Contributed by jose on from the strange-but-true dept.
Yes, it's true. I guess you just have to see it to believe it. Personally, I'm still very fond of the ports tree, and would rather put effort there, but hey, if you can't live without portage ...
(Comments are closed)
By janus () janus at errornet dot de on http://janus.errornet.de
Portage is the pure sh*t in my eyes... unstable, based on python, unstable, is touching config files without asking you, inconsistent, unstable and... i think it's unstable.
I abosolutely dislike this project. If there's any way to avoid this inconsistent stuff for *BSD, please tell me.
And for the proponents: you've be warned, you should use the good old, stable and consistent ports like you ever did.
Comments
By Anonymous Coward () on
(I'm not trolling, I don't think there are going to be any Linux zealots here to respond)
Comments
By Anonymous Coward () on
I still think the apt-get and dpkg tools in Debian are a better thing,.. no need to wait for hour long compiles.
And the ability to let is NOT touch any config files is a great thing in apt.
Comments
By Anonymous Coward () on
Comments
By janus () janus at errornet dot de on http://janus.errornet.de
For example setup your favorite pam-config and then re-emerge base-layout... hope you've got a backup.
Comments
By Anonymous Coward () on
By cycloon () on
Howerver, it definitely backs up your configs and even prints out infos about everything it does in this step.
By Anonymous Coward () on
Yeah, because I love having to install all of Gnome to get something simple like GAIM to work. Or vise-versa, not having Gnome support in GAIM, even if I have it installed.
Binary packages suck. Give me source packages any day, but without the hassle of figuring out every single damn option available, working out every last dependency of a dependancy, and without having to write a bunch of patches just because a package wants a config file to be in a different place.
> And the ability to let is NOT touch any config files is a great thing in apt.
I want the default configuration files installed. If it exists, gentoo won't touch it untill you give it a command to tell it what to do with them.
By not-a-zelot :p () on www.tuxslare.org
Well, actually, that's not true. It will save the new config files as ._something, and ask you later if you want to upgrade for the new ones, keep them as is, show you a diff between the files, or merge them in a interactivly manner.
Comments
By janus () janus at errornet dot de on http://janus.errornet.de
Comments
By Anonymous Coward () on
By Anonymous Coward () on
Check out the hours.
JIZ
I posted before then you did.
NAH NAH NAH NAH
Do grow up!
By rankor_industries () on
I like that its based on python. Python is just the glue that makes things work, so speed issues really do not apply. Python, by its design, tends towards more correct code and is generally IMHO easier to debug. That is just fine for me.
And your comments about the config files are incorrect. config file editing is done interactively and you can choose which line to keep, which to overwrite, etc..
Finally, if you don't like it, don't use it. Nobody is going to force you to. I probably wont use it as the ports tree works well enough for me. If someone finds it useful for them however, more power to them and I applaud the developers for increasing the amount of choice available.
By Anonymous Coward () on
Right? Wrong!
Choice = good. This gives a user choice in order to choose the road to Rome.
And all these efforts to get Debian base (APT/DPKG etc.) working on NetBSD, Amiga, Hurd are useless? Ok, sure.
Now go away and don't tell us what we should do, you are too archistic and your point should therefore be ignored since this new feature doesn't give you a disadvantage at all!
Comments
By Anonymous Coward () on
Comments
By Anonymous Coward () on
By Jeremiah Cornelius () half-trolling:jeremiah@nur.net on mailto:half-trolling:jeremiah@nur.net
Do you think that if people were interested in choice, that they'd be running BSD?
Security - yes
Consistancy - yes
Choice - er, well... That's why there are 3 free one's, and some you can even buy, right?
If I want choice, I run Debian or Gentoo.
Bonus joke for those who've been here:
Hey! I have an Idea! Let's remotely rev our farm of 30 OBSD boxes up two minor versions this weekend!
By Python () on http://forums.gentoo.org/viewtopic.php?t=75060
I suggest you either (1) discuss this with the Portage developers and/or Gentoo community in a fruity and constructive manner and that you *do* the needed research. My Portage on my Gentoo doesn't overwrite my config files at all. It never did either. Neither did apt/dpkg/aptitude/dselect on my Debian box. Or (2) you go port it to C or C++. Or (3) you STFU and leave me and other people alone with our Portage on our *BSD. It is MY box and MY choice and the license(s) and opensource structure allow me -and others- to run this software. It doesn't make much sense to comment it down as 'one of the worst ports', and for me it isn't gonna stop me from using it, although that's just my opinion.
By Anonymous Coward () on
Spoken by a true ignorant bastard! Way to go, you should submit this posting along with your Resume to Microsoft's FUD department.
By Anonymous Coward () on
By ?? () ?? on ??
I do not suppose you could autobuild a self booting distro cd using ports and it's tools.
With portage and its latest tool 'catalyst' it
will be easy to do it on Linux. I wonder what will you say about portage if the catalyst starts to work for *BSD also :>
I think: sit in the corner and listen, kid!
By Anonymous Coward () on
By Anonymous Coward () on
Sorry, but this really scares me. OpenBSD packages/ports are targeted for the following platforms: alpha, i386, m68k, powerpc, sparc, sparc64, and vax. (And probably more in the future). AFAIK, portage only supports i386, powerpc, sparc, and sparc64.
If you've never installed a *BSD before, you might find the partitioning/slicing to be a bit confusing. I wouldn't recommend using a production hard drive!
You dualboot production systems? Umm...
At this point you should have a modestly functional portage. Let's see if it works:
"Modestly functional" to me means that I can install most of my favorite applications. It doesn't mean applying Linux-specific patches to software for use on OpenBSD.
This seems entirely premature for an announcement or even a webpage. portage doesn't even work on darwin/MacOSX yet, so there's no idea what other kind of portability issues are lurking in the dark.
I think competition is healthy and I wish them the best of luck, but I just don't think this is going to be a viable alternative.
Comments
By RC () on
Mac OS? It has a hard time working on Linux as it is.
By AC () on
Wtf are you talking about? Portage is written in Python, so it should run on any platform that has a compatible version of a Python interpreter. I think you're confusing the platforms the Gentoo distribution chooses to support, with the actual potential support of Portage. You see, ebuilds (The install scripts) for each package needs to be written by people, and to support other architectures means that the other architectures need to be taken into account in each of the ebuilds. Gentoo simply chooses a subset of the potential number of architectures to support in their ebuilds. Nothing is stopping anyone from adding support for other architectures, and it's usually not even that hard. If you take a look at an ebuild script (they're written in Bash) it's pretty easy to figure out how they work just by eyeballing one. Theoretically, the limit for architecture support is whatever the kernel supports, with the added stipulation of course, that the package itself suppots such a target platform.
You dualboot production systems? Umm...
Who said anything about dualbooting? The author was merely giving advice that the conventions for partitioning a hard disk drive under *BSD are different than those for Linux, and that those who aren't experienced with *BSD should be wary of losing data.
"Modestly functional" to me means that I can install most of my favorite applications. It doesn't mean applying Linux-specific patches to software for use on OpenBSD.
Where does it say anything about Linux-specific patches? That step is obviously a bootstrapping stage intended to merely get Portage onto the system. The tarball includes python, bash, bzip2, rsync, and GNU versions of make, patch, sed, and tar (all of which have a "g" prepended so that they don't conflict with the BSD versions). It's obvious that these stand-alone versions of the utilities are meant for exclusive use by Portage, and aren't meant to be used by anything or anyone else. Nothing on your system is being "patched"
This seems entirely premature for an announcement or even a webpage. portage doesn't even work on darwin/MacOSX yet, so there's no idea what other kind of portability issues are lurking in the dark.
It's obvious that the author would like for this to get more testing; how else would you suggest he get more testers? And what, exactly, does OS X have to do with ANYTHING? If Portage runs on OS X without a hitch, does that mean it will run on *BSD perfectly as well? You seem to think that OS X has a lot more in common with the other BSDs than it does.
Comments
By Anonymous Coward () on
You realize that vax doesn't have shared libraries right? So I have to make another ebuild for every entry in portage because it can't cope with architectures which don't have shared libraries..?
Where does it say anything about Linux-specific patches? That step is obviously a bootstrapping stage intended to merely get Portage onto the system. The tarball includes python, bash, bzip2, rsync, and GNU versions of make, patch, sed, and tar (all of which have a "g" prepended so that they don't conflict with the BSD versions). It's obvious that these stand-alone versions of the utilities are meant for exclusive use by Portage, and aren't meant to be used by anything or anyone else. Nothing on your system is being "patched"
I'm talking about the ebuilds themselves. All the patches in the ebuilds are Linux specific. Portage was created with portability being an afterthought and it shows.
It's obvious that the author would like for this to get more testing; how else would you suggest he get more testers? And what, exactly, does OS X have to do with ANYTHING? If Portage runs on OS X without a hitch, does that mean it will run on *BSD perfectly as well? You seem to think that OS X has a lot more in common with the other BSDs than it does.
OSX has a lot to do with it. It is the second platform they targetted portage for, and portage is still useless on OSX. They should finish support for OSX first so they can discover and fix portability issues before branching out into other OSes IMHO. I know OSX is radically different from BSD because I run it. I'm suggesting they finish OSX so they can see the shortcomings of portage and fix and extend it, rather than leaving it half baked, and continue making other half baked ports to FreeBSD & OpenBSD.
Comments
By AC () on
First of all, I'd assume that the compiler/linker wouldn't (couldn't) build shared libraries for an architecture where shared libraries aren't supported in the first place. After all, if it could do that, then the support is already half there. Even if that were not the case, it's a simple matter to force static binaries to be built. This can be done by modifying the global compiler flags. In this case, add "-Wl,static" to the CFLAGS variable in the file /etc/make.conf. The contents of the CFLAGS variable are used as the default arguments to the compiler for every merge (which can be overrided if need be). The compiler, upon receiving the -Wl,static argument, will pass it onto the linker at the linking stage, which will then resolve library dependencies and produce static binaries. This is one small change in a single file that will result in changes across all ebuilds.
I'm talking about the ebuilds themselves. All the patches in the ebuilds are Linux specific. Portage was created with portability being an afterthought and it shows.
"All" is a pretty big word you're using there. Are you implying that the majority of programs need a third party patch to make them run on Linux? In fact, I'd say the majority of ebuilds don't apply any third party patches at all. And the majority of patches that are included are either bugfixes that haven't been merged into the main tree yet, or patches that modify the behaviour of the program in some trivial, non system-dependant way. In fact, I don't think I've ever seen a patch in any way, shape, or form, that's "Linux specific" in the way you put it. And with a properly written patch, it wouldn't matter whether or not it was applied to a non-Linux system anyway. The patch itself would simply say, "Wait, this isn't a Linux system, I should just not do anything."
They should finish support for OSX first so they can discover and fix portability issues before branching out into other OSes IMHO.
Well, this has been a classic discussion of ideologies for some time, in many other systems. Basically, the goal is to find flaws in Portage, and the question is , "How do we find bugs in Portage?" Your suggestion is that they do a "depth-first-search", whereby they complete the port to one platform before moving onto the next, the goal of the search being bugs. However, to your dismay, they seem to be taking on a breadth-first-search, where they port to many different platforms at once. Which is better is very dependent on the situation at hand, and of course each has its own advantages. Breadth-first-search, while having the effect of scattering developer resources, has the advantage that bugs discovered on one platform may not be discovered on the other until much further into its development. In fact, if memory serves me correctly, Theo DeRaadt once stated the same benefit in porting OpenSSL to other platforms besides OpenBSD. It forces them to discover erroneous assumptions made about the platform, that would have been exceedingly more difficult to discover, had the port not been done. NetBSD is another example, where porting the kernel to dozens of architectures has no doubt helped them to discover bugs that they didnt' even know existed in certain architectures. Basically,it's a philosophical debate which may be constrained by the practical consideration of the amount of developer resources available.
By Shingu () tshingujp@yahoo.co.jp on mailto:tshingujp@yahoo.co.jp
Comments
By Anonymous Coward () on
Comments
By janus () janus at errornet dot de on http://janus.errornet.de
Comments
By blkwolf () on http://blkwolf.dyn.ca
And while there has been the occasional albit rare minor lib issue or specific emerge issue for some app, I have had 0 downtime to any portage related issue on any of 5+ gentoo servers over the past year almost two.
I'm not going to say portage is perfect or anything but in my experience at least, I've found it to be far from the unstable mess you want to make it out to be.
I for one welcome the choice to be able to run portage on my other favorite OS.
By Bernard () on http://obsd.at
Comments
By OurPresentIsOurPiss () peasent@gmx.net on mailto:peasent@gmx.net
By Masta () on
Comments
By TalkForYourSelf () on
YOU do not need. YOU. It is YOU. And I do not need YOU to expres MY opinion which is _not_ YOUR opinion - far from it, actually.
While i'm at it: there's no point on flaming Linux or GNU/Linux. It doesn't make sense imo. You're babbling something about Linux and Gentoo. What that actually matter? We're only talking about a part of Gentoo: Portage.
Comments
By Anonymous Coward () on
Exactly the same thing is happening to Linux; it is a bunch of crud but marketing got a hold of it and now it is the bestest thing.
So Linux == Windows - (a few years)
Comments
By Anonymous Hero () on
Tell me, how can _that_ become the same bloat, moneydrive, exploiting warmachine like Windows is in just a few years?
Maybe a few distributions will head that way. We'll see. For now, the communities existing, the freedom as in beer of distributions, and the thing which drives it all: the freedom of speech, the free software, are speaking directly against your paranormal visions.
So, now... you just go use BSD. And have sex with Theo. Or whatever. Then we, the People, can use whatever we want, and can use and/or support 'Portage on *BSD', without YOUR authorian "no you don't, it sucks, it comes from Gentoo" or "it's Python" whinings? Boohoo! I couldn't care less.. such an authorian behaviour drives directly against the fundemantals of free software and free speech. If you don't like such i suggest you use something which more towards your own behaviour which is reflected by your true philosophy: perhaps something more propierty...8)
By Anonymous Coward () on
By Strog () on http://www.pkgsrc.org
Portage was a good idea but it seem to be pulled around a bit right now. They need to get/keep focus if they want to make something good of it. I'll stick with ports/pkgsrc for me but to each his own.
By Anonymous Coward () on
Comments
By Anonymous Coward () on
I take it you can quote 'the authors of Gentoo'?
You see, the author day i heard Theo babbling about a backdoor in OpenSSH. Believe me, this is the truth. I can't quote him, though.
..
Comments
By Anonymous Coward () on
side note; why waste development time when you find it inferior?
By Anonymous Coward () on
Before making such a sizeable accusation, see if you can grep the code for a backdoor. If it's there, submit it to independent security folks.
If it isn't (which I'm willing to bet dollars to donuts is the case), then why are you wasting our time with such drivel?
Comments
By Anonymous Coward () on
Ofcourse it's BS. It's an anology in practise.
Comments
By Anonymous Coward () on
But instead of being vague, you should probably make your point instead of leaving us to guess at your meaning.
By Anonymous Coward () on
March on OpenBSD project
By Anonymous Coward () on
By Anonymous Coward () on
Get over yourselves and be happy that we such a large selection of amazing software to choose from that is made freely available. And shame on the boys from gentoo for being similar in mindset to the trolls commenting here.
Comments
By Anonymous Coward () on
Show me more correctness in a operating system that has the same functionality than OpenBSD, and guess where I'll move to? Unfortunatley, I'll doubt you'll have one in mind, or find one to present for your argument.
Comments
By Anonymous Coward () on
Does Openbsd beat linux for security? You bet your ass it does. Does this mean linux or any other operating system doesn't have a place, or can't be useful? Not likely. That is as sadly narrow-minded as the crap Stallman is constantly spewing in the linux community. The source is open here, but I see some of the minds aren't.
Comments
By Anonymous Coward () on
Can you elobarate?
By Anonymous Coward () on
Ever wonder why that is the case? Ignorance will not allow you to spend time on anything that freely gives the results to the competitors, which the BSD license allows. Yet, those same companies will *use* BSD licensed software.
However, I agree with the quoted statement. Linux is *far* more appropriate for SMP; OpenBSD SMP is in its infancy. Now ask yourself, if OpenBSD SMP was mature enough, what would reason use? Right, correctness wins everytime. Some evidence: OpenSSH and pf. Both have replaced crap software. OpenSSH is everywhere, and pf has become something not to ignore. Given the complexity of OpenSSH and pf, I'll wager OpenBSD SMP will be in the same league.
"The League of Extraordinary Software!" :P
Comments
By Anonymous Coward () on
What is the reason that Openbsd SMP isn't at that level yet? Well it hasn't been a developer priority, which is absolutely fine. Openbsd isn't a good choice for an average user as a desktop, and I say "average" user before I get a dozen people saying it works fine for them. Some day it might be, and if so great!
The point which you have missed repeatedly is that damn sandbox is big enough for everyone to play in, and that includes linux, and all the bsds. However it seems some people would prefer to push the idea of using only one operating system, regardless of whether there are other options that might fit a particular situation better. That idea is as silly as someone building a house and insisting that everyone only use one particular tool. I don't know about you, but I wouldn't buy a house from someone who built it only using screwdrivers.
Is my point a little clearer? Or should I just toss you a Phillips and let you go on your merry way?
By Anonymous Coward () on
RTFC (read the fscking comment)
By Anonymous Coward () on
Comments
By Anonymous Coward () on
The point of popularity is exactly what it is; a lot of people have jumped on the Linux bandwagon. The popularity of OpenBSD is minute compared to that of the Linux community, yet you'll hear time and time again how Linux needs to take some pointers from the success of OpenBSD. That success is not being the most popular, it's the success of being correct and not deviating or compromising that goal.
By djm () on
More interesting is Openwall Linux's take on the ports-style "build from pristine source with patches" philosophy - they build the entire OS from a CVS tree of RPM spec files and patche (and have a great attitude to security).
Gentoo is derivative and caters to the 16-year old Slashdot crowd who just want compile everything with compiler optimisations so they can feel l33t. No thanks.
Comments
By Anonymous Coward () on
By Anonymous Coward () on
Comments
By Anonymous Coward () on
By Anonymous Coward () on
Comments
By Anonymous Coward () on
Comments
By Anonymous Coward () on
By Anonymous Coward () on
OpenBSD should be taken seriously based on technical merit, period. Not on what people think of the community. I'm not here to make anyone feel comfortable by entertaining them with my incredible charisma and character. :P People pay shrinks for that.
By Anonymous Coward () on
By Anonymous Coward () on
Seems to make more sense to piggypack off our current ports system than to introduce a whole new system.
By heywood () on
Comments
By Anonymous Coward () on
Never heard of that. It should be noticed that not all Linux distro's are the same, that not the whole community GMTA's, that not the mentality is the same at all, and so on.
So i'm looking forward to your examples and proof that portage is a 'dirty hack'.