[Fwd: Red Hat Linux 9 -- Get it Early]

[Eric "Ronin" Krieger]

Just and fyi for you red hat-ers out there...

Eric


-----Forwarded Message-----

We appreciate your support of Red Hat Network and wanted to alert you
to a special service that we are extending to paying Red Hat Network
customers such as yourself. For the past couple months we've gathered
feedback and listened to our customers. We've heard that one of the
things you want most is early access to Red Hat Linux ISOs.
Well, we've responded. Starting March 31st at 9am Eastern, you can
start downloading Red Hat Linux 9 ISOs -- a week before they will be
generally available in retail stores or via Red Hat FTP. For more
http://www.redhat.com/mktg/rh9iso/
You also asked for improved technical support. Beginning March 31st,
you will notice an improved Red Hat Network FAQ, and you will also
notice that Red Hat Network technical support (with a Service Level
Agreement) is being extended to paying customers.
We hope that you enjoy these added benefits of being a Red Hat Network
subscriber and that you will continue to provide us with feedback in
the future.
Thanks for all of your continued support of Red Hat.
--the Red Hat Network team

[Ken Long]

I wonder why they jumped from 8.0 all the way to 9? 8.0 hasn't been out
all that long has it?

Ken

Post by Eric "Ronin" Krieger
Just and fyi for you red hat-ers out there...

[Ed Brown]

Trying to catch up to MacOS X?

More seriously, my guess is that it will include the 2.6 kernel, which
would certainly call for a new OS major version. Does anybody know for
sure if that's right?

-Ed

Post by Ken Long
I wonder why they jumped from 8.0 all the way to 9? 8.0 hasn't been out
all that long has it?
Ken

Post by Eric "Ronin" Krieger
Just and fyi for you red hat-ers out there...

_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

[Steve Browne]

Post by Ed Brown
Trying to catch up to MacOS X?
More seriously, my guess is that it will include the 2.6 kernel, which
would certainly call for a new OS major version. Does anybody know for
sure if that's right?
-Ed

The most recent kernel I see being tested is 2.5.65; nobody is
discussing a 2.6 yet. For the latest, check the newsgroup
fa.linux.kernel.

Steve
Stephen B. Browne
***@ix.netcom.com
"Veritas contemno captio."

[Peter Espen]

Post by Steve Browne

Post by Ed Brown
Trying to catch up to MacOS X?
More seriously, my guess is that it will include the 2.6 kernel, which
would certainly call for a new OS major version. Does anybody know for
sure if that's right?
-Ed

The most recent kernel I see being tested is 2.5.65; nobody is
discussing a 2.6 yet. For the latest, check the newsgroup
fa.linux.kernel.

The way the linux kernel versions work is even numbers are official
releases and odd numbers are development branches. So, when 2.5.x
is released it will be released as 2.6.x and the new development
branch will most likely become 2.7.x.

[C. Ulrich]

Post by Peter Espen
The way the linux kernel versions work is even numbers are official
releases and odd numbers are development branches. So, when 2.5.x
is released it will be released as 2.6.x and the new development
branch will most likely become 2.7.x.

I was originally going to note that 2.6.0 was still a long way off, but then I
checked the timestamps on kernel.org to make sure I wasn't going to spout
complete B.S.

Good thing I had look, then, because I note that the timestamp for the 2.0
kernel is June 8, 1996. (I'm assuming that the timestamps correspond to
release date.) The next stable series, 2.2 was on Jan 25, 1999. And finally,
the 2.4 seems to have been released on Jan 04, 2001. We draw the conclusion
that the average lifetime of a stable kernel series is 2 to 2 1/2 years. We're
at about 2 years and 4 months right now since the last stable series introduction.

A look at the various patchlevels could provide some insight as well. Unless I
read the directory wrong, the 1.3 devel series has 100 patchlevels even. The
2.1 series has 132. The 2.3 series had only 51 patchlevels BUT 2.4.0 had 12
"test" patchlevels plus one "prerelease" before 2.4.0 final was released for a
logical total of 64 patchlevels. In 2.5, we're at 66 patchlevels.

Conclusion: we're at the point now where 2.6 could be released at any time.
The only problem we face is that developers haven't indicated that they're
going to do so yet. :) (Unlike 2.4, which has 7 months of "test" patchlevels.)

Finally, for those new to kernel versioning, bear in mind that the devel
kernels traditionally don't branch until the stable series has seen a thorough
round of bugfixes and patchlevels.

--Charles Ulrich

[Ed Brown]

Ted Ts'o is scheduled to speak at the Usenix conference in June on the
"current release", 2.6, and the development kernel, 2.7:

from the website:
Although the material will focus on the latest release version of the
Linux kernel (v. 2.6), it will also address aspects of the development
kernel codebase (v. 2.7) where its substance differs from 2.6.

It may not be included in redhat 9.0, but it can't be far off.

-Ed

Post by C. Ulrich

Post by Peter Espen
The way the linux kernel versions work is even numbers are official
releases and odd numbers are development branches. So, when 2.5.x
is released it will be released as 2.6.x and the new development
branch will most likely become 2.7.x.

I was originally going to note that 2.6.0 was still a long way off, but then I
checked the timestamps on kernel.org to make sure I wasn't going to spout
complete B.S.
Good thing I had look, then, because I note that the timestamp for the 2.0
kernel is June 8, 1996. (I'm assuming that the timestamps correspond to
release date.) The next stable series, 2.2 was on Jan 25, 1999. And finally,
the 2.4 seems to have been released on Jan 04, 2001. We draw the conclusion
that the average lifetime of a stable kernel series is 2 to 2 1/2 years. We're
at about 2 years and 4 months right now since the last stable series introduction.
A look at the various patchlevels could provide some insight as well. Unless I
read the directory wrong, the 1.3 devel series has 100 patchlevels even. The
2.1 series has 132. The 2.3 series had only 51 patchlevels BUT 2.4.0 had 12
"test" patchlevels plus one "prerelease" before 2.4.0 final was released for a
logical total of 64 patchlevels. In 2.5, we're at 66 patchlevels.
Conclusion: we're at the point now where 2.6 could be released at any time.
The only problem we face is that developers haven't indicated that they're
going to do so yet. :) (Unlike 2.4, which has 7 months of "test" patchlevels.)
Finally, for those new to kernel versioning, bear in mind that the devel
kernels traditionally don't branch until the stable series has seen a thorough
round of bugfixes and patchlevels.
--Charles Ulrich
--
http://bityard.net
_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

[jdavis]

Post by Steve Browne

Post by Ed Brown
Trying to catch up to MacOS X?
More seriously, my guess is that it will include the 2.6 kernel, which
would certainly call for a new OS major version. Does anybody know for
sure if that's right?
-Ed

The most recent kernel I see being tested is 2.5.65; nobody is
discussing a 2.6 yet. For the latest, check the newsgroup
fa.linux.kernel.
Steve
Stephen B. Browne
"Veritas contemno captio."
_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

yes but 2.5.65 *is basically* 2.6 just as 2.3.* was 2.4.*

[wes robbins]

how do you stop a bsd box from having a
nightmare'ishly long connect time for ssh.
I did not test any other ports....

=====


__________________________________________________
Do you Yahoo!?
Yahoo! Platinum - Watch CBS' NCAA March Madness, live on your desktop!
http://platinum.yahoo.com

[M. Warner Losh]

In message: <***@web41508.mail.yahoo.com>
wes robbins <***@yahoo.com> writes:
: how do you stop a bsd box from having a
: nightmare'ishly long connect time for ssh.
: I did not test any other ports....

Get a faster box? I've definitely noticed a huge slowdown with my ssh
connections to a box that has ssh2 keys/connections by default. Maybe
ssh -1 might help? Alternatively, maybe it is a DNS issue and the
name server is timing out?

Warner

[C. Ulrich]

Post by M. Warner Losh
Alternatively, maybe it is a DNS issue and the
name server is timing out?

This is a problem I have on my small network (that I've been too lazy to fix).
I have a box that's a firewall / dialup gateway. All of the other computers
have the IP of that box set as their TCP/IP gateway, /etc/resolv.conf on each
machine contains the DNS servers of my ISP, and /etc/hosts has the names of
all my local machines. When I'm connected to the net, ssh connections are
instantaneous and take an eon otherwise. No other services do this, so I'm not
sure why ssh is the exception.

I have seen it mentioned that not building IPv6 support can help some cases,
but I honestly don't remember if I tried it.

Comparing the output of "ssh -v" in both cases shows no obvious differences.

Okay, now I'm interested.
--Charles Ulrich

[Ed Heron]

Isn't there supposed to be somplace where the DNS resolv order would be?
It sounds like this machine is using DNS to resolve and timing out before
using the hosts file. I looked at the man page for resolv.conf (under
GNU/Linux) and didn't see an order option there. Shouldn't we be able to
set the order to hosts then DNS?

----- Original Message -----
From: "C. Ulrich" <***@securenym.net>
To: <***@nmlug.org>
Sent: Wednesday, March 26, 2003 12:38 PM
Subject: Re: [NMLUG] bsd question

Post by C. Ulrich

Post by M. Warner Losh
Alternatively, maybe it is a DNS issue and the
name server is timing out?

This is a problem I have on my small network (that I've been too lazy to fix).
I have a box that's a firewall / dialup gateway. All of the other computers
have the IP of that box set as their TCP/IP gateway, /etc/resolv.conf on each
machine contains the DNS servers of my ISP, and /etc/hosts has the names of
all my local machines. When I'm connected to the net, ssh connections are
instantaneous and take an eon otherwise. No other services do this, so I'm not
sure why ssh is the exception.
I have seen it mentioned that not building IPv6 support can help some cases,
but I honestly don't remember if I tried it.
Comparing the output of "ssh -v" in both cases shows no obvious differences.
Okay, now I'm interested.
--Charles Ulrich
--
http://bityard.net
_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

[Gary Sandine]

Post by Ed Heron
Shouldn't we be able to
set the order to hosts then DNS?

It's done in /etc/nsswitch.conf, but default behavior is usually files
before dns. e.g.

hosts: files nisplus nis dns

[Ed Heron]

Yes. When I look in my /etc/nsswitch.conf file I see:
hosts: files dns

but at
http://alert.udfcd.org/help/tcpip/user_guide/utils/resolv.conf_f.html it
talks about:
lookup
This keyword is used by the library routines gethostbyname() and
gethostbyname(). It specifies which databases to search, and the search
order. The legal space-separated values are:
bind
Use the Domain Name server by querying named.
file
Search for entries in /etc/hosts.
If the lookup keyword isn't used in the system's resolv.conf file, then the
assumed order is bind file. If resolv.conf doesn't exist, the only database
used is file.

----- Original Message -----
From: "Gary Sandine" <***@laclinux.com>
To: <***@nmlug.org>
Sent: Wednesday, March 26, 2003 1:22 PM
Subject: Re: [NMLUG] bsd question

Post by Gary Sandine

Post by Ed Heron
Shouldn't we be able to
set the order to hosts then DNS?

It's done in /etc/nsswitch.conf, but default behavior is usually files
before dns. e.g.
hosts: files nisplus nis dns
_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

[C. Ulrich]

I'll reply to several emails at once interst of saving time.

Charles,
Do you mean connections to other hosts on the network take a long time
when the internet connection is not up, or to hosts on the internet (or
both)?
-Ed

Sorry it wasn't too clear; I meant hosts on the network when the internet
connection is not up.

It's done in /etc/nsswitch.conf, but default behavior is usually files
before dns. e.g.
hosts: files nisplus nis dns

Mine says:

hosts: files dns

hosts: files dns
but at
http://alert.udfcd.org/help/tcpip/user_guide/utils/resolv.conf_f.html it
lookup
This keyword is used by the library routines gethostbyname() and
gethostbyname(). It specifies which databases to search, and the search
bind
Use the Domain Name server by querying named.
file
Search for entries in /etc/hosts.
If the lookup keyword isn't used in the system's resolv.conf file, then the
assumed order is bind file. If resolv.conf doesn't exist, the only database
used is file.

Interesting... but the man page on my system doesn't mention "lookup" (or
anything like it that I can tell). On a whim, I tried it anyway and no luck.
While I was there, I commented everything out of my resolv.conf and noticed
that the problem disappeared. This tells me that it *does* have to do with
hostname resolution somehow, *but* it's restricted to ssh since I have no
problems whatsoever with ping and apache under the same conditions.

I'm running Slackware 8.1, btw.

Wes, maybe you can try some of the things that were suggested. If your results
match mine then at least we know we have the same problem. :)

--Charles Ulrich

[Ed Brown]

I don't know if any apps still look to /etc/host.conf for the resolve
order, but if that file doesn't look like:
order hosts,bind
then you might try changing it.

As an aside, having once run a setup like you describe, I found it
helpful to run a caching nameserver to reduce unnecessary dialups.

-Ed

Post by C. Ulrich
I'll reply to several emails at once interst of saving time.

Charles,
Do you mean connections to other hosts on the network take a long time
when the internet connection is not up, or to hosts on the internet (or
both)?
-Ed

Sorry it wasn't too clear; I meant hosts on the network when the internet
connection is not up.

It's done in /etc/nsswitch.conf, but default behavior is usually files
before dns. e.g.
hosts: files nisplus nis dns

hosts: files dns

hosts: files dns
but at
http://alert.udfcd.org/help/tcpip/user_guide/utils/resolv.conf_f.html it
lookup
This keyword is used by the library routines gethostbyname() and
gethostbyname(). It specifies which databases to search, and the search
bind
Use the Domain Name server by querying named.
file
Search for entries in /etc/hosts.
If the lookup keyword isn't used in the system's resolv.conf file, then the
assumed order is bind file. If resolv.conf doesn't exist, the only database
used is file.

Interesting... but the man page on my system doesn't mention "lookup" (or
anything like it that I can tell). On a whim, I tried it anyway and no luck.
While I was there, I commented everything out of my resolv.conf and noticed
that the problem disappeared. This tells me that it *does* have to do with
hostname resolution somehow, *but* it's restricted to ssh since I have no
problems whatsoever with ping and apache under the same conditions.
I'm running Slackware 8.1, btw.
Wes, maybe you can try some of the things that were suggested. If your results
match mine then at least we know we have the same problem. :)
--Charles Ulrich
--
http://bityard.net
_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

[C. Ulrich]

Post by Ed Brown
I don't know if any apps still look to /etc/host.conf for the resolve
order hosts,bind
then you might try changing it.

That's what mine says.

Post by Ed Brown
As an aside, having once run a setup like you describe, I found it
helpful to run a caching nameserver to reduce unnecessary dialups.

Are you talking about a dial-on-demand type system? My connection is
controlled manually: I wrote a Tcl server script that listens to a particular
port for commands to dial or hang up the modem and then calls on chat and pppd
to do the dirty work. The (rather simple) client is written in Tcl/Tk and just
sends commands to the server and displays the output of chat, pppd, and the
firewall script.

Thus, caching nameserver probably wouldn't be a huge benefit. :)

--Charles Ulrich

[wes robbins]

the dns seems to have done the job.
Thank you everyone.

On another topic...To bad I'm "R0adKill" in
1st person games. I could be rich.....

game gambling.....
http://www.miami.com/mld/miamiherald/business/5423630.htm

=====


__________________________________________________
Do you Yahoo!?
Yahoo! Platinum - Watch CBS' NCAA March Madness, live on your desktop!
http://platinum.yahoo.com

[Ed Heron]

----- Original Message -----

Post by C. Ulrich

Post by Ed Heron
hosts: files dns
but at
http://alert.udfcd.org/help/tcpip/user_guide/utils/resolv.conf_f.html it
lookup
This keyword is used by the library routines gethostbyname() and
gethostbyname(). It specifies which databases to search, and the search
bind
Use the Domain Name server by querying named.
file
Search for entries in /etc/hosts.
If the lookup keyword isn't used in the system's resolv.conf file, then the
assumed order is bind file. If resolv.conf doesn't exist, the only database
used is file.

Interesting... but the man page on my system doesn't mention "lookup" (or
anything like it that I can tell). On a whim, I tried it anyway and no luck.
While I was there, I commented everything out of my resolv.conf and noticed
that the problem disappeared. This tells me that it *does* have to do with
hostname resolution somehow, *but* it's restricted to ssh since I have no
problems whatsoever with ping and apache under the same conditions.
I'm running Slackware 8.1, btw.

One could assume that if there wasn't a name server listed in resolv.conf,
then only the hosts file would be available.

You could probably eliminate your problem by running a local name server.
Set it up to resolve certain domains like the tld bob. or platapus. and
forward all other requests to your ISP's name servers. Since your 'fake'
tlds don't exist, they would not be masking any 'real' ones. Or you could
have one 'fake' tld like local. and put that in your search list. That way
if you tried to resolve bob, it would append local on it and try to resolve
bob.local. and your local server would correctly return an IP address. ( I
quote fake and real in relation to top level domains - tlds - because the
ones we have now, though not completely arbitrary, are still more
restrictive than necessary. I'm sure they were set up in an environment that
assumed our networks would not be quite so )

If you use Bind 9, I could probably give you a good idea of what your
config could be on your internal name server.

[C. Ulrich]

Post by Ed Heron
One could assume that if there wasn't a name server listed in resolv.conf,
then only the hosts file would be available.
You could probably eliminate your problem by running a local name server.
Set it up to resolve certain domains like the tld bob. or platapus. and
forward all other requests to your ISP's name servers. Since your 'fake'
tlds don't exist, they would not be masking any 'real' ones. Or you could
have one 'fake' tld like local. and put that in your search list. That way
if you tried to resolve bob, it would append local on it and try to resolve
bob.local. and your local server would correctly return an IP address. ( I
quote fake and real in relation to top level domains - tlds - because the
ones we have now, though not completely arbitrary, are still more
restrictive than necessary. I'm sure they were set up in an environment that
assumed our networks would not be quite so )
If you use Bind 9, I could probably give you a good idea of what your
config could be on your internal name server.

That was one solution that crossed my mind... been wanting to run a nameserver
anyway just to familiarize myself with the whole process, but I haven't had
the time and it doesn't look like I will anytime soon.

Another thing: deleting the default route (the one to the server) in the
routing table also eliminates the problem. But that of course needs root
priviledges, which is not something that's easy to automate (securely). Oh
well, it's really not that huge a problem in the long run. I can always access
my machines via IP if I'm in a hurry.

Thanks Ed, and everyone else who has offered suggestions.

--Charles Ulrich

[M. Warner Losh]

In message: <002901c2f3d4$68338b60$***@EdLaptop>
"Ed Heron" <***@heron-ent.com> writes:
: Isn't there supposed to be somplace where the DNS resolv order would be?
: It sounds like this machine is using DNS to resolve and timing out before
: using the hosts file. I looked at the man page for resolv.conf (under
: GNU/Linux) and didn't see an order option there. Shouldn't we be able to
: set the order to hosts then DNS?

/etc/nsswitch.conf

Warner

[Alex Rice]

Post by wes robbins
how do you stop a bsd box from having a
nightmare'ishly long connect time for ssh.
I did not test any other ports....

It could also be configuration on the sshd you are connecting to. See
options -i, -k for possible slowdowns.

Alex Rice, Software Developer
Architectural Research Consultants, Inc.
***@ARCplanning.com
***@swcp.com

[Steve Browne]

Post by Ken Long
I wonder why they jumped from 8.0 all the way to 9? 8.0 hasn't been out
all that long has it?
Ken

Because Mandrake jumped to 9 and is about to release 9.1. The higher
the number the better it must be, you know...

Steve
Stephen B. Browne
***@ix.netcom.com
"Veritas contemno captio."

[James Hamilton]

Isn't it all about the numbers when you talk to the masses?

Post by Ken Long
I wonder why they jumped from 8.0 all the way to 9? 8.0 hasn't been out
all that long has it?
Ken

Post by Eric "Ronin" Krieger
Just and fyi for you red hat-ers out there...

_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

[smbinyon]

8.0 was an abomination, so they went to 9.0. But with the so many
updates in oss, it would be difficult to say if what they have done
warrants a +1.0. It could be a 10 or just shy of. Maybe look it at this
way: since the last time they released 8.0 (think it was about 6 mths
ago) does oss deserve a whole point or just a 10th of a point. Or they
just put 1.0 worth of updates into this release and save the 0.1 for
9.1. Actually the marketing department took over Configuration
Management and this is what happens. They have to sell and make money.

A good discussion about the effect of RHCE on this was interesting in
the RH list. This makes me wonder the worth of those Certs, with all the
updates, the value in updating the cert would be tax right off for sure.

What kernel are they going to use?

~smbinyon
redHat User.

Post by Ken Long
I wonder why they jumped from 8.0 all the way to 9? 8.0 hasn't been out
all that long has it?
Ken

Post by Eric "Ronin" Krieger
Just and fyi for you red hat-ers out there...

_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

[Brian Gonzales]

An abomination? I thought it was a nice upgrade from 7.x.

Post by smbinyon
8.0 was an abomination, so they went to 9.0.

[John Fleck]

Post by Brian Gonzales
An abomination? I thought it was a nice upgrade from 7.x.

Post by smbinyon
8.0 was an abomination, so they went to 9.0.

Yeah, I'd like to see the justification for that claim as well. I
thought RH 8.0 worked well, and showed some significant improvements in
desktop consistency and usability over previous versions.

Cheers,
John

[Brian Gonzales]

With RH8, I got better USB, and a nice installer that ID'd my graphics
card. I like my new Debian install, but it took 3 times to get it right.
:)

1st method: blind install, no instructions
2nd: installed too much crap
3rd: Installed just the stuff I like (except Galeon, still working
that).

Post by John Fleck

Post by Brian Gonzales
An abomination? I thought it was a nice upgrade from 7.x.

Post by smbinyon
8.0 was an abomination, so they went to 9.0.

Yeah, I'd like to see the justification for that claim as well. I
thought RH 8.0 worked well, and showed some significant improvements in
desktop consistency and usability over previous versions.
Cheers,
John

[jdavis]

I switched distros because of rh8.
First, theres that stupid redhat on the start bar in gnome. Then, they
removed Xconfigurator or renamed it.. And they went with apache2.
I used to have a list of gripes...but the anger is passing and the
healing process has begun...with Debian :)

I think that RH7.3 was RH best effort...and Gnomes too(1.4). Then new
gnome2 i feel is slow and kinda buggy, plus by default it wants to use
metacity wich is way less configurable and does not have edge flipping
for desktops.

Alot of my dislike for RH8 has to do with gnome2. But now seeing Gnome2
on a debian box with out that ugly red fedoria..Im warming up to gnome2.

$.02,
jd

Post by Brian Gonzales
With RH8, I got better USB, and a nice installer that ID'd my graphics
card. I like my new Debian install, but it took 3 times to get it right.
:)
1st method: blind install, no instructions
2nd: installed too much crap
3rd: Installed just the stuff I like (except Galeon, still working
that).

Post by John Fleck

Post by Brian Gonzales
An abomination? I thought it was a nice upgrade from 7.x.

Post by smbinyon
8.0 was an abomination, so they went to 9.0.

Yeah, I'd like to see the justification for that claim as well. I
thought RH 8.0 worked well, and showed some significant improvements in
desktop consistency and usability over previous versions.
Cheers,
John

[John Fleck]

Post by jdavis
Alot of my dislike for RH8 has to do with gnome2. But now seeing Gnome2
on a debian box with out that ugly red fedoria..Im warming up to gnome2.

FYI, the Fedora on the menu button was easily changeable. :-)

Cheers,
John

[jdavis]

Post by John Fleck

Post by jdavis
Alot of my dislike for RH8 has to do with gnome2. But now seeing Gnome2
on a debian box with out that ugly red fedoria..Im warming up to gnome2.

FYI, the Fedora on the menu button was easily changeable. :-)
Cheers,
John

in KDE yes... but could you tell me how, for fun, in gnome

[John Fleck]

IIRC (I'm at work now and can't check) it's in /usr/share/pixmaps - you just
sub in whatever image you want.

Cheers,
John

P.S. You can also easily sub in Sawfish for Metacity.

John Fleck, science writer
Albuquerque Journal, PO Drawer J, Albuquerque NM, 87103
[phone]
***@abqjournal.com (w), ***@inkstain.net (h)
http://www.abqjournal.com/scitech/, http://www.inkstain.net/fleck/

"Always be amazed."
- Fritz Thompson
----- Original Message -----
From: "jdavis" <***@taproot.bz>
To: "NMLUG" <***@nmlug.org>
Sent: Thursday, March 27, 2003 9:05 AM
Subject: Re: [NMLUG] [Fwd: Red Hat Linux 9 -- Get it Early]

Post by jdavis

Post by John Fleck

Post by jdavis
Alot of my dislike for RH8 has to do with gnome2. But now seeing Gnome2
on a debian box with out that ugly red fedoria..Im warming up to gnome2.

FYI, the Fedora on the menu button was easily changeable. :-)
Cheers,
John

in KDE yes... but could you tell me how, for fun, in gnome
--
jd
Bad spellers of the world untie!
"I can't tell if I have worked all my life or
if I have never worked a single day of my life"
Miguel de Icaza
_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

[jdavis]

Post by John Fleck
IIRC (I'm at work now and can't check) it's in /usr/share/pixmaps - you just
sub in whatever image you want.
Cheers,
John
P.S. You can also easily sub in Sawfish for Metacity.
John Fleck, science writer
Albuquerque Journal, PO Drawer J, Albuquerque NM, 87103
[phone]
http://www.abqjournal.com/scitech/, http://www.inkstain.net/fleck/

thats not what i would call easy..redhat doesnt want you doing that...
and i tried subing sawfish for MetaCity and it was really really slow.
Dont know why, but it was.

[James Hamilton]

Post by Brian Gonzales
With RH8, I got better USB, and a nice installer that ID'd my graphics
card. I like my new Debian install, but it took 3 times to get it right.
:)
1st method: blind install, no instructions
2nd: installed too much crap
3rd: Installed just the stuff I like (except Galeon, still working
that).

Interesting, once the system is up if you have a broadband connection of some
sort you should be able to recover from just about any problem. Reinstall ==
bad :-)

Post by Brian Gonzales

Post by John Fleck

Post by Brian Gonzales
An abomination? I thought it was a nice upgrade from 7.x.

Post by smbinyon
8.0 was an abomination, so they went to 9.0.

Yeah, I'd like to see the justification for that claim as well. I
thought RH 8.0 worked well, and showed some significant improvements in
desktop consistency and usability over previous versions.
Cheers,
John

--
_______________________________________________
NMLUG mailing list
http://www.nmlug.org/mailman/listinfo/nmlug

[smbinyon]

Post by John Fleck

Post by Brian Gonzales
An abomination? I thought it was a nice upgrade from 7.x.

Post by smbinyon
8.0 was an abomination, so they went to 9.0.

Yeah, I'd like to see the justification for that claim as well. I
thought RH 8.0 worked well, and showed some significant improvements in
desktop consistency and usability over previous versions.
Cheers,
John

Well, I'm happy that the email got out and I'll get rid of my nasty
habit of top-posting...

RH8.0 took the big step and decided to take the best of Gnome and the
best of KDE and put into one. Of course one must determine what defines
'best'. This caused some strain in the community, ether you liked or
you didn't. There were those who thought RH was now promoting just one
type of desktop and therefore hurting the KDE and/or Gnome projects.
(There was an article about this, but one can do a google search on Red
Hat Bluecurve and see the negative and positive responses). As a regular
lurker on the Red Hat list, there was even a few comments on Bluecurve
(+/-'s). So from the book Dune, to some Bluecurve was the Alia of
desktops. I thought Alia was one of the hero's of the Dune.

~smbinyon

[Alex Rice]

Post by smbinyon
Actually the marketing department took over Configuration
Management and this is what happens. They have to sell and make money.

I read a good article about why redhat was justifying the jump to 9.0
from a new-features standpoint. The main thing is a new permissions &
ACL features, and metadata filesystem to support it. This ACL stuff is
apparently is a big deal. As if chmod-ing your files with an octal code
is not powerful enough!

Sorry, lost the URL of the article.

Alex Rice, Software Developer
Architectural Research Consultants, Inc.
***@ARCplanning.com
***@swcp.com

[Aaron Birenboim]

Post by Alex Rice

Post by smbinyon
Actually the marketing department took over Configuration
Management and this is what happens. They have to sell and make money.

I read a good article about why redhat was justifying the jump to 9.0
from a new-features standpoint. The main thing is a new permissions &
ACL features, and metadata filesystem to support it. This ACL stuff is
apparently is a big deal. As if chmod-ing your files with an octal code
is not powerful enough!

seems like a technology a bit behind the times.
Its not as needed these days, where you rarely have
a large number of users on a system.

However, for people with Fileservers, it can be a real benefit.
Especially when you are working projects with many people.
I often lament the loss of the ACL stuff we had on VMS.
I guess NT has ACL features, but I rarely see them used.