases and vice-versa)

Re:Yes, Gentoo is a mess

[agaffney]

Do you realize that you're telling the Gentoo x86 Release Coordinator that he's wrong about what constitutes a release in Gentoo? The "releases" in Gentoo are really nothing more than media refreshes. The new profiles may give you a slightly different set of default USE flags, but in most cases, you will get the exact same versions (the latest) when you 'emerge -uDN world'.

Re:Yes, Gentoo is a mess

[unixj]

It's not all bad. Does anyone remember how a couple months ago it would take 30 mins to emerge --sync, with the disk crunching the whole time (at least on my machine)? Now it takes like a minute. That was a pretty sorely needed change.

But I do agree that Gentoo seems more intricate and less stable than 3 years ago. Some of this intricacy must be related to increasing flexibility and power? But I don't use my box for anything different than I did 3 years ago, and I'm becoming increasingly glum at the thought of spending the whole frickin weekend updating my machine and debugging problems with the update! I am this -><- close to just buying a Mac and roughing it out with the beautifully integrated suite of built-in apps.

Re:Yes, Gentoo is a mess

[oddfox]

Wow, you're funny.

Re:Yes, Gentoo is a mess

[dhasenan]

Then a fork sounds like a good idea. If good politicians are involved, perhaps it could be arranged that one team manages unstable and the other stable/base; you could keep one repository and one installer, but switch between the two with minimum hassle.

This allows separate development of the two and simple reintegration once the base is stable and the new whiz-bang features can safely be added. And the magic of portage means it'd be relatively easy. (Though it'd also be rather simple for most distributions; Portage would probably make it easier to mix and match, though.)

Re:Yes, Gentoo is a mess

[Anonymous Coward]

You seem to describe what Debian (to some extent) already is...

Re:Yes, Gentoo is a mess

[smash]

OT - whats all the fuss & hype about Ubunto/Kubunto ? I've install it a couple of times and I can't see any compelling reason for those distros compared to anything else like debian (apart from an easy install) or SuSE ? They just feel like a cut-down debian for some weird reason.

As said, it's because they ARE a cut down version of debian.

Or rather, (more specifically) a sane collection of default packages pulled from the debian unstable tree.

The big thing with *ubuntu is that if you run "ubuntu" it's a known quanitity - you *will* have at least a given base set of packages installed, that have been tested to work together.

If you want something more complete/customised, then use Debian... Personally I've been using Kubuntu any time I use Linux recently (i used debian exclusively between 1997 and 2001, nowadays I'm swinging more towards FreeBSD instead of Linux) because of the liveCD (run it on any machine that doesn't otherwise have linux installed) and because you can be up and running with a fairly complete, usable desktop in well under an hour.

I did the whole compiling/upgrading from source thing back in the mid-late 90s with slackware :D I'd rather spend my time working on my own code, instead of buggerising around compiling somebody elses.

Re:Yes, Gentoo is a mess

[tknd]

I'm a gentoo user and I agree. I upgraded my compiler, kernel, and all of my packages a while back and the worst part of it all was that I was afraid that something in the process of the upgrades was going to break the system especially if I didn't do things right. I got through it cautiously with a few gotchas but nothing I couldn't fix myself.

Prior to that, I had problems building an mplayer version that was marked stable with certain use flags. There were also issues where I would emerge a package that was marked stable then a few weeks later it would be hard masked and attempt to unmerge itself and install some other version which failed.

Now I simply don't emerge until I'm ready to deal with the maintenance headaches because I'm afraid something or another is going to break. But I know at some point, something in my system needs to be changed or I'm going to want to install package X which depends on new version of package Y which needs compiler version Z.

I'd be much more interested in a system where I'm either not afraid to install an experimental or bleeding edge package (even if it's marked stable) as long as I'm given the option to roll back to my previously working configuration. Even better yet, insert another package marker 'stable-mature' meaning the package has been available for some duration of time (say 3 months or more) and has not had any build issues reported during that period.

The whole fixing things part is great and fun when you have plenty of time. But for other things I just want it to work, and I want to be 99% sure that it's going to work even though it's 6-month old technology. I noticed the only time I'm really interested in an upgrade is when either the new version provides functionality that I need or am interested in using or the new version includes a bug fix that I find important. Otherwise it seems like more headache and experimentation than what it's worth.

Re:Yes, Gentoo is a mess

[coolcold]

Is it possible to fork the code base into one experimental and another stable?

given gentoo is GPL (no?), is it possible to bring create another distro with the current code base to make it more stable?

Re:Yes, Gentoo is a mess

[starmang]

Sounds to me like Gentoo developers need to read The Pragmatic Programmer, then decide where they want Gentoo to go. Possibly create two teams, one for the stable development and fixing current problems, and another for the shiny magpie features. Gentoo can't expect their project to continue without direction. The problem here is that there is no direction, no goals and because of this the project is going off in different directions. This is due to lack of communication, and lack of scope. When this happens software spirals out of control and developers are left wondering what it is they were trying to do.

It's not hard to sort out - go back to basics.
Firstly, decide where you want Gentoo to be a year from now.
Secondly, sort out the existing problems before adding new features (problems) to the list.
Thirdly, create a sub team from the current dev team (not really a fork) to work on shiny features. Infact, there should probably be several teams designated to certain parts of Gentoo.
Once you have scope and know what you're working on it's alot easier to get the job done.

Re:Yes, Gentoo is a mess

[Anonymous Coward]

Oh the irony of a Debian user criticising Gentoo. FYI: I don't use either and consider both to be hopeless for production systems.

I recently had the misfortune to get the job of setting up a Debian stable box for a pushy developer on the other side of the world. Ancient software loaded with bugs with shitty hardware support that freaks out when you try to mix PCI graphics cards and PCI network cards. GUI network control tools THAT DON'T FUCKING WORK.

I got so caught up in the labyrinthine Debian hell that I eventually wiped it an started again... in the meantime I had Fedora Core 5 running on it in less than 15 minutes. But hey, Mr. Debian Zealot insisted on Debian... so I had to restart my Debian nightmare again.

You know... Debian stable is old software... and in open source... old = buggy unless it's heavily supported by a commercial organisation (something like RHEL).

Re:Yes, Gentoo is a mess

[petrus4]

Anyone trying to have technical discussion is called out and accused of flaming by the once great Seemant (who has not done any development himself for years) and his horde of fanboy minions (most noticably, Jakub)

This sounds to me like the same old story with regards to any form of participation in the terminal sociological disease we call "the Linux community" in general.

Eric Raymond, Richard Stallman, and Bruce Perens are the broader community's resident egomaniacs. RMS in particular considers anyone who comes anywhere near Linux to automatically become his bitch by default. The other two aren't quite as bad, but they're not much better. From what I'm seeing here, it seems the seperate communities of some distros have the same type of "celebrity" problem...people whose output of actual work ceased years ago, but who insist on staying around and trying to demand that those who actually *are* working bow down and worship them on the basis of *past* accomplishments. Stallman is the single biggest example of what I mean, here...you need to go back *at least* 15 years to point to any of his programmatic contributions, and yet he still hangs around now, shooting his mouth off, demanding credit for things that don't belong to him, and causing nothing but problems generally.

I'm working on something almost entirely alone as a way of avoiding this type of garbage. If I need to deal with anyone, I talk to the people who are doing the various sub-projects' actual gruntwork, direct on Freenode...the proverbial people in the trenches. They're the ones who really build Linux, day in and day out, and they generally get zero credit for it. I remember what ESR once said about this, and I'm going to expand on it:- If you're not producing actual code, but are simply looking to build your own ego or a clique, then kindly sit down and shut the fuck up so that the rest of us who *are* doing something useful can concentrate.

The various social, psychiatric, and neurological disabilities of a number of people associated with Linux by themselves constitute the operating system's main problem...nothing else. Said people need to get over themselves, and above all, quite honestly disappear if they're not willing to do anything genuinely constructive.

That is who gets my own respect, though...people doing actual work. Linux's "celebrities" are formally invited to go and perform anatomically impossible acts with various sharp-edged gardening implements, as far as I'm concerned...and that goes triple for RMS.

Re:Yes, Gentoo is a mess

[Sloppy]

emerge -B shadow && emerge -C pam-login && emerge -1k shadow

That was one of the most infuriating and inexcusable events in Gentoo history, IMHO. If that kind of shit ever happens again, I'm switching to .. oh crap. Well, something. Anything. Gentoo is on double-secret probation for that one.

Re:Yes, Gentoo is a mess

[Anonymous Coward]

Are you sure your misguided opposition to this is not sour grapes that it was "your" idea and someone else took the initiative to try to implement it? As developers, should not everyone look to the betterment of the disto and not nickpick on whose idea it was/is or whose territory we're walking on?

If your ego gets in the way of Gentoo moving forward, maybe its time to move on. Your contribution to Gentoo is valued and very respected, but sometimes we all need an ego check.

[17:28:02] robbat2 - I think that the most interesting stuff to come out of seeds won't be their stage4 binaries (which they've admitted will be for only one or two arches) - but rather their much smaller profiles and meta-package stuff
[17:28:45] wolf31o2|mobile - see... this is what pisses me off... I've been trying to do exactly this sort of thing for ages and I've been denied at every step
[17:28:57] wolf31o2|mobile - suddenly, somebody "comes up" with this idea, and they're praised for it
[17:29:09] wolf31o2|mobile - it's fucking bullshit
[17:29:13] wolf31o2|mobile - and I'm letting it rest at that
[17:30:09] robbat2 - let em do it, and then give them prior art, and be done with it. end of subject.
[17:30:54] wolf31o2|mobile - I have nothing to do with stopping them
[17:31:07] wolf31o2|mobile - I'm honestly to the point of complete apathy for pretty much everything Gentoo at this point

Re:Yes, Gentoo is a mess

[Viduliya]

Sadly, I had to switch out of Gentoo this week. Fixing problems with my Gentoo machine was getting in the way of completing my work. I do hope that your efforts will lead to a "good, stable, reliable, and flexible system that is capable of meeting the demands put upon it". The way things used to be over a year ago. I would consider switching back to using Gentoo only when I can be absolutely convinced that some developers experiments are not going to be done on my machine. I am not willing to be a lab rat. Thank goodness there are still people who thing straight in Gentoo. There is still hope for the distribution. I hope it gets back on track.

Re:Yes, Gentoo is a mess

[HorsePunchKid]

My favorite was when bash 3.1 was marked ~x86 before any of the init scripts had been updated to work with it. I rebooted at some point to find that (seemingly) almost none of the scripts in /etc/init.d or a number of other places were functional anymore. I found after a lot of wasted time that there was some problem with declaring multiple local variables on one line; something like that, anyway.

As I recall—and my memory is hazy (like the moors of Scotland)—the 3.1 ebuild was withdrawn from ~x86 perhaps the next day, but if you happened to run your update in that window, you were screwed. Never heard why that happened; didn't really care, since I managed to muddle through fixing it on my own.

(And no, I don't need a lecture on running a system with A_K=~x86...)

Re:Yes, Gentoo is a mess

[Anonymous Coward]

No offense, but we can't fix stupid.

...

But you knew that, didn't you?

I hope—despite your email address—that you're not a Gentoo developer or in any way representative of the the Gentoo development team. I've found most of the people I've interacted with in my experiences with Gentoo to be very helpful, and I wouldn't want your post to stain that reputation.

Re:Yes, Gentoo is a mess

[jlowe]

I have used gentoo exclusively since 2001. But, you should stop talking because you are coming across as an ass, which isn't going to help gentoo's case at all.

Re:Yes, Gentoo is a mess

[agaffney]

But, you should stop talking because you are coming across as an ass

I'm well aware I'm coming across as an ass, but I'm getting tired of people spreading false information about Gentoo.

Re:Yes, Gentoo is a mess

[arth1]

Do you realize that you're telling the Gentoo x86 Release Coordinator that he's wrong about what constitutes a release in Gentoo?

That would perhaps be a clear indication that the coordinators don't speak the same language as the users and fail in getting their vocabulary through?

Re:Yes, Gentoo is a mess

[hawaiian717]

Does anyone remember how a couple months ago it would take 30 mins to emerge --sync, with the disk crunching the whole time (at least on my machine)?

I do. Only on my machine, a 160MHz PowerPC 603ev, it took hours. As I recall, it had something to do with the KDE ebuilds, and it took a while to get released to stable, but a Portage update ultimately fixed it.

Re:Yes, Gentoo is a mess

[Anonymous Coward]

| I'd be much more interested in a system where I'm either not afraid to install an experimental or bleeding edge package
| (even if it's marked stable) as long as I'm given the option to roll back to my previously working configuration.

But you *are* able to roll back easily. You just need to know your tools (which many gentoo users don't).

Too many users which *shouldn't* be using gentoo (because they lack the knowledge how linux works) are using it these days, which is a shame.

Sunrise disaster

[urbanradar]

For those who have no idea what exactly the "sunrise disaster" in the summary is supposed to mean, like I did, here's the link: Project sunrise.

Re:Sunrise disaster

[paskie]

And what exactly is disastrous about it? According to the timeline it is alive and nothing hints about any disasters...

Re:Sunrise disaster

[Anonymous Cumshot]

What about gentoo games? Or portage-ng? Or even the new website layout? That's right. They all failed. Just keep adding more trash to the pile, guys.. The gentoo project seems to be like the child with AD(H)D in the linux world.

[this is good]

[AtariDatacenter]

I personally benefited from the Ubuntu 'server' install CD. One of the options was clearly labeled something to the effect of, "INSTALL LAMP SERVER". (Linux, Apache, MySql, Php) In no time at all, I went from bare metal to up and running Drupal. I can't tell you how much of a time saver it was. (And out-of-the-box pretty secure, unnecessary daemons all disabled.)

Re:[this is good]

[noamsml]

What idiot modded this funny? This is obviously a serious comment, not a joke.

Re:[this is good]

[*BBC*PipTigger]

What idiot modded this !funny? This is obviously a !serious comment, !not a joke. ;)

The grandparent reads like a commercial (and is thus funny whether it was serious or not). It's akin to "I'm computer illiterate. I'm convinced all computing devices hate me personally and break just to spite me. I had no idea what I was doing. Ubuntu saved my business in no time! Thanks Ubuntu. ... Act now and Ubuntu can be your personal savior too. It slices. It dices. It even fries eggs." It's like a goofy quip. The P in L.A.M.P. also regularly stands for Perl or Python instead of (or alongside) PHP.

-Pip

QA vs. Q&A

[crow]

"Q&A" is Question and Answer.

"QA" is Quality Assurance.

Both are applicable. The former is for people who are tired of being told to RTFM. The latter is for people who are tired of things not working as they should when the do RTFM.

Re:QA vs. Q&A

[tomstdenis]

True and Gentoo sucks on both accounts most of the time.

The only reason I use Gentoo [other than USE flags] is that I *like* tinkering with my box [e.g. hacking to install packages].

Most people aren't as inclined as people like me which is why OSS sucks.

Tom

Re:QA vs. Q&A

[MrAnnoyanceToYou]

Um.... OSS Sucks, and yet you still use it? Um.... Look. You're using the most user-unfriendly chunk of OSS there is. You're complaining it's user-unfriendly. I've installed both Ubuntu and the old Mandrake, I've done it for myself and non-techies, it was easy, I'm actually pretty happy overall. I'm not a demon-monster-coder. I can code enough to do it for a living.

Your continual 'OSS Sucks' comments are mildly offensive, seeing as you're judging an entire community based upon your experience with a small piece of it. They show you personally lack in certain areas of consideration.

OSS doesn't have any real problems, it's individuals within the system. The system / movement / whatever you want to call it itself mutates all the time. It's a software project that has lasted and lived in the same area for over ten or fifteen years, and managed to keep up with modern technology. There aren't a whole lot of codebases like that out there.

I'm sorry you think OSS sucks as a whole. But I don't agree and flaming the crap out of all the people that don't is just trolling.

Re:QA vs. Q&A

[tomstdenis]

Um, I write OSS [which I release as public domain]. So I'm not strictly an outsider.

But when I get comments like "your code has comments in it, god bless you" from random people on a near daily basis ... that's saying something.

I mean honestly try and decipher the kernel, or GCC or any other project [firefox?]. You'll see a trillion lines of goobly gook code that lacks organization or proper forethought. Is there good OSS out there? Yes. Is it the rule? No, it's the exception.

OSS is much like drivers on the road, most are shity but they work just enough not to get into crashes. And when the software is stable it's usually still a mess underneath anyways.

Tom

Re:QA vs. Q&A

[Anonymous Coward]

OSS is much like drivers on the road, most are shity but they work just enough not to get into crashes. And when the software is stable it's usually still a mess underneath anyways.

This is different from commercial software how exactly?

By the way, if you want the right to complain you pay for the privilege. These people are releasing their work free to the world, with no monetary compensation. Their technique may be bad but so what? If you want an assurance of quality, pay for it. OSX thrives on that market, and frankly you sound like you should be using it. Free means no guarantees, and some of us are willing to make that bargin. You obviously aren't, and shouldn't. I agree with you about the state of code out there, but since its free it doesn't bother me. If I want better code, I can either fix it up myself or pay someone to.

If your personal coding standards are high even for free code, I applaude you and wish more people were like you. HOWEVER, I know that I have no right to deride others when they do not code this way, because I have not paid them anything. I suggest your tone and attitude are not going to achieve even your own end, which I assume is to get people to write better code. Try instead suggestions, pointers to lessons on techniques, and polite discussion of the potential benefits. Attitude NEVER helps, and usually hurts.

Re:QA vs. Q&A

[MrAnnoyanceToYou]

I am strictly an outsider. I've never written a single line of OSS code, mainly because I don't think I'm a good enough coder. But if you think that OSS code is worse than corporate code, well, you've got another think coming. I can't imagine it's anywhere near as weird or self indulgent in any of the ways some of the crap I see is. Not in the slightest. I've considered writing damning or blessing e-mails in response to the way people code, but quite often there's no way to get a hold of them and/ or they don't care.

Code is like science fiction. 90% of it is crap. But since OSS is so big and old, the 10% sliver has become a larger percentage of what people actually run.

Re:QA vs. Q&A

[bit01]

Closed source code is every bit as poor as OSS. And worse. At least with OSS you get to see it all and fix it if you want.

Your comments, complaining about OSS instead of software in general, make you a software bigot.

---

Vista: Billions of marketing words and no delivered product.

Re:QA vs. Q&A

[Dan Ost]

True and Gentoo sucks on both accounts most of the time.

Over the years I've made extensive use of both the Gentoo forums and
the Gentoo Bugzilla and have found both to provide quick and high
quality feedback. Perhaps all these years I've just been asking easy
questions, but it seems to me that Gentoo Q&A is excellent.

I do periodically run into QA problems, but they are always short-lived
and the solution is generally on the forums or bugzilla before I've
noticed the problem.

Re:QA vs. Q&A

[sarathmenon]

Now, who modded the parent funny? We're definitely eeeing strange streaks of humor here.

Re:QA vs. Q&A

[Anonymous Coward]

> OSS is much like drivers on the road, most are shity but they work just enough not to get into crashes.

It is a lot like driving actually: most people rate their own skills as "above average" when in fact they're anything but.

There are alternatives to Gentoo

[SaidinUnleashed]

I'll take the opportunity to pimp Source Mage Linux (http://www.sourcemage.org). SMGL is far simpler, easier, and faster to set up than Gentoo. The system management scripts are fast, and work astoundingly well, and the devs are always in irc and love to help. Just an all-around nicer bunch of guys and a better distro than Gentoo's seen in a few years.

Re:There are alternatives to Gentoo

[Darkael]

Sourcemage is a nice little distro with some neat ideas, but unfortunately the lack of interest about it and its consequences (lack of community support, lack of packages, lack of documentation...) is a boner killer and makes Gentoo a better choice for a source-based distro. That's too bad, because Gentoo could benefit from some competition.

Re:There are alternatives to Gentoo

[Zach978]

I find the Source Mage community VERY helpful. The IRC room seems to never leave me hanging, and has always been very friendly and helpful. I think they have a pretty large library of packages (grimiore in SMGL terms), and if a package isn't there it's pretty easy to create. There is even a script to do it for you.

quill -f -- will try to grab from freshmeat and generate spell (SMGL for package)

Or run with no arguments for a guided spell generation, then of course submit your package!

Re:There are alternatives to Gentoo

[dogwarrior]

Here's another satisfied SourceMage user. :)

My first source based Linux installation was SourceMage, which I decided to try after the Gentoo 2006.0 installer failed to accept the keymap I told it to use. I was quite happy with SourceMage but I wanted to try Gentoo because it's more popular. So, when the Gentoo 2006.1 installer came out, I decided to give it a go. This time it accepted my keymap and I got Gentoo successfully installed. Still, I found Gentoo installation a real PITA when compared to SourceMage's straightforward installer.

Then I read Gentoo documentation and found out that I needed to consult some man page in order to set USE flags. This brought me dark memories from the time when I fought with FreeBSD's Ports system. And, sure enough, I learned that Gentoo's Portage was inspired by FreeBSD's Ports. I never had to edit any config files to set dependencies in SourceMage. SourceMage just asks which dependencies I want before installing packages and then it remembers my selections.

At this point I started to realize what a fool I had been in giving up SourceMage just because Gentoo is more popular. I had been very happy with SourceMage and now I began to see how much simpler it is to set up a SourceMage system than a Gentoo system. Why should I continue to fight with Gentoo when SourceMage wanted to make things easier for me?

With these thoughts I decided to say goodbye to Gentoo and reinstalled SourceMage. And I don't think I'll give Gentoo another try as long as SourceMage is still around. SourceMage is just more intuitive and less of a hassle than Gentoo.

These experiences make me wonder why Gentoo is so popular? Is it just because people don't know about the alternatives? Or is it because everyone knows that Gentoo is the most popular source distro and people automatically assume that the most popular must always be better than the less popular alternatives?

Re:There are alternatives to Gentoo

[SaidinUnleashed]

I have no idea what you're smoking, but SMGL has awesome community support and participation. Within 30 minutes of kernel 2.6.18 being released last night, it was put into the testing distribution. Other packages like KDE, Gnome, etc see similar speed in submission. Zach978 already stated the facts about "the lack of" packages, so I'll leave that issue. As far as documentation, the only parts of SMGL that are SMGL-specific are the sorcery scripts, and they are documented decently enough. All the spells (packages) install vanilla versions of whatever application from the source downloaded from the project's own site. So if the project is documented in it's homepage, there you will find the documentation. It should not be expected for every distro to create redundant documentation for every piece of software that one may use, because that makes for an awful lot of time wasted that could be spent improving the distro. Instead, maybe you should just expect docs for the unique aspects of that distro.

Re:There are alternatives to Gentoo

[shutdown -p now]

It's all well and good, but how many packages are there in the SourceMage ports tree (whatever it is called there), and how fast are they updated when new versions are released upstream?

Re:There are alternatives to Gentoo

[Darkael]

I have no idea what you're smoking, but SMGL has awesome community support and participation
Well, I'm sorry but a small IRC channel is not exactly my definition of an "awesome" community.

Within 30 minutes of kernel 2.6.18 being released last night, it was put into the testing distribution.
2.6.18 is currently in the stable branch of Gentoo, and this is probably the case for many other distros, so what is your point?

Zach978 already stated the facts about "the lack of" packages, so I'll leave that issue.
I'm sorry again, but I don't think "do it yourself" is a satisfactory answer to that problem.

So if the project is documented in it's homepage, there you will find the documentation. It should not be expected for every distro to create redundant documentation for every piece of software that one may use, because that makes for an awful lot of time wasted that could be spent improving the distro
You are just looking for excuses for the lack of documentation around Sourcemage. People always need good and preferably centralized documentation for their distro. I'm not talking about redundant documentation, but at least something to get you quickly started on a particular topic and how to get past the most common problems related to Sourcemage (and please don't tell me that there aren't any problems in Sourcemage)

Re:There are alternatives to Gentoo

[dogwarrior]

how many packages are there in the SourceMage ports tree (whatever it is called there), and how fast are they updated when new versions are released upstream?

SourceMage calls packages "spells" and a collection of packages is called a "grimoire". http://wiki.sourcemage.org/Codex I have four grimoires installed: test, stable, games and z-rejected. "test" (which I mainly use) has new packages/spells while "stable" contains mostly the same spells that are in "test", only these have been more tested and bug-fixed and, hence, the spells in "stable" tend to be older than those in "test". Then there's the "games" grimoire that includes, you guessed it, games. And "z-rejected" has non-free stuff and binary packages. So let's see how many spells are there in these grimoires:

$ gaze grimoire stable | grep Total
Total spells: 4268

$ gaze grimoire test | grep Total
Total spells: 4488

$ gaze grimoire games | grep Total
Total spells: 314

$ gaze grimoire z-rejected | grep Total
Total spells: 141

That's more than in most smaller distros, I guess. Anyway, I find that pretty much everything I need is already there. But you also asked how fast are spells/packages updated in SourceMage. The simple answer is: quite fast. :)

The more elaborate answer is that if you want new stuff fast, then the test grimoire is the way to go ("scribe add test"). You may meet occasional bugs but reported bugs are usually fixed quite swiftly. I checked the DistroWatch page for SourceMage and most apps seem to be the latest versions. It says that OpenOffice is still version 1.1.5 in the "test" grimoire but actually there's a binary package for version 2.0.3 in z-rejected. GCC in test grimoire is version 4.0.3 and GLIBC is version 2.3.6. Also, SourceMage hasn't yet moved to the modular X.Org -- they use version 6.9.0. Most other programs should be the latest versions -- in the "test" grimoire, that is.

One word of warning, though: When I rebuild glibc from the "test" grimoire after the installation, the package manager (called "sorcery") told me that "Glibc spell doesn't know your architecture!" (my CPU is pentium3). I should probably become a responsible citizen and report this bug, but actually I'm quite lazy and I haven't yet got around to do it. Anyway, I edited /var/lib/sorcery/codex/test/libs/glibc/BUILD and commented every other architecture except my computer's (GLIBC_ARCH=i386). After that small bump on the road, my ride with SourceMage has been pretty smooth. :)

Re:There are alternatives to Gentoo

[dogwarrior]

Gentoo's excellent documentation has helped me a lot in configuring my SourceMage system. :)

Package management and most other aspects that are specific to SourceMage are well documented here: http://wiki.sourcemage.org/

This is how you manage init services in SourceMage:

telinit list
telinit enable|disable <service>
telinit run <service> status|start|stop|restart
telinit move <service> <runlevel>

Hooray for zealots!

[Anonymous Coward]

Well gee, Tom, then maybe you should stop recommending a buggy piece of crap, you fucking useless waste of space. On second thought, don't. The more crappy opinions you spout, the faster other people will relise what a fucktard you truly are.

Love,
Mom.

Re:Hooray for zealots!

[tomstdenis]

Kids....

I feel pity for people like you. Who can't simply co-exist with others because you feel you have to mock and insult others.

Friends buy you pints at the pubs.

Foes just ignore you.

Keep that in mind.

Oh, and Gentoo has USE flags which NO OTHER LINUX DISTRO HAS [or to the same degree] which is why I use it. It isn't because I'm fan. Yeah, building 900 packages, that's fun! really!

Tom

Re:Hooray for zealots!

[kelnos]

I feel pity for people like you. Who can't simply co-exist with others because you feel you have to mock and insult others.

Hello, kettle? Yes, this is pot. I'd just like to inform you: you're black.

Seriously, have you even read the other posts that you've written?

Suck my Seeds

[Anonymous Coward]

I'm Pro-Linux to the Gavrial "Pro" Strates.

Gunna hire the mob to wack Steve "Rim" Jobs

On the back of pimp Cadillac, stuff a man muffin 'till his sad sack spats.

Over the hill for the Gentoo Seed Cluster that segfault of my Still!

Fo' Schizzle my nizzle, I'm the Consulting overlord in the hizzle!

Call me, a man, secretary for the redmond clan: Moving the users to the distant land:

An' drainin' the world from the Linux debate,
'cause I the money I don't have is what I hate for another to make!

Watchin' the stills as made of the pimp'd rig from Bill's:
I'm the corporate stinga that like to manhole a cunnalinga:

The Bruce Perens, enter the Sodomite a care-barons, for the swearin' to uphold the GPL underneath yuh.

Logged-in users aren't forced to preview their comments.
Create an Account! To confirm you're not a script,
please type the word in this image:

• purred

Pointless...

[God of Lemmings]

I chose Gentoo specifically for the ability to build packages
with the features I wanted, and not to have to depend on packages
used in someone elses system layout and choices.

Seeds is a waste of time and resources. If I wanted a distro based
around someone elses packages I would have chosen Suse or Debian.

Idiots...

Squash key bugs first!

[starseeker]

I use and like Gentoo Linux, primarily because it is a distribution that lets me install virtually anything, including odd obscure scientific software, with a minimum of fuss. Additionally, many times when things work, they REALLY work because the distribution doesn't get in the way.

But I'm considering trying KUbuntu for my next go-around. In addition to the new software compile requirements gradually outrunning my computer's hardware, I must agree that the smoothness of massive universal upgrades just hasn't felt "as clean" of late. The most important environments for my linux box I will usually wind up building myself anyway (Maxima, Axiom, BRL-CAD, various Lisp packages) and for the rest of it I'm less interested in building for hours upon end for minor upgrades. Particularly if there is a decent chance of introducing problems.

Conceptually, I like the idea of a system that can build itself from source code - there's something clean about it, and also self sufficient. If a system can build itself, it means most everything on the system is pretty solid as far as having what it needs in place. But waning horse power and a focus on things other than endless system tweaking may motivate me to shift.

Originally, I loved that Gentoo let me turn on exactly what I needed to get my hardware to work well, and that was my primary motivation for using it. I still love its documentation, and that I suspect may someday outlive the main Gentoo project itself. But I think it might be time to check out the alternatives again, and lower my monthly power bill ;-)

If people aren't coders...

[Anonymous Coward]

...all they can do is submit bug reports, and use what is offered. If the coder part of the community doesn't want bug reports from normal random users-then don't take them! Don't ask for them! Keep the buzilla thing closed off, a total wall, blessed developers only. Take it further, don't distribute it except internally to yourselves! That's one way to solve the "complaints" problem!

GSD

[Umbral Blot]

Gentoo Seed Destiny?

This is how Free Software dies.

[RLiegh]

So, let's recap. Debian is having to pay developers in order to get a working distribution out of the door anywhere near on-schedule; NetBSD is embroiled in a scandal surrounding the undue influence of Wasabi on the core team -when it's not flayling wildly trying to cope with its' other management problems, and now it emerges that gentoo has been stuck in a political quagmire for years holding back even the most frivilous of changes (forget any major ones).

We've reached the point where all-volunteer, non-commercial unix-style Operating Systems are drowning in personality conflicts; and the only technical strides and achievements are coming largely from private companies (Sun, Redhat).

This quaint social experiment of altruistic development has shown two things: as much as you may dislike corporate culture, corporate structure and the incentive of a paycheck are what is needed to gain any sort of professional-quality software going out of the door on a regular basis.

Remove the structure, remove the incentive and before long you're left with nothing more than quibbling dorks and software packages like gentoo which half of the time are badly broken because no one can be bothered to work on them.

Re:This is how Free Software dies.

[Stalyn]

We've reached the point where all-volunteer, non-commercial unix-style Operating Systems are drowning in personality conflicts; and the only technical strides and achievements are coming largely from private companies (Sun, Redhat).

It should be noted that the majority of people working for RedHat/Novell/Intel on OSS projects were OSS developers first and then did good work which got them noticed by the corporate structure. They were then hired to do what they were already doing, of course now they have managers to deal with.

Also I think it's a little preemptive to conclude OSS has failed just because some Gentoo devs got into a cock-size contest. It happens in the corporate world too, but behind closed doors.

Re:This is how Free Software dies.

[kestasjk]

Bear in mind that you're talking about distributions of set of free software. The beauty of our current free software ecosystem is that it's divided into so many small chunks and dependencies. These small parts don't suffer from personality conflicts and other problems that arise with very large, influential projects.

Re:This is how Free Software dies.

[RLiegh]

>>We've reached the point where all-volunteer, non-commercial unix-style Operating Systems are drowning in personality conflicts; and the only technical strides and achievements are coming largely from private companies (Sun, Redhat).

>It should be noted that the majority of people working for RedHat/Novell/Intel on OSS projects were OSS developers first and then did good work which got them noticed by the corporate structure.

In short, we had the F/OSS boom which has led to a subsequent brain-drain, where the brightest and the best have left for the greener fields of corporate sponsorship. Nothing wrong with this, of course; without corporate sponsorship GNOME, GCC and many of the advanced features of the Linux kernel would either have never happened or would have happened many years later.

But a volunteer-run project cannot compete with the dollars and stability offered by the likes of Redhat or IBM, so we have a situation where the best and brightest (understandably) abandon the volunteer-oriented space in favor of a more favorable work enviroment.

>They were then hired to do what they were already doing,

In short, given a financial incentive to keep going past the point where the project lost interest for them, or -looking at it another way- to work on features which they find uninteresting but are needed by end users.

This is an incentive which volunteer projects -by their nature- do not have, which as a result ends up making volunteer run projects less featureful, less accessible and often times less innovated than their corporate owned competitors.

>of course now they have managers to deal with.

Which help to keep the corporate projects focused and keep releases going out the door in a predictable time frame. These same managers also function to prevent cock-size contests from getting in the way of reaching the projects' stated goals.

Now, I would agree with you that a spat between a few gentoo devs is hardly the end of FOSS; but if you read about the extent of the damage done to the NetBSD project by being tied to Wasabi, and then look at the fact that debian is having to pay people in order to release on time, the writing appears to be on the wall for your volunteer-run OSS. In fact there won't be a future for volunteer-based OSS model because the volunteer based model is dying (as evidenced by troubles with NetBSD-core, debian etc).

Re:This is how Free Software dies.

[RLiegh]

These small parts suffer from the same personality conflicts from which they originate (I'll point you to the circumstances surrounding the formation of the OpenBSD project, which you should already be familiar with if you're holding this conversation).

Larger projects manage to side-step personality conflicts and have infrastructure in place which prevents said conflicts from detracting from the project (not to say the project won't have other faults -NIH and Beaureaucratese are two that readily come to mind).

Privately-sponsored projects have people who manage not just the code but also the egos; which allows those projects to go forward with achieving goals instead of fighting and then splintering into a million dead-or-dying redundant sourceforge projects.

Re:This is how Free Software dies.

[acechase]

We've reached the point where all-volunteer, non-commercial unix-style Operating Systems are drowning in personality conflicts; and the only technical strides and achievements are coming largely from private companies (Sun, Redhat).

This quaint social experiment of altruistic development has shown two things: as much as you may dislike corporate culture, corporate structure and the incentive of a paycheck are what is needed to gain any sort of professional-quality software going out of the door on a regular basis.

I just want to point out the fact that this does not mean open source development as a business model is bad. But yeah, it's gotta be a "business model", i.e. it's gotta bring home the bacon. I work in Ocean Science and I know quite a few open source developers who spend all their time writting open source code, but they get paid for it. They're working in a small enough market that the big fish don't bother going after it, or because they know more about the domain so they have an advantage. But operating systems aren't like that, the OS market is one of the tastiest treats out there, and all the fish want it. All the quality open source products I can think of have businesses paying to help support the development costs (firefox, eclipse, etc), but operatings systems don't lend themselves as well to that type of sponsorship (especially because the market is so flooded).

Re:This is how Free Software dies.

[drew]

this must be the perfect example of the old physics experiment rule:
"When you need to ensure a close fit to a line, make sure that you use only two data points."

OK, so you mentioned three open source projects that are having trouble right now, and two commercial companies that aren't.

What happens when you add Ubuntu, KDE, FreeBSD, or Firefox to your list. (OK, nix Firefox. Bad example).

Or for that matter, look how much corporate structure and financial incentive have helped Microsoft to get Vista, IE 7, and Office 12 out the door on time.

Any group can be run badly, and any group can be run well if there is enough interest and leadership. This incident reflects poorly on the leadership and members of the Gentoo project, nothing more and nothing less.

Re:This is how Free Software dies.

[RLiegh]

Yep, that is exactly my point -it has to be a business. The volunteer-run projects are either collapsing from their on inertia (eg the debian project), mis-management (NetBSD) or are going nowhere at the speed of light (look at most of the projects on sourceforge).

It takes serious resources to get things done; and those resources don't come for free, as you excellently pointed out.

Re:This is how Free Software dies.

[Spit]

The large projects that succeed all have an arbitrator who has final call on any conflicts: Linux has Linux and OpenBSD has Theo. Sun and Redhat have executives. The decision maker is the final step required in a project to prevent ego-paralysis. It is telling that OpenBSD and Linux are two of the most innovative and complex projects around.

Note: Linux is a kernel.

Re:This is how Free Software dies.

[RLiegh]

Really quickly here since I"m on my way out the door....
>What happens when you add
>Ubuntu,
Sponsored by Mark Shuttleworth, or rather, his company (canonical?): therefore proves my point about the necessity of governship

>KDE,
Has a large amount of developers from Trolltech, Trolltech has a lot invested in the KDE project simply because it is *the* showcase for the qt library.

>FreeBSD
Anyone who tried the 5.x series can tell you what happened to FreeBSD. Matt Dillon also would probably have some valuble insights as to the health of the FreeBSD project, though he may have his hands too full managing its' successor -DragonflyBSD- to comment.

Re:This is how Free Software dies.

[RLiegh]

Typed too fast, I meant:
>Sponsored by Mark Shuttleworth, or rather, his company (canonical?): therefore proves my point about the necessity of corporate sponsorship

Re:This is how Free Software dies.

[wall0159]

What a silly arguement. Let's try turning it around to show just how silly it is:

OS/2 was a commercial failure and was disbanded, Netscape became bloated and unusable, and ended up being disbanded as a commercial entity. therefore commercial software is inherantly problematic.

I mean... come on! There are many examples of working OS/GNU/BSD/etc projects. To cite two, and from these conclude that 'Free Software' has 'reached the limits' is disingenious. (Note that you could list a million failed F.S. projects and it wouldn't change a thing - I can show projects that are working well, and that invalidates your arguement.)

Note to mods: think about what you read. Don't just mod something 'Insightful' because it's written with conviction!

Re:This is how Free Software dies.

[wall0159]

...And another thing: since when has 'Free Software' and a 'corporate culture' been mutually exclusive??

Re:This is how Free Software dies.

[ChaosDiscord]

Christ, what about this article is drawing out the trolls?

This is not "how Free Software dies." Yes, if you want certain guarantees like deadlines met, particular features added, or particular bugs quashed, yeah, being able to pay someone helps. But was this not true for the last 21 years? Yet somehow we ended up with Linux, Perl, Apache, the GIMP, Gnome, KDE, Konqueror and more. Some projects stumbling doesn't mean the whole thing is doomed. And sometimes projects hire people not because no one will work on the hard parts, but because they want someone to be able to dedicate more of their time so thing can progress faster.

Free Software isn't going anywhere. Indeed, that Free Software has been adopted by commercial endeavors like Red Hat, Novell, IBM, and Apple is just a further sign of its success. They're building great things on the foundations provided that that "quaint social experiment."

Perhaps a point could be made that community build software doesn't scale well, that it is bad at providing the unity necessary to ship a full distribution. But that's not the conclusion you came to. You concluded "This is how Free Software dies." That's silly.

Re:This is how Free Software dies.

[EnderWiggin99]

Exactly.

OSS oozes natural selection. If your project doesn't scratch an itch, it dies. If it becomes deadlocked, it evolves (different focus) or reproduces (forks).

Contrast this to the commercial product that has already 'won' genetically...Windows. When genetic 'perfection' (foodchain or market dominance) is achieved, there's really no reason to evolve any further; there are diminishing returns with increasing variations. Genes remain stale. Genuine progress in not made.

With OSS, there are projects devoted to every itch under the sun. Where one doesn't succeed, another does because it is genetically superior (technically superior) or has a greater drive for evolution (driven developer). Throw a bunch of these projects together into a framework like a Linux distribution and you can build a platform with incredible flexibility and incredible potential for genetic variation to serve the next big itch.

Consider this (admittedly bad) example. There are many different type of fish in the sea, physically adapted to thrive in their own environmental situation. Big fish, small fish, fish-eating fish, plant-eating fish, etc. There is only one type of human (relatively speaking). This type is 'good enough' to ensure survival, and we have no need for drastic variations because we're already on top. But if (when) the environment changes, we'll likely face extinction long before the fishies do.

Re:This is how Free Software dies.

[Anonymous Coward]

This quaint social experiment of altruistic development has shown two things: as much as you may dislike corporate culture, corporate structure and the incentive of a paycheck are what is needed to gain any sort of professional-quality software going out of the door on a regular basis.

Corporations are vastly inefficient organisations which get nothing really worthwhile done, as most of the time is getting wasted in useless things. As the organisation grows, so does the communication complexity, eventually ending up with dumb giants.

Last time I checked, the Microsoft quality was not. Terrible crap, and hobbyists have outdone them already. Or think about Adobe Acrobat: the download is hundreds of megabytes, it's slow, it wants to enforce tracking capabilities on you. Free solutions exist which are really small and work just as well, if not better.

Remove the structure, remove the incentive and before long you're left with nothing more than quibbling dorks and software packages like gentoo which half of the time are badly broken because no one can be bothered to work on them.

No, occasionally people argue, it's normal. The incentive is always there: to write software which scratches your personal itch. That one won't go away. The structure is useless. It's all about communication, and that's easy over the Internet.

I think people like you just do not understand. What you see from corporations is ONLY the outside. It's the image they try to give you. But try going to the inside of the corporation and you'll see what a big bubbling mess it is. Try working for them. You'll see. The future is not that of strict structures and corporations, it's about distributed development and freedom. The movement is only beginning!

Re:This is how Free Software dies.

[advocate_one]

Christ, what about this article is drawing out the trolls?

they're not trolls... they're anti-OSS shills planted by Microsoft. Their job is to sow discontent. Where ever there is a potential for rational discussion about the pros and cons of OSS vs proprietary software you will find them.

Re:This is how Free Software dies.

[labratuk]

Hey, dude, I've been running a gentoo system for four years. The same system. It started at gcc 2.95.3 and XFree 4.2.0, now it's running gcc 4.1 and xorg 7.0. It's still working as good as ever and being upgraded near flawlessly.

I also run debian systems, which work well.

The truth is that Free software pundits need something to write about. There are people sitting down and actually putting the slog in to maintain packages to make everything 'just work'. There are also people coming up with crazy projects and making press releases. The second group give the pundits something to write and conject about.

In reality, in the corporate world, these press release crazies would be in management and would have the power to force engineers to start work on their new crazy idea and abandon dull maintenance.

Us Free software guys aren't going anywhere, thanks.

Re:This is how Free Software dies.

[HikingStick]

An astute assessment of the situation. I do believe there is hope for some non-corporate projects, but from what I have seen, the ones that will succeed are those that are supported by a deeply committed community. I forward the opinion that those most likely to succeed are those where that commonality of community existed before the onset of the project. There appear to be many similarities between the OSS community (at-large) and the world of non-profits in the corporate world. As a long-time volunteer for a national youth service organization, I know that successes on the local level usually grow out of a small group of concerned people who get together to do something about it. In places where other local arms of the organization try to emulate the success of the one, often by recruiting volunteers to fill specific pre-defined needs, they rarely have the same rate of success. Pulling together disperate personalities and egos for a common good takes careful leadership and good planning, not to mention conflict-resolution skills comparable to those of a good hostage negotiator. As soon as those volunteers become employees, however, the common motivation of the paycheck can get most people to set aside many personal issues for the sake of the assigned task. The risk there, however, is that the original goal, whether it be developing the best, most user-friendly Linux distro or having the fastest growing chapter of a non-profit, may move from being a visionary dream to being just a job.

Re:This is how Free Software dies.

[archen]

Pretty much what I was going to post. It is also similar to what Charles Hannum stated about NetBSD. People seem to be forgetting that all of these projects are driven by people and people NEED leadership. Someone has to take the bull by the horns and keep everyone focused. Just having someone act as a traffic cop is not enough because a project is just treading water at that point.

Crazy as this might sound, this looks like human civilization playing itself out again, this time through software projects. You move from single people, to a tribe with elders - but that only scales so far before you end up needing something like a king. It's sort of scary when you think about where that road might take us though.

Re:This is how Free Software dies.

[Big_Al_B]

Corporations are vastly inefficient organisations which get nothing really worthwhile done

Could you, at your convenience, provide me contact information for the presumably unincorporated manufacturers of all the hardware, including chips, boards, and display and user-interface technology that you use to operate and/or develop free software? I'm interested.

If it's not too much additional bother, please provide additional information on the unincorporated manufacturers of your furniture, appliances, stereo, TV, clothing and automobile.

Thanks in advance!

Re:This is how Free Software dies.

[quag7]

I'm not sure your post is fair. Despite the problems you list, free operating systems have come *this* far. There are quite a few very usable, (in fact, some quite polished) linux distributions which have been developed by volunteers.

Whatever Gentoo's organizational problems, no one should get the idea that the distro itself is falling apart. Frankly, if I didn't see stories like this, as a Gentoo user, I wouldn't know something was seriously wrong.

I assume you were using hyperbole when you said that Gentoo packages were broken "half of the time." Well, I just re-emerged world - that's every package on my system, because I'd updated to a new gcc version.

The results were 6 out of 743 packages didn't compile. When I tried to compile them again, 2 worked, leaving 4 out of 740 packages with problems - not exactly a disaster; that's just over half of a percent broken packages.

This was approximately the same as on my other Gentoo box which uses a different arch (the above is x86_64, this other one is x86).

This has been my experience consistently - would I like a 100% working stable branch? Yes. It's worth working toward. Is my system an unstable shambles? On average I reboot every 60 days or so when I upgrade my kernel. I wouldn't call that...rickety.

You call a decade and a half (speaking of Linux) of volunteer Linux development "This quaint social experiment of altruistic development"?

I'm sorry, but I have disagree - that is a seriously skewed perspective. If I read your post, I would conclude that Linux hadn't risen above, oh, MS-DOS in terms of usability and sophistication.

I've been using a Gentoo Linux desktop for 5 years and Debian production servers at work for three. I don't mean to suggest I'm an expert, but I do have a fair amount of experience with these "quaint social experiments," and it could be that you're just letting off some steam, but if not, I find it hard to agree with your assessment of the situation.

In my opinion the reality is that the volunteer development process is (and always has been) imperfect, and can probably be improved - in fact, this model has been around long enough that now is a good time for a re-assement of social dynamics of such a system.

Nor is it purely altruistic - developing software (I assume) gives a volunteer developer a sense of personal satisfaction, credibility, esteem from others, and even something to put on a resume. Users could stand to be a tad more courteous and grateful to developers, this much is true, and I don't want to speak for developers but I find it hard to believe they consider themselves religious clerics slavishly serving users out of some sense of self-sacrifice. I am personally deeply thankful for their efforts. Perhaps some developers could add to this discussion in terms of their motivations.

People have really been unfair to the GPL, volunteer developers, and so on - immoderate and sometimes mean-spirited in their criticism. Let's not forget what this "quaint social experiment" has in fact yielded. I know - I use products of that process every day.

I hope people smarter at social dynamics and project management than I am can improve the process, but, well, Gentoo and Debian work for me, professionally and as a hobbyist.

To suggest that this system is some kind of a lark or failure simply contradicts the basic reality of the situation. I'm surprised you feel as you do, as I assume by your post that you use or have used one or several free UNIX-like OSes. Your assessment of the situation could not be further from my own.

Lastly, and I'll be a bit of a broken record with this, Gentoo is not for everyone. It is not intended as a side-by-side replacement to something like SLED though with some work it can certainly work that way. Gentoo is for tinkerers, developers, and people who are interested in the OS itself, rather than simply what can be accomplished with the OS. It's definitely for me but I don't think less of anyone who couldn't be bothered with Gentoo.

If anything, I hope to see the problems addressed simply so life will be easier on developers, and encourage them to stick around (and new ones to contribute).

Am I the only one?

[Stalyn]

Who has a gentoo system that works perfectly fine? Plus my system is entirely from the unstable branch.

Re:Am I the only one?

[/dev/trash]

I have three. Id say 2 of them work great.

Re:Am I the only one?

[RLiegh]

Running from any unstable source tree comes with it's own risks and necessitates a certain skill level (which, as a poster with a three-digit UID you doubtlessly have). That level of skill should not be necessary for building a stable tree. Finding problems with dependencies (routinely being unable to build package Y because library C has not be integrated) on a stable tree should never happen (and yet routinely does with gentoo); that kind of problem should only happen in unstable.

I have no doubt that you can find people who have the technical capacity to force gentoo into working through a series of kludges and nursing. That says less about the questionable health of the project and more about the willingness of users to come up with what should be un-needed workarounds (at least, shouldn't be needed in the stable tree) just to get a working system.

It's not the fans of a system that you judge it's quality, it's by the reasons put forth by it's detractors. Bring on the 'me toos' all day; but what they have to say is less indicative of the health of gentoo than are the ones who say 'tried it, never again!'.

Re:Am I the only one?

[Sterling Christensen]

Same here. Maybe it's the stable branch that's messed up.

Gentoo seems to have been steadily improving in a way that I'm satisfied with, and what few problems I've encountered have been easily resolved by a simple forum search. I don't understand what everybody's complaining about.

Re:Am I the only one?

[temojen]

I'm on stable (except for media-libs/libdv media-video/cinelerra-cvs and media-video/kino), and it works for me.

Re:Am I the only one?

[Azghoul]

I've been running servers on Gentoo for a couple years now, no problems like the ones people are having currently. I only update a few bits at a time, none of that emerge world silliness. My desktops at work and home are ~x86 and they run fine, except I can't figure out why my X cursor theme is ignored inside the gecko rendering area in firefox on one of them and not the other......

Okay, Gentoo has some sorting out of leadership personalities to do, and there will be some shakeout, no big deal.

Re:Am I the only one?

[HorsePunchKid]

No, you're not. I've had Gentoo on my elderly Thinkpad as my primary personal computer for something like three years now, and it works perfectly fine. I've had occasional problems but nothing I couldn't solve. I, too, am running from the unstable branch.

Re:Am I the only one?

[Dan Ost]

I run mostly stable and haven't had any non-trivial problems.

Really, I don't see any of the problems that people seem to be
complaining about here (I maintain 6 gentoo boxes with different
configurations).

Re:Am I the only one?

[the_greywolf]

I do. Keyword amd64 with a handful of ~amd64 packages. (I keep a tight watch on my /etc/portage/package.keywords file and rarely use my /etc/portage/package.unmask file except for packages I need.) I'm running a number of 32-bit binary packages without a hitch, and it's as stable as bedrock.

The only trouble I've run into is that I still haven't learned how portage handles 32-bit non-binary packages on a 64-bit environment. (I.e., I want to slot a couple libs for both 32-bit and 64-bit, and there is no binary package - and even less documentation on the process - one of the few things at this state.)

Parent is FUD

[whoever57]

'm tired of trying to talk people into Gentoo only to find out that the tree is half-f***ed all the time [like packages marked stable requiring other libs NOT IN THE F***** TREE YET].

This is simply not true. If there are problems, the parent has wildly exaggerated them.

I have been engaged in massive updates to my system recently, prompted by:

1. GCC update (and the effect on libstdc++)

2. OpenSSL (requiring a re-build of all packages depending on OpenSSL)

3. A large number of package updates recently.

Now, running the updates using revdep-rebuild for libcrypto and libssl has not got without any problems, but they have usually been solved by upgrading to the latest stable version of the package. So, I know from personal experience that the situation is simply not as bad as the parent claims. I have several systems, from servers to desktops, so I have a good range of packages.

Re:Parent is FUD

[tomstdenis]

I just finished building a Core2 box LAST WEEK. It required me to add new [undoced] USE flags to get Pango/cairo/etc installed correctly [flags I may add that my Opteron box STILL doesn't have].

Clearly the "xft" flag [iirc] should be either documented or defaulted as I didn't see it required elsewhere [and strictly speaking I was emerging Gnome so Gentoo should set the flags required for Gnome regardless].

Tom

Gentoo works fine for me

[Anonymous Coward]

I've been on Gentoo since RH EOL'd RH9. I switch all my desktops and servers to it and 95% of the time everything works just great. I emerge sync and -uDN almost every day and it's a very rare day when portage is broken. portage is the real beauty of Gentoo. It's the Lincoln's Axe of distros, and that's a good thing. Y'know Lincoln's Axe, right? How old is this axe? 150 years old, but of course no one piece is because it got a new handle, then a new axe head, then a new handle, then a new head...

I maintain 6+ Gentoo boxes this way and I've never had to re-install for any reason other than a h/w failure. About twice a year I have to do more than just emerge -uDN world to get a stable system again, but I will gladly pay that for never having to do a painful full-system upgrade.

Political problems aside (and what organization doesn't have those, be they FOSS or commercial?), it's a great, largely stable distro.

Re:Gentoo works fine for me

[Progman3K]

Totaly with you on this.

I've been running Gentoo for about 3 years. Before I was running Windows.

Gentoo is so much more stable, responsive, easy to maintain and has the best forums I've ever seen.

Like you, the only time(s) I've ever had problems on the machines I have installed it on was always due to hardware or running out of disk space! :-)

It is time to send them some money.

Re:Gentoo works fine for me

[eechuah]

I totally agree. I've used Ubuntu, Debian and currently use Gentoo on my mythbox. My experience is that when you want to do something on Debian/Ubuntu that "they" didn't think of, it's supremely difficult. Also there is very little documentation. Compare this to Gentoo's wiki and forums. You can find almost anything there; my current system has less "hacks" to make it work than it had when I was on Debian/Ubuntu.

I find that Gentoo's tree is the most well maintained among the few; it contains the most packages (I don't have to point sources.list to someone else's website that is sometimes down...) I'm really pretty happy with the way Gentoo works.

Re:Gentoo works fine for me

[Anonymous Coward]

Same here, I've been using quite some distros over the years, but for about 3-4 years I'm stuck at gentoo, it's just too good. I run "emerge -puvD world" about once a week, and I very rarely have problems. Other than that, it's zero maintenance.

Re:Gentoo works fine for me

[CanadaDave]

I totally agree as well. I had 5 boxes at one point, now 4, all run Gentoo. I share your experience exactly.

Re:This is how Free Software dies.-- (not)

[Halvy]

There is NOTHING *Interesting* about the flat out false assumptions you make.

It is important for people to realize that M$ with apparently alot help from the likes of /. management are not only spreading FUD, but M$ is doing all it can to bring viruses, discontent and worse to the OSS community.

However as long as we do what we have done all along, M$ and it's heirs cannot win.

-- The Jewish Mafia must be and will be stopped at all costs, period.

Great idea anyway

[Progman3K]

It does address a criticism often directed at Gentoo.
So now, no more waiting to get the tool you need.
Good idea, devs!

Well

[MmmmJoel]

There's hope. Technically, at least.

Why do you need Gentooo?

[bayankaran]

I use and like Gentoo Linux, primarily because it is a distribution that lets me install virtually anything, including odd obscure scientific software, with a minimum of fuss. Additionally, many times when things work, they REALLY work because the distribution doesn't get in the way.

I dont find any (practical) difference between Gentoo and an distro like SUSE. And all the fuss seems to be about how difficult its to use Gentooo!!!

Re:Why do you need Gentooo?

[trupoet]

It's called compiling things like PHP with what you need instead of what the distribution thinks you need so that you have to manually recompile from source the package and all dependencies from scratch.

Hmm compile from scratch and search out all src dependencies manually or simply:

USE="imap" emerge php

I take the latter, Kk thx

Debian Upgrade

[McGiraf]

"s/old release/new release/g"

bah, me i use s/stable/stable/g on servers and s/unstable/unstable/g on desktops when there is a new release of Debian.

I'd say migrate to Debian

[Anonymous Coward]

Spend more time on fucking Q & A. I'm tired of trying to talk people into Gentoo only to find out that the tree is half-fucked all the time [like packages marked stable requiring other libs NOT IN THE FUCKING TREE YET].

I honestly do not understand why you need to spend so much effort to get this distro going. You may want to consider Debian as an alternative; I've been running Debian testing and I don't remeber the tree being broken even once.

I give up..

[B5_geek]

I have just spent the last 2 days trying to install Gentoo. First I tried the newest 2006.1 LiveDVD. It wouldn't get past gpm (general purpose mouse), so I disabled gpm, and it got stuck on the next section.

I went to IRC, #gentoo@freenet.org and the sage advice I got was: um, yeah 2006.1 is bjorked, try 2005.1.

So I did. I popped in the LiveCD, let it boot and upon once complete I had a CLI. (surprised me, all other LiveCD's I have used actually booted to a GUI) Not a problem, I can handle this. I followed the directions in the handbook exactly. Everything went smoothly untill it came time to reboot (after setting up grub).

Reboot. Grub panics because it can't find what it needs. I got the edit menu and try to fix it. No luck.

So I go through the whole process again. This time I even went so far as to make my partitions the exact same size so that everything would be verbatim. reboot, same grub panic.

Third try; I avoid the Stage 3 install and do everything live via the online handbook.
It works! Glorious Rapture I can now boot to a CLI. The handbook on the CD is DIFFERENT AND WRONG. The online handbook is accurate and worked.

So now it's time to start installing apps. MC and rar were the first to be installed, portage was complaining about using an old profile, so I switched it manually. It still didn't like it, so thanks to help on IRC I emerged eselect and was able to change my emerge profile. I test it with a couple other small apps, and errors are all gone.

Now I need a web-browser, so I can google for answers to questions that I have. emerge lynx
Emerge now throws up some access violation. Next I try links, same error.

I think to myself, I'll get back to those later. so I emerge fluxbox (expecting to get xorg too, but I didn't despite flux's obvious dependencies).

Flux installs with no errors.
startx -> nothing
ok, so now emerge xorg-x11, and I get another Access Violation. I toss in a knoppix CD, get online to google these access Violations, turns out that it is (possibly) due to a font conflict between 2 differnet packages that need to be installed (that both need the same font).

I quit. Back to Debian for me. Apt I missed you.
I have tried:
redhat, mandrake, suse, slackware, DSL, puppy, linspire, debian, ubuntu, and now gentoo.
They have all caused me grief. But I still love debian.

Re:I give up..

[oddfox]

What is so broken about 2006.1?

Portage 2.1.1 (default-linux/amd64/2006.1, gcc-4.1.1, glibc-2.4-r3, 2.6.17-gentoo-r8 x86_64)

The GPM thing is rather odd, I've never had any problems with GPM except I sometimes have to restart the service when it decides it just wants to not work, but preventing an install? Very strange indeed because this would be a problem with the GUI installer and something I would think they'd know about. Did you search on bugzilla at all?

GRUB panics don't appear out of thin air, so if your GRUB was doing that there was some problem in the configuration. Are you aware with how GRUB works? The way it addresses partitions/hard disks is pretty weird at first.

I don't know about discrepancies between the online handbook and the one provided on the CD/DVD, but I would guess that the CD/DVD has the installation docs, not the Gentoo Handbook (There is a difference), and I have a hard time myself finding the installation docs on the website, they seem to be pushing the Handbook since around 2005.0. Different and wrong seems weird too because, well, the Gentoo installation process hasn't changed much at all since I've been using it consistently when I got sick of Slackware back before 10.0 came out. An old/incorrect profile right after installation? Sounds like there was an issue with how the system was setup, like an old stage3 tarball for possibly an incorrect architecture was obtained. Can't find anything out about your access violations either, but I'll take your word on it because hey, I've been hit by that problem before once or twice but the issue was resolved in a timely manner and I did my searches on bugzilla and the forums.

Seriously, I'm not trying to rag on anyone, yourself included, when I say this, but I really don't think Gentoo was the right distribution for you if you have so many problems with it just getting the system up and running. It's great that it gives you so much control, but you need to understand what you're doing and have the ability to do some problem solving.

P.S. -- While I have been using Gentoo consistently for quite some time now, this current installation is not the same one that I had back in the day, especially since I moved over to amd64 when I bought my processor. However, I can attest to the fact that the guidelines for how to perform an installation has not changed. The only difference you will notice in the Handbook today and the installation docs of the past is that stage1 and stage2 installations are not recommended anymore, and maybe some updates about devfs/udev and the addition of genkernel directions. Of course, I didn't look any further in the Handbook than the installation, since I kinda just learned as I went with portage and the like, as well as lurking on the forums for hot tips on things. Also, Debian is a great distribution and deserves all the props it gets, especially for how staunch they are on the positions they do take.

Re:I give up..

[neersign]

I, too, have been installing gentoo for the past 2 days. This is my second gentoo install, so I thought it would go a little smoother, and it did (both using 2006.1). However, it was not without bumps, and it is definitely not the most elegant of install procedures. Both times I have used the online documentation and had no problems. The first time, i used the Gentoo Linux x86 Quick Install Guide (even tho I was installing amd64, not x86). This time I used the Gentoo Handbook. I think that both documents are well written and if you read them carefully, you will not have any problems getting a base system installed "quickly". For me, it took a long time (well over an hour) because I read the sections as I am installing and it takes me a little longer to understand everything because I'm still newer to Linux. Personally, I would like to see a ncurses based install process much like Slackware's (or others) where I have to be at the keyboard to set up some initial settings and select packages, but then I can walk away while everything is downloading and installing.

All in all, I would say that the LiveCD+CLI install method of Gentoo did not cause me any less or more distress than any other *nix install I have used (i have not used Ubuntu yet because after using apt in Debian, I have no wishes to use a binary-based distro). Still, I would choose Slackware's or FreeBSD's install process any day of the week.

Re:I give up..

[the_greywolf]

What is so broken about 2006.1?

The installer apps are horribly broken and rarely work right - not to mention the fact that it encourages too many people to take an oversimplified approach to the install. Gentoo isn't SuSE, and they shoudl stop trying. The installer is a mistake. It needs to be scrapped.

GRUB panics don't appear out of thin air, so if your GRUB was doing that there was some problem in the configuration. Are you aware with how GRUB works? The way it addresses partitions/hard disks is pretty weird at first.

IMO, GRUB is a broken, overengineered peice of shit and I have no qualms about saying it.

Ooh, neat. It can read a filesystem and load a kernel all by its lonesome! Unlike silly LiLo, which has to have the kernel's drive sector locations to load it!

All that is for naught when it's so easily confused by a SATA controller. Works like a charm on the first try when it's an all-IDE system. But when you throw SATA (or SCSI) into the mix, everything gets moved around. All of a sudden, the second SATA drive is hd(0) and the third IDE drive is suddenly hd(1).

I mean, seriously. It tries too hard, and it's too confusing to make it work consistently across multiple systems. I like the features, don't get me wrong. But the whole configuration system needs to be rethought and made consistent for once.

...And the Multiboot spec is just a pipe dream.

(FWIW, I've used both LiLo and GRUB, and have successfully installed both on multiple systems. I've had much better luck with LiLo, and greatly prefer it over GRUB for a number of reasons - not the least of which is the configuration file. LiLo just makes more sense, it works, and it's more mature and stable.)

Re:I give up..

[oddfox]

On your first point, I elaborated elsewhere on my own personal blog yesterday about the Gentoo GUI Installer and what I feel about it.

First, the GUI installer stole my lunchmoney thing. I would counter that you shouldnt be using software thats only at 0.4 (There are some exceptions to be made where its something thats nowhere near as complex as an installer app), but thats your choice, thats your perogative if youre unwilling to simply do things manually and have it take a little bit longer. A stage3 to installed Gentoo LiveCD image doesnt take any more than a day on modern hardware. Two to three on older hardware, and longer on extremely old hardware that you just plain shouldnt be running a source-based distro on if its your main machine. The stage3 is the Gentoo LiveCD image install minus X, KDE, GNOME and the branding. Out of those three, X takes the least amount of time to compile, GNOME takes a bit more and KDE is the big kahuna that I always save for last, sometimes at a really low nice level so it doesnt interfere, but that hasnt been a problem for a while since the 2.6 kernel just responds better, especially with the CFQ scheduler that the gentoo-sources kernel has defaulted to for a while now. If the GUI installer gives you problems, dont be afraid now to do a manual install because quite frankly its stuff youll be doing anyways once the system is up and running.

I agree they should stop recommending the installer, immediately, in all documentation including the Gentoo Handbook. It's not tested well enough and has caused enough grief in the community. I understand they probably did it for wider testing, but in my opinion that's not a good enough reason for the flak it's drawn.

I don't want to get into an argument about GRUB because frankly I understand where you're coming from about it, but I didn't really try to imply GRUB was perfect for every situation, which is why LiLo is supported on Gentoo and most other distributions. Peace.

Yo, Tom!

[Anonymous Coward]

Tom,

You might try FreeBSD. Although it's not immune from code out-of-sync'ness, the team at FreeBSD ride them doggies a LOT harder than they seem to over on your side o' the creek. You'd be welcome.

Re:Yes, Gentoo is a mess-- (NOT)

[Halvy]

..with the kind of fuckups that OS X and Sunrise were..

Huh??

If you look closely, you'll see that Gentoo has not actually done anything for about two years now.

Repeat after me: COMPIZ!!

Nuff sed.

-- Nooo I am NOT trying to get into the good graces of /. management =:]

Well, guys, I have a solution.

[typicallyterrific]

I switched away from Gentoo maybe two months ago, to Ubuntu, after two years of Gentoo.
I was goddamn tired of updating over night, rebooting in the morning and finding that maybe a half dozen things broke between stable tree updates.

Please, for the rest of us, just fork already. It'll save everyone a lot of headaches.
Portage is too good of a system to be allowed to be made irrelevant.

Re:Well, guys, I have a solution.

[Stalyn]

did you do a revdep-rebuild? It's not really Gentoo's fault if after you recompile libraries their dependencies start to fail.

Re:Well, guys, I have a solution.

[Anonymous Coward]

did you do a revdep-rebuild? It's not really Gentoo's fault if after you recompile libraries their dependencies start to fail.

WHY I DON'T USE GENTOO -- reason number #99813

Re:Well, guys, I have a solution.

[Stalyn]

more like...

WHY I CAN'T USE GENTOO -- reason #662 -- I don't know what libtool is.

Gentoo never portrayed itself as a noob distro. If you are willing to learn some things then try out Gentoo. If not stick with Ubuntu or whatever.

Re:Well, guys, I have a solution.

[Anonymous Coward]

WHY I CAN'T USE GENTOO -- reason #662 -- I don't know what libtool is.

Reason #1 -- l33ter than thou users, who don't realise how clueless they really are.

Re:Well, guys, I have a solution.

[Stalyn]

Yes, I use Gentoo. So? I don't see any problem. I embraced my Gentoo distro long ago and I am happy together with my workstation (that is a stage 1 box!). I have a fucking lot of ebuilds in and outside of portage and my CFLAGS are pretty optimized and solid.

But thanks anyway asshole. Go and use your stupid Debian while I EMERGE new ebuilds.

Re:Well, guys, I have a solution.

[Anonymous Coward]

I embraced my Gentoo distro long ago and I am happy together with my workstation (that is a stage 1 box!).

Stage 1 eh... woah. Is that like category A or HIGH LEVEL... or other such impressive sounding labels.

I have a fucking lot of ebuilds in and outside of portage and my CFLAGS are pretty optimized and solid.

Really... wow. Hey everybody... his CFLAGS are optimized and solid. I'm like, *so*, jealous right now.

But thanks anyway asshole. Go and use your stupid Debian while I EMERGE new ebuilds.

Debian? You think I use Debian (or a Debian derived distro)... how amusing. Debian is nearly as big a swamp as Gentoo. P.S. I bet you were wanking while you wrote that.

Re:Well, guys, I have a solution.

[Stalyn]

YHBT YHL HAND

Just in case you like Perl too...

Perl is like being molested by your uncle. There's something off about him, but everyone regards him very highly, so you trust him, and then on a family camping trip out at Montauk Point he takes advantage of you. Years later, you accept and acknowledge what happened, but you still refuse to believe that he's scarred you, because that would put him in control, not you, and the last thing you want is a molester in control of your life -- but your denial doesn't make it the truth. You want to believe that deep down inside, Perl is a good person, and you see that Perl has very redeeming qualities, but you sit down to try and program Perl and all you can think of is that camel's hard, throbbing cock.

Re:Well, guys, I have a solution.

[Anonymous Coward]

Yes... I was thinking as I posted my anonymous comment that I might be being trolled. You fucking idiot.

Yeah.

[Bill, Shooter of Bul]

Very simular to my expereince even though it was 2004. I managed to persist in getting a working system, kde and everything. BUt it got to be too much portage was almost always broken for what I wanted. I had to google for missing packages to get stuff to install. Im using SUSE right now, but its giving me crap as well. I haven't had time to mess with it, I think I can fix it. If not then I'll give Debian another try. Suse was the only one I could get other than Gentoo to work on this box. Its not there fault really, there's some serious hardware problem with my DVD drive. I'll get a new one before trying another install.

Go back to being an organized Linux From Scratch

[Anonymous Coward]

Gentoo was easy enough when it was a shell install with a kick ass handbook. Let the GUI/framebuffer installers fork.

Rocks and Glass Houses

[Anonymous Coward]

Debian can talk shit about Gentoo's problems after they mend fences with Ubuntu, fix the administrative bottlenecks that made Sarge take so long from freeze to release, stop hemmhoraging developers to other projects with higher productivity and lower bullshit political overhead, and generally find more constructive things to do than engage in year-long license flamewars (like, maybe develop and maintain a Linux distro, or something). Until then, comments about getting one's own house in order are best kept on the debian-navel-gazing mailing list.

What was the disaster?

[gknoy]

(As the parent said, and needs to be modded up ...)

Not being familiar with it, what was disasterous about Project Sunrise? Is it a giant pile of multiple WTFs, like the "You aren't supposed to report bugs, you're supposed to FIX them" bit that I only just noticed?

Re:What was the disaster?

[Viduliya]

Exactly... I can not see what is disastrous about project Sunrise. Someone please clarify what the "Sunrise disaster" mean.

... and what is an "overlay"?

[gknoy]

I feel like a complete noob for asking this, but ... what's an "overlay", in the context that the Gentoo people are using it? I feel like i kindof get it, but an far from grokking it, let alone being able to discuss it intelligently. :)

Re:... and what is an "overlay"?

[Kaseijin]

I feel like a complete noob for asking this, but ... what's an "overlay", in the context that the Gentoo people are using it?

Portage stores package information in a directory tree updated via rsync, which overwrites local modifications. An overlay is a separate directory maintained by the administrator. This capability has been used to kludge a third-party repository system, since Portage lacks direct support.

Why thank you,

[Anonymous Coward]

I think I will.

meh

[Anonymous Coward]

Unless he's taking over the project, it won't do any good. As an ex-Gentoo dev, I can tell you that the project is poorly managed, and most of it is due to a few egomaniacs with way too much power on devrel (developer relations). Personality clashes led to some of the best devs, most notably Ciaran McCreesh, being kicked out or leaving in disgust.

I love Gentoo's approach to Linux, and I wish I could still contribute. But the project is overmanaged in all the wrong areas (trying to make all the developers play nice), and undermanaged in areas that need some whip-cracking, like junior devs announcing absurdly ambitious projects long before they're ready.

To quote ciaranm:

Which is why so little gets past the drawing board... Too many people sabotaging the planning and saying they're doing it "to prevent flamewars", and too many people not doing the planning properly at all.

They need Seeds for the Macs

[Nice2Cats]

What Gentoo should do is produce a standardized base version for the Mac Intel line -- MacBooks, MacPros, iMacs, MacMini. These machines have standardized hardware, and the iMac especially doesn't lend itself to fooling around with the internals. Buy a Mac, pop in the Gentoo CD, fool around with some of the minor points, and then you have an instant computer.

Re:They need Seeds for the Macs

[Anonymous Coward]

...or you get a Mac and just leave it as it is: then you have an instant computer.

I'm in

[Anonymous Coward]

given gentoo is GPL (no?), is it possible to bring create another distro with the current code base to make it more stable?

Indeed it is. Get someone good to lead it, and I'll gladly join up and start maintaining ebuilds again. ciaranm took a step in that direction with Paludis, but it's just a package manager. We need a whole new distro.

oh and...

[Anonymous Coward]

Half the beauty of Gentoo is that you don't need "experimental" and "stable" branches, you don't need to "port" to every different architecture. The ARCH flags and profiles take care of all this very simply when used properly. The problem is that more and more, they're not being used properly. Dependencies get screwed up and you can't even do a fresh install anymore without getting a portage error.

I really want to contribute to a Gentoo-like distro again, as long as it's being managed by people who know what they're doing and know that technical debate is critical. Somebody please start it, and leave Gentoo to the self-obsessed power-mongers and clueless newbies who should have never been allowed to become devs.

Comment removed

[account_deleted]

Comment removed based on user account deletion

Gentoo is for posers

[Anonymous Coward]

Weee, I can compile stuff!

I'm making it faster because my CPU optimizes code better than someone elses!

Re:Gentoo is for posers

[Anonymous Coward]

Weee, I can compile stuff! I'm making it faster because my CPU optimizes code better than someone elses!

very true.. they are the ricers of the linux distro world

Fixing another's work

[TheNinjaroach]

Unfortunately chances are that when you submit a bug, the person reviewing it isn't the person responsible for the bug. I find it to be a responsibility to fix bugs that I have created. If you can point one to me, then great you saved me some time finding it myself.

The real pain in my opinion is fixing the work of someone else. First of all, I didn't create the bug. Second and more importantly, I didn't create that piece of the software - I'm not familiar with it. If it's fairly complex then I must spend a lot of time adapting to their coding style, understanding how exactly they accomplish the task, identify where the bug comes in and THEN fix it. I'd feel better re-writing the whole piece but that isn't progress, I'll maybe save a few KB of source but I'm re-inventing that wheel.

MOD PARENT UP

[Anonymous Coward]

+1, camel cock.

I've reported problems to Gentoo

[Slashdot Parent]

I've reported gentoo bugs using their bug system. Nothing ever gets fixed, so I stopped reporting and came up with my own fix.

It's much nicer to report bugs to folks who genuinely want to receive the reports and want their packages to work.

You copied this from funroll-loops.org

[dildo]

Ha! I love it when people copy their comments from this site!

Grand archive of fools and their foolish comments on Gentoo...

Debian over ubuntu?

[Slashdot Parent]

I'm curious why you chose Debian over Ubuntu. The reason I ask, is that I was thinking of switching from Debian to Ubuntu, but if you had a bad experience with it, then I won't bother. :)

Re:Debian over ubuntu?

[B5_geek]

I have noticed some quirks with NFS and Debian/Ubuntu together. But mostly, I went (back) to Debian was to keep my system identical (gcc, libs, etc) to all my production servers. I don't need the newest java, or flash, or "game-of-the-week" so I don't care about Ubuntu as much.

I have setup/installed Ubuntu on about 17 other systems for family/friends and $ clients $. I stongly suggest it for everybody.

Ubuntu: Stable and Easy like Debian, just newer & shiny.

It's not OSS...

[SanityInAnarchy]

...It's everything that gets too big without scaling.

Look at the Portage tree. Once upon a time, I'm sure bandwidth was at a premium and it made sense to just use rsync. Now, I'm guessing that rsync is slowing things down more at both ends, and ultimately costing more money, than the Debian way of doing things -- just split it into 10-20 small downloads, and use HTTP.

But that's a technical thing, and it's been worked on. Regenerating metadata takes much less time now, for one thing.

What bugs me is, in the early days, I could send a mail to something like gentoo@gentoo.org and get a response, even before I tried out the distro. Later, when I tried it out, just about any question in irc.freenode.net#gentoo would either get me an immediate answer, or someone willing to spend 10-20 minutes on my own problem.

Debian, on the other hand, was downright unhelpful, down to their IRC channel. RTFM all around.

Nowadays, there's Ubuntu, which is both extremely helpful with support and extremely easy to figure out even without support. I had a boss teach me Debian, so learning curve isn't a huge deal anymore. And maybe it's just that I'm asking tougher questions now, but I frequently will ask something in #gentoo and get absolutely no response, ever. Worse, I've submitted a few bug reports, got a couple of users saying "me too", but the bug still hasn't been assigned after a couple weeks of just sitting there.

I used to think OSS was more responsive than corporations, and it used to be true, but I'm guessing Apple will respond to my bug report before Gentoo does.

The community can also be downright hostile now. Usually it's nice to newbies, but I remember an incident where I reported a bug in documentation -- there was a mention of Reiser4 as alpha, untested, unstable, unreleased, or some other such nonsense. I was suggesting that the wording be changed so it's at least factually accurate -- Reiser4 has actually been released, even if Hans Reiser is the first to admit that until it's hammered on by thousands of users, it probably won't be stable enough for production servers. The person handling the bug kept saying "That's your opinion," and closing the bug. I think it got pretty personal, but even the majority of Slashdot seems to be able to distinguish between facts and opinions. Whether you think Vista is even Alpha quality now is irrelevant, it would be innacurate to call it anything other than Release Candidate, except in jest, or when making it clear that it's your opinion that it's Alpha, and MS actually calls it RC.

I think we need a truly community distro. I'm trying to remember who gave me this idea, it was over IRC, but we need a distro and a package management system that reflects the structure of the Internet.

troll/flame some more

[Anonymous Coward]

You're a douche.

A little hint

[Cinquero]

What is Gentoo and what makes it different from other distros?

I think it is different because it uses portage.

They should factor it out and separate the portage effort from the Gentoo distro which includes forums, stage tarballs, ISOs, DVDs, LiveCDs and all sorts of derived projects.

For me, portage is not much more than a configuration option translater -- effectively implementing a common and standardized package configuration mechanism for all the packages in the portage tree. I don't see any use in tightly binding it to Gentoo Linux. It could also be used by all other distro manufacturers to build their binary packages.

-1 flamebait?

[spiritraveller]

christ almighty it was just a joke... maybe kind of a lame joke... but not flamebait.

jeez