Gentoo/FreeBSD is a Unix-like operating system developed by Gentoo Linux developers in order to bring Gentoo Linux design principles such as Portage to the FreeBSD operating system. Gentoo/FreeBSD is part of the larger Gentoo/BSD project. Read the interview here.
Interview with Diego Petteno, Gentoo/*BSD Developer
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
What’s the point of this again? Yes, I read the linked interview, but I don’t see how replacing FreeBSD’s clean textfile init with Gentoo’s symlinked mess helps matters. Replacing ports with the extremely buggy clone that is portage doesn’t seem to do anything productive, either.
Then again, the interviewee admits that he’s a Gentoo fanatic who only began using FreeBSD to destroy it. No surprise there.
“replacing FreeBSD’s clean textfile init with Gentoo’s symlinked mess helps matters”.
*cough* fanaticism without facts sucks.
Gentoo doesn’t use symlinks for init scripts- you’re thinking of one of the other linux distros.
“Replacing ports with the extremely buggy clone that is portage doesn’t seem to do anything productive, either.”
Uh huh. So… python codebase instead of make hackery is viewed as worse, and pkg ports being bash rather then make is viewed as
A) a clone (wow)
B) somehow worse?
Yes, similar chunks (options vs IUSE, build from src), but that’s akin to stating linux is a clone of bsd (more so same common design pts- posix) because bsd came first and linux implementation benefited from being able to study an existing attempt at it to see what works/doesn’t work. Portage owes ports for the concept and base metadata; that said it’s hardly a clone considering the differences.
“Then again, the interviewee admits that he’s a Gentoo fanatic who only began using FreeBSD to destroy it. No surprise there.”
Fanaticism here isn’t from the gentoo side of things. Portage isn’t ports, and linux isn’t bsd. Different from what you know doesn’t equate to worse, just different (evaluating the differences however may mark it as worse
Edited 2006-04-12 04:45
*LOL*
How funny to see such a flaming post being modded up. Looks like *bSD and MS-zealots are on the street… Or just plain anti-linux zealots.
Considering how much *BSD has in common with Linux there shouldn’t be this hatred (mostly against linux rather than the other way).
Considering how much *BSD has in common with Linux there shouldn’t be this hatred (mostly against linux rather than the other way).
Most of this hatred you speak of comes from the total disregard of the BSD philosophy everytime a project like this comes up. It’s not Linux, but people keep trying to turn it into Gentoo, or Ubuntu, or some other crap and start weird dedicated projects to “improve on BSD” rather than just helping to improve FreeBSD itself. In doing so they tend to completely destroy every advantage of the base system.
I use PcBSD as an example. If you are too daft to read the documentation and then figure out how to use the base installer to get a system up and running, frankly you have no business in FreeBSD. Go use Fedora, ubuntu, Suse, or something else designed in such a way that new people are welcomed and can easily learn. I think I speak for most BSD people when I say we get very tired of know-nothing people asking us stupid questions that are clearly explained in the easily accessible documentation. It’s like when someone asks you to help them find something online and it ends up being the first search result on Google that has the answer they are looking for. You’ll only actually tell them so many times before you start sending them the http://justf–kinggoogleit.com/ link.
It gets to the point where people stop even answering questions in the mailing lists and just say “search the damn archive you lazy son of a bitch” because we’ve had to explain it so many times (and yet we’re described as friendly compared to the OpenBSD mailing lists). Ports vs. Portage is the same things. I’m about to end up doing it yet again. I’d say I personally have about 10 more times of having to explain this before I end up just saying “screw you you lazy know-nothing prick” like other people do, because every response I get will be yet another question that has already been answered 30 times.
It’s not Linux, but people keep trying to turn it into Gentoo, or Ubuntu, or some other crap and start weird dedicated projects to “improve on BSD” rather than just helping to improve FreeBSD itself. In doing so they tend to completely destroy every advantage of the base system.
Quick questions.
1) Would FBSD be interested in converting all of their ports over to ebuilds, since some folks find ebuilds simpler then the ports equiv? (shoe on the other foot, would gentoo be willing to convert over to ports because some find ports simpler then ebuilds?)
2) Would FBSD be interested in maintaining a repository of ports that work across *bsd (more then just fbsd), and linux?
Ironic thing is that folks could swap linux/bsd from your statement and complain about FBSD if they attempted what G/*BSD is attempting. See the catch-22?
I’d say I personally have about 10 more times of having to explain this before I end up just saying “screw you you lazy know-nothing prick” like other people do, because every response I get will be yet another question that has already been answered 30 times.
Frankly, you just ignore them. Either someone will be willing to respond and help them nicely, or someone will fire off telling them to get bent/die and/or rtfm.
Former (even if no one responds) requires less effort, latter is just firing off at some poor sod (some have it coming, but that doesn’t mean all do).
That said… that’s not a bsd vs linux thing, that’s a “what type of community do we want to build/maintain” thing. There are linux distros that fire off at the newbs just as quickly (maybe a function of age?)…
Personally I wonder what those folks are thinking. Yes the guy making the request needs to google, but that same pita may turn out to be someone who’s just hitting a brain fart, someone who if not berated for the sake of frustration might turn to be a useful user/dev contributer.
Berating the dude (ignoring whether or not being a jackass is valid) is dumb from the angle it drives off potential contributors. Granted, if they don’t listen after telling them repeatedly how to find out info (eg, they’re just being lazy), strong words are useful then, but as first choice? Just seems self-defeating.
Note this isn’t pie in the sky “love/hump your neighbor” rhetoric either, have wanted to backhand more then my share of new users, just have also seen cases where that same initial moron in my eyes turned out to not be a moron and made valuable contributions.
Edited 2006-04-12 11:29
Ironic thing is that folks could swap linux/bsd from your statement and complain about FBSD if they attempted what G/*BSD is attempting.
Not to be overly rude, but that’s a rather pointless assertion to make. It doesn’t really apply to anything.
As a side note: a lot of the people who work on ports also work on pkgsrc (a branch off of ports), which works across a great many more things than *BSD and Linux, so the answer to your second question would seem to be “yes, they would”. I think that, in itself, demonstrates that really ports is a better design by nature as it’s already been put to use on 12 unrelated systems, compared to the struggle they are going through to get Portage onto a second. That’s 12 compared to almost 2.
Now then… you have, it seems, missed the point. The Gentoo/*BSD teams are not bringing portage to another system, they are bringing the other systems to portage. They are not making portage such that it will work on *BSD, they are making BSD such that it will work with their pre-designed software packager. Not only that, but they are doing it to add features that are already there.
Now, that in itself is not generally a problem. In the linux camp it is done all the time. The problem is in the fact that *BSD is a system, not a kernel. Once you start pulling parts of the userland out and swapping them with whatever fits your fancy it’s not *BSD anymore.
Yes, what’s in a name and all that. However, they are fundamentally destroying the underlying design philosophy of the system in what they are doing. The systems are stable because of how they are built. The reasons behind their choosing *BSD is stated as being that they are stable. If you destroy the base design, the stability goes with it.
So here’s the thing. Linux vs. BSD *is* a community thing, it’s *only* a community thing. The basic designs underlying the systems stem from the differences in mindset of the people who use and develop them. The things people keep trying to do are contrary to the BSD mindset. Firstly, you should not have a packaging system that depends on anything other than the base system. Secondly, the packaging system should not break if one updates the package database (has been known to happen with year+ old installs of gentoo requiring hand updating of emerge tools). Third, the base system must be constant or you cannot control stability. Giving 4 different options of binutils, 3 of which are “just patched enough to work on FreeBSD” but no more is not offering stability. Did I mention the packaging system shouldn’t depend on any odd outside software yet? The simplicity in the design is the reason ports is easily ported.
As was stated in the interview, the original intent was to be able to swap out kernels and userlands independently of each other. That’s a linux thing, and outside of “seeing if it can be done” it really has no practical gain from *BSD. It also has absolutely nothing to add to the BSD camp. It can however add to the already existing confusion surrounding them and lead people to poor first impressions of other software systems. Again I say, they should either: Make ports better if they don’t like it, or not try to make other things into linux. If they make portage portable, well, that’s great. Opensolaris could really use a decent package management solution since blastwave and openpackage have crap in the way of available software. They are not doing that, however. They are not even attempting to do that. This becomes evident by the statement “it would be as hard to get ports running on Gentoo/FreeBSD as on Linux.” Ports is a pretty simple system, if you are changing the underlying features that much there is something wrong. That is what I have issue with.
Last side note: I don’t have problems with new users, everyone starts somewhere. I have issues with people falling out of place. If you are too lazy to look something up, don’t ask someone to do it for you. If you aren’t willing to look into something the way it is, don’t think it needs to be changed.
As a side note: a lot of the people who work on ports also work on pkgsrc (a branch off of ports), which works across a great many more things than *BSD and Linux, so the answer to your second question would seem to be “yes, they would”.
Getting the pkg manager to run on alt OS’s isn’t a huge thing (have seen the bsd specific patches for portage, actually fairly minor) that said modifying the pkg repository to work on said alt OS’s is the real work.
I think that, in itself, demonstrates that really ports is a better design by nature as it’s already been put to use on 12 unrelated systems, compared to the struggle they are going through to get Portage onto a second. That’s 12 compared to almost 2.
Running on an os doesn’t not mean having ports available for an OS however.
Like I said, porting the pkg manager across unixen is actually easy enough (osx being the usual bad apple for it)- pkg manager running on that OS isn’t all that useful (nor quite worth boasting about) if their aren’t pkg repos for the 12 os’s.
I’ve yet to see a pkg-src/ports repo that claims linux support- would definitely be interested if there were such a beast.
This becomes evident by the statement “it would be as hard to get ports running on Gentoo/FreeBSD as on Linux.” Ports is a pretty simple system, if you are changing the underlying features that much there is something wrong. That is what I have issue with.
Harping the same point again, but pkg manager running on OS != pkgs available for the OS, thus the simplicity/cleaness/whatever of the manager doesn’t matter all that much if there aren’t any pkgs for it to work with.
Response probably would be “well they should just work with ports pkgs instead” which… is dumb. FBSD would never convert to ebuilds if asked (they like their format), same goes for gentoo (each format/manager does have it’s own lil benefits/deficiences after all).
say we get very tired of know-nothing people asking us stupid questions that are clearly explained in the easily accessible documentation.
It gets to the point where people stop even answering questions in the mailing lists and just say “search the damn archive you lazy son of a bitch” because we’ve had to explain it so many times (and yet we’re described as friendly compared to the OpenBSD mailing lists). Ports vs. Portage is the same things. I’m about to end up doing it yet again. I’d say I personally have about 10 more times of having to explain this before I end up just saying “screw you you lazy know-nothing prick” like other people do, because every response I get will be yet another question that has already been answered 30 times.
You just gave a perfect example of why I use Gentoo. With people like you using BSD who the hell would be attracted to the platform? Now I rarely visit the Gentoo forums anymore because I am very comfortable with the system but when I did need to ask questions I was sure I would NEVER get an answer like yours.
You don’t like being told to do your own work? O.o
Also, that last bit was really intended for something longer I was working on at the time, which was why I included the “about to do it again” part.
And of course nobody words things quite so harshly in actual responses (except perhaps for Theo), in fact *BSD communities are generally spoken of as being far friendlier than Linux communities, which are usually described as “elitist”. Everyone thinks it, however, when they see someone being an obvious tick on the side of the community after people spend 100s of hours making easily accessible, easily readable documentation. It is not unreasonable to expect someone to pull their own weight.
And if you ask a question in the forums that is answered in 4000 other threads that are all indexed and archived for easy searching and nobody tells you to get bent then there’s something wrong with those people… or maybe they are just used to the laziness. Either way it’s not a good thing.
You don’t like being told to do your own work? O.o
That’s not it at all. I do my own work but even if I didn’t, ripping me a new one for asking a question is a little absurd. There is no harm in being polite. Just because you get annoyed at the same question doesn’t actually give you the right to be a prick. That’s your own attitude problem.
And of course nobody words things quite so harshly in actual responses (except perhaps for Theo), in fact *BSD communities are generally spoken of as being far friendlier than Linux communities, which are usually described as “elitist”. Everyone thinks it, however, when they see someone being an obvious tick on the side of the community after people spend 100s of hours making easily accessible, easily readable documentation. It is not unreasonable to expect someone to pull their own weight.
You are right. It is not unreasonable to expect someone to pull their own weight. It is also not unreasonble to expect respect when asking a simple question.
And if you ask a question in the forums that is answered in 4000 other threads that are all indexed and archived for easy searching and nobody tells you to get bent then there’s something wrong with those people… or maybe they are just used to the laziness. Either way it’s not a good thing.
So what you’re saying is that someone isn’t normal unless they have elitist, prick attitudes when responding to simple questions that have been answered before. I prefer the gentoo forums way which is a link to the manual and a polite suggestion to search the documentation for the answers.
I use PcBSD as an example. If you are too daft to read the documentation and then figure out how to use the base installer to get a system up and running, frankly you have no business in FreeBSD. Go use Fedora, ubuntu, Suse, or something else designed in such a way that new people are welcomed and can easily learn.
So you want people to feel unwelcome unless they’re willing to read a swathe of documentation first? You want FreeBSD to have less users, and thus less developpers? And finally, you want FreeBSD to be less usable unless you can memorise cryptic commands?
That explains a lot.
So you want people to feel unwelcome unless they’re willing to read a swathe of documentation first?
You’re clouding the issue in a huge way here. If the alternative to reading a ‘swathe’ of documentation is asking a ‘swathe’ of questions that were answered yesterday and every day for the past however long, then yes, there is a problem.
Poor documentation is one of the most annoying things about an operating system so when maintainers actually go to the trouble of writing good docs, they damn well want people to use them. Of course, people can’t be expected to understand that instantly, but explaining it is often much more productive in the long run than continuing to treat everyone with kid-gloves. Seeing people get flamed (and occasionally getting flamed myself) is what made me realize how much more efficient reading documentation can be, both for me and for the people that already know the answers.
I think that most of the people that you find ‘mean’ actually turn out to be quite pleasant and welcoming if you take the effort to do a few things to make their day a bit easier.
“So you want people to feel unwelcome unless they’re willing to read a swathe of documentation first?”
You’re clouding the issue in a huge way here. If the alternative to reading a ‘swathe’ of documentation is asking a ‘swathe’ of questions that were answered yesterday and every day for the past however long, then yes, there is a problem.
Actually, I was responding to GF’s assertion that PCBSD was a dumb idea because people who can’t install FreeBSD using the existing tools have no place using a BSD in the first place.
I’m not positioning ‘ask swathe of questions’ as preferable above ‘read swathe of documentation’, it’s the ‘easy-to-use install tools that take the work out of the user’s hands, not requiring advanced systems-administration knowledge’ I’m preferring.
I’m currently in the process of reading the DFlyBSD book, which is basically the FreeBSD book with minor edits. Assuming the FreeBSD install proces is comparable to the DFlyBSD install process, I’ve come to the decision that regular BSD maintenance/installing is not for me. However, I AM interested in a stable, reliable and mature FOSS OS with a plentyufull amount of usefull software, that’s not Linux. In other words: PCBSD. The original poster would apparently rather have me continue using Windows.
PS: what’s currently holding me back is uncertainty about wether/how I can use my wireless USB dongle under PCBSD; I can’t find the answer on the PCBSD site..
You know dude, you aren’t helping out our community by acting like this. I’ve had some easy ass question thrown at me in my time. But instead of acting like a dick, I just pointed them to the docs. Alot of newbies don’t know how great the handbook really is. We were all newbies once, and I highly doubt you never asked a stupid question on the mailing-lists. Instead of going off half-cocked, maybe you should think about what you’re going to say. That way you won’t make the rest of us FreeBSD users look bad. I don’t really care if gentoo wants to make gentoo/FreeBSD. It’s not going to hurt anything. Everybody should just calm down, ports isn’t better than portage, FreeBSD isn’t better than gentoo or vice versa. It’s all a matter of taste, I like ports and FreeBSD, but that’s my prerogative. The zealotry on both sides needs to stop because it doesn’t help anybody.
You just made my day much better
You know “dude”… well, I guess you don’t… but for all those who didn’t pick up on it yet, any comments made previously were in response to people being jerks to the community. I don’t expect people to have to read 400 pages of information to use a system, however the base FreeBSD handbook, all the way up to the parts on the kernel, was at last check like 20 pages. 20 Pages of triple spaced bold size 18 font. You can read it in less time than it takes to do a full install on an average computer system. O.o
That way you won’t make the rest of us FreeBSD users look bad.
Oh, I’m sorry, I didn’t realize that everyone here was so closed minded that they would assume my saying something meant everyone who happens to pick up the OS felt the same as I did. Way to overgeneralize you bigot.
When I said “I think I speak for us all when I say ‘spend the 30 seconds it takes to search the archive before you ask a stupid question'” I hardly think that was in any way injurious to the community as a whole. The rest of my calling people jerks or lazy ass zombies does not in any way reflect upon anyone else, I never made the claim that it was a commonly held belief. It is an unfortunate fact of life that 95% of people are in fact slugs that simply drain sustenance off of other people by asking the same stupid questions over and over again, most people, including myself, tend to be tolerant and point them to the archives, or, if it’s a short explanation, I explain what they wanted to know with a short note that they check the archives the next time. I only make a stink out of it here because it’s a public forum and maybe the retards that ask those questions will see it and think “hey, maybe I should check the archives so that guy that was yelling at everyone on OSNews doesn’t reply to me with something nasty and call my genealogy into question.”
Also, no, I have never in my life asked a stupid question on the mailing list that there was an answer for elsewhere, and certainly not in an archived text searchable list. If you did then you can join everyone else on my list of people I highly despise for being too lazy to press a search button. Not that you care.
A bigot, that’s me I’m intolerant of other peoples opinions and beliefs. Trust me when you start calling people who ask easy questions daft, know-nothings,and lazy SOB’s it does make us look bad. I do hope people on here are open minded enough not to listen to the ramblings of a zealot like yourself. But there are going to be people who aren’t. Showing a little class isn’t a bad thing. So you have never asked a stupid question huh? For this we have only your word. Anyways, I’m not going to continue to argue with you, it won’t do anybody any good.
Looks like *bSD and MS-zealots are on the street… Or just plain anti-linux zealots.
Or just those who don’t spend their lives sucking Stallinman’s cock.
Wooot!?
Eeewwww… don’t even bring that picture forth… eewww
I don’t want to know.
What’s the point of this again? Yes, I read the linked interview, but I don’t see how replacing FreeBSD’s clean textfile init with Gentoo’s symlinked mess helps matters. Replacing ports with the extremely buggy clone that is portage doesn’t seem to do anything productive, either.
Then again, the interviewee admits that he’s a Gentoo fanatic who only began using FreeBSD to destroy it. No surprise there.
I know this has been replied to already but your post is so misleading I had to reply. The Gentoo INIT is actually symlinked as opposed to what another poster said but not in any way like other distros. It is not a mess. All the INIT scripts reside in /etc/init.d/ and all their confs reside in /etc/conf.d/. All the scripts have dependencies built in so you never have to manually assign INIT scripts to any order at all. The only symlinking that is done is for the different runlevels, namely boot and defualt. They are located at /etc/runlevels/boot and /etc/runlevels/default. The INIT scripts are symlinked from /etc/init.d/. This system basically allows you to add and remove using rc-update without having to figure anything more out other than whether you want it to be a boot service or a default service.
As for portage, calling it a buggy clone is a huge injustice. It is not a clone at all. Portage is very different than ports and I would hardly call it buggy. I’ve had very little issues with portage itself and I’ve been on Gentoo for 3 years. It is also a much more capable and configurable package manager than ports. USE flags alone make it worth it to me. Who wants to configure a package everytime you install?
The Gentoo INIT is actually symlinked as opposed to what another poster said but not in any way like other distros. It is not a mess. All the INIT scripts reside in /etc/init.d/ and all their confs reside in /etc/conf.d/. All the scripts have dependencies built in so you never have to manually assign INIT scripts to any order at all. The only symlinking that is done is for the different runlevels, namely boot and defualt. They are located at /etc/runlevels/boot and /etc/runlevels/default. The INIT scripts are symlinked from /etc/init.d/.
That is indeed a mess.
As for portage, calling it a buggy clone is a huge injustice. It is not a clone at all.
Gentoo’s founder boasts that he ripped off ports for portage on Gentoo’s homepage.
USE flags alone make it worth it to me. Who wants to configure a package everytime you install?
FreeBSD’s ports system has had such flags for years — long before portage was even a gleam in Dan Robbins’ eye. The problem with port’s flags is that they aren’t enforced to the level that they are in portage. That is, one package maintainer decides to use “WITHOUT_X = yes” whereas another uses “WITH_X = no”, which leads to users listing dozens of redundant flags in the system-wide make.conf. If the Gentoo/FreeBSD team really wanted to make things better, they’d devote their resources to promoting the standardization of these flags in FreeBSD and across the open-source world.
Gentoo’s founder boasts that he ripped off ports for portage on Gentoo’s homepage.
That’s not true. What he does state, is that he was inspired by his experiences with ports. Therefore he created a solution for Gentoo which had some of the same functionality, but working better (according to his needs). So it’s not a clone. It’s a different solution to some of the same problems.
The Gentoo INIT is actually symlinked as opposed to what another poster said but not in any way like other distros. It is not a mess. All the INIT scripts reside in /etc/init.d/ and all their confs reside in /etc/conf.d/. All the scripts have dependencies built in so you never have to manually assign INIT scripts to any order at all. The only symlinking that is done is for the different runlevels, namely boot and defualt. They are located at /etc/runlevels/boot and /etc/runlevels/default. The INIT scripts are symlinked from /etc/init.d/.
That is indeed a mess.
I beg to differ. The symlinks are only used by the system so facilitate parsing with simple tools that just run all symlinks in a directory. Simple tools that are not only fast, but do one thing well.
As long as people use rc-update — Gentoo’s init manager — to configure stuff, its not a mess, and distro mantainers can push big changes and not break compatability (They can just modify rc-update to respect/migrate the old format)
Well, if it works fine, why is it even considered a mess?
Compared to a single text file with *_enable=YES lines to enable/disable daemons, and *_flags=whatever lines to configure them, it’s a huge mess.
The simple fact that you need to use a specialised program to “manage” the startup order should be the first clue that it is over-engineered.
You don’t need an application to manage it. You can do it manually if you’d rather do that. But why, when you have such tools available?
You’re missing his point.
If someone even thinks writing an app to manage the init order is a good idea then the init system is overengineered. That you can do it manually is moot and pointless. When we say the Linux init system is a mess we mean it is a mess compared to RCng in Free/NetBSD. RCng is so easy that you’d never even come up with the idea that an app to manage it would be neat. Runlevels just make me laugh every time I look at them (though technically you CAN have runlevels in FreeBSD but why, oh why, would you want to?
It depends whether you’re a power user or a newbie.
A newbie running FreeBSD (let’s say PC-BSD) would no doubt want an application to wrap around it, rather than having to delve down. Personally I agree that things could be much easier than the runlevel system – but it’s not messy. Just a bit more complicated. Whether or not it’s overengineered depends on what one wants to do with it. There are limitations in simplicity, even though KISS is generally the best principle.
That is indeed a mess.
Care to explain why or do you just want to make unsubstantiated claims for no reason other than to stoke the fire?
Gentoo’s founder boasts that he ripped off ports for portage on Gentoo’s homepage.
He does no such thing. Daniel Robbins explains that he liked ports and decideded to implement something like it. He did but it operates very differently than any of the BSDs.
FreeBSD’s ports system has had such flags for years — long before portage was even a gleam in Dan Robbins’ eye. The problem with port’s flags is that they aren’t enforced to the level that they are in portage. That is, one package maintainer decides to use “WITHOUT_X = yes” whereas another uses “WITH_X = no”, which leads to users listing dozens of redundant flags in the system-wide make.conf. If the Gentoo/FreeBSD team really wanted to make things better, they’d devote their resources to promoting the standardization of these flags in FreeBSD and across the open-source world.
I wouldn’t compare the two systems at all. USE flags on Gentoo are much more robust and flexible than BSD. Gentoo’s USE flags are very simple yet fine-grained. Does anyone actually use the mess that BSD’s make.conf is or do they configure at time of compile?
Maybe befaore talking about the port system you should learn to read the documentation, as this is explained in a number of articles on the net (if you’re not able to understand the official documentation correctly).
In FreeBSD nobody uses make.conf for ports, but pkgtools.conf, this means something like:
MAKE_ARGS = {
‘multimedia/mplayer-*’ => ‘WITH_GUI=1 WITH_FREETYPE=1′,
}
this means that you can configure make args on a per port basis.
make.conf is used for core options, so kernel and core subsystems.
Everything else is the same old Linux hype: a few ignorant informations make a brand new system “more robust and flexible” than one that’s been available and in production ’till 1998.
Regarding the other consideration on make flags I couldn’t agree more, but it’s not Gentoo or FreeBSD fault, developers should use the same flag for the same option.
this means that you can configure make args on a per port basis.
make.conf is used for core options, so kernel and core subsystems.
Everything else is the same old Linux hype: a few ignorant informations make a brand new system “more robust and flexible” than one that’s been available and in production ’till 1998.
Tell me how that is simpler than Gentoo? It’s not. Tell me how it is more flexible when it is not. Gentoo is obviously more flexible because their USE system is much more organized and universal. Maybe you should learn how portage works before making such claims. Gentoo’s USE system is much more robust. Portage is an integrated part of the operting system, ports is not. That means you have access to USE flags in many different ways. You can rebuild all packages that changed USE flags. You can see what flags have been changed everytime you emerge an application. You can query what USE flags were used to build an application. There are many more examples like this. ports is just a build system. Portage is an integrated part of the operating system.
Please, you have already proven to totally ignore the basis of the port system, every claim you make is so ignorant that the only possible answer is RTFM or pay me for a consult. And, as a matter of fact, a person who’s saying that a system written in python is “more integrated” than a system based on standard Unix software like make (1) and cvs is at a level of zealotry that the first “consult” he needs is a psychological one.
Please, you have already proven to totally ignore the basis of the port system, every claim you make is so ignorant that the only possible answer is RTFM or pay me for a consult. And, as a matter of fact, a person who’s saying that a system written in python is “more integrated” than a system based on standard Unix software like make (1) and cvs is at a level of zealotry that the first “consult” he needs is a psychological one.
Wow you are clueless. Somehow you want to believe that just because make and cvs have been around for a while that they become a part of the operating system while python cannot be. THAT is ignorant. Do you even know what constitutes an operating system? Gentoo is an operating system, Linux is not. Gentoo the operating system uses python. In fact portage and by extension python is Gentoo. This means that portage ties into everything in Gentoo and allows you to access packages and their information through portage. Portage’s abilities are leaps and bounds above make and cvs. It’s fine if you don’t want to believe it but that doesn’t make it any less true.
If it makes you feel better to change the definition of operating system go right ahead but I don’t think most people will agree with you.
Personally, I think that eMagius has a fair point. But, it’s interesting to see the strong feelings on both sides. That’s what OSnews articles are for, I guess.
The point, of course, is because it scratches some itches, and it’s fun. I personally prefer OpenBSD to FreeBSD, so I started by modifying portage to work on OpenBSD. Why? Because I like the OpenBSD userland, kernel, and libc, but I prefer Gentoo’s init system and package manager. (Incidentally, a seldom-mentioned advantage of portage is the ease of writing an ebuild, since the average ebuild is nearly a direct copy of the shell commands that one would use to manually install a package. Make files have a significantly harsher learning curve.)
Another point of this project is better, more portable software. We have a single portage tree that supports both Linux and *BSD (FreeBSD, initially). Any patches required are universally applied (on both Linux and *BSD systems), and we work with upstream to get the patches incorporated.
Nowhere here is there any attempt to destroy any of the *BSDs, although the hyperbole is certainly entertaining.
This would bring the best of both worlds into a single distribution: rock-solid stability of *BSD systems along with the superior convenience of Gentoo portage over FreeBSD ports. This is of course arguable, but I’ve tried both extensivelly in the past and stick with Gentoo mainly because of Portage. I’m definatelly gonna keep an eye on this project.
same thing here.
I’m back to gentoo because I highly prefer portage over ports. But still, I prefer BSD kernel for it’s better documentation and easy configuration. (I never suffer from instability with my linux system – but fBSD feel lighter though). A union of both world is my wish. But as shared before, still personal choice.
I think Ports suffer from the lazy unix(ibm?) saying : ” if it’s not broken, dont fix it ” (or something like that) – I prefer the german proverb : “Nothing is good but always better” Then, lets go, move and innovate – make my life easier so I can be productive on more serious things. With good developpement pratice and design, software can become more friendly but yet powerfull (and reliable).
Can you tell me portage advantages over ports? I really don’t know anything about portage (and Linux in general).
I’m still not understanding the fixation on how much better portage is, I’ve used both, quite honestly ports is faster and easier and the tools available make it almost on par with the feature of portage from what I’ve seen. Of course, I’m no portage fanatic here, so maybe I’m missing some nitty gritty features. Then again, I don’t care. Ports works just fine.
well from your point of view none ofcourse.
but apperently not all share that viewpoint or the project would not have been started.
personaly i prefer the gentoo init scripts and i like portage way better than ports.
im not saying that one is better than another but that
it is about taste would be kind of boring if evryone liked the same thing. stop beeing so narowminded
A lot of zealots know ports doesn’t have many (if any) advantages compared to portage, so they jump to arguments like “omg what if the database becomes corrupt”
Which is pretty much impossible since there is no real database (at least in the terms they’re thinking). They should consider taking a look in /var/db/pkg sometime.
You can fool portage into thinking a package is installed simply by creating a directory in the appropriate section. I’ve done it a couple of times when playing with builds I wasn’t supposed to
Edited 2006-04-12 07:38
…Why do some one want to remove FreeBSD’s clean design to use messy Linux code instead? It’s like using a Windows Server kernel with a Windows 98 file tree…
Please tell me.
Why do some one want to remove FreeBSD’s clean design to use messy Linux code instead?
Read the article, it’s effectively creating ports for bsd kernel/userland (pkg’ing them in the portage tree), and modifying ebuilds in the portage tree to behave properly with bsdisms (de-gnu-ification as the article states).
Ironic thing? For the most part, it *is* straight fbsd just using a different pkg manager. It’s not trying to mismash the fbsd kernel onto gnu glibc, nor linux kernel on libc; it’s adding ports to the tree for *bsd userland, getting their toolchain ported (and behaving), and modifying ebuilds that are linux specific to behave properly for linux and bsd.
“…Why do some one want to remove FreeBSD’s clean design to use messy Linux code instead?”
– Well for the FreeBSD’s clean design it seems that some BSD developpers would not agree with you.
Examples ? Sure !
Dragonfly, OpenBSD, NetBSD all forked from the orginal BSD branch… Because in a way or another they wanted different things to be added/removed. That’s the power of OSS ! You’re free to add/remove/modify whatever you want ! *BSD developpers know that since their BSD license is one of the less restrictive OSS licence.
IMHO Portage is very good tool (not perfect) and many unix-like OS can benefit from it so, from my point of view this Gentoo/*BSD thing is a good thing and I’ll look forward to this project.
“It’s like using a Windows Server kernel with a Windows 98 file tree…”
– Don’t be so rude with Windows XP ! 😉
(just a joke)
“Please tell me.”
– Hope I did, at least partially…
Dragonfly, OpenBSD, NetBSD all forked from the orginal BSD branch… Because in a way or another they wanted different things to be added/removed. That’s the power of OSS ! You’re free to add/remove/modify whatever you want ! *BSD developpers know that since their BSD license is one of the less restrictive OSS licence.
FreeBSD and NetBSD were developed at the same time by two independent groups, using the sources for 4BSD. They were not forks of the 4BSD code, but independent ports to the i386 architecture.
DragonFlyBSD forked from FreeBSD 4.x.
OpenBSD forked from NetBSD 1.something, I believe.
Nobody is putting linux code into FreeBSD. Nobody is ruining the clean design.
It’s merely another package manager, that’s all. It doesn’t change the underlying architecture as such.
Besides that, the design of FreeBSD isn’t all the clean compared with non-*nixes. You can always find something cleaner, something better, in some areas. This is true for all aspects of life.
Nobody is putting linux code into FreeBSD. Nobody is ruining the clean design.
It’s merely another package manager, that’s all. It doesn’t change the underlying architecture as such.
Have you read up on the project at all? Have you even read the interview? Gentoo/FreeBSD explicit purpose is to change the underlying architecture and bring in Linux-isms such as the mess of symlinks. This isn’t a project like the NetBSD-sponsered pkgsrc, where pkgsrc was made as portable as possible so that it could run well on Debian, Windows, Solaris, etc. — it’s a project to gut FreeBSD and mangle its corpse to fit around Gentoo.
From Gentoo’s own FAQ ( http://www.gentoo.org/proj/en/gentoo-alt/bsd/fbsd/index.xml ):
1. Is Gentoo/FreeBSD just a portage on top of FreeBSD system?
No, Gentoo/FreeBSD aims to provide a complete FreeBSD-based system using Gentoo design principles. This means that it’s going to use the Gentoo init system, administration utilities and toolchain support.
I read the article yes. And I’ve read the documents on gentoo.org.
And no. There is not going to be linux code in Gentoo/*BSD.
There is however going to be Gentoo-specific code, but this is not related to Linux, nor to GNU/Linux.
I’m curious about the differences between ports and portage. Why does ports seem to be more up to date than portage. For instance, unless I’m misreading something on the Freebsd and Gentoo websites, FreeBSD is using KDE 3.5.2 while Gentoo is using KDE 3.2. I’m not a big authority on either distro. I’ve done a total of one Gentoo install on my PC and one install of NetBSD nonroot on Solaris 8. I found NetBSD to be a bit more straight forward, but the comparison isn’t entirely fair since “Gentoo for nonroot Solaris 8” isn’t available and I didn’t install a PC.
Please no flamewars. This is a legitimate question on my BSD and Gentoo make different choices?
I don’t know really about ports/FreeBSD but on Gentoo website you’ll find that the last stable relase of KDE is 3.4.3 and version 3.5.2 is available but still in testing phase.
Take a look there:
http://packages.gentoo.org/packages/?category=kde-base;name=;offset…
Yeah, it’s a pity that Portage’s stable packages are lagging behind more and more these days. I use both Gentoo Linux and FreeBSD on different servers, and it always bites me that Gentoo still has lighttpd 1.3, MySQL 4.1, PostgreSQL 8.0 and Rails 1.0 marked stable. The popular FreeBSD ports are always up-to-date and don’t seem to be bugged by the compatibility issues.
Back in the Gentoo 1.4 days things were always up to speed and that’s what got me excited about the distro in the first place. It’s a pity that Gentoo lost that agility.
I think the problem is that they lost focus. Too many script kiddies jumped on board and not enough mature focus driven developers manage it these days. I may be wrong but I think if they simply sit down and develop a structured process for Quality Assurance and accountability they would be back to the direction they start off heading in. Cutting edge without the bloat.
Honestly there’s not a lot of difference in the underlying function. People like to claim portage as being something other than a clone, it’s not really. Yes, they changed the scripting, in some ways it’s better in some ways it’s worse, but in function the two systems have more or less the same functionality available, you just have to go about any specific task in a slightly different way.
Make.conf is more or less identical in the two in function, gentoo took the FreeBSD WITH_X_ORG use values and put them into the make.conf file, but other than that all configuration there is the same.
Portage seems to allow easier installed package management, FreeBSD supplies the pkg-upgrade utility to do those tasks, usually it works, once in awhile it will get confused and core-dump (happened to me when they first switched to xorg)
I believe (I haven’t really checked recently, but it used to be true) that portage tends to have more up-to-date binary packages available. That was the original intent when they announced they would be adding the remote-packaging thing anyway. The FreeBSD people will constantly update the binary packages, but the new builds only work on FreeBSD-Current, so everyday users don’t really benefit from it, but they make binary builds everytime a new release comes along, so it’s never too far between.
As to any differences in how up-to-date things are, it comes down to three things. The number of people assigned to any project, the amount of code you have to change, and whether it passes the testing phase.
You can currently get KDE 3.5.2 on Gentoo also, it’s just not “stable” yet so it probably doesn’t show up in the main list.
Portage does have the added feature of keeping outdated libraries around if you want for things with odd dependencies, which is good and bad. The Ports people tend to fix such cases by doing things like leaving gettext at an old version for a year until all the software in the ports tree supports the new one, but all in all that works out too.
Who brought that guy out of the pits of hell? The number of people who have no tolerance of patience for *normal* users just astounds me.
I tried PC BSD and I thought they actually had some great ideas. I don’t knock anyone who’s giving a new idea a red hot go. The distro’s not as polished as Suse but then it’s nearly as old either. The PBI application installation format is a breath of fresh air. Why attack it? If you don’t like it, don’t use it.
It seems that Portage tends to be a unix-like SAL which is a good thing because it could be a very good way to unify the installation process of OSS on several differents OS such as Linux, OS X, xBSD among others.
I don’t see why some people get scared of Portage since it’s a powerfull and flexible tool. It could used aside of the original package manager used on your distro/OS.
That was just a though…
Select kernel (freeBSD, openBSD, netBSD, linux)? _
Select bootloader (lilo, grub)? _
Select installation drive (/dev/hda, /dev/sda1, etc)? _
Select cron (vixie-cron, fcron, dcron, anacron)? _
Select shell (bash, zsh, ash)? _
Select X server (none, X11, etc)? _
Select desktop (enlightenment, KDE, Gnome)? _
Installing/emerging, please wait………..
Your boring Unix clone is installed. Press enter to reboot. Reminder the OS installed was ???? – please write this down in a safe place (you probably won’t be able to tell after you reboot).
perhaps with “uname” ?
Sounds bland, but bland is good.
The underlying system doesn’t really matter. What does matter is that it works properly, and can perform the tasks you want it to do.
But yes, it’ll be veeeerrrry boring. But luckily we’re not forced to use Gentoo/GNU-Linux or Gentoo/*BSD. We have the freedom to choose, so it’ll only get boooooring if you decide to go that way
Yeah I know – they are all good and will get even better..
I just wish people were more interested in *alternative* OSs, rather than having thousands of developers working on “reinventions” of the same tired old Unix design over and over and over…
It’s like an ice cream shop where you can get chocolate or vanilla (or plain, or a vanilla and plain hybrid, or cream, or creamy vanilla, or creamy plain, or a creamy vanilla and plain hybrid, or milk, or any of the wide variety of milk and vanilla/plain/cream combinations), and perhaps for something different, a “chocolate vanilla” (which won’t work in industry standard ice cream cones).
I want some turkish delight, some honeycomb or some mint flavours, but the few people who are interested in doing something new have too few resources to acheive anything, possibly because of repetitive plain/vanilla/cream/milk flavours sucking all the life out of the ice cream industry…
You might want to take a look at Syllable, Haiku or Zeta, SkyOS, and OS/2 (eComStation).
I’m a FreeBSD user of 9 years and a Gentoo user of one. I much prefer FreeBSD but I believe the portage system gets the nod. Try doing a portupgrade of Gnome on FreeBSD followed by the same on Gentoo then get back to me. The ports system will be interrupted countless times with steps that have to be done manually ad infinitum.
That’s odd… I just did a ‘portupgrade -r -a’ on three separate machines. It upgrade hundreds of ports, including gnome, and the only interruption ws on one machine (my laptop) because I had unplugged it from the network and it couldn’t download the distfiles necessary.
No, it isn’t. If it where there wouldn’t be a separate script for upgrading Gnome at http://www.freebsd.org/gnome/
You obviously did a minor upgrade. Try going from 2.12 to 2.14 and see what happens.
Nothing, as I run the separate script for upgrading it.
It does work, it’s documented, what’s wrong with using a separate script 2/3 times in a year for upgrading Gnome ?
Well if other packaging system doesn’t need the separate script then they’re better..
This project seems fine. I personally don’t mind when I see projects like this take place because they either take off or they don’t. No real harm is done.
I say use what tool is best for the job. I find the *BSD’s to be a lot cleaner than most (not all GNU/Linux distro’s are considered here) GNU/Linux Distro’s. I also find the documentation a lot better under the *BSD world… maybe not exactly better but more centralized and organized. There is a lot of great documentation out there for the GNU/Linux’s as well. Gentoo is a great example in this regard.
AS for the whole Portage vs Ports debate. Well I personally like bits and pieces of both and wish I had some aspects of apt-get mixed in there with them as well.
One thing I like about Ports compared to Portage is that it’s a lot easier to get rid (deinstall/uninstall same thing…) of orphaned packages, that is packages not being used by anything anymore and shouldn’t be installed. This is not to say you can’t do this exact same thing in the Gentoo/Portage world, I just find it a lot easier under FreeBSD. For example:
I can install a program called pkg_cutleaves under FreeBSD and it will run through the packages that aren’t directly linked to being used by other programs (dependancies) and it will remove them after I have gone through the list. There are a few other programs to do the same thing, but I simply like this program best. Also recrusive dependency checking is built-in to the Ports system by default. So it wont let me accidently uninstall a program that will/might break another package that relies on it.
Under Gentoo I have to make sure everything I have installed was compiled with my specified set of USE flags to have a small script called depclean work to its full potential. What this means is, depclean will do just what pkg_cutleaves did for me in FreeBSD but with a little more hassle as it will be reliant on how everything was compiled in the first place (which USE flags where used) This is a none issue anyway, as I am already use to waiting for things to compile, hence my reason for using Gentoo and FreeBSD in the first place.
Gentoo also has another cool program call revdep-rebuild which is great for using after a depclean process as it will check for missing dependencies and make sure that all of your installed apps will be fully functional after a system cleaning.
So you see the functionality between the two Ports and Portage, is very much the same with each having there ups and downs. I personally miss some aspects of how easy Debian made updating packages… but thats a different story.
One more point. I love that FreeBSD is a FULL OS, though this isn’t in my opinion that big of a difference. The only thing here is that the userland and the kernel are all developed by the same people which keeps everything very tightly integrated. Something lacking in many distro’s with the GNU tools, though you basically get the same functionality, FreeBSD or GNU/Linux. The main difference is that the base system is _NEVER_ touched by the Ports system. This is a plus and a minus all at the same time. It makes it easier and safer to mess with apps installed from Ports as you will always have a bootable OS no matter what messes up. In Gentoo (GNU/Linux world basically) things are a lot more modular and each app in the system is treated separately. This makes it easier to patch security fixes as you only need to recompile that one small app and not the while base system like in FreeBSD… though FreeBSD has other methods as well, such as binary security updates which ease this very much.
Ok now that I’ve rambled on forever… I just don’t see why people have to worry so much about their preferred OS. No one is making you change it, stay with it and it will not be affected by projects like this. It shouldn’t matter to you how some particular developers out there spend there time and interest. Just because it’s not spent on something you feel will be beneficial to your OS of choice doesn’t mean you should trash it.
I think its great to see that FreeBSD and other *BSD’s are getting all of this attention, beneficial or not, because there is no such thing as bad publicity right?
Just my 2cents (Canadian)