The name XFree86 was derived from the X386 project which preceeded it - XFree sounds like a logical successor.
Its short enough that its config file can be named after the project -/etc/xfree.conf. Which sure beats capital X capital F eight six capital S lowercase u t u p dot cfg.
All your apps are in path all the time, so you can run or make a link to a program just by typing its name. If you're not an admin user, admin apps (`system binaries') won't appear in your path.
You can use multiple partitions for your install, but only one/, which must always contains the/bin and/sbin directories - is needed to recover your system in case of emergency.
/usr/local is the local admins playground. If you're doing the right thing and using packaged apps, the only thing here will be a few custom scripts.
Yeah,/opt sucks. Most of the FHS list seems to agree. Its there cause old proprietary Unix apps want it to be, and no other reason.
Linux has the LSB. But not all Unix OSs do. Many of the apps you've mentioned installed across a variety of different Unix OSs, and hence when building from unpackaged source (which should never be done, but anyway) might suggest a different location.
There's two main issue: lack of standards support (initscripts should always go in/etc/init.d, documentation in/usr/share/doc) and lack of comprehensiveness of those standards (eg, there's no standard for removable media, because Rusty Russel removed the agreed location from the FHS before publication on a personal whim).
Remember Linux/unix is a MULTIUSER system , its not single user like Windows.
People who haven't used Windows recently but feel free to demonstrate their lack of knowledge AGRESSIVELY IN ALL CAPS.
Windows has been multiuser for ages. It also does local mutiuser a lot easier than linux does (yes, even more intuitive as `ctrl alt f1, login, password, start --:1').
Alas, neither up2date or apt-get or yum can be used to install individual files downloaded to disk without having to make a local repository and updating its indexes.
As part of RHLP, redhat-config-packages will be updated to work with remote repositories, but it currently lacks a command line client so scripting's out.
Oh, and rpm itself has the --aid command which can install any required dependencies necessary from a local directory. But the documentation consists of searching on google for it - you'd never know if someone didn't tell you.
The resolver is a set of routines in the C library (resolve(3)) that provide access to the Internet Domain Name System.
In English, that would be `The Linux DNS client is called the resolver. This is its configuration file'
On a normally configured system, this file should not be necessary. The only name server to be queried will be on the local machine
That's right kids, all of your machines should be running local DNS servers. Ever get the feeling that this documentation is unmaintained? (and yeah, I'm doing something about it, but that's irrelevant to the fact its currently annoying).
Oh, and slave name servers saving zone files to tell expiry times in millions of seconds, when `1D' would suffice.
actually last i checked i was modded up, so somebody agrees with me.
But not because you've made the fantastic point of `actually, you're wrong'./. moderation upwards is certainly no indication of logic. People here are morel likely to support breaking IT policy as much as they're likely to complain about the `I Love You' `virus' under Windows then tell people Linux doesn't get `viruses'.
and if you dont care why did you reply ? Because you infuriate me.
you seem to have missed the 18,000 times i have said "its no longer under the radar"
We know you think you're right. We also don't care. Score -1 redundant.
you wouldnt have to worry about them installi...redhat would support any other issues you have if you bought a support contract
You seem to have missed the `under the radar' part of the topic. Users modifying their PCs in environments large enough for this kind of action is a dumb idea and likely a violation of their security policy. Whether that's cracking an admin account on their Windows box or wiping it to install Linux, its still a Very Bad Idea.
as for streamlined management well you could simply run a local up2date server with cronjobs as neccasary, and run ssh locally on the clients so that when (and this will be very rare) there is an issue you can just ssh into the box and fix it.
Again, under the radar. It takes time for people to learn how to run Linux.
you have a policy that states what IT supports (evolution, mozilla, gaim etc) and whenever somebody asks for help with something not supported you point and say "No".
Actually, there has been - an analyst leaked that the Read Copy Update code in the Linux kernel source is identical to that of Unix and there's a fairly good chance this is actually the case. GIYF.
Re:Non castrated RedHat...
on
Ark Linux
·
· Score: 3, Interesting
I can't believe peopel are still spinnign this shite. Care to provide some supporting arguments, or at least refute those in my sig?
It is, however, a standard and I think its also a defacto one (in that most Linux systems use it, and the market share of those that don't is fairly small no matter who you talk to - eg, Netcraft or IDC). Many of the freatures on the autopackage homepage already exist in RPM. Surely you'd be better off adding whatever features you find lacking in RPM (such as suggested / recommended dependencies) to the format rather than trying to uproot the existing standard packaging system (a very ambitious goal)?
Every package in the current Red Hat Linux is signed using GPG, and IIRC this has been the case for a few years now. Most other Linux distros also sign their packages.
Saying "All source should be Open" implies he means Open Source. He doesn't - he just means the source code should be available. See later in the original article when he tells readers to "Note that I am not advocating open source licensing for commercial software. ".
I think the poster you're responsing to was using proprietary in the true, dictionary definition - in context, Unix is based on open systems where APIs are well known and shared between vendors.
Technically, proprietary does not mean closed source. Closed source does.
I don't really care to much either way, but you're arguing with eachother using two different definitions of the term.
RPM is the standard Linux package file format. If your distro aims to be Linux Standards Base compliant, it must have a mechanism of installing such files.
Preferably a full RPM implementation, but systems like alien or even (I guess) rpm2cpio are acceptable.
Sounds like a good card. I'm assuming it has Linux drivers, otherwise it wont work on my system and I will have to buy a 4600.
No reason to fear from NVidia, they've produced Linux drivers for all their cards since befoe they were on shelves since the GF2Ultra, but does anyone have any info?
Also, many of these bugs are actually to get Mozilla to render a page "Correctly" when the page is written totally wrong, I.E. not W3.org valid, like slashdot.org, only worse.
A document viewer's primary aim is to view documents - to support web standards may also be an aim, but its not the primary one. Most documents aren't written in W3C HTML, but Mozilla should and must render these documents correctly to function as a practical web browser. If it fails to render non compliant data (i.e., most of thwe web) its a bug in that it prevents Mozilla from being used my most of its target audience.
And too bad it only takes 1 service pack: they're cumulative in nature. Install Win2k, and if your install media wasn't updated to SP3 already, apply SP3 yourself.
It's a valid point, but the cost in switching from one to the other is very, very small compared to the cost of an entire upheaval of the Debian project to convert it to RPM.
Its nice to see the discussion has moved from `Dpkg has advantages over RPM and RPM has none' to different arguments abotu the amoutn of effort required and the cleanliness of the RPM source code. Not that anyone has acknowledged I have a point on those matters, they've just changed topic for Mysterious Reasons.
Do you honestly think that the work of the Debian maintainers would be reduced if the world adopted RPM as the one and only package format?
Did I say the world? As I said, I'm talking about Linux distributions used in large scale environments. Hobbyists should use whatever OS they want. Businesses and Universities Government departments often have stricter needs.
Who says they're not? Have you ever actually used alien?
That's really fucking rude. Of course I have, otherwise I wouldn't be talking about it. Have you? If you have, then you'd know that, as I said, alien strips dependency information. Without dependencies there's not much management happening now is there?
Gentoo's package management system makes an interesting change from binary package-based systems like RPM and dpkg, and think Gentoo should be applauded for pushing Linux in a new direction. However, under your view of what you seem to think the LSB should become, they would be forever denied the opportunity to become LSB compliant due to an arbitrary constraint on internal implementation details, and that's just wrong.
Indeed, that's a useful idea. But why is it necessary to break a standard to do this? Why not implement a frontend installs software and resolves dependencies by downloading and rebuilding source packages rather than binaries? It wouldn't be particularly hard to do. Better yet, everyone can benefit, because its compatible with the existing system. Setting higher standards encourages this kind of behavior. Encouraging people to take advantage of a good idea without breaking an existing system is very, very right.
It's a standard than can be adopted by all of the Linux community, not just the subsection of it that happens to use RPM
Hehe. By most estimates, your `subsection' means `most installed systems'. Standards don't stifly innovation, but the standard generally does lag behind the bleedign edge. That's a good thing. Nobodies provided a reason why source based dependency resolution could work within RPM, nor suggests / recommends.
The LSB is designed to encourage innovation and diversity in distributions
No. Its primary goal is to standardize that platform.
and not to force a view that "any distro is LSB compliant as long as it looks and smells like Red Hat".
Did I say that? I don't think I did. Red Hat has very little to do with the LSB, mainly because RH's people were wary that their involvement and size in the early stages would discourage smaller distros.
Slashdot Mirror
Got a URL? I think you might be thinking of X386.
Mike
Yeah,
Linux has the LSB. But not all Unix OSs do. Many of the apps you've mentioned installed across a variety of different Unix OSs, and hence when building from unpackaged source (which should never be done, but anyway) might suggest a different location.
/etc/init.d, documentation in /usr/share/doc) and lack of comprehensiveness of those standards (eg, there's no standard for removable media, because Rusty Russel removed the agreed location from the FHS before publication on a personal whim).
There's two main issue: lack of standards support (initscripts should always go in
People who haven't used Windows recently but feel free to demonstrate their lack of knowledge AGRESSIVELY IN ALL CAPS.
Windows has been multiuser for ages. It also does local mutiuser a lot easier than linux does (yes, even more intuitive as `ctrl alt f1, login, password, start --
Alas, neither up2date or apt-get or yum can be used to install individual files downloaded to disk without having to make a local repository and updating its indexes.
As part of RHLP, redhat-config-packages will be updated to work with remote repositories, but it currently lacks a command line client so scripting's out.
Oh, and rpm itself has the --aid command which can install any required dependencies necessary from a local directory. But the documentation consists of searching on google for it - you'd never know if someone didn't tell you.
In English, that would be `The Linux DNS client is called the resolver. This is its configuration file'
That's right kids, all of your machines should be running local DNS servers. Ever get the feeling that this documentation is unmaintained? (and yeah, I'm doing something about it, but that's irrelevant to the fact its currently annoying).
Oh, and slave name servers saving zone files to tell expiry times in millions of seconds, when `1D' would suffice.
nice retort. I can see my arguments are well and truly refuted.
:D
don't call me names, you idiotic little bitch.
actually last i checked i was modded up, so somebody agrees with me.
But not because you've made the fantastic point of `actually, you're wrong'.
and if you dont care why did you reply ?
Because you infuriate me.
you seem to have missed the 18,000 times i have said "its no longer under the radar"
You've said that exactly 0 times in this thread.
actually your post is pretty much just FUD.
We know you think you're right. We also don't care. Score -1 redundant.
you wouldnt have to worry about them installi...redhat would support any other issues you have if you bought a support contract
You seem to have missed the `under the radar' part of the topic. Users modifying their PCs in environments large enough for this kind of action is a dumb idea and likely a violation of their security policy. Whether that's cracking an admin account on their Windows box or wiping it to install Linux, its still a Very Bad Idea.
as for streamlined management well you could simply run a local up2date server with cronjobs as neccasary, and run ssh locally on the clients so that when (and this will be very rare) there is an issue you can just ssh into the box and fix it.
Again, under the radar. It takes time for people to learn how to run Linux.
you have a policy that states what IT supports (evolution, mozilla, gaim etc) and whenever somebody asks for help with something not supported you point and say "No".
Indeed, how it should be done.
The point is the simple and ease of use
;^)
What could be simpler and easier than alt-f2, (username), enter, (password), enter, startx-space-dash-dash-space-colon-one?
Actually, there has been - an analyst leaked that the Read Copy Update code in the Linux kernel source is identical to that of Unix and there's a fairly good chance this is actually the case. GIYF.
I can't believe peopel are still spinnign this shite. Care to provide some supporting arguments, or at least refute those in my sig?
It is, however, a standard and I think its also a defacto one (in that most Linux systems use it, and the market share of those that don't is fairly small no matter who you talk to - eg, Netcraft or IDC). Many of the freatures on the autopackage homepage already exist in RPM. Surely you'd be better off adding whatever features you find lacking in RPM (such as suggested / recommended dependencies) to the format rather than trying to uproot the existing standard packaging system (a very ambitious goal)?
...as standard.
/usr/X11R6/lib/X11/fonts/TTF/luxi*
They're called Lucidux or Luxi, and they're installed into
Use them. They'll become the standard fonts for Linux web browsers once the rest of the distros get XFree 4.2
I'm having trouble resolving the last part of your post with your sig. Don't we already have a technical standard for Linux packaging?
Every package in the current Red Hat Linux is signed using GPG, and IIRC this has been the case for a few years now. Most other Linux distros also sign their packages.
Saying "All source should be Open" implies he means Open Source. He doesn't - he just means the source code should be available. See later in the original article when he tells readers to "Note that I am not advocating open source licensing for commercial software. ".
I think the poster you're responsing to was using proprietary in the true, dictionary definition - in context, Unix is based on open systems where APIs are well known and shared between vendors.
Technically, proprietary does not mean closed source. Closed source does.
I don't really care to much either way, but you're arguing with eachother using two different definitions of the term.
RPM is the standard Linux package file format. If your distro aims to be Linux Standards Base compliant, it must have a mechanism of installing such files.
Preferably a full RPM implementation, but systems like alien or even (I guess) rpm2cpio are acceptable.
Sounds like a good card. I'm assuming it has Linux drivers, otherwise it wont work on my system and I will have to buy a 4600.
No reason to fear from NVidia, they've produced Linux drivers for all their cards since befoe they were on shelves since the GF2Ultra, but does anyone have any info?
But there's NO EXCUSE not to use a Mac
Er, yes there is. Right now you can probably play more games udner Linux than MacOS.
Also, many of these bugs are actually to get Mozilla to render a page "Correctly" when the page is written totally wrong, I.E. not W3.org valid, like slashdot.org, only worse.
A document viewer's primary aim is to view documents - to support web standards may also be an aim, but its not the primary one. Most documents aren't written in W3C HTML, but Mozilla should and must render these documents correctly to function as a practical web browser. If it fails to render non compliant data (i.e., most of thwe web) its a bug in that it prevents Mozilla from being used my most of its target audience.
And too bad it only takes 1 service pack: they're cumulative in nature. Install Win2k, and if your install media wasn't updated to SP3 already, apply SP3 yourself.
It's a valid point, but the cost in switching from one to the other is very, very small compared to the cost of an entire upheaval of the Debian project to convert it to RPM.
Its nice to see the discussion has moved from `Dpkg has advantages over RPM and RPM has none' to different arguments abotu the amoutn of effort required and the cleanliness of the RPM source code. Not that anyone has acknowledged I have a point on those matters, they've just changed topic for Mysterious Reasons.
Do you honestly think that the work of the Debian maintainers would be reduced if the world adopted RPM as the one and only package format?
Did I say the world? As I said, I'm talking about Linux distributions used in large scale environments. Hobbyists should use whatever OS they want. Businesses and Universities Government departments often have stricter needs.
Who says they're not? Have you ever actually used alien?
That's really fucking rude. Of course I have, otherwise I wouldn't be talking about it. Have you? If you have, then you'd know that, as I said, alien strips dependency information. Without dependencies there's not much management happening now is there?
Gentoo's package management system makes an interesting change from binary package-based systems like RPM and dpkg, and think Gentoo should be applauded for pushing Linux in a new direction. However, under your view of what you seem to think the LSB should become, they would be forever denied the opportunity to become LSB compliant due to an arbitrary constraint on internal implementation details, and that's just wrong.
Indeed, that's a useful idea. But why is it necessary to break a standard to do this? Why not implement a frontend installs software and resolves dependencies by downloading and rebuilding source packages rather than binaries? It wouldn't be particularly hard to do. Better yet, everyone can benefit, because its compatible with the existing system. Setting higher standards encourages this kind of behavior. Encouraging people to take advantage of a good idea without breaking an existing system is very, very right.
It's a standard than can be adopted by all of the Linux community, not just the subsection of it that happens to use RPM
Hehe. By most estimates, your `subsection' means `most installed systems'. Standards don't stifly innovation, but the standard generally does lag behind the bleedign edge. That's a good thing. Nobodies provided a reason why source based dependency resolution could work within RPM, nor suggests / recommends.
The LSB is designed to encourage innovation and diversity in distributions
No. Its primary goal is to standardize that platform.
and not to force a view that "any distro is LSB compliant as long as it looks and smells like Red Hat".
Did I say that? I don't think I did. Red Hat has very little to do with the LSB, mainly because RH's people were wary that their involvement and size in the early stages would discourage smaller distros.