Slashdot Mirror

In some political way, yes.

Holy balls, Eben Moglen is a deep thinker: Visionary Keynote

>>> 1. No fecking media support!

Distributor's policy.
For RedHat, please see: http://www.redhat.com/legal/patent_policy.html

>>> 2. Why the hell do I have to install a new kernel? Why?

These are build numbers. Many are test builds which are not released. These are not always changes bugs but lots of times performance and hardware compatibility adjustments.

See http://rpm.pbone.net/index.php3/stat/22/idpl/28456 36/com/changelog.html for an example of changes.

Windows also does this, it's called Windows Update. But most of it's fixes and patches tend to be in other applications. The windows kernel rarely changes because Microsft is not actively developing or maintaining it. Also hardware support does not need to be programmed into the kernel in the same way it does in Linux. But this has tradeoffs in terms of security and stability.

>>> 3. Point 2 also breaks my nvidia drivers.
>>> I don't want to re-compile new drivers everytime there's a new 'patch'. For the love of god, why?!

You generally shouldn't have to, unless the kernel changes version or some other change breacks ABI (application binary interface) compatibility. That's where two compiled programs no longer work together and need to be recompiled to function. Sometimes syscalls get removed or changed in linux and then you need to recompile to get them working again.

Again windows doesn't have this issue because windows never changes (which may or may not be a good thing).

>>> 4. X-Windows. What a mess. Why do I have to tell it my x & y refresh rates for my monitor?

It *DOES* just 'know'. For quite a while now. It's call EDID, short for "Extended display identification data". It's an open hardware standard that allows the video card to query the monitors capability. If things are configured properly, X along with your video driver will query the card and will know which scanrates are valid. Check your /var/log/Xorg.log.0 for details. You should no longer have any 'ModeLines' in your /etc/xorg.conf file.

>>> 5. Lack of decent file-browser. The best I've come across is Nautilus
>>> in a mode that resembles Windows Explorer. It'll do for now...

Not every enjoys "spacial mode", the default on some distros. It was introduced as the default for Gnome in 2.10 I beleive to very much mixed review. It can be disabled via gconf (similar to windows registry). See: http://www.larsen-b.com/Article/133.html

>>> as far as I'm aware, offers no context-sensitive menus for applications
>>> (like the Winamp "Play in Winamp" right-click menu on folders.

In fact, it actually does. In fact it's got "Open With" and a feature to allow you to specify which program to launch with, which is remembered and presented in the menu and associated with the file type by it's mime-type (file extentions don't matter), which is superior with windows, imho.

From the right-click context menu on any file, choose Properties, and go to the "Open With" tab.

>>> Actually, I think that's largely it. In all, Linux has, and is
>>> continuing to be great fun to play with. So many cool tools -
>>> yum being one of them. I'll stick to Linux @ home; it can only
>>> get better, but I'd be interested to know what people think of
>>> the above points - any suggestions maybe? I want this to work after all...

Glad to hear it. I wish more people shared your supportive and constructive attitude. The fact that desktop linux is where it is now is even amazing considering it's being built and shaped by so many people and specialized into so many different areas of use. It's still relatively new in comparis

I have worked with both dpkg and rpm, and there is no question: rpm is vastly superior to dpkg, when it comes to building packages, checking what package a file belongs to, or verifying the installed software (can't do it with dpkg).

• building packages Lets see, to build a package we just run apt-get build-dep foopkg; apt-get install build-essential fakeroot; apt-get source foopkg; cd foopkg-*; fakeroot debian/rules binary;. Hrm. That wasn't so hard...
• checking what package a file belongs to Is the package installed? Ok, dpkg -S foofile; Not installed? apt-get install apt-file; apt-file update; apt-file search foofile; Not running Debian? Visit packages.debian.org and search for a file.
• verifying installed software cd /; md5sum -c /var/lib/dpkg/foopkg.md5sums|grep -v OK. Too hard? Install debsums and use it intsead.

• Sun Microsystems Solaris and Linux Training
• IBM AIX Training
• IBM Linux Training
• HP's HP-UX Certification Training
• HP's Tru64 UNIX® Training
• Red Hat Training
• Novell SUSE Linux Training
• HP's NonStop UX Training
• Apple's Mac OS X Server Training
• And, if you're really sick... SCO's SCO UNIX Training

https://www.redhat.com/training/training_special.h tml?sc_cid=MSI_GTRRTKB_4

Red Hat released a purely security related kernel update on 2006/05/24:

      https://rhn.redhat.com/errata/RHSA-2006-0493.html

I would be very surprised if your kernel did not have known security issues that you are unaware of. Whether or not the various security issues apply to your environment is another question.

robert

What documentation issue?

There are boatloads of documentation available. Ever hear of The Linux Documentation Project? Plus, most distributions offer lots of very good documentation. Why there was a Slashdot story just two days ago about the excellent Ubuntu documentation. There are no fewer than 600 books available about Red Hat distros available for sale on Amazon. Not to mention that Red Hat Enterprise Linux itself includes lots of lots of documentation and most of it is available on the Web gratis. Plus the hundreds of open source apps that include very good documentation with their package. Have you actually read the documentation and free books available on the Samba website? It's darned good!

Any perceived documentation issue is Laura DiDiot's head.

What the hell has Red Hat ever done for the Linux community?

Maybe you should look at http://sources.redhat.com/.

OTOH, I agree with you somewhat. There's no fucking way I'm doing beta testing for RedHat after they pulled the old bait and switch on the Linux community. ("Oh, did you like your supported free version of our OS? Well guess what? It's gone, sucker. Oh, but you can do our beta testing for us by running our unstable bleeding-edge version.")

Regardless, RedHat has done a HELL of a lot for the Linux community, and for the FOSS community in general. Even Cygwin is a redhat project these days.

What the hell has Red Hat ever done for the Linux community?

Maybe you should look at http://sources.redhat.com/.

OTOH, I agree with you somewhat. There's no fucking way I'm doing beta testing for RedHat after they pulled the old bait and switch on the Linux community. ("Oh, did you like your supported free version of our OS? Well guess what? It's gone, sucker. Oh, but you can do our beta testing for us by running our unstable bleeding-edge version.")

Regardless, RedHat has done a HELL of a lot for the Linux community, and for the FOSS community in general. Even Cygwin is a redhat project these days.

hmm. price per seat...
I just took the Red Hat System's Maintenance and support class. in this class Red Hat shows you how to buy only one support contract (that is all you buying) and use a satellite server to get updates for all the rest. They have a desktop per(50) seat license for the desktop software, but in comparison it cheep. This is for large corporations who do not understand buy one use on all. nothing stops you from doing this.

also you have enterprise and centOS that are build on red hats source.

and Fedora 5, well the new SeLinux with a soon to be PL 4 rated system, is very useful for government systems. and built in ZEN for virtualization, you have now opened up a whole new way to run servers and securing them.

There change went from a per version license to a time support license. This helped companies be assured that they could buy it and it would be supported. They felt they could write code and not have it change next week. Also there support is for free upgrades during the support time. But you still use the software when you stop paying.

Unlike Microsofts...... where when you stop paying you must .. well let them tell you...
"At the end of the subscription the rights to use the software end and the customer must either uninstall the software or renew its Enterprise Subscription enrollment."


the whole thing..
Enterprise Subscription Agreement is designed for organizations that prefer to standardize their Microsoft software throughout their organization based on the Microsoft enterprise products (Microsoft Office Professional, Microsoft Windows Professional desktop operating system upgrade, and Core Client Access License) at prices based on a three-year agreement term. At the end of the subscription the rights to use the software end and the customer must either uninstall the software or renew its Enterprise Subscription enrollment.

So I don't mind paying $2,400 for a copy of Red Hat, put it on all my boxes (over 50 at this site) and use a satellite server to get updates. Call for support for any problems..

you get
Web and phone-based comprehensive support
24x7
1 hour response
Unlimited incidents
1 year Red Hat Network 1



also if your a student you can get all this for $50 http://www.redhat.com/rhel/details/academic/indivi dual/

Several years ago I tried to set up my Linux box as an internet router/gateway, using IPTABLES and what-not. I failed just through sheer lack of time to commit to learning all the stuff I'd need to know to do it properly. About that time, the first "Cable/DSL routers" came on the market, and made moot my need.

Now, however, it is very easy to configure the various widgets that you'd need for this task because tools exist on (e.g.) Fedora to make it so.

For myself, I'm glad I can put the effort into learning more in-depth some of the things I can do with Linux, and yet those things I find tedious, or don't have the time to do, have "easy-to-use" tools handy.

More FreeBSD hype for a barely working feature. DTrace is yesterday's news anyway: SystemTap is the new hotness.

Give me a moment, and I'm sure I'll think of one.

I don't know of a large company that still lets most employees install software, have admin rights, or do anything like that.

http://www.redhat.com/about/careers/ :-)

Could Sun be using this to eventually get to the desktop, or at the very least, allowing companies to run a complete linux system. Solaris server, Ubuntu clients for the employees?

In some ways RedHat/Fedora might be a better fit for this with Stateless Linux which would fit nicely into that sort of model.

Jedidiah.

Buy our "free" product for only $9.95!

From what I've read, I think that might be because some developers on the Linux camp have been a factor of irritation when they produce Linuxisms is C code, foresaking portability.You see this phenomenom mentioned in what regards GNOME in the article. It almost sounds as if GNOME developers are a clique that don't give a shit about other projects.

Another famous little war was Linuxers resistance (glibc maintainers, to be exact) resistance against the safer strlcpy and strlcat functions from OpenBSD's libc:
See these amazing threads that illustrate prejudice against the OpenBSD developers. After 2 or smth years, they finally gave in and the OpenBSD functions are part of glibc. But here's my sample:

Here's a Debian developer calling on GNOME developer's biased and prejudiced views against OpenBSD's innovation for safer C programming:
http://lists.debian.org/debian-devel/2002/03/msg00 305.html

Here's the guy that sends the patch for glibc: http://sources.redhat.com/ml/libc-alpha/2000-08/ms g00052.html

Here's the amazing answer from the glibc's maintainer Ulrich Drepper, a real insight into strong software engineering principles. No wonder Linux boxes got so rootkitted:
This is horribly inefficient BSD crap. Using these function only
leads to other errors. Correct string handling means that you always
know how long your strings are and therefore you can you memcpy
(instead of strcpy). http://sources.redhat.com/ml/libc-alpha/2002-01/ms g00002.html

Theo's take:
TdR: They're still not in glibc. They're everywhere else. They're in Solaris. We invented them two years ago. They're showing up in vendor operating systems. We made a convincing argument why these things are necessary. http://www.ddj.com/184404914

Look at CERT's list for "glibc" vulnerabilites here. Please draw comparisons with BSDs. Answer honestly: who's got bragging rights?

From what I've read, I think that might be because some developers on the Linux camp have been a factor of irritation when they produce Linuxisms is C code, foresaking portability.You see this phenomenom mentioned in what regards GNOME in the article. It almost sounds as if GNOME developers are a clique that don't give a shit about other projects.

Another famous little war was Linuxers resistance (glibc maintainers, to be exact) resistance against the safer strlcpy and strlcat functions from OpenBSD's libc:
See these amazing threads that illustrate prejudice against the OpenBSD developers. After 2 or smth years, they finally gave in and the OpenBSD functions are part of glibc. But here's my sample:

Here's a Debian developer calling on GNOME developer's biased and prejudiced views against OpenBSD's innovation for safer C programming:
http://lists.debian.org/debian-devel/2002/03/msg00 305.html

Here's the guy that sends the patch for glibc: http://sources.redhat.com/ml/libc-alpha/2000-08/ms g00052.html

Here's the amazing answer from the glibc's maintainer Ulrich Drepper, a real insight into strong software engineering principles. No wonder Linux boxes got so rootkitted:
This is horribly inefficient BSD crap. Using these function only
leads to other errors. Correct string handling means that you always
know how long your strings are and therefore you can you memcpy
(instead of strcpy). http://sources.redhat.com/ml/libc-alpha/2002-01/ms g00002.html

Theo's take:
TdR: They're still not in glibc. They're everywhere else. They're in Solaris. We invented them two years ago. They're showing up in vendor operating systems. We made a convincing argument why these things are necessary. http://www.ddj.com/184404914

Look at CERT's list for "glibc" vulnerabilites here. Please draw comparisons with BSDs. Answer honestly: who's got bragging rights?

howabout here... or here... or even here!

OSS developers are not like Microsoft developers.

Spot on. OSS developers are often employed by companies such as redhat, suse and IBM. Sometimes they need to focus on boring features proposed by their respective marketing divisions, but these horrible tasks are divided between the competing companies so that the developers can spend more of their time doing whatever the hell they like, i.e. working on neat features and rewriting the core.

I had a similar problem with my Gateway WX6421 with Fedora Core 5 x86_64 (Fedora bugzilla 188283, https://bugzilla.redhat.com/bugzilla/show_bug.cgi? id=188283). DHCP works and I can connect, however whenever I put it under any type of load, the driver locks up and stops responding. I can recover by stopping the network, rmmod sky2, insmod sky2, and restarting networking. Under Linux, I can only use this laptop with wireless (not entirely a bad thing, but not as secure).

2.6.16 SMP also locks up on my Supermicro dual XEON box (Fedora bugzilla 188291, for now I'm just using the UP kernel).

These are the types of problems that must be fixed. An upgrade should not break known, working hardware especially causing a hard lockup.

I agree with the previous comment. Memory that has gone bad, memory being run at a speed it can't cope with, overclocked machines, undercooled machines, overstretched PSUs etc all cause weird and intermittent problems (see this Linux kernel developer's note on hardware that results in weird problems). If you are binary driver free and still having strange intermittent lockups I'd definitely test the memory for a day with Memtest and start swapping bits of hardware out.

Yes, it's a good idea.

But don't waste time on bugs that only affect legacy hardware.

It would also be a good idea for some effort to be spent on consolidating, corrrecting, and updating the various lists of "Hardware supported by Linux". There are lots of such lists on the web, for example:

• Hardware Howto
• Linux Hardware Support
• LinuxQuestions HCL

- not to mention the distro-specific compatible hardware lists maintained for Redhat, Mandriva,Suse, and others.

We need one correct, maintained list, not dozens of nearly-correct, usually out-of-date lists. And it seems to me that the list should depend only on the kernel version, not on the distro.

It's important and interesting to point out that Red Hat is eliminating the Fedora Foundation.

See http://directory.fedora.redhat.com/wiki/Mod_nss for more details.

You can send feedback to developers via the Mozilla NSS newsgroup: http://groups.google.com/group/mozilla.dev.tech.cr ypto?lnk=lr&hl=en

There's nothing disingenuous about this. We released it as a press release on our own site:

http://www.redhat.com/about/news/prarchive/2006/fa a.html

Everything about it shouts "press release", including the SEC warnings at the bottom and the press contact information. As is typical with press releases, it was picked up and run all over the place. That's what press releases are for. Anything that comes from Business Wire is a press release.

If you think it's dishonestly masquerading as "real news," that's your mistake.

The idea behind Address Space Randomization is to put program code at a different address each time it starts. This way, an exploit can't know where the return address pointer should point to.

Working with Microsoft consultants

I have to wonder, how much were they willing to spend on Linux consultants (which, in fact, do exist-- http://www.novell.com/ http://www.enterux.com/ http://www.redhat.com/services/consulting/ --just to name a few big ones)? It seems like there is a reluctance to spend any money on Linux. Heck, it's free software, right? Too many companies seem to look to Linux as a way to reduce costs by not spending money for the software or expertise to make it run. Too few seem to look at the longer term savings brought about by increased flexibility and better control.

Maybe if they spent the same amount on Linux consultants to show them how to get what they wanted (and maybe to help them figure out what they needed) as they spent on MS cosultants, their experience would have been a little different.

But to play devils advocate for a bit, the free like beer "download editions" for most Linux distros have left a larger market of Linux admins to hire from. We are a mixed shop but most of the "new" guys into out Linux/UNIX deparment tend to prefer Linux becasue that is what they know. Everytime we lose an old Solaris guy, we get a new Linux guy. I don't think the days ahead are going to get any easier for Sun.

Not so much. Nobody here seems to realize that even the nv driver is maintained by NVIDIA and contains obfuscated source. Not that useful without hardware documentation.

There was a human error in the updates pushed for FC4/FC5 update repository in the last night. Fedora's people are fixing it.

More information about it:
https://www.redhat.com/archives/fedora-devel-list/ 2006-April/msg00898.html

Temporary fix:
yum update --exclude=\*evolution\*

http://www.redhat.com/docs/manuals/ccvs/

JBoss might be a different product and different market but it makes me wonder if JBoss with end up like CCVS. Red Hat purchased another opensource project/product a while ago called CCVS( Credit Card Verification System ) and converted it to their proprietary license before later killing the product couple years later. They told their existing customers they'd be supported til the end of their contract by a 3rd party( mainstreetsoftworks.com ) and that MainStreet Works had a replacement product( also proprietary ).

If you've ever looked for GNU/Linux based CC processing software, you know how long and unsuccessful the search was/is.

There's definately a larger market for JBoss but the results could be the same in the long run if Red Hat can't market the product to profits. They are not a friend to Open Source when they do these kinds of things and it also shows/helps Microsoft when they do this... IMO.

LoB

Every single distro does its own thing and there is no standardization whatsoever.

I humblely disagree.

Filesystem Hierarchy Standard (FHS)

The filesystem standard has been designed to be used by Unix distribution developers, package developers, and system implementors. However, it is primarily intended to be a reference and is not a tutorial on how to manage a Unix filesystem or directory hierarchy.

Gentoo FHS
RedHat FHS
Suse FHS&LSB

And for binary distros there is Linux Standard Base (LSB)

The LSB specification is made up of several components, known as modules. The base specification consists the of Core, Graphics and CXX (C++) modules. The specification is further extended with the Desktop set. Each module might be subdivided into a common document plus architecture-specific documents (in some cases the subdivision is not needed). A complete binary standard for a particular processor architecture consists of the set of necessary common documents plus the matching set of architecture-specific documents.

Latest LSB Spec 3.1.0

Quote: "In any case, getting a precompiled kernel with everything, including the kitchen sink, compiled in so it will run on and use as much hardware as possible, and checking its memory usage, is hardly a usefull test. Fedora doesn't know what you're going to use the kernel on, so it includes all possible support code, and that of course means bloat; if you're using the kernel on a $100 laptop with known hardware, you can make it much smaller."

Yes, I do realise this. The thing you're looking at is an average system. Tell me, how many Linux distributions allow you to fully customise your kernel at install time? mmm? None that I'm aware of (and even if there was one, it'd be only for the geeks amongst us, not the mainstream users). The test does a pretty good job of showing the average distribution, running averagely.

Quote: "Besides, what difference does running or not running X make on kernel memory usage ? Total system memory usage, yes - but even an X-less Fedora system is going to be running lots of daemons, just in case someone happens to need them. Or is Fedora smart enough to only load the display driver when it's needed ?
"

I simply removed X out of the equation to make it easier to run on the example PC that I chose. A modern Linux distribution like Fedora Core 5 with X wouldn't most probably run on a Pentium 100 with 64mb RAM. Hell, let's look at the system requirements for Fedora Core 5:

http://fedora.redhat.com/docs/release-notes/fc5/#i d3099264

"Recommended for text-mode: 200 MHz Pentium-class or better "

and

"Minimum RAM for text-mode: 128MiB "

Now this is the text mode specs I might add. Compare this to Redhat 6.2 specs (I still have the boxed version in my room hehehe, God knows where my 5.2 box is lol):

x86 architecture
500mb hdd space
16mb RAM

The box doesn't state if that's for a full graphical interface or not, but I'll give it the benefit of the doubt and say it's for a text mode install!

Now - let's note that the original article says "Linux is NOT fat". I've been kind to it and went with the kernel, but if I were to go the whole hog, with X and either KDE or Gnome...I can honestly tell you from personal experience that Debian Sarge (running kde 3.5 from Sid) ran a shitload slower than my XP installation does.

You can really tell the Linux zealots are out in force.

Dave

i started hacking on linux around 7 years ago. rh 7.2 was the word. kernel compilation was quite easy, a few items to say N and some to say M to, to get your oracle and apache and modperl running.

I'll tell you who installs Linspire. ESR. Or at least he should.

Fedora is not trying to be a desktop OS. If it was, flash, java and mp3 would ship out of the box.

Also, yes, it's an RTOS, though maybe not the most sophisticated one out there. The idea seems to be to provide something that's very small, lightweight, and fast, with absolutely minimal interrupt handling times, etc.

In case you're interested, here's a link to a book, Embedded Software Development with eCos, which describes the thing.

Simple QA at best is all you can hope for, not a year of open beta.

No, you surely can't

For the sake of completeness, here is a link to the *full text* of the email that was sent to the fedora-lists with the Foundation announcement.

https://www.redhat.com/archives/fedora-list/2006-A pril/msg01022.html

People are giving up on Redhat.

...

I decided that Fedora has gone beyond bloated and sucky, and that if I were to ever prefessionally [sic] recommend any Linux flavors, they'd be Gentoo and the free Redhat Enterprise clones (Whitebox, etc.).

I am only guessing by professional, you mean business. So let me get this straight--you are looking to professionally recommend a Linux distro and you are complaining about a distro that is openly NOT a professional distro not being professioinal enough. How does that make sense? See Red Hat's brief explanation here:

https://www.redhat.com/apps/download/

Here is the link to the email Redhat sent out. https://www.redhat.com/archives/fedora-announce-li st/2006-April/msg00016.html

To say that the article writer has a bias against Redhat would be an understatement. Even when Redhat is transparent they are still lambasted. People want to hate Redhat, but without Redhat we would be much worse off in the Linux world. It's time people admit it.

The source code for the OLPC software is at: http://download.fedora.redhat.com/pub/fedora/proje cts/olpc/ This is the specific software that they will be using, not just a generic Red Hat distribution

Developers conference: "We are planning an international-developers conference to be held most probably in Singapore at the end of the year. As presently envisioned, the invitation-only meeting of a thousand or more developers would last from three to five days and focus on open-source, localization and educational-software issues." I'm sure if you were interested and contacted some people (check the "Contacts" page on the wiki and the "Getting Involved" page below) and had a genuine interest in helping that you could receive an invitation.

What work to do?: The wiki page "Getting involved in OLPC" http://wiki.laptop.org/wiki/Getting_involved_in_OL PC has lots of ways to help. If you can't find something here, I don't know what to say...

Best wishes and I hope it works out for you.

oh if only there were a way!

Errr... okkkk... you must be a debian user (which uses exim) - as an example the following use sendmail as their default MTA:

Solaris
HP-UX
RedHat
Suse
Slackware

0. Are they powerful enough to run XP / 2K?
If not, then how well do you think 98 will fare over time (when is it EOL'd?)


If they can't run 2k / XP they won't run a modern Linux dist either. Recommended specs are near enough the same - eg. http://www.microsoft.com/windowsxp/pro/evaluation/ sysreqs.mspx http://fedora.redhat.com/docs/release-notes/fc4/#s n-hardware-reqs.

Remember, XP dates from 2001 (and win 2k is earlier), for a PC to be non-XP-capable it would probably be >5yrs old. Consumer PCs that old are not going to be "potentially used for years" unless you are doing a lot of work fixing and just finding rare spare parts. Most charities (at least here in the uk) simply won't take systems that old.

Sure, you _can_ run Linux on older systems. The last one I threw away (since non of the charities would take it) was a 486 with 16M, ran Slackware up to about 1.2 quite happily - but upgrading to Redhat 2.x was always a regret as it slowed it down so much (glibc, elf binaries, pic shared libs - vs. the old lean-mean a.out jumptable stuff). Good luck with getting any uptodate apps to run, even if you could get it to build, I suspect just a modern browser (ie. Firefox) would kill that system, let alone something like OpenOffice.

Worried about EOL ?

MS still support win 2k (from, um 2000). Try finding patches for Linux vendors from that era. Redhat have EOL'd everything up to RHL9 (2003) - and even fedora "legacy" supports only from 7.3 (2002 vintage).

https://rhn.redhat.com/rhn/errata/details/Details. do?eid=3971 (rhn login required)

ummmm .... whether you are subscribed to the beta channels or not ...:

https://www.redhat.com/archives/nahant-list/2006-M arch/msg00049.html

That is when the ISOs for the offical release came out, as well as the RPMS on RHN were made available. Now if you are talking about the Beta release of u3 ... that is a different story. But that was a totally different set of RPMS. (yes, I have access to the beta channel)

There seems to be some people who think CentOS does not Red Hat ... that could not be further from the truth. They have a re-distribution policy that we more that meet the requirements of. We have worked with them in the past and will continue to work with them in the future to address any issues that they have with CentOS.

Mod parent down.

In one sentence, it states that CentOS uses the "source packages published by Red Hat", and in the next sentence it says that RHEL is "distributed only in uncool binary form".

The source code to RHEL is fully available.

ftp://ftp.redhat.com/pub/redhat/linux/enterprise/4 /en/os/i386/SRPMS/

Check out this page. But yeah, I see what you are saying.