| View previous topic :: View next topic |
| Author |
Message |
Sith_Happens
Veteran
Joined: 15 Dec 2004
Posts: 1807
Location: The University of Maryland at College Park
|
 Posted: Thu Jun 16, 2005 12:35 am Post subject: Re: white menues in firefox |
 |
|
| widu wrote: | | Bob P wrote: | i'm running jackass, kde 3.3 and the latest versions of firefox and thunderbird, and i don't have the problem.  |
The silly problem was them related (gnome-2.10). So no real need to recompile all the stuff. |
Ah, gtk quirks. Sounds about right. Although the same problem could occur in KDE since no matter what firefox uses gtk.
_________________
"That question was less stupid; though you asked it in a profoundly stupid way."
I'm the brains behind Jackass! | Tutorials: Shorewall |
|
| Back to top |
|
 |
vipernicus
Veteran
Joined: 17 Jan 2005
Posts: 1462
Location: Your College IT Dept.
|
|
| Back to top |
|
|
scharkalvin
Guru
Joined: 31 Jan 2004
Posts: 331
Location: south florida
|
| Posted: Thu Jun 16, 2005 3:00 pm Post subject: K6 on a Pentium tarball? |
|
|
I have a K6-300 machine.
Can I use the Pentium iso image and after loading the tarball
safely change the -march= and -mtune= settings to "K6"?
Granted I would have to re-emerge the system TWICE after that
to get any improvement in the toolset, but at least my kernel and any apps I build
will be specific to the K6 (i think the K6 adds some better pipelining and
mmx support over the Pentium. Actually maybe I should start with a
pentium-mmx iso image?) |
|
| Back to top |
|
|
Sith_Happens
Veteran
Joined: 15 Dec 2004
Posts: 1807
Location: The University of Maryland at College Park
|
| Posted: Thu Jun 16, 2005 6:01 pm Post subject: Re: K6 on a Pentium tarball? |
|
|
| scharkalvin wrote: | I have a K6-300 machine.
Can I use the Pentium iso image and after loading the tarball
safely change the -march= and -mtune= settings to "K6"?
Granted I would have to re-emerge the system TWICE after that
to get any improvement in the toolset, but at least my kernel and any apps I build
will be specific to the K6 (i think the K6 adds some better pipelining and
mmx support over the Pentium. Actually maybe I should start with a
pentium-mmx iso image?) |
Sure, you could do that without any problems. I'm not sure what tarball would be best to start with from a K6, but something tells me that at 300 Mhz, a PII is probably the best place to start. Post the output of cat /proc/cpuinfo though and we'll see.
_________________
"That question was less stupid; though you asked it in a profoundly stupid way."
I'm the brains behind Jackass! | Tutorials: Shorewall |
|
| Back to top |
|
|
vipernicus
Veteran
Joined: 17 Jan 2005
Posts: 1462
Location: Your College IT Dept.
|
|
| Back to top |
|
|
96140
Retired Dev
Joined: 23 Jan 2005
Posts: 1324
|
| Posted: Thu Jun 16, 2005 7:51 pm Post subject: |
|
|
| Just for a little while. Please be patient. I'm positive it'll be back up and running before too long; there's probably maintenance being done. |
|
| Back to top |
|
|
madmango
Guru
Joined: 15 Jul 2003
Posts: 507
Location: PA, USA
|
| Posted: Thu Jun 16, 2005 8:05 pm Post subject: |
|
|
Hm. Still down. 
_________________
word. |
|
| Back to top |
|
|
Sith_Happens
Veteran
Joined: 15 Dec 2004
Posts: 1807
Location: The University of Maryland at College Park
|
| Posted: Thu Jun 16, 2005 8:07 pm Post subject: |
|
|
| madmango wrote: | | Hm. Still down. |
I turned over server duties to Bob at the end of May, so I've sent him a PM about the server being down. Just give him some time and I'm sure he'll figure out whats wrong and fix it ASAP.
_________________
"That question was less stupid; though you asked it in a profoundly stupid way."
I'm the brains behind Jackass! | Tutorials: Shorewall |
|
| Back to top |
|
|
vipernicus
Veteran
Joined: 17 Jan 2005
Posts: 1462
Location: Your College IT Dept.
|
| Posted: Thu Jun 16, 2005 8:52 pm Post subject: |
|
|
| Sith_Happens wrote: | | madmango wrote: | | Hm. Still down. |
I turned over server duties to Bob at the end of May, so I've sent him a PM about the server being down. Just give him some time and I'm sure he'll figure out whats wrong and fix it ASAP. |
Also, is there any way you guys could have an HTML version of the guide made up?
_________________
Viper-Sources Maintainer || nesl247 Projects || vipernicus.org blog |
|
| Back to top |
|
|
96140
Retired Dev
Joined: 23 Jan 2005
Posts: 1324
|
| Posted: Thu Jun 16, 2005 9:01 pm Post subject: |
|
|
| That's been discussed previously in the Jackass! threads and in the Stage 1/3 Guide threads. I would suggest you look there for explanations straight from Bob P. |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Thu Jun 16, 2005 10:09 pm Post subject: |
|
|
| vipernicus wrote: | | is there an html version of the jackass installation guide? that way I can use links2 as a reference. |
no HTML manual has been made for Jackass! because the Jackass! tarballs behave exactly like generic Gentoo tarballs. if you want to use an HTML-based manual, just use the Gentoo Installation Handbook, as Jackass! is 100% Gentoo-compatible. 
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Thu Jun 16, 2005 10:15 pm Post subject: |
|
|
| Sith_Happens wrote: | | madmango wrote: | | Hm. Still down. |
I turned over server duties to Bob at the end of May, so I've sent him a PM about the server being down. Just give him some time and I'm sure he'll figure out whats wrong and fix it ASAP. |
well, the server problem was a simple one: the DSL cable had come loose and this caused the modem to DHCP into a new IP address.  as it turns out, the server was up, its just that the DNS servers didn't know where to find it.
the problem is fixed now. 
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Thu Jun 16, 2005 10:49 pm Post subject: Re: K6 on a Pentium tarball? |
|
|
| Sith_Happens wrote: | | scharkalvin wrote: | I have a K6-300 machine.
Can I use the Pentium iso image and after loading the tarball
safely change the -march= and -mtune= settings to "K6"?
Granted I would have to re-emerge the system TWICE after that
to get any improvement in the toolset, but at least my kernel and any apps I build
will be specific to the K6 (i think the K6 adds some better pipelining and
mmx support over the Pentium. Actually maybe I should start with a
pentium-mmx iso image?) |
Sure, you could do that without any problems. I'm not sure what tarball would be best to start with from a K6, but something tells me that at 300 Mhz, a PII is probably the best place to start. Post the output of cat /proc/cpuinfo though and we'll see. |
The K-6 chips are 586 CHOST boxes, so don't use a Pentium II tarball or you'll run into problems. The best tarball to start out with would be the Pentium-MMX tarball.
The Jackass! webserver actually runs on a K6-266 that uses the Jackass! Pentium-MMX tarball. The kernel is even configured for a Pentium-MMX CPU.
I kind of hate to say it, but having done it, I don't think that it was worth all of the effort to rebuild a Pentium-MMX version of Jackass! into a K-6 version of Jackass! I say this for a couple of reasons:
1. PERFORMANCE
In spite of the theoretical arguments, my personal experience has been that rebuilding the toolkit for K-6 doesn't offer enough of an improvement over the toolkit for Pentium-MMX to make the effort worthwhile. In spite of the theoretical arguments, real world performance doesn't meet the expectations that have been brought about by the K-6 hype. In my case, I'm using my box in a server role, and not as a desktop box in a window manager type of environment.
In a server role, rebuilding for K-6 offered NO measurable improvements over the build for Pentium-MMX.
Theoretically, rebuilding might offer a minimal incremental improvement if you're trying to turn that old KD-300 into a multimedia box, but the advantages would be minimal. If you're going to be banging on that old K6-300 to try to make it do multimedia, it might be worth the effort, but I doubt it.
2. RELIABILITY
K-6 toolkits are finicky, and just aren't as reliable as the Pentium toolkits. I hate to say that too, but the truth is that I ran into ALOT of problems building K6 boxes for Jackass! and when developing the Stage 1/3 Install, and all of these problems VANISHED when I ran Pentium-MMX toolkits instead. As a result, I always run Pentium-MMX toolkits and kernel configurations on my K6 boxes.
3. EFFORT
Rebuilding on old hardware takes a very long time. Personally, I don't see that its worth the effort of going through the trouble of recompiling for days on end, only to get a system that's got a theoretically marginal performance improvement at the cost of decreased reliability.
Just my two cents... 
Aside: this should explain why we didn't bother to build AMD-specific versions of Jackass! for the antique AMD boxes. IMHO the best bang for the buck is to just run with the Pentium-MMX Jackass! tarball. OTOH, if you really want to do it, I'd recommend a Stage 1/3 Install for your K6-box, or just rebuilding from the Pentium-MMX Jackass! installation.
Good Luck!
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
scharkalvin
Guru
Joined: 31 Jan 2004
Posts: 331
Location: south florida
|
| Posted: Fri Jun 17, 2005 1:51 am Post subject: |
|
|
| Quote: | | I kind of hate to say it, but having done it, I don't think that it was worth all of the effort to rebuild a Pentium-MMX version of Jackass! into a K-6 version of Jackass! |
Looking at the GCC documentation, the K6 and Pentium-MMX are real close.
I think the K6 is actually a better performer than the Pentium, running
Pentium code, so I probably would just install the P-MMX tarball and
leave it at that (though I might change the flags to K6 and see how
the kernel runs compiled that way). |
|
| Back to top |
|
|
Sith_Happens
Veteran
Joined: 15 Dec 2004
Posts: 1807
Location: The University of Maryland at College Park
|
| Posted: Fri Jun 17, 2005 1:36 pm Post subject: |
|
|
| scharkalvin wrote: | | Quote: | | I kind of hate to say it, but having done it, I don't think that it was worth all of the effort to rebuild a Pentium-MMX version of Jackass! into a K-6 version of Jackass! |
Looking at the GCC documentation, the K6 and Pentium-MMX are real close.
I think the K6 is actually a better performer than the Pentium, running
Pentium code, so I probably would just install the P-MMX tarball and
leave it at that (though I might change the flags to K6 and see how
the kernel runs compiled that way). |
Read this if you want to compile the kernel with your own cflags.
_________________
"That question was less stupid; though you asked it in a profoundly stupid way."
I'm the brains behind Jackass! | Tutorials: Shorewall |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Fri Jun 17, 2005 5:08 pm Post subject: |
|
|
| scharkalvin wrote: | | Quote: | | I kind of hate to say it, but having done it, I don't think that it was worth all of the effort to rebuild a Pentium-MMX version of Jackass! into a K-6 version of Jackass! |
Looking at the GCC documentation, the K6 and Pentium-MMX are real close.
I think the K6 is actually a better performer than the Pentium, running
Pentium code, so I probably would just install the P-MMX tarball and
leave it at that (though I might change the flags to K6 and see how
the kernel runs compiled that way). |
IIRC, when i compared the cpuinfo output on the two chips, the significant difference amounted to the addition of 3Dnow instructions for the K6, and that was about it.
just to elaborate a bit on my other post: the fact that I had problems with some ebuilds when I did the K6 builds a few months ago doesn't mean that those problems still exist. for all i know, they could have been fixed in the portage tree and everything might be fine now. but having given up on K6-specific toolkits a few months ago, my information could very well be out of date. its entirely possible that the problems aren't problems any more, and that you could be completely successful in doing the build.
but i still think that a complete K6 system rebuild from Pentium-MMX would take a lot of time, and that the narrow differences between the two instruction sets isn't likely to give much payoff.
so if you have the time and you want to try an experiment, go for it. you might find out that all of the problems are fixed, and that things actually work better for you. that would be good to know. even if the project fails, you can still fall-back to reinstalling the Jackass! Pentium-MMX toolkit., and you'll only have to do the lengthy compilation once instead of twice, like you'd have to do with a Stage 3 install..
good luck.
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
scharkalvin
Guru
Joined: 31 Jan 2004
Posts: 331
Location: south florida
|
| Posted: Sat Jun 18, 2005 1:40 pm Post subject: |
|
|
| Quote: | | IIRC, when i compared the cpuinfo output on the two chips, the significant difference amounted to the addition of 3Dnow instructions for the K6, and that was about it. |
Actually only the K6-2 and K6-3 have 3dnow. The K6 is just MMX.
My machine is the older, 66mhz FSB K6.
Story:
K6 was a better Pentium clone with some of the PentiumPro's advanced pipelining
added, but NOT compatible with the PP. Early specs on the K6 power requirements
matched the Pentium up to 233mhz. BUT the 266 and 300 mhz chips needed
sligthly higer voltages. SO most mb's wouldn't work with the fastest K6 chips.
Azza published some info on their web site on how to modify their socket 7 mb
to work with the K6-266/300 chips. Like a fool, I took up my soldering iron and
tried to replace two smt resistors and lift a pin on a tiny smt regulator ic to make it
work. I melted part of the cpu socket (close spacing and large soldering iron!) in the
process. I also had to use a binocular microscope to see what I was doing!
BUT IT WORKED!!!!!!!! :- >
The K6-2 increased to FSB from 66 to 100mhz (to match Intel's PII) and got the clock
up to 550mhz. The K6-3 added level 2 cache to the chip (again to match Intel) but
the increased chip size and extra heat made the yeild rather poor. AMD never got
the chip to work better than 450mhz and pulled production as soon as their first
K7 (athlon) chips were available. The K6-2 however remained in production for
at least another year or two.
(can you tell I'm an AMD fan?)
PS still trying to get my Athlon64 up with 2005.0. kde+cups now giving me fits. |
|
| Back to top |
|
|
Stephonovich
n00b
Joined: 20 May 2004
Posts: 72
Location: United States
|
| Posted: Sun Jun 19, 2005 2:51 am Post subject: |
|
|
The portage snapshot in the Athlon-XP release breaks xfsprogs for some reason. It downloads fine, then attempts to re-download from SGI's site, and then dies with the message 'failed to download'. The file is intact, and can untar just fine, if you wish to manually install it. Re-syncing the portage tree fixed this.
_________________
Who needs reincarnation when you've got parallel universes? |
|
| Back to top |
|
|
96140
Retired Dev
Joined: 23 Jan 2005
Posts: 1324
|
| Posted: Sun Jun 19, 2005 3:46 am Post subject: |
|
|
| Are you sure it wasn't just a connectivity problem, either on the distfiles site or on your machine? |
|
| Back to top |
|
|
Stephonovich
n00b
Joined: 20 May 2004
Posts: 72
Location: United States
|
| Posted: Sun Jun 19, 2005 3:56 am Post subject: |
|
|
| nightmorph wrote: | | Are you sure it wasn't just a connectivity problem, either on the distfiles site or on your machine? |
I'm fairly certain it wasn't a problem with the sites. I tried manually getting the files with wget and it worked fine, from multiple servers, including SGI's. It could have been a problem with my machine, I suppose. There is this, which may or may not be relevant.
_________________
Who needs reincarnation when you've got parallel universes? |
|
| Back to top |
|
|
96140
Retired Dev
Joined: 23 Jan 2005
Posts: 1324
|
| Posted: Sun Jun 19, 2005 4:12 am Post subject: |
|
|
| After reading the other thread you provided, it seems that there is some possibility that someone upstream screwed up that XFS ebuild. If someone else's mistake did end up in a Jackass! snapshot....I'll do some further investigation, but Bob P and Sith_Happens are the actual tarball maintainers. You might be the first Jackass! installer to use XFS, or at least the first one I've heard of who runs into this error. At the time, the Jackass! Portage snapshot was chosen because of its overall stability. I would hate to think there was a flaw lurking in it this whole time ... |
|
| Back to top |
|
|
Sith_Happens
Veteran
Joined: 15 Dec 2004
Posts: 1807
Location: The University of Maryland at College Park
|
| Posted: Sun Jun 19, 2005 12:30 pm Post subject: |
|
|
| nightmorph wrote: | | After reading the other thread you provided, it seems that there is some possibility that someone upstream screwed up that XFS ebuild. If someone else's mistake did end up in a Jackass! snapshot....I'll do some further investigation, but Bob P and Sith_Happens are the actual tarball maintainers. You might be the first Jackass! installer to use XFS, or at least the first one I've heard of who runs into this error. At the time, the Jackass! Portage snapshot was chosen because of its overall stability. I would hate to think there was a flaw lurking in it this whole time ... |
I tested the Pentium MMX tarball with XFS, but I probably emerge --sync'ed before I actually emerged xfsprogs. That would be my suggestion, that or install a new portage snapshot manually from one of the mirrors.
_________________
"That question was less stupid; though you asked it in a profoundly stupid way."
I'm the brains behind Jackass! | Tutorials: Shorewall |
|
| Back to top |
|
|
vipernicus
Veteran
Joined: 17 Jan 2005
Posts: 1462
Location: Your College IT Dept.
|
| Posted: Mon Jun 20, 2005 2:02 pm Post subject: |
|
|
Nice release. I used Gentoo from November04-April05, used the stage1 gentoo default install, really liked portage, but found Gentoo to be slower than most Linux distros, and ditched it for ArchLinux until recently when I tried your Jackass Install method. After some research, I installed your P4 tarball, set my package.keywords to where I couldn't stupidly upgrade my GCC 3.4.3, set up ext3 w/ dir_index&full_journal, and started installing. And I must say my system is now blazing, and I have not had one package fail on me yet. Thank you.
_________________
Viper-Sources Maintainer || nesl247 Projects || vipernicus.org blog |
|
| Back to top |
|
|
Geministorm
n00b
Joined: 10 May 2005
Posts: 50
Location: Melbourne, FL USA
|
| Posted: Mon Jun 20, 2005 2:11 pm Post subject: |
|
|
Bob P.,
Could you add a date to the documentation and downloads on the webpage please?
Thanks.
_________________
Abit NF7-S (Realtek RTL8201, RTL8801b, ALC650, SiL3112a)
1GB (2x 512MB) DDR400, dual-channel
AMD 2500+ (Barton) XP (@10x200)
200GB Maxtor SATA
120GB Maxtor U-ATA
45GB IBM U-ATA
Plextor PX-712 DVD+/-RW
AOpen 16x DVD
OS: Gentoo/WinXP |
|
| Back to top |
|
|
Bob P
Advocate
Joined: 20 Oct 2004
Posts: 3355
Location: Jackass! Development Labs
|
| Posted: Tue Jun 21, 2005 3:27 pm Post subject: |
|
|
| Stephonovich wrote: | | The portage snapshot in the Athlon-XP release breaks xfsprogs for some reason. It downloads fine, then attempts to re-download from SGI's site, and then dies with the message 'failed to download'. The file is intact, and can untar just fine, if you wish to manually install it. Re-syncing the portage tree fixed this. |
our objective in choosing a portage snapshot for the Jackass! tarball construction was to select a portage tree that contained no errors in the files that comprise The Jackass! Toolkit. we made no effort whatsoever to test every package in the portage tree.
we extensively tested Jackass! to assure that it worked properly when you used the packages we provided and when you followed our directions. if you choose to deviate from the directions, well, there's no way for us to predict what packages you'll install or whether or not they will work. i hope that you realize that its just not possible for us to test the entire portage tree.
the purpose of our Jackass! Portage Snapshot is to create a working basic system that would successfully complete a Stage 1/3 Install to build a stable set of Jackass! tarballs. We did extensive testing about that, and Jackass! works.
testing the portage tree to ensure a trouble-free Jackass! toolkit was a monumental undertaking. it required a small army of helpers, and the result is a portage tree that is included on our CD that is GUARANTEED to be free of defects in both the toolkit and in the packages used in the Installation Guide. as the support threads will show, its a truly rare day for Gentoo when the portage tree is totally free of errors for all of the major toolkit components. we've solved that problem with the snapshot that we include on the CD.
if it turns out that your portage snapshot has some other ebuild that fails, well, you've just fallen subject to the slings and arrows of outrageous fortune. blame the ebuild maintainer, and resync to get the new ebuild. but when you do, bear in mind that you should not rebuild your toolkit unless you like to take chances. this is a fine example of why everyone should be using hielvc's update script to update world packages and not toolkit packages.
_________________
.
Stage 1/3 | Jackass! | Rockhopper! | Thanks | Google Sucks |
|
| Back to top |
|
|
|
Unsupported Software
