Unfortunately this won't work. The way Hyperion, Freenode's IRCD, is designed, server passwords not used as such get passed directly on to whoever happens to be using the nickname defined in the config as the 'identify service'. In Freenode's case, this just causes a PRIVMSG to be sent from your nick to NickServ, whichever server he happens to be using, with the identify command and password. It's no harder to hijack than a regular/msg. The same goes for the 'raw' nickserv commands, which are similarly translated to PRIVMSG.
This is compounded by the fact that due to the way Hyperion's server-hide works, it is in theory impossible for normal users to know which server another client is using, so '/msg NickServ@services.' doesn't work either.
By far RedHat's biggest failing IMO is the lack of support for ReiserFS - JFS and XFS would be nice for others, but the former is all I really care for. I like having a filesystem that genuinely allows for atomic disk transactions without any noticeable performance hit. But as has already been stated, RedHat aren't interested in supporting it. It's a real shame, but something we have to live with for now.
And they're not supporting it for good reason-- its extended attribute implementation is horrifically broken, and so it won't even mount on an SELinux system. IMHO (and a great many people share the same view), the increased security from SELinux is more important than the slight speed gain, especially at the expense of much higher CPU usage.
If you look at the official developer list there seem to have been about 100+ developers ever. But so many people who aren't developers work on it.
Probably a better estimate is the number of user accounts on dev.gentoo.org, which is currently 374. Once you discount system accounts and retired devs, that's still in excess of 300 AFAICT.
Last I knew, Vida used a hacked up version of Anaconda to install the gentoo base system. The Gentoo installer will install with portage. That's quite a difference.
Disclaimer: I haven't used either, so I might be slightly inaccurate on the details.
What's planned for 2005.0 is an experimental test release of the new graphical installer. It'll be there for people who want to test it, but don't go relying on the thing.
I think I know how to break a Gentoo system, thanks. One of those ways is setting USE / ACCEPT_KEYWORDS on the commandline. Another is being careless with unmerge / depclean. Another is being overly reckless with per-package USE flags.
I'm not talking about USE=mysql bringing in mysql as a dependency. I'm talking about the xorg-x11 ebuild building the Xprint server code only if xprint is in USE. The behaviour of xprint has changed slightly recently, so it might have been fixed now, but building a client app with USE=xprint and the server without was easy breakage.
I know better than most how portage works, and it's quite possible to break things with per-package USE, at least until the new dep resolver comes around with the ability to depend on a use flag being enabled in a particular package.
USE flags are very powerful - if and only if - you set them on a per package basis.
However, there are a couple that are only safe to use if you set them system-wide (Xprint was one, until its behaviour got changed to work more nicely with Firefox binaries). What happens when a library doesn't get built because a USE flag isn't enabled in one package, but you then enable the same USE flag in another package, which causes it to link to said non-existant library? I'd say that setting them per-package is more likely to make you forget/not realise something, and screw up.
What you see and what you get is a bash prompt on the cd boot. It is worse than slackware on the install
I'd have said "better than slackware on the install". I realise I'm in the minority here, but I like a simple, text-based install process. After Gentoo's, my favourite installers (of those I've used) are Debian's old-style one and [Free|Open]BSD. They do what they need to, and don't get in the way.
Oh, and the other major plus with Gentoo's install: you don't need an install CD to do it; any Linux system will do. I haven't yet found another distro or OS that lets you do that.
Even here, the situation is improving, from a Linux point of view. Look at UT2004 and Doom 3. Two of the big releases this year, and one has Linux binaries (and even an install script) on the same CDs as the windows version, and the other is promising Linux binaries to download. Linux gaming is going mainstream. Slowly, but it's happening.
Yup, and they didn't even write that. They bought it (insanely cheaply, IIRC) and marketed it. There's one thing they have figured out and are extremely good at, and that's marketing.
If you're a software developer, you can do a lot worse than sourceforge. The shell server doesn't seem to have a compiler, but you can opt-in for Compile Farm access, which should more than satisfy your compilation needs. Pine isn't there, but Mutt is, and i haven't checked IRC.
Slashdot Mirror
Unfortunately this won't work. The way Hyperion, Freenode's IRCD, is designed, server passwords not used as such get passed directly on to whoever happens to be using the nickname defined in the config as the 'identify service'. In Freenode's case, this just causes a PRIVMSG to be sent from your nick to NickServ, whichever server he happens to be using, with the identify command and password. It's no harder to hijack than a regular /msg. The same goes for the 'raw' nickserv commands, which are similarly translated to PRIVMSG.
This is compounded by the fact that due to the way Hyperion's server-hide works, it is in theory impossible for normal users to know which server another client is using, so '/msg NickServ@services.' doesn't work either.
So basically what you're saying is that Gentoo sucks when you don't know how to use it. Big surprise there.
I hear you learn lots about Linux that way.
No, you learn lots about following instructions that way. Of course, it's still useful knowledge that a lot of people could benefit from gaining.
By far RedHat's biggest failing IMO is the lack of support for ReiserFS - JFS and XFS would be nice for others, but the former is all I really care for. I like having a filesystem that genuinely allows for atomic disk transactions without any noticeable performance hit. But as has already been stated, RedHat aren't interested in supporting it. It's a real shame, but something we have to live with for now.
And they're not supporting it for good reason-- its extended attribute implementation is horrifically broken, and so it won't even mount on an SELinux system. IMHO (and a great many people share the same view), the increased security from SELinux is more important than the slight speed gain, especially at the expense of much higher CPU usage.
If you look at the official developer list there seem to have been about 100+ developers ever. But so many people who aren't developers work on it.
Probably a better estimate is the number of user accounts on dev.gentoo.org, which is currently 374. Once you discount system accounts and retired devs, that's still in excess of 300 AFAICT.
-1 Missed the point.
OTOH, the comment in question was a fairly useless one that added nothing to the discussion... Not to say that this one is any different, of course.
so I could still idle on IRC and such
You know, the 2004.3 livecds (or at least the x86 and amd64 ones) do include irssi for that...
Last I knew, Vida used a hacked up version of Anaconda to install the gentoo base system. The Gentoo installer will install with portage. That's quite a difference.
Disclaimer: I haven't used either, so I might be slightly inaccurate on the details.
What's planned for 2005.0 is an experimental test release of the new graphical installer. It'll be there for people who want to test it, but don't go relying on the thing.
It's funny as long as you're not paying for it. A lot of us are.
I think I know how to break a Gentoo system, thanks. One of those ways is setting USE / ACCEPT_KEYWORDS on the commandline. Another is being careless with unmerge / depclean. Another is being overly reckless with per-package USE flags.
I'm not talking about USE=mysql bringing in mysql as a dependency. I'm talking about the xorg-x11 ebuild building the Xprint server code only if xprint is in USE. The behaviour of xprint has changed slightly recently, so it might have been fixed now, but building a client app with USE=xprint and the server without was easy breakage.
I know better than most how portage works, and it's quite possible to break things with per-package USE, at least until the new dep resolver comes around with the ability to depend on a use flag being enabled in a particular package.
USE flags are very powerful - if and only if - you set them on a per package basis.
However, there are a couple that are only safe to use if you set them system-wide (Xprint was one, until its behaviour got changed to work more nicely with Firefox binaries). What happens when a library doesn't get built because a USE flag isn't enabled in one package, but you then enable the same USE flag in another package, which causes it to link to said non-existant library? I'd say that setting them per-package is more likely to make you forget/not realise something, and screw up.
I'm still waiting for the Hello Cthulhu MMORPG.
What you see and what you get is a bash prompt on the cd boot. It is worse than slackware on the install
I'd have said "better than slackware on the install". I realise I'm in the minority here, but I like a simple, text-based install process. After Gentoo's, my favourite installers (of those I've used) are Debian's old-style one and [Free|Open]BSD. They do what they need to, and don't get in the way.
Oh, and the other major plus with Gentoo's install: you don't need an install CD to do it; any Linux system will do. I haven't yet found another distro or OS that lets you do that.
Provided you have a NUMA-aware operating system, that is.
Like Linux 2.6, you mean?
it makes it harder to distinguish between stable and development
(Overly) simple answer: 2.6-mm is development. 2.6 vanilla is stable.
Even here, the situation is improving, from a Linux point of view. Look at UT2004 and Doom 3. Two of the big releases this year, and one has Linux binaries (and even an install script) on the same CDs as the windows version, and the other is promising Linux binaries to download. Linux gaming is going mainstream. Slowly, but it's happening.
Have I suddenly awoken in the fabled "Year of Linux"?
Yes, you have. Welcome to the Year Of Linux, number 3.
Yup, and they didn't even write that. They bought it (insanely cheaply, IIRC) and marketed it. There's one thing they have figured out and are extremely good at, and that's marketing.
If you're a software developer, you can do a lot worse than sourceforge. The shell server doesn't seem to have a compiler, but you can opt-in for Compile Farm access, which should more than satisfy your compilation needs. Pine isn't there, but Mutt is, and i haven't checked IRC.
Poor poor POOR UI design there, Bill...
Well, at least it's what we've come to expect.
...when Firefox crashes whenever it tries to load the page. Don't get me wrong, I love Firefox, but it isn't perfect yet.
If that's what you care about, use BSD.
Six times 9 is 54. Write 54 in base 13, and you have 42. (4 x 13 = 52; 52 + 2 = 54)