Re: [OT] how to clean grime off old computer MB?
On Sun, Mar 09, 2008 at 01:44:33PM -0400, Douglas A. Tutty wrote: On Sat, Mar 08, 2008 at 11:05:34PM -0600, Russell L. Harris wrote: I have a couple of new-to-me old computers. They've been well used in what looks like a normal office environment and they're a bit grimey inside; not just dust that blows away. I figure that I should clean that off so the dust doesn't act like a thermal insulator but I'm unsure what to use, since air alone isn't doing it. I don't want to remove e.g. the CPU from its socket. (P-133, socket 7). You need to remove all socketed devices, and especially the CPU. Cleaning presents a danger to the aluminum electrolytic capacitors. Some chemicals quickly penetrate the seals and ruin the capacitors. Yeah, that's my concern. Boards are usually cleaned of flux after the soldering with circuit-board cleaner, which presumably is safe for the newly-soldered caps. Newly manufactured boards are cleaned with Hydrogen Hydroxide. This has been the standard board cleaning agent for quite a while. As it isn't very effective on grease, it doesn't help you much, though it wouldn't hurt to try. -- Rob signature.asc Description: Digital signature
Re: Nice GUI/CLI Password Manager for Linux
On Fri, Jan 25, 2008 at 01:25:15PM -0500, Joey Hess wrote: Amit Uttamchandani wrote: Recently moved from Mac to Debian Linux. I am looking for a nice and powerful FLOSS password manager similar to Keychain on Mac OS X. I preferably would want a CLI tool...so I could remote login using SSH and look at some passwords that I have forgotten. vim + gpg vim can be configured to automatically use gpg to decrypt *.gpg files when they're read and re-encrypt thenm when saving. The decrypted data never touches the disk (though encrypting your swap partition too wouldn't hurt). Dump the following in your .vimrc: Transparent editing of gpg encrypted files. By Wouter Hanegraaff [EMAIL PROTECTED] augroup encrypted au! First make sure nothing is written to ~/.viminfo while editing an encrypted file. autocmd BufReadPre,FileReadPre *.gpg set viminfo= We don't want a swap file, as it writes unencrypted data to disk autocmd BufReadPre,FileReadPre *.gpg set noswapfile Switch to binary mode to read the encrypted file autocmd BufReadPre,FileReadPre *.gpg set bin autocmd BufReadPre,FileReadPre *.gpg let ch_save = ch|set ch=2 autocmd BufReadPost,FileReadPost*.gpg '[,']!gpg --decrypt 2 /dev/null Switch to normal mode for editing autocmd BufReadPost,FileReadPost*.gpg set nobin autocmd BufReadPost,FileReadPost*.gpg let ch = ch_save|unlet ch_save autocmd BufReadPost,FileReadPost*.gpg execute :doautocmd BufReadPost . expand(%:r) Convert all text to encrypted text before writing autocmd BufWritePre,FileWritePre*.gpg '[,']!gpg --default-key=80BF97AA --default-recipient-self -ae 2/dev/null Undo the encryption so we are back in the normal text, directly after the file has been written. autocmd BufWritePost,FileWritePost*.gpg u augroup END I had trouble with this occasionally, probably when the encrypted file had invalid byte sequences for the charset. The unencrypted file had only 7 bit characters. Symptom was inability to read the .gpg file written from within vim. Note the 'set bin' pair for the write path too. augroup encrypted au! First make sure nothing is written to ~/.viminfo while editing an encrypted file. autocmd BufReadPre,FileReadPre,BufNewFile *.gpg,*.asc set viminfo= We don't want a swap file, as it writes unencrypted data to disk. autocmd BufReadPre,FileReadPre,BufNewFile *.gpg,*.asc set noswapfile Switch to binary mode to read the encrypted file. autocmd BufReadPre,FileReadPre *.gpg set bin autocmd BufReadPre,FileReadPre *.gpg,*.asc let ch_save = ch|set ch=2 autocmd BufReadPost,FileReadPost*.gpg,*.asc \ '[,']!sh -c 'gpg --decrypt 2 /dev/null' Switch to normal mode for editing autocmd BufReadPost,FileReadPost*.gpg set nobin autocmd BufReadPost,FileReadPost*.gpg,*.asc let ch = ch_save|unlet ch_save autocmd BufReadPost,FileReadPost*.gpg,*.asc \ execute :doautocmd BufReadPost . expand(%:r) Convert all text to encrypted text before writing autocmd BufWritePre,FileWritePre*.gpg set bin autocmd BufWritePre,FileWritePre*.gpg \ '[,']!sh -c 'gpg --default-recipient-self -e 2/dev/null' autocmd BufWritePre,FileWritePre*.asc \ '[,']!sh -c 'gpg --default-recipient-self -e -a 2/dev/null' Undo the encryption so we are back in the normal text, directly after the file has been written. autocmd BufWritePost,FileWritePost *.gpg,*.asc u autocmd BufWritePost,FileWritePost *.gpg set nobin augroup END -- Rob signature.asc Description: Digital signature
Re: Movies, household network and 54g limits... (maybe...)
On Fri, Jan 11, 2008 at 10:03:41AM -0800, johnny wrote: The graphs say it is clearly not bandwith fault. Reading around: It is well known that the medium access control (MAC) layer is the main bottleneck for the IEEE 802.11 wireless LANs. In particular, the hidden transmitter problem really bites 802.11b/g. If computer A , Computer B, and access point P can see each other, things work better. If A and B can see P, but not each other, problems arise. A starts transmitting. B doesn't see A, and also starts transmitting. P gets junk because there are overlapping transmissions. Neither A nor B see the collision, and have to wait for a timeout before retrying. This is a large window, because B could start transmitting any time during A's packet. This is also a problem when A can see B, and they start transmitting at the same time, but these occurrences are less frequent since the window is much smaller. -- Rob signature.asc Description: Digital signature
Re: Compatibility with ASUS P5E-VM HDMI
On Thu, Dec 27, 2007 at 07:55:06PM -0500, Douglas A. Tutty wrote: On Mon, Dec 24, 2007 at 12:17:46PM -0700, Rob Sims wrote: On Thu, Dec 20, 2007 at 10:19:21PM -0500, Douglas A. Tutty wrote: On Thu, Dec 20, 2007 at 02:02:46PM -0500, Patrick Zaloum wrote: I am shopping components for a (will-be) Debian server and am currently considering the P5E-VM HDMI from Asus with the G35 chipset. Detailed info can be found here http://usa.asus.com/products.aspx?l1=3l2=11l3=584l4=0model=1912modelmenu=1 Could anyone please help me determine if I might encouter compatibility issues with linux (etch) especially at the level of the SATA RAID controller, video and essentials. As it is a server, i don't particularly care about the sound, hdmi, dual display, firewire.. but the rest yes. No MB SATA RAID controller will be true raid. Other than that, the only people that will know will be people who have tried. Careful with those absolutes. My ASUS P5W DH Deluxe has hardware RAID. I haven't used it under Linux, but as the RAID appears as a single SATA device, it should be compatible. However, the P5E-VM HDMI does appear to be software RAID. If you haven't tried it under Linux, then under what OS does it appear as a single SATA device? As I mentioned in the post, Linux sees the controller, but there are no disks for it to talk to. There's a difference between used and detected. Linux _detects_ something on the SATA port in question, at 3.0Gbps, but eventually finds no drive there, so I can't _use_ the controller. Truly disconnected SATA ports have a different, faster indication to the kernel. In addition, I did some research and found several statements saying the same thing: the hardware RAID is attached to an SATA port. The board uses a Silicon Image 4723 Hardware RAID controller, found here: http://www.siliconimage.com/products/product.aspx?id=64 The product brief is worth reading. From the latest kernel sources (2.6.24-rc6+ - the patch was posted 2007-10-24, http://www.mail-archive.com/[EMAIL PROTECTED]/msg11641.html): /* On ASUS P5W DH Deluxe, the second port of PCI device 00:1f.2 is * hardwired to on-board SIMG 4726. The chipset is ICH8 and doesn't * support PMP and the 4726 either directly exports the device * attached to the first downstream port or acts as a hardware storage * controller and emulate a single ATA device (can be RAID 0/1 or some * other configuration). * * When there's no device attached to the first downstream port of the * 4726, Config Disk appears, which is a pseudo ATA device to * configure the 4726. However, ATA emulation of the device is very * lame. It doesn't send signature D2H Reg FIS after the initial * hardreset, pukes on SRST w/ PMP==0 and has bunch of other issues. * * The following function works around the problem by always using * hardreset on the port and not depending on receiving signature FIS * afterward. If signature FIS isn't received soon, ATA class is * assumed without follow-up softreset. */ I'm guessing that the chip is really the 4723, as there are only two ports on the board and the 4726 supports five, and the referenced bug reports say 4723 also. Either way, it's SATA connected. -- Rob signature.asc Description: Digital signature
Re: Compatibility with ASUS P5E-VM HDMI
On Thu, Dec 20, 2007 at 10:19:21PM -0500, Douglas A. Tutty wrote: On Thu, Dec 20, 2007 at 02:02:46PM -0500, Patrick Zaloum wrote: I am shopping components for a (will-be) Debian server and am currently considering the P5E-VM HDMI from Asus with the G35 chipset. Detailed info can be found here http://usa.asus.com/products.aspx?l1=3l2=11l3=584l4=0model=1912modelmenu=1 Could anyone please help me determine if I might encouter compatibility issues with linux (etch) especially at the level of the SATA RAID controller, video and essentials. As it is a server, i don't particularly care about the sound, hdmi, dual display, firewire.. but the rest yes. No MB SATA RAID controller will be true raid. Other than that, the only people that will know will be people who have tried. Careful with those absolutes. My ASUS P5W DH Deluxe has hardware RAID. I haven't used it under Linux, but as the RAID appears as a single SATA device, it should be compatible. However, the P5E-VM HDMI does appear to be software RAID. My JMB363 works for PATA/SATA; the P5E-VM uses a JMB368, likely compatible. My hardware RAID controller causes a long delay at boot when the RAID (connected to an SATA port on the ICH7) is probed. The Atheros ethernet has got to be better than the Marvell I've got, though the atl1 driver is relatively new. Everything else appears to be pretty vanilla. -- Rob signature.asc Description: Digital signature
Re: Why ext3 doesn't need defragmentation ?
On Fri, Nov 30, 2007 at 02:53:32PM -0800, Alvin Oga wrote: David Brodbeck wrote: On Nov 30, 2007, at 9:45 AM, Stefan Monnier wrote: No. The NTFS file system does not need defragmentation. all file systems can use a defragmentor ...description of interleaving deleted... the defragmentor can be used to move sectors around to optimize reading the whole file w/o waiting for the next revolution This is called interleaving, and has nothing to do with fragmentation. Defragmentors do nothing about interleave. If one were to interleave at the filesystem level, performance would quickly get poor because of the huge number of pointers to chunks of the file. Interleaving is done at the hardware format level, and because the controller is integrated with the media, is completely hidden from the host. Note that in a sequential read case, performance would be half or worse for interleaved sectors, as you'd suffer through at least two rotations to read each track. - how the defragmentor displays used and unused sectors can make a big difference in the pretty pic you see vs the actual performance what you see the defragmentor showing would be a continuously allocated file instead of scattered across various sectors within a track or having to move the heads to a different tract to get to the next 512byte there's only 512bytes per sector 63 sectors per track and any number of cylinders depending on your disk size I'll bet you can't find a hard drive made in the past decade with 63 sectors per track. Hard drives have a number that varies with the circumference of the track. The 63 spt is an artificial construct for backwards compatibility. lba ... maps all the cylinder/heads/sector into other whacky numbers ( lba blocks ) A continuous range from 0 to (# sectors - 1) is whacky? Compared to using physical c/h/s numbers that bear no relation to reality? -- Rob signature.asc Description: Digital signature
Re: How to access RS-232 console from terminal?
On Wed, Oct 03, 2007 at 09:29:40PM -0700, Amit Uttamchandani wrote: I know this may seem so obvious but I have no experience with this at all. Basically I want to access the console interface of a network interface device using RS-232. ... And I usually need to transfer files via the 1K Xmodem protocol. Do I need an external software for this? Have a look at minicom. -- Rob signature.asc Description: Digital signature
Re: DHCPD giving IP to wrong machine
On Thu, Jul 26, 2007 at 04:18:39PM -0700, Clarence W. Robison wrote:
Content-Description: Mail message body
On 26 Jul 2007 at 16:05, Clarence W. Robison wrote:
I have an entry in my dhcp3 dhcpd.conf which says that host xyz with
certain MAC address should receive a fixed ip address. The server does
not respect that entry and gives the IP address to another host with a
different MAC address. I don't quite understand why it, dhcpd, should do
that. Is normal behavior?
OPPS, the message left before I could paste snippets of the conf file.
dhcpd.conf -- # #
Global Options pid-file-name /var/run/dhcpd.pid; lease-file-name
/var/lib/dhcp3/dhcpd.leases; log-facility local1; ignore client-updates;
ddns-update-style none; option domain-name-servers XXX.XXX.XXX.3,
XXX.XXX.XXX.223; default-lease-time 3600; max-lease-time
14400; authoritative; subnet XXX.XXX.XXX.0 netmask
255.255.255.192 { # Default Options
option routersXXX.XXX.XXX.1;
option subnet-mask255.255.255.192;
option domain-name.XX.XXX;
option time-offset-25200; # Mountain Standard Time
option ntp-serversXXX.XXX.XXX.3, XXX.XXX.XXX.58;
range dynamic-bootp XXX.XXX.XXX.22 XXX.XXX.XXX.60;
Here, you tell dhcpd it's OK to hand out 22 - 60 to non-matching hosts,
so it does just that. Fixed address and range assigned addresses should
not overlap.
host xxx {
hardware ethernet 00:13:20:2d:31:d1;
fixed-address XXX.XXX.XXX.22;
}
--
Rob
signature.asc
Description: Digital signature
Re: Losing lines with HPOJ 7410
On Wed, Jul 11, 2007 at 09:35:07AM -0700, Todd A. Jacobs wrote: I have an HP OfficeJet 7410, and am using the hplip drivers with the following lpoptions: Default hpoj/draft PageSize=Letter PrintoutMode=Draft.Gray Duplex=DuplexNoTumble I'm losing lines off the top and bottom. If I try: seq 1 78 | lp numbering starts at 3 at the top, ends at 62 on the bottom, and starts at 65 on the flip side. How can I correct this? I found the text filter's lack of respect for margins annoying also. Here's my config: $ cat /etc/cups/lpoptions Default oj7310 page-top=36 page-bottom=36 page-left=18 page-right=18 cpi=16 lpi=10 wrap=true scaling=100 -- Rob signature.asc Description: Digital signature
Re: (OT) Looking for Cell Phone/PDA Recommendations
On Wed, May 23, 2007 at 09:59:36PM -0700, Mark Phillips wrote: On Wednesday 23 May 2007 20:59, Rob Sims wrote: How about the Wing? I'm reasonably happy with the predecessor, the MDA. You need to download Java if you need it (I did it to try an app, so I know it works). I plan on trying the Wing soon. The manual is on-line under the support tab. According to the manual, Java is now built in. The only missing feature appears to be an IrDA port. Used it once. I looked at the Wing, and it is very thick - really much to thick to carry on a belt and even in a pants pocket. It is also very heavy. From a feature standpoint, it is a super phone/pda. But, IMO, it fails as a portable phone. Just too much to carry around all day. When I put it in my pocket in the store 2 things happended - three sales guys started to follow me and I felt as if I were listing to starboard as I walked around the store, it was so heavy and bulky in my pocket...;-) I use the belt clip case that came with the MDA, which is even thicker than the Wing (1 vs 0.7), and find it not to be an issue. I'm a big guy, though. Another point about phone usage is that without an earpiece, you can hit the screen with your cheek, though that hasn't been a real problem like it was on the iPaq 6315 (talk about bulky...) I also think they took away the touch screen, but I may be wrong on that point. I'm fairly sure it's still there. They are available in T-mobile stores if you want to try one out. Let me know what you think. Whatever you end up with, I can really recommend the Garmin GPS10. It includes software for Windows, Windows Mobile, and Palm. It does a great job of routing, and integrates with the address book. The GPS iself works fine with Linux. If you find something thinner, and not an iPhone, let us know. -- Rob signature.asc Description: Digital signature
Re: (OT) Looking for Cell Phone/PDA Recommendations
On Wed, May 23, 2007 at 04:37:20PM -0700, Mark Phillips wrote: I need to buy a cell phone/pda, and I am looking at the T-Mobile Dash ( http://www.geek.com/hwswrev/pda/dash/index.htm). However, I use Debian, KDE/Gnome and Kontact, so the Windows Mobile OS is not very appealing. 1. Is there a way to synchronize my Kontact contacts, calendar, email, etc. with Windows Mobile devices? You could try synce-kde; I have not tried it. I sync my Windows Mobile contacts calendar to XP. Email is via imaps to my self-hosted server. 2. If not, any recommendations for a phone/pda that will synchronize with Kontact? In addition, I would like to have: * 1.3-2 megapixel camera * wireless ear piece * ability to locally download Java Midlets that I create * compatible with US T-Mobile network * QWERTY keyboard or touchscreen keyboard How about the Wing? I'm reasonably happy with the predecessor, the MDA. You need to download Java if you need it (I did it to try an app, so I know it works). - 2 MP camera (though the 1.3 MP on the MDA is quite poor) - Bluetooth works with all earpieces I tried (Jabra, Motorola, Logitech), Garmin GPS 10, ActiveSync to PC, and using the phone as a modem. 3rd party freeware to redirect system sounds to bluetooth works. Logitech headset best in wind, and cheap. - slide-out keyboard; on-screen keyboard and character recognition as well. - PocketPuTTY for ssh; password auth only. Terminal emulation works with mutt well. - 802.11b (MDA has g enabling registry hack; Wing, b/g) - Display poor in bright light (MDA) - tcpmp makes a great media player - Any mp3 file for a ringtone (per contact, too) I plan on trying the Wing soon. The manual is on-line under the support tab. -- Rob signature.asc Description: Digital signature
Re: PayPal - Limited Access
On Thu, Mar 22, 2007 at 10:49:37AM +0100, googan wrote: Having read all of the above I feel rather foolish to have accepted this limited access email to be true... So, whithout following links I went to the UK version of paypal and logged in as normal... only to face a page that told me that my account had indeed had limited access and that I had to go to the resolution centre and fill out some security details... as I had logged into my account i thout this must be the case so i followed the link and it took me to a page where it was requesting bank info and credit card info... then the penny dropped so now i find myself here after contacting paypal... they said it was a soof/phishing email, but still when i log onto paypal through no link, but typing it in mysef and logging on i still get this limited access page..??? now, Has my system been infected or even my account??? your thoughts please Your DNS may have been hijacked. There are many ways to do this; the current fad is to alter your firewall/access point settings to point to the attacker's DNS servers. Check what DNS servers your router is serving. If that's not it, more information is required: what OS are you running, what browser are you using, what exact URL you typed, etc. -- Rob signature.asc Description: Digital signature
Re: Ethernet card Marvell
On Wed, Feb 28, 2007 at 10:09:30AM +0100, Bruno Costacurta wrote: I have a Marvell Ethernet card (on motherboard) model 88E8053. It is not recognized by Etch (however the CD I tried to install from is rather old +- 1 year). (note: under Fedora 6 it is recognized and works fine with driver sky2) What is the current situation / support about Marvell 88E8053 under Etch ? As my Etch CD is rather old do I need to burn a new one ? If so where can I check which hardware is supported in .iso Etch to avoid unecessary download ? Do I need to make some 'home-computing' like download driver and/or source (via a supported Ethernet card) card to build my own module..etc.. ? You need the sky2 module. The Etch (2.6.18) module should work well enough to get a newer kernel and compile it. To get 2.6.18 in Etch, you will need a newer install disk. Yesterday's daily build definitely has sky2. You will need to obtain a 2.6.21-rc1 kernel or better (or at least a newish kernel + sky2.c and sky2.h version 1.13) because there are some serious bugs in earlier versons. Use make-kpkg (in package kernel-package) to build your kernel and you'll stay inside the Debian package management system. -- Rob Nothing is finished until the paperwork is done. signature.asc Description: Digital signature
Re: Patch for Daylight Savings Time (DST))
On Tue, Jan 30, 2007 at 08:44:19PM -0500, Miles Fidelman wrote: Oleg Verych wrote: 30-01-2007, Oleg Verych [EMAIL PROTECTED] пишет: 30-01-2007, Ron Johnson: On 01/30/07 13:58, William Chipman wrote: Has there been a patch to adjust the start / end date changes for daylight savings time in the US? Which branch are you running? So... does anybody have an answer to the original question? Personally, I'm running Sarge, and am also interested in whether such a patch exists. $ zdump -v /etc/localtime | grep 2007 /etc/localtime Sun Mar 11 08:59:59 2007 UTC = Sun Mar 11 01:59:59 2007 MST isdst=0 gmtoff=-25200 /etc/localtime Sun Mar 11 09:00:00 2007 UTC = Sun Mar 11 03:00:00 2007 MDT isdst=1 gmtoff=-21600 /etc/localtime Sun Nov 4 07:59:59 2007 UTC = Sun Nov 4 01:59:59 2007 MDT isdst=1 gmtoff=-21600 /etc/localtime Sun Nov 4 08:00:00 2007 UTC = Sun Nov 4 01:00:00 2007 MST isdst=0 gmtoff=-25200 Just take the current security updates for libc6. -- Rob signature.asc Description: Digital signature
Re: [OT] Dilbert cartoon featurng Linux
On Sat, Jan 27, 2007 at 01:08:20AM +, s. keeling wrote: Max Hyre [EMAIL PROTECTED]: Sven Arvidsson wrote: On Thu, 2007-01-25 at 11:16 -0700, ChadDavis wrote: Has anyone seen the Unix dilbert where a crusty old, Unix guy comes up= to dilbert, flips him a dime, and says,Here kid, go buy yourself a real computer. http://ozguru.mu.nu/Photos/2005-11-11--Dilbert_Unix.jpg :-) Odd, the strip says ``go buy yourself a /better/ computer''. I would have sworn a might oath that the original said ``real computer'', as ChadDavis says. That's how Neil Stephenson quotes it in ``In the Beginning was the Command Line''[1], and I think it's better that way. He's suggesting a nickel would get him a better computer than what he's using. That's a pretty damning indictment. A nickel can't buy a computer in any way, shape or form. Are you kidding? I can get for free (out of a dump) a computer that 30 years ago people would have killed to own; that 50 years ago governments would have killed to own. Of course, 30 or 50 years ago, nobody was running Windows, so let the arguments about better continue... -- Rob p.s - on the original subject, http://www.bfmartin.ca/finder/ really helps. signature.asc Description: Digital signature
Re: random quirkyness
On Tue, Jan 23, 2007 at 04:48:43PM -0600, Mike Myers wrote: On 1/23/07, Andrew Sackville-West [EMAIL PROTECTED] wrote: Mike, please don't cc: me as I subscribe to the list. thanks. I apologize, I'm using gmail and it did that automatically for some reason. On Tue, Jan 23, 2007 at 03:27:40PM -0600, Mike Myers wrote: On 1/23/07, Andrew Sackville-West [EMAIL PROTECTED] wrote: On Tue, Jan 23, 2007 at 02:41:40PM -0600, Mike Myers wrote: Hi all, I'm still trying to adjust from Gentoo's way of doing things (do it manually) to debian's (apt-something) way. So far everything has been great, but i'm having trouble finding docs on a couple of issues I'm having. Both of them seem related to modules. First one is with the nvidia driver. It seems like everytime my debian box is rebooted, I have to re-apt-get nvidia-glx before I can use xorg. Also, GDM doesn't seem to like my 1440x900 (widescreen) resolution and I can't seem to do anything about it other than just not use GDM (not that ditching it is a big deal). I can't speak to the resolution issue, but the xorg issue should not be happening. when you re-apt-get it, does it download it again and appear to be actually reinstalling it? I wonder if your xorg.conf is not getting updated correctly and you are correcting for it by reinstalling each time. how about a copy of your xorg.conf for us to look at as well as dpkg -l | grep nvidia I'm pretty sure it's not related to xorg, since it works fine after running 'apt-get --reinstall install nvidia-glx', even with a widescreen. It's only after a reboot that I must run that, as long as I want to use the nvidia driver. If I use the 'nv' driver, then of course there's no issue there, but that driver sucks. Just to oblige you, here's the contents: (hopefully it looks sane enough to read) so, what exactly happens when you *don't* apt-get install nvidia-glx? what output do you get? It tries to start and then fails, saying it can't find the nvidia module, even though it's loaded. So I get a blank screen. Does it say it can't find the _kernel_ module? Are you using udev? If yes to both, add nvidia to /etc/modules. If this works, the problem is that 'nvidia' is not auto-loaded by the kernel. The device nodes are created by udev upon module insertion. If the device nodes are manually created (by installing the nvidia package, for example), accessing the device node DOES trigger loading of the module. With udev, /dev is on a tmpfs, hence is discarded at shutdown. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT] CPU and GHz
On Mon, Jan 22, 2007 at 10:13:29PM +0100, Misko wrote: I am just curious! Todays computers are running on high clock frequencies. If I have CPU that runs on (let say) 1 GHz what parts of hardware are actually running on this speed? (except crystal :) I understand that memory chips are much slower than CPU so every time that CPU need data that is not in cache it must slow down. I also read somewhere that CPU is internally dividing clock impulses and run on slower speed that it is advertised. So where is this high speed used? Actually, the crystal generally runs more slowly than the CPU clock. Generally, a CPU has a phase-locked loop clock generator on the chip for use by the core logic. A smaller multiple is used for the interface logic. In short, the high frequency is used by the arithmetic units, internal cache, and instruction pipelines, all internal to the CPU chip. The CPU does not normally slow down when a cache miss occurs, it idles until the data comes from memory, but the clock doesn't vary. The CPU can be told to go into a low power state in which the clock frequency is reduced. -- Rob signature.asc Description: Digital signature
Re: local network with twisted rj45
On Thu, Jan 18, 2007 at 04:36:28PM -0500, Evan wrote: As a standard I always purchase red crossover cables. For me, this eliminates the confusion. Many 100 Base-T and most? all? 1000 Base-T support MDI/X; ordinary cables will work to connect two network cards. http://www.iol.unh.edu/services/testing/ge/training/#Auto-Crossover_White_Paper I once bought some straight through cables that turned out to be crossover, and didn't notice until the cables were used for a non-ethernet application... -- Rob signature.asc Description: Digital signature
Re: PCI-E Gb LAN controller support
On Sun, Jan 14, 2007 at 11:15:37AM -0500, Douglas Tutty wrote: I don't understand PCI-E..on-board. To me, either it's an adapter in a PCI-E slot or its integrated on the motherboard. My Asus M2N-SLI Deluxe has two gigabit ethernets on-board integral to the nVidia chipset. Etch installer saw them OK. Almost all peripherals are attached via one of the standard busses, whether or not there's actually a socket in between. On-board Ethernet controllers can be attached by ISA, PCI, PCI-X, PCI-Express, or even USB, though virtually all on-board Ethernet today is PCI or PCI-E. -- Rob signature.asc Description: Digital signature
Re: PCI-E Gb LAN controller support
On Fri, Jan 12, 2007 at 12:05:14PM -0700, Paul E Condon wrote: My new ASUS motherboard has PCI-E Gb LAN controller on-board. I can't get the Etch installer to recognize it. Should I be able to? Or is it too new for Etch? ( I don't really need 1 Gb ethernet, but I would like to be able to use the PCI slot for something else than an old LAN card. ) Using my psychic powers to determine what the chip is, try the sky2 driver in 2.6.18 (somewhat unstable - better in 2.6.19) or the sk98lin driver available from the Marvell or Syskonnect web site. If that doesn't work, provide a motherboard model or lspci output. -- Rob The truth of a proposition has nothing to do with its credibility. And vice versa. signature.asc Description: Digital signature
Re: DVD-RAM questions
On Tue, Jan 09, 2007 at 01:25:20PM -0800, Paul Johnson wrote: UDF was designed for the Write-Once, Read-Many DVD-ROM/RAM/RW. Just use udf. Any compuer with a DVD drive can read udf; ext2/3 is designed for read-write filesystems and only works on Linux. UDF was designed for write once and RW situations, initially for Magneto-Optical drives. Later, sequential write once ability for CD-R was added to the existing random write once ability. The focus of UDF design was interchangeability. There is nothing in UDF that would have better performance characteristics over ext2 on DVD-RAM. You wouldn' be able to use ext2 on -R media, but that's not the issue. Further, the Linux UDF fs is buggy. -- Rob signature.asc Description: Digital signature
Re: Root privilege (SOLVED)
On Tue, Jan 09, 2007 at 02:12:33PM -0800, Paul Johnson wrote: Now, please explain how I can use sudo to tranfer X credentials .. It does this automatically. sudo xproggie It does not: $ sudo xlogo Password: X11 connection rejected because of wrong authentication. It might work if the X app can find and read the .Xauthority file as root (you've set up sudo to retain $HOME, and $HOME is local), but if root can't find .Xauthority (due to environment variable sanitization done by updated sudo) or if root can't read the .Xauthority file ($HOME over NFS), X programs fail. -- Rob signature.asc Description: Digital signature
Re: mounting the minimum
On Wed, Jan 03, 2007 at 01:11:20AM +0200, gustavo halperin wrote:
Hello
I have some kernel-modules questions:
After the linux SO was started, can I check which modules of
/etc/modules was successfully mounted and aren't in use?
Second, if I know the modules that aren't in use, Is OK if I remove
them from /etc/modules and mount them after the boot time using some
script or manually as root ?
Just try rmmod modname; it won't remove modules in use. It's OK to
manually load modules later, but why would you want to?
Last question: there are any way to have a /etc/modules file
corresponding to each kernel and her modules ?
Looking at /etc/init.d/module-init-tools:
KVER=$(uname -r)
KMAJ=${KVER%${KVER#*.*[^.]}}
KMAJ=${KMAJ%.}
if [ -e /etc/modules-$KVER ]; then
MODULES_FILE=/etc/modules-$KVER
elif [ -e /etc/modules-$KMAJ ]; then
MODULES_FILE=/etc/modules-$KMAJ
else
MODULES_FILE=/etc/modules
fi
--
Rob
signature.asc
Description: Digital signature
Re: SCSI hard disk connected to IDE interface ???
On Mon, Dec 25, 2006 at 11:06:26PM -0300, Alejandro wrote: People, I have a Debian PC with a hard disk with the following features: IDE interface ide: 1 description: IDE interface product: 82801FB/FW (ICH6/ICH6W) SATA controller This is a Serial ATA interface, not IDE. Disk: description: SCSI disk product: SAMSUNG SP0812C vendor: ATA physical id: 0.0.0 bus info: [EMAIL PROTECTED]:0.0 logical name: /dev/sda This is a Serial ATA interface, not SCSI. The kernel drivers used are in the SCSI stack. The ata_piix driver is making your SATA controller look like a SCSI card to the higher level SCSI disk driver. The kernel is moving to all IDE/SATA devices behaving this way. Is it possible to have a SCSI disk connected to a IDE interface ??? Not without some horribly hacked black box in between. And if I want to put a seconadry similar disk, do I have to treat it as IDE or SCSI in order to see it on my system ? Physically or virtually? Physically, you need SATA (IDE will likely work too; most x86 motherboards still have an IDE port). On the software side, the kernel will automatically see the new SATA drive with the existing drivers. -- Rob Q:What's tiny and yellow and very, very, dangerous? A:A canary with the super-user password. signature.asc Description: Digital signature
Re: Sid: Unable to ls after update a few minutes ago
On Mon, Dec 18, 2006 at 04:10:38PM -0500, Curt Howland wrote: Up to date Sid, in fact updated about an hour ago. First I noticed that the KDE desktop crashed. A neat skull and crossbones cursor, thanks guys. Anyway, I exited and restarted Xwindows, and KDE refused to come up. Cannot initialize kinit, check your installation (paraphrased from memory). So I start Windowlab, which works fine, Kmail, Konqueror, sure, but then I try to look at the local files: The Process For File Protocol Died Unexpectedly. I open an Xterm, type ls -al, and get the following: - $ ls -al ls: relocation error: /lib/libacl.so.1: symbol getxattr, version ATTR_1.0 not defined in file libattr.so.1 with link time reference - http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=403585 Apperently this fix as not trickled to x86. If you happen to have the old .deb laying around: dpkg -i /var/cache/apt/archives/libattr1_2.4.32-1_i386.deb will fix it. -- Rob signature.asc Description: Digital signature
Re: New hardware under Etch ?
On Sun, Dec 17, 2006 at 11:48:34AM +0200, Tshepang Lekhonkhobe wrote: An off-topic question: why is this Debian kernel limited to 4GB and why do there have to be an option instead of just setting the whole thing to 64GB? There is a slight performance penalty for the 64G support. Another off-topic question: Is there any one out there running Linux on over 64GB of RAM, and if so, is there such support in the kernel or did they have to modify the src themselves? Yes - using a generic kernel, I'm running Debian on a 128GiB box. The limit of 64GiB is a hardware limit for i386, so no kernel options or hacks will help. IA64 supports 1TiB+. Be aware that when using 32 bit code on i386, no process can get more than 3GiB user memory. 32 bit code on x86_64 or IA64 kernels will get close to 4GiB. You need a 64 bit app to get greater than 4G/process. -- Rob signature.asc Description: Digital signature
Re: Etch w/o icedove
On Sat, Nov 25, 2006 at 05:13:16PM -0800, Marc Shapiro wrote: I have Etch installed and have been keeping it up to date with occaisional updates using aptitude. I decided to see what it would want to do if I asked for a dist-upgrade. The answere -- it wants to install icedove. I do not need icedove. I have the upstream firefox 2.0 installed. I also have Debian's 1.5.x package installed, but I am using the upstream 2.0. I certainly have no need for icedove. I have tried 'aptitude hold icedove' 'aptitude hold firefox' but neither made any difference. I then tried 'echo icedove hold | dpkg --set-selections' and 'echo firefox hold | dpkg --set-selections' still with no change. I am quite happy with the firefox that I have and do not want to change that. icedove is the thunderbird package; firefox is packaged as iceweasel. -- Rob signature.asc Description: Digital signature
Re: wired or not wired, that's the question
On Sat, Oct 07, 2006 at 02:29:25AM -0700, wimpunk wrote: CJ van den Berg wrote: On Wed, Oct 04, 2006 at 03:16:42PM -0700, wimpunk wrote: I'm running debian testing on my laptop and I have a question about choosing a configuration. I use my laptop on wired and wireless environments. At this moment I switch manually but I want to get it done automaticly: if there's a wire plugged in (with signal) try to use it and if no signal, try to get wireless up. Has anyone a suggestion on how to do it? aptitude install network-manager-gnome or aptitude install network-manager-kde NetworkManager does exactly what you want out of the box. With a cool GUI and everything. ... I know I should check the doc's but it would be nice if it worked at startup... http://live.gnome.org/NetworkManagerToDo It's targeted for the 0.7 release. -- Rob signature.asc Description: Digital signature
Re: df and du disagree, or How to get 1.1GB of data onto a 650MB CD
On Sat, Sep 23, 2006 at 11:50:53PM +0100, [EMAIL PROTECTED] wrote: Original Message From: Rob Sims [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: Re: df and du disagree, or How to get 1.1GB of data onto a 650MB CD Date: Fri, 22 Sep 2006 16:41:09 -0600 I also have an empty directory: poota:~# ls -l temp/ total 0 Then I mount the iso, loopback style: mount image.iso temp/ -o loop What is the file system? Is it a compressed file system? Also, ls -l won't show you all files in temp; try ls -al, though that isn't the problem here. Hi Rob. The filesystem that provides the mount point is ext3. As for the iso file, file says that it's just ISO 9660 CD-ROM filesystem data. Once the iso is mounted (using loopack), /proc/mounts says /dev/loop0 /root/temp iso9660 ro 0 0 so I assume the iso file really is just straight iso9660. Sounds like there may be some hard linked files. You won't be able to see that directly, but if you find some identical files, I'd suspect they share data space on the disc. mkisofs apparently will create such discs if it detects hard links on the source file system. -- Rob signature.asc Description: Digital signature
Re: df and du disagree, or How to get 1.1GB of data onto a 650MB CD
On Fri, Sep 22, 2006 at 09:14:37AM +0100, [EMAIL PROTECTED] wrote: Hi all. I have a strange problem: I have an iso image file which (according to ls) is 569M in size: poota:~# ls -lh image.iso -rw-r--r-- 1 root root 569M 2006-09-12 15:12 image.iso I also have an empty directory: poota:~# ls -l temp/ total 0 Then I mount the iso, loopback style: mount image.iso temp/ -o loop What is the file system? Is it a compressed file system? Also, ls -l won't show you all files in temp; try ls -al, though that isn't the problem here. -- Rob signature.asc Description: Digital signature
Re: Michelle Konzack's sex (was: Email programs that work.)
On Fri, Sep 08, 2006 at 03:45:20PM +0200, Kim Christensen wrote: On 9/8/06, Derek Martin [EMAIL PROTECTED] wrote: I do know Michelle from mutt-users, where I could swear I'd seen a post (from her) which referred to her as male. I remember specifically thinking that it was odd for a man to have the name Michelle. Obviously I was mistaken... FWIW, I work with a guy named Erin, and another named Kim (not an Asian family name, but a European first name). :-P Hey! My first name is Kim, the middle one's Michelle - and yeah, I'm male So this isn't you: ;) http://9news.com/company/bios/christiansen.asp ...and yes, I noted the difference in the last name. -- Rob Established technology tends to persist in the face of new technology. -- G. Blaauw, one of the designers of System 360 signature.asc Description: Digital signature
Re: Email programs that work.
On Mon, Aug 28, 2006 at 10:13:39PM +0300, Micha Feigin wrote: But I have messages separated into about a dozen folders (I know of people who use more). I want to know which one has new mail, I want to read the mail in the important folders before mailing lists, usually I want to know if I have important mail now and clean up mailing lists at the end of the day, etc. All these work much better with a folders pane. In mutt, 'y' gives me a list of folders and a message count per folder. The list is defined by the mailboxes parameter. When I hit 'c' I am prompted with a default that is the first folder in the mailboxes list with new mail. This lets me address the important folders first. This even works across multiple IMAP accounts. In addition to mailboxes, I have the following set to enable this behavior: set spoolfile=imaps://server1.com/inbox set folder=imaps://server1.com/INBOX. # imap settings set imap_user=user1 set timeout=15 set mail_check=60 # Priority: inbox on server1, inbox on server2, Debian folder on server1, ... mailboxes ! imaps://[EMAIL PROTECTED]/inbox =Debian =Misc alternates [EMAIL PROTECTED] [EMAIL PROTECTED] # Default on sid, not sarge macro index,pager y change-folder?toggle-mailboxes show incoming mailboxes list I usually just hit 'c' and enter, and I'm in the 'right' folder much more quickly than choosing it from a folder list. If I do want the folder list, it is only a keypress away. -- Rob BOFH excuse #23: improperly oriented keyboard signature.asc Description: Digital signature
Re: Recursive copy to udf-formatted dvd-ram freezes system
On Thu, Jul 27, 2006 at 09:06:11AM +0200, Johannes Wiedersich wrote: Eckhard Kosin wrote: The cp -rv showed up a file name with an unprintable character (the file originated from DOS, long time ago). Thanks. I renamed the file and now the system doesn't freeze, but the writer doesn't stop to write (more than 15 min for 293 MB ;-(. No hint from syslog, the only method to stop the writer is to switch it off and after a new switch on no data have been written to the DVD. It seems that UDF is rather unstable, at least in the AMD64 branch. I guess as a 'universal disk format' for all operating systems, it is bound to be less feature rich than something tailored to linux. It was specifically developed with RW-media in mind and probably isn't at the top of fs performance for RAM-media. UDF is incredibly feature rich. The kernel implementation appears buggy; I can crash the kernel using a loopback device. I've intended to go through the code for quite some time now. -- Rob You can observe a lot just by watching. -- Yogi Berra signature.asc Description: Digital signature
Re: Wireless : best policy / pratice to start it
On Wed, Jul 26, 2006 at 04:37:50PM -0600, Rob Sims wrote:
I just discovered network-manager{,-kde} which I now have working on my
...
I did need to add a stanza to /etc/dbus-1/system.d/knetworkmanager.conf:
policy group=netdev
allow own=org.freedesktop.NetworkManagerInfo/
allow send_destination=org.freedesktop.NetworkManager/
allow send_interface=org.freedesktop.NetworkManager/
/policy
...which is corrected in today's sid upload.
--
Rob
signature.asc
Description: Digital signature
Re: Wireless : best policy / pratice to start it
On Mon, Jul 24, 2006 at 08:51:50PM +0200, Bruno Costacurta wrote:
I'm looking for the best policies / practices to start wireless networking.
Briefly I use wireless in following locations:
-home:
wpa via wpa_supplicant
-public (bars..etc..):
no wep/wpa
Both works fine. Currently I start a console logged with user root to call
ifup with related params.
Should I used root to call ifup ?
I would like to keep a manual connection as I do not want to automatically
connect.
I just discovered network-manager{,-kde} which I now have working on my
orinoco based and ipw2200 based laptops with kernel 2.6.17. Makes
managing connections painless. Simply add users who need permission to
manage the interface to the netdev group and run. I don't use WPA (WEP
only), but the package depends on wpasupplicant, so WPA should be
supported.
I did need to add a stanza to /etc/dbus-1/system.d/knetworkmanager.conf:
policy group=netdev
allow own=org.freedesktop.NetworkManagerInfo/
allow send_destination=org.freedesktop.NetworkManager/
allow send_interface=org.freedesktop.NetworkManager/
/policy
...to match the corresponding stanza in NetworkManager.conf. The package is
not in sarge.
Connects automatically to trusted nets, but not untrusted.
--
Rob
signature.asc
Description: Digital signature
Re: transcode package for stable?
On Wed, May 03, 2006 at 12:52:33PM +0100, John Stumbles wrote: Anyone know if there's currently a debian stable package of transcode? ftp://ftp.nerim.net/debian no longer seems to have it. My sources.list has: deb ftp://ftp.nerim.net/debian stable main Didn't the stable/testing/unstable aliases get dumped by Marillat quite some time ago? Replace stable with sarge and retry. -- Rob If not completely satisfied, return for full refund of purchase price. signature.asc Description: Digital signature
Re: Udev rule help
On Sat, Apr 29, 2006 at 10:32:39PM -0600, Jeremy Nickurak wrote:
Roberto C. Sanchez wrote:
Have you check this out?
http://www.reactivated.net/writing_udev_rules.html
I have. The problem is that the set of information required to
uniquely identify these devices spans two SYSFS levels, and according
to that document, I can only match on one level per rule. I have tried
it this way anyways, and failed.
KERNEL==event* SYSFS{bInterfaceProtocol}==01
SYSFS{bInterfaceSubClass}==01 SYSFS{bInterfaceClass}==03
SYSFS{bNumEndpoints}==01 SYSFS{bAlternateSetting}== 0
SYSFS{bInterfaceNumber}==00 SYMLINK+=input/event_kbd
KERNEL==event* SYSFS{bInterfaceProtocol}==02
SYSFS{bInterfaceSubClass}==01 SYSFS{bInterfaceClass}==03
SYSFS{bNumEndpoints}==01 SYSFS{bAlternateSetting}== 0
SYSFS{bInterfaceNumber}==01 SYMLINK+=input/event_mouse
Probably overspecified, but it works for me.
--
Rob
signature.asc
Description: Digital signature
Re: ATTN: Barbara Oncay
On Thu, Apr 13, 2006 at 02:54:31PM -0700, Paul Johnson wrote: On Wednesday 12 April 2006 12:41, Barbara Oncay wrote: I have been trying to unsubscribe with no success. I have sent more than 10 e-mails with no response. I have used the cancel link but it only sends me to a link where I am directed to subscribe to a link. TELLL ME HOW TO UNSUBSCRIBE OR I WILL CONTACT THE FCC. There are others who can't seem to unsubscribe. Follow the instructions at the end of every list message, including this one. Also, don't threaten people who have the means to find out where you work and publicize your stupidity in the international press. I think everyone in this thread missed the real issue. I'm willing to bet that NONE of the messages she wants to stop mention how to unsubscribe from debian-user. Remember the subject? emusing... A shot in the dark led to: http://www.emusing.com/, which has a form in the upper left corner: Get FREE FLASH Animations By Email Daily! [___] Subscribe Now I'll bet a beer that this site is the source of her annoyance, not Debian. Does the Debian lists site even have a cancel link? I could only find an Unsubscribe button. The unsubscribe link at the bottom of http://www.emusing.com/ leads to: Not Found The requested URL /lyris_unsubscribe.asp was not found on this server. Apache/2.0.52 (CentOS) Server at www.zdmedia.com Port 80 On the topic of the appended sig, messages have leaked through this list without the unsubscribe footer, but none I looked at in this thread were missing. -- Rob A CONS is an object which cares. -- Bernie Greenberg. signature.asc Description: Digital signature
Re: question about my .muttrc and mutt
On Tue, Feb 28, 2006 at 05:35:53PM -0700, Paul E Condon wrote: Thanks for the help. I had not known of these, but ... On my computer the two installation Muttrc files are identical (no differences discovered by diff) and there is no Muttrc.d directory in either installation. I can see in /usr/lib/mutt evidence that Muttrd.d is a legitimate part of some installations of mutt (a script that references it), but its not there for me. I didn't see any config information in /usr/lib/mutt (not surprising, there shouldn't be any according to policy) The different behavior remains a puzzle. I have no site-specific stuff, so I have no Muttrc.d either. It was just an opportunity for differences. Is your login name the same on both systems? If you want to check the actual installation for diffs: for f in `dpkg -L mutt`; do if [ -f $f ]; then md5sum $f; fi; done muttmd5 Transfer muttmd5 to the second machine. md5sum -c muttmd5 -- Rob A witty saying proves nothing. -- Voltaire signature.asc Description: Digital signature
Re: question about my .muttrc and mutt
On Tue, Feb 28, 2006 at 03:13:51PM -0700, Paul E Condon wrote: Additional information in this 'reply': On Tue, Feb 28, 2006 at 02:30:13PM -0700, [EMAIL PROTECTED] wrote: At the end of yesterday evening, I had a second installation of Sarge that contained the same software as the original. But with the same /home. But its not so. Mutt behaves differently in the two. For example, in one installation my posts to this list are listed as To: debian-user... and in the other they are listed as From: pecondon ... Why? My reading of the man page is that mutt configuration is Your earlier mail had no From: line. all done in ~/.muttrc and that there is not a master config file in /etc. Am I mistaken? Where is another that mutt config is stored? What is the source of this difference in behavior? /etc/Muttrc and /etc/Muttrc.d/* -- Rob She ran the gamut of emotions from 'A' to 'B'. -- Dorothy Parker, on a Kate Hepburn performance signature.asc Description: Digital signature
Re: debian-user and mutt...
On Wed, Feb 22, 2006 at 02:04:24AM +0100, Michelle Konzack wrote: Am 2006-02-09 15:27:22, schrieb Digby Tarvin: There is nothing incorrect in my headers as far as I know, but I do use a 'Reply-to:' to make sure that mail goes to my mailbox rather than just back to wherever I happened to originate a message from... This is an example of what Reply-to is actually for. It is not munging when done by the original poster. Everything should work as expected; from mutt, 'r' should go to the original poster (as specified by the Reply-to: header) and 'L' still goes to the list. Do not do this!!! How can someone reply to the list if you are munging the Headers? In mutt: 'L' (I just tried it on my sid box) Note: I have setup my mutt to ignore the Reply-To: Header. How can a user get a personal response if you ignore his request for responses to go to a different address than the one used to send from? Ignoring the Reply-to: header is as bad as a mailing list munging it. -- Rob signature.asc Description: Digital signature
Re: why my IP in sarge differ from my IP in win2k
On Sun, Feb 19, 2006 at 04:42:02PM -0800, Serena Cantor wrote: Both use auto-config, ( I guess it's DHCP), below is output of ipconfig by win2k: Connection-specific DNS Suffix . : IP Address. . . . . . . . . . . . : 221.137.89.103 eth0 Link encap:Ethernet HWaddr 00:03:0F:FA:E7:72 inet addr:221.137.93.88 Windows uses the machine name as the default identifier when making a DHCP request; dhclient and others in Debian use the MAC address. For dhclient, see the send dhcp-client-identifier config in /etc/dhclient.conf. DHCP servers use this identifier to grant persistent addresses; if Windows usually gets the same address each time, and Linux does also, this is likely the cause. Can I use this connection to set up a web site? You can direct traffic there using dyndns.org or some similar service; typically you don't get a static IP unless you ask for it. Whether or not inbound ports in general or port 80 specifically are blocked by your ISP or country is another matter. Set up a server and try it, and read your ISP's policies. Here, cable modem policies prohibit any servers, but web servers are not blocked. -- Rob Arithmetic: An obscure art no longer practiced in the world's developed countries. signature.asc Description: Digital signature
Re: Bash commands
On Sat, Oct 15, 2005 at 02:47:46PM -0700, Freddy Freeloader wrote:
$a is unset.
What does:
echo $a
print?
What does:
set | grep -e ^PROMPT -e ^PS
print?
--
Rob
b=${a/23/BB}
echo b = $b
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Bash commands
On Thu, Oct 13, 2005 at 08:18:05AM -0700, Freddy Freeloader wrote:
Hi all,
I downloaded the Advanced Bash Scripting Guide the other day an have
started to work my way through it. I'm fairly new to bash so I get more
than a little confused when the output I get is nothing similar to what
the ABS Guide says it should be.
Here is what has me confused at the moment.
b=${a/23/BB}
echo b = $b
Now the ABS guide says that where I'm setting b it should be
substituting BB for 23. It also says that the output of 'echo b - $b'
should be: b = BB35
However, what I get as output is as follows:
[EMAIL PROTECTED]:~$ echo b = $b
b = total 520716
drwxr-sr-x 2 ffreeloader ftp48 2005-10-13 07:50 script
-rw-r--r-- 1 ffreeloader ftp 532692172 2005-10-12 09:38 server_2003.zip
What does the last line containing 'a=' look like? b=${a/23/BB} sets
b = a if a contains no '23.'
--
Rob
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: daisy-chaining internet connectivity
On Mon, Aug 08, 2005 at 01:30:30AM -0600, Bob Proulx wrote: * Configure the second wired network card with a static IP address. Use a different subnet than you have on our WAN side. That is, if you have a 10.* address use 192.168.1.1 on the LAN. If your outside address is already 192.168.* then use 10.0.0.1 on your LAN side. Just make sure they are different. When using this setup, the router box (with both WAN and LAN interfaces) must be set up to do Network Address Translation for the wired network. iptables can be used for this. * Install and configure DHCP for your LAN interface. This is probably the most complicated part of the problem. But not overly so. Or use static IPs in /etc/networking/interfaces. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: HP Proliant DL360 G3 memory issue
On Wed, Jun 01, 2005 at 04:38:39PM -0500, Greg Gilmour wrote: I currently have an HP Proliant DL360 G3 with two 2.8 GHz Xeon processors, SmartArray 5i, and 1024 MB of memory installed currently. I wanted to install more memory so I purchased and installed 2 GB more to make a grand total of 3 GB of system memory (PC2100 DDR ECC) -- 2x512 MB sticks and 2x1024 MB sticks. Upon boot-up the BIOS detects all 3 GB, but once I boot into Linux it only registers as 900264 KB. I've tried checking the BIOS on the server You need a kernel with high memory support. In the kernel configuration, it is: Processor type and features High Memory Support (4G) -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: HP Proliant DL360 G3 memory issue
On Thu, Jun 02, 2005 at 11:38:18AM -0500, Greg Gilmour wrote: On Thu, Jun 02, 2005 at 09:18:50AM -0600, Rob Sims wrote: On Wed, Jun 01, 2005 at 04:38:39PM -0500, Greg Gilmour wrote: I currently have an HP Proliant DL360 G3 with two 2.8 GHz Xeon processors, SmartArray 5i, and 1024 MB of memory installed currently. I wanted to install more memory so I purchased and installed 2 GB more to make a grand total of 3 GB of system memory (PC2100 DDR ECC) -- 2x512 MB sticks and 2x1024 MB sticks. Upon boot-up the BIOS detects all 3 GB, but once I boot into Linux it only registers as 900264 KB. I've tried checking the BIOS on the server You need a kernel with high memory support. In the kernel configuration, it is: Processor type and features High Memory Support (4G) Hello Rob, I have tried a kernel rebuild with this option (a snipit of my config is below): CONFIG_MPENTIUM4=y CONFIG_HIGHMEM4G=y CONFIG_MTRR=y CONFIG_SMP=y CONFIG_MTRR=y CONFIG_SMP=y But, I am still getting the same results. ... Suggestions would be greatly appreciated. First, don't top post. Second: Look in /var/log/dmesg. On a 2.6.12-rc4, 4G kernel, 1G phys RAM, I have some lines very near the top: 127MB HIGHMEM available. 896MB LOWMEM available. On another box, 2.4.23, 64G kernel, 6G phys: 5248MB HIGHMEM available. 896MB LOWMEM available. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Audio CD Burning Perplexities
On Sat, May 07, 2005 at 04:06:43PM -0400, Thomas H. George wrote: I have on the hard drive 20 wav files (Signed 16 Bit Little Endian, Rate 44100 Hz, Stereo). ln -s filename.wav trackxx where xx runs from 01 to 20 and then burning the cd with the command cdrecord -v dev=ATA:1,1,0 -pad -audio trackxx A file with a name like track01 must contain raw data: ...MSB left, LSB left, MSB right, LSB right, MSB left and so on. The track should be a multiple of 2352 bytes. If a filename ends in .au or .wav the file is considered to be a structured audio data file. The result was 20 tracks of noise although CD Player knew the name of the tape and the titles of songs (The 20 files were captured from tape using Jackd and Ardour and I assume CD Player obtained this information from the Internet as my system was online). Yes, you had header junk and the rest of the data were byte-swapped. Use track01.wav instead for your links. And I assume your cdrecord line ended with track?? instead of trackxx. Replace this with track??.wav. I have not seen an audio player that won't handle track at once recording; however, if that's the issue, look into the -dao and -sao options, or use cdrdao. You can also try a lower recording speed. The ideal speed may vary with media. Unless a disc from a very specific list (per drive) of known media is used, the laser strategy is not optimized, resulting in a poor quality recording. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ati radeon 9600 problem
On Fri, Dec 03, 2004 at 11:27:25AM +1100, Stuart Prescott wrote: The fglrx .deb file produced by alien creates a file conflict so dpkg then refuses to upgrade libmesa etc later. Instructions in other posts will fix things for you. But for the benefit of others rather than following the ATI instructions, do the following (from http://www.nanonanonano.net/linux/debiannx7010 :) # alien fglrx-4.3.0-3.7.6.i386.rpm # dpkg-divert --package fglrx --divert /usr/X11R6/lib/libGL.so.1.2.libmesa-gl \ /usr/X11R6/lib/libGL.so.1.2 # dpkg -i fglrx_4.3.0-4.7_i386.deb I have found this doesn't work. libGL.so.1.2 is protected from libmesa-gl upgrades, but libGL.so.1 gets screwed up: -rw-r--r-- 1 root root 681072 Sep 28 06:09 /usr/X11R6/lib/libGL.a lrwxrwxrwx 1 root root 12 Oct 10 08:02 /usr/X11R6/lib/libGL.so - libGL.so.1.2 lrwxrwxrwx 1 root root 20 Nov 3 20:47 /usr/X11R6/lib/libGL.so.1 - libGL.so.1.2.distrib -rwxr-xr-x 1 root root 798658 May 11 2004 /usr/X11R6/lib/libGL.so.1.2 -rw-r--r-- 1 root root 437344 Sep 28 06:09 /usr/X11R6/lib/libGL.so.1.2.distrib The distrib version is libmesa-gl's. I've tried also diverting libGL.so.1, but it doesn't help: lrwxrwxrwx 1 root root 12 Oct 10 08:02 /usr/X11R6/lib/libGL.so.1.distrib - libGL.so.1.2 I guess the problem is that fglrx packages (via alien) don't have libGL.so.1 in them. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: DHCP on two interfaces
On Saturday 13 November 2004 05:49 am, Viktor Rosenfeld wrote:
I have to local interfaces, eth0 (192.168.1.1) and wlan0
(192.168.2.1). There's also an isdn interface (ippp0) connected to
the outside world.
I want DHCP to serve IPs and stuff on the two local interfaces. The
problem now is with the routing. Depending on which network the
client is on (LAN or WLAN) he'll need a different IP of the server as
a default gateway. (I assume that no client will be on both local
networks at the same time.) How can I achieve that, without
hard-wiring the mac addresses of the interfaces into the dhcp.conf
file?
# Serve wireless net
subnet 192.168.2.0 netmask 255.255.255.0{
range 192.168.2.128 192.168.2.192;
option routers 192.168.2.1;
option domain-name-servers dns1, dns2;
}
# Serve wired net
subnet 192.168.1.0 netmask 255.255.255.0{
range 192.168.1.128 192.168.1.192;
option domain-name-servers dns1, dns2;
option routers 192.168.1.1;
}
--
Rob
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Redirecting stdout and stderr into a file
On Wednesday 10 November 2004 09:36 am, Otto Wyss wrote: Blake Swadling [EMAIL PROTECTED] wrote: but how can I redirect both together? grep 21 logfile Sorry does not work, stderr comes still on the terminal. command logfile 21 -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: number of files in directory?
On Tuesday 02 November 2004 07:24 pm, Maurits van Rees wrote: Find is actually a lot faster. I timed it on my system, as root in the root dir: mauritsvanrees:/# time ls -1RA | egrep -v ^$ | egrep -c -v :$ 216348 real4m10.773s user0m14.210s sys 0m19.960s mauritsvanrees:/# time find . -regex .* | wc -l 216334 real1m5.149s user0m5.560s sys 0m3.900s So `find' is roughly four times as fast. The difference in the number of files is probably because some files were removed during the operation, though I'm not completely sure of that. And `find' also counts the current directory, so it finds one more than the other solution. I don't agree with your time results: [EMAIL PROTECTED]:/home/robsims: time ls -1RA | egrep -v ^$ | egrep -c -v :$ 126062 real0m26.526s user0m0.665s sys 0m0.658s [EMAIL PROTECTED]:/home/robsims: time find . -regex .* | wc -l 126067 real0m1.651s user0m0.617s sys 0m0.258s [EMAIL PROTECTED]:/home/robsims: time ls -1RA | egrep -v ^$ | egrep -c -v :$ 126062 real0m0.664s user0m0.469s sys 0m0.189s [EMAIL PROTECTED]:/home/robsims: time find . -regex .* | wc -l 126067 real0m1.193s user0m0.591s sys 0m0.177s So yes, the second run was much shorter than the first, but only because it was second, not because it was inherently faster. Beware the effects of caching when benchmarking. You can shorten the ls command line: time ls -1RA | egrep -c -v -e ^$ -e :$ with no significant change in execution time, but fixing the find line: [EMAIL PROTECTED]:/home/robsims: time find . | wc -l 126067 real0m0.295s user0m0.119s sys 0m0.177s ...helps quite a bit. My count differents correlates to one . entry and four directories ending in a colon. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: weird constant hissing sound coming from one speaker
On Friday 24 September 2004 09:49 am, H. S. wrote: I am not sure this will work since if I boot in Windows the speakers work perfectly well. It is only if I boot into Debian that the hiss comes back during the boot process someplace (which as I mentioned is around when lp0 starts). Note that the hiss just not some background white noise. It is very pronounced. It gets attenuated when some sound plays (e.g. when I do a sound test in KDE). Otherwise it is like an FM radio, but coming only from my left speaker. Right speaker is perfectly okay as it was before. Use kmix, alsamixer, etc. to check your levels on various (especially unconnected) inputs - mic, line in, etc. Set them to zero/muted. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ATI 3D acceleleration and dpkg-divert
On Sunday 19 September 2004 10:53 am, Nicolas de Sereville wrote: Following the ATI Radeon Linux How-To, I managed to get the 3D acceleration with the previous xserver version (4.3.0.dfsg.1-6) but overwriting the MESA libGL library with the ATI one, through a dpkg --force-overwrite -i fglrx.deb. Everything worked well but I had a lot of troubles upgrading the xserver due to the libGL library. So I decided to make use of dpkg-divert as suggested on this web page (http://g-tec.co.at/ati_alt.html): dpkg-divert --package fglrx --add /usr/X11R6/lib/libGL.so.1.2 dpkg -i fglrx_4.3.0-4.12_i386.deb Where does /usr/X11R6/lib/libGL.so.1 point now? On my system, it keeps getting set to the diverted library. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Duplicate Entries in /proc/scsi/scsi
On Sunday 18 July 2004 08:19 pm, rankine wrote: Hi *, I decided to downgrade from kernel 2.6.x to 2.4.x The problem is that a weird problem is happening since then. In /proc/scsi/scsi I get a lot of duplicate entries. Namely: Attached devices: Host: scsi0 Channel: 00 Id: 00 Lun: 00 Vendor: ASUS Model: CRW-5224ARev: 1.37 Type: CD-ROM ANSI SCSI revision: 02 Host: scsi0 Channel: 00 Id: 00 Lun: 01 Vendor: ASUS Model: CRW-5224ARev: 1.37 Type: CD-ROM ANSI SCSI revision: 02 Host: scsi0 Channel: 00 Id: 00 Lun: 02 Does anyone know what is going on? You are mixing the kernel parameter CONFIG_SCSI_MULTI_LUN=y (config question Probe all LUNs on each SCSI device) with defective CD drives that identify a drive at each LUN rather than only on the first. Re-compile your kernel with CONFIG_SCSI_MULTI_LUN=n. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: postfix trusted network
On Tuesday 06 July 2004 08:29 am, Frank Kaldewey wrote: Hi postfix has to relay mails from a webserver(win2k) in LAN to WAN. webserver 192.168.1.150 mailserver/postfix 192.168.1.100 I edit in etc/postfix/main.cf mynetworks = 192.168.1.0/28, 127.0.0.0/8 Neither 192.168.1.100 nor 192.168.1.150 are in 192.168.1.0/28. Try mynetworks = 192.168.1.0/24, 127.0.0.0/8 instead. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How does one get ssh to not wait?
On Saturday 12 June 2004 07:22 pm, Dan Jacobson wrote: How does one get ssh to not wait? ssh somewhere ! touch file sleep 333 rm file ! echo I want control to arrive at this line without waiting 333! I know this has been solved other ways, but I'm thinking the generic syntax is useful too (in the case where a program doesn't have a background mode): ssh -T somewhere ! touch file sleep 333 rm file ! echo I want control to arrive at this line without waiting 333! The -T makes ssh not complain about stdin not being a terminal. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: TMDA and other challenge-response systems considered harmful
On Wednesday 02 June 2004 11:25 am, John Hasler wrote: monique writes: At least that method of circumvention is a serious legal offense ... If so why have none been prosecuted for it? http://www.cnn.com/2004/TECH/internet/05/27/tech.spam.reut/ -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: DHCP slow renewal, actually times out but mysteriously still gets an IP
On Friday 21 May 2004 12:55 pm, Stalks wrote: # I first restarted the dhcp3-server process May 21 14:24:13 nooblet dhcpd: Wrote 0 deleted host decls to leases file. May 21 14:24:13 nooblet dhcpd: Wrote 0 new dynamic host decls to leases file. May 21 14:24:13 nooblet dhcpd: Wrote 0 leases to leases file. # ipconfig /release May 21 14:24:53 nooblet dhcpd: DHCPRELEASE of 81.168.82.220 from 00:0c:6e:70:29:33 via br0 (not found) # ipconfing /renew (start) May 21 14:25:00 nooblet dhcpd: DHCPDISCOVER from 00:0c:6e:70:29:33 via br0 May 21 14:25:00 nooblet dhcpd: DHCPOFFER on 81.168.82.220 to 00:0c:6e:70:29:33 via br0 May 21 14:25:05 nooblet dhcpd: DHCPDISCOVER from 00:0c:6e:70:29:33 via br0 May 21 14:25:05 nooblet dhcpd: DHCPOFFER on 81.168.82.220 to 00:0c:6e:70:29:33 via br0 May 21 14:25:13 nooblet dhcpd: DHCPDISCOVER from 00:0c:6e:70:29:33 via br0 May 21 14:25:13 nooblet dhcpd: DHCPOFFER on 81.168.82.220 to 00:0c:6e:70:29:33 via br0 May 21 14:25:30 nooblet dhcpd: DHCPDISCOVER from 00:0c:6e:70:29:33 via br0 May 21 14:25:30 nooblet dhcpd: DHCPOFFER on 81.168.82.220 to 00:0c:6e:70:29:33 via br0 May 21 14:26:06 nooblet dhcpd: DHCPDISCOVER from 00:0c:6e:70:29:33 via br0 May 21 14:26:06 nooblet dhcpd: DHCPOFFER on 81.168.82.220 to 00:0c:6e:70:29:33 via br0 May 21 14:26:06 nooblet dhcpd: DHCPREQUEST for 81.168.82.220 (0.0.0.0) from 00:0c:6e:70:29:33 via br0 May 21 14:26:06 nooblet dhcpd: DHCPACK on 81.168.82.220 to 00:0c:6e:70:29:33 via br0 # ipconfig /renew (end, finally get an ACK) Maybe a firewall is getting in the way or another device is acting as a DHCP server also. What does the output of iptables -nvL look like? -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: '8' character displays incorrectly on console
On Monday 05 April 2004 02:33 am, Nicholas Pattison wrote: I have an interesting issue on my recent install of Debian. There are a few quirks, but the one that is getting a bit annoying is when the console trys to display an '8'. About half the time, it will display fine, and look like and 8. About the other half, it will look like a lowercase 'o' (the top half appears to be missing). As I'm pretty much clueless as to what this could be. Everything displays fine in X, but not the console. When it displays correctly in one console, it works in all consoles. When it displays incorrecly in one console, it displays incorrectly in all. It will also magically go back and forth between displaying characters correctly and incorrectly. I have found no specific pattern that recreates this issue. This happens on both the 2.6.3 and 2.6.4 kernel. This sounds very much like flaky hardware. In console mode, fonts are rendered by your video card. The fonts are rendered from a table in hardware, ROM, or RAM. Since it changes over time and doesn't get worse, it appears that the source is good - the 8 couldn't heal otherwise. It doesn't matter if the source is ROM or RAM. It appears to be when the graphics chip is accessing that particular area of the lookup table that it gets all zeros instead of a pattern. Try freeze spray on various parts of the card to see if you can make the problem appear or disappear. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Which package creates /etc/environment?
On Friday 19 March 2004 01:03 pm, Phil Edwards wrote: On Fri, Mar 19, 2004 at 12:12:57PM +, Robert Harris wrote: Phil Edwards wrote: Because 'dpkg -S' can't find it. And more to the point, what's looking at that file? It isn't my From my /etc/pam.d/login: # This module parses /etc/environment (the standard for setting # environ vars) and also allows you to use an extended config # file /etc/security/pam_env.conf. # (Replaces the `ENVIRON_FILE' setting from login.defs) auth required pam_env.so -- Rob
Re: ssh through NAT firewall host ID problem
On Wednesday 10 March 2004 08:23 am, Colin Watson wrote: On Wed, Mar 10, 2004 at 10:17:02AM -0500, Gregory Seidman wrote: On Wed, Mar 10, 2004 at 12:02:08PM +0200, Micha Feigin wrote: } I have two computers at home running sshd which I can get to through my } firewall using NAT on two different ports. } The problem is that when connecting from the remote host to the } different servers I start getting errors about wrong rsa key and it } won't connect until I delete the known_hosts file. } } How can I bypass this? This is a common complaint, and has been discussed on the OpenSSH mailing list. It seems to be seriously low priority to them. As I understand it, it's difficult because the known_hosts format would need to be changed. Couldn't ssh just use the specified host name instead of the resolved hostname? I could set up my .ssh/config: host machineA Hostname firewall Port 1234 host machineB Hostname firewall Port 1235 Seems like an easy fix, and wouldn't affect non-aliased entries in known hosts. The first connection to aliased machines would be flagged, but that's no big deal. -- Rob (who is suffering the same issues) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: External DVD-burner
On Thursday 04 March 2004 09:20 am, Hamid Gh wrote: On Wed, Mar 03, 2004 at 02:45:19AM -0500, Hamid wrote: I have an external DVD burner which is connected to ieee1394 port. ohci1394 and sbp2 are the missing keys to the puzzle. The device name would be something like /dev/sda I tried different /dev/sda* but none worked. Is there any message in the kernel log that says which /dev I should use ? DVD devices are CD devices, not disks. Try /dev/scd0. You need sr_mod on top of sbp2. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: scsi-ide-cdrw
On Saturday 17 January 2004 08:52 am, Thomas Sommer wrote: Am Sat, 17 Jan 2004 03:20:10 +0100 schrieb Roberto Sanchez: Thomas Sommer wrote: Hi, after reading a lot of howtos, newsgroups, etc. i constandly have the following problem with my cdrw-drive (when i try to mount it): dusk:/# mount /dev/cdrw mount: /dev/scd0 ist kein gültiges blockorientiertes Gerät (english: is not a valid block device) scsi-ide-module is installed proper, and i even can burn cds with the drive. This means that ide-scsi and sg are working. ... /dev/cdrw - /dev/scd0 What is the output of lsmod or your .config? i use a kernel that doesn't support loadable modules. i integrated the sd und ide-scsi modules to the kernel Add sr_mod to your configuration (CONFIG_BLK_DEV_SR=y) -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Basing new file permissions on current dir perms
On Friday 09 January 2004 05:06 am, Alex Malinovich wrote: Lets say I have a directory called mydirectory. The permissions are as follows: drwxrwxr-x root mygroup ... What I want, is a way to force the default permissions for new files in this directory to be: -rw-rw-r-- myuser mygroup chmod g+s mydirectory Note that changing the directory's group will clear the sticky bit. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ATI Radeon driver + Linux2.6?
On Monday 05 January 2004 05:30 pm, Björn Johansson wrote: I've recently installed a 2.6.0 kernel, and I have a problem with the graphics in X. The problem is that I can't succesfully recompile the fglrx driver, so the system uses MESA instead of the graphic libraries provided by the official ATI package. Is it possible to use the Ati FireGL drivers with 2.6? Check for instructions (and a patch) on www.rage3d.com, in the Linux forum. I got 3.2.8 to compile with those instructions. Initially, I had errors when starting kdm (and many other apps) that caused X to exit; the logs showed entries like Can't register SecondaryColorxxxEXT. Fixed by removing /usr/X11R6/lib/libGL.so, which was not provided by the ATI drivers; libGL.so.1 was. Also, you must be actively running the desired kernel to compile the module using ATI's script. Run the installation from a text-mode window with the modules radeon and radeonfb not loaded. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ATI Radeon driver + Linux2.6?
On Tuesday 06 January 2004 10:39 am, Roberto Sanchez wrote: The ATI 3.7.0 driver works out of the box. The open source DRI drivers also work great. Does the open driver support DRI and dual heads simultaneously? DRI doesn't work with Xinerama. The proprietary driver provides a single framebuffer mode to allow dual head and DRI. If the open driver does do this, where do I get it? The ones in the X package seem to have the DRI/dual head exclusion. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [NAL] Re: OT: SCSI better than IDE?
On Monday 29 December 2003 09:39 am, Bijan Soleymani wrote: On Mon, Dec 29, 2003 at 10:36:49AM +, Rus Foster wrote: Interesting, Paul. So are you saying that SCSI hard drives are simply not worth the extra money or that the advantages of SCSI are in the interface and not in the media? IMHO SCSI disks last longer so its worth paying extra there. Also the interface for SCSI is just better as the controller handels the disk rather than IDE which is done more on the CPU This is no longer true. Most current IDE controllers support various DMA schemes. Like ATA-100, ATA-133, etc. So with IDE the controller now bypasses the CPU too. This IS still true - there's more of a difference than DMA. SCSI drives allow multiple outstanding requests _per drive_, let alone per bus. This means that the drive can choose to complete the requests in optimum seek order, which cannot be determined by the host, which does not have access to internal drive geometry. Also, while the drive is transferring a block, it is already seeking to the next request. With IDE, the data transfer must complete before the next request even reaches the drive. The ATA standard describes queuing/disconnect behavior; it just isn't widely implemented. Now it may be true that SCSI drives have better performance, but right now I think that's just because they're better drives (mechanically, and electronically) rather than because of the controllers. They're also *much* more expensive. Even when the same mech is available with either SCSI or IDE, the SCSI price is higher. It's a simple matter of volume - SCSI volumes are low, so costs and prices are high. When a low volume drive (very high performance and cost) is made, there really isn't any savings to go IDE, and the target market for such drives usually desires SCSI anyway. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Driver Download and Installation ATI Proprietary Linux Driver 3.2.8 Download--for those interested
On Thursday 09 October 2003 07:16 pm, Haralambos Geortgilakis wrote: Hi yall, subject says what wants said. Follow the yellow brick url too http://www.ati.com/support/drivers/linux/radeon-linux.html?type=linuxprodType=graphicprod=productsLINUXdriversubmit.x=18submit.y=10 I tried this driver with a custom 2.6.0-test6 kernel. The XF86config-4 file is that generated by fglrxconfig. I selected single framebuffer mode to enable a dual headed single screen, 1600x1200x24. The card is a Radeon 8500. I'm running unstable, x-window-system 4.2.1-12.1 X appears to try to start a few times, and I'm dropped back to a text console. There are no errors in XFree86.0.log. However, kdm.log contains: [fglrx] API ERROR: could not register entrypoint for SecondaryColor3bEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3bvEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3dEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3dvEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3fEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3fvEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3iEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3ivEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3sEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3svEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3ubEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3ubvEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3uiEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3uivEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3usEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColor3usvEXT [fglrx] API ERROR: could not register entrypoint for SecondaryColorPointerEXT As a further check, I tried running xdm (normal for me is kdm), and it starts X and runs. I can't get past xdm, guessing that the KDE start has the same problems as KDM. A Google search on SecondaryColor3bEXT yielded four non-useful hits. These API ERROR strings appear in /usr/X11R6/lib/modules/dri/fglrx_dri.so: -rwxr-xr-x1 root root 5908219 Sep 21 20:44 fglrx_dri.so KDE works when I use the radeon driver, but I get no DRI. Where do I go from here? I didn't find a user forum for the ATI drivers, just a no-response bug reporting link. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 2.4.20 autofs: non-toplevel mounts broken?
On Tuesday 29 July 2003 11:39 pm, Bob Proulx wrote: Greg Folkert wrote: Modules not being loaded is typically the problem. That or the rpc/portmap/...blah not compiled in. So load them ... in /etc/modules Great idea! Which modules would you suggest? I can't guess at what additional module would be needed. Googling reveals a large amount of information, way too much, but nothing that points me to a solution. autofs4; the autofs module doesn't have the hierarchical support. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: RAID controller
On Thursday 19 June 2003 06:36 am, Folin Nicolas wrote: I have a hardware raid controller in my computer. My Linux don't recognize it at all. How can I detect it and access to my harddrives ? It's only 'accessing the harddrive' that interest me, not the hard-raid (I don't want to use raid0 or raid1) I have a sil 680 based raid controller, and all of my linux installation works perfectly (video, scsi, soundcard, X11) except this 'ATA controller'. This is a software raid solution. It's the BIOS on the card that provides RAID features. Kernel 2.4.18 does not have support for the cmd/sil 680; 2.4.20 does (it's in the cmd64x driver). 2.5.69 has a separate driver (sil680). I don't know yet whether it's the card or the Linux driver, but I get corrupted data in PIO mode and a hang on the first access after setting DMA mode when using this card. I haven't yet tried 2.5.69 with this card. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 3 Question:Kernel 2.5.69, sensors and mplayer
On Wednesday 11 June 2003 02:04 pm, Tinus Kotzé wrote: 1) I am running debian unstable and tried kernel 2.5.69. when I install it however, it comes passed the point saying Loading kernel, then the bios check. Then after the Uncompressing kernel it freezes. I disable the boot logos and also double checked my memory and processor settings, but to no avail. Somebody by any chance had the same problem? I am running an Athlon XP1800+, Gigabyte 7va-c mb, 512mb ddr333(not registered),geforce mx2. Current kernel is 2.4.20(recompiled) Same symptom here (hang after uncompress), Tyan 2460, 2x 1.4GHz Athlon MP, 1G ECC memory, Radeon 8500. There appears to be some access on the IDE bus. Tried gcc-2.95 and gcc-3.2.3. Trying 386 target tonight instead of Athlon/Duron/K7. Your motherboard even uses a different chipset; it appears the common hardware is the processor itself. I use grub, so my messages are slightly different before the freeze. Setup works fine with 2.4.18 Debian and 2.4.20 custom. -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: bash question: reporting a variable named within a variable
On Thursday 03 April 2003 08:01 pm, Bob Proulx wrote:
Craig Dickson wrote:
echo $var=$(eval echo \$$var)
That works. Personally I prefer to eval the entire line. This way
you only use one layer of processing rather than the two in the above.
for var in FOO BLAH ; do
eval echo $var = \$$var;
done
The following works only in bash, but is smaller and faster:
for var in FOO BLAH ; do
echo $var = ${!var};
done
--
Rob
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: mkisofs problem with woody
On Wednesday 19 March 2003 08:29 pm, Rich Price wrote: I am running Debian 3.0 and have installed this mkisofs package: Package: mkisofs Priority: extra Section: otherosfs Installed-Size: 724 Maintainer: Eduard Bloch [EMAIL PROTECTED] Architecture: i386 Source: cdrtools Version: 4:1.10-7 The command: mkisofs -r -graft-points \ -x /var/image\ -o /var/image/cdromimag1 \ primus/bin=/bin \ primus/boot=/boot\ primus/etc=/etc \ primus/lib=/lib \ primus/home=/home\ primus/root=/root\ primus/sbin=/sbin\ primus/usr=/usr \ primus/var=/var Try: mkisofs -r -graft-points \ -x /var/image\ -o /var/image/cdromimag1 \ primus/bin/=/bin \ primus/boot/=/boot\ primus/etc/=/etc \ primus/lib/=/lib \ primus/home/=/home\ primus/root/=/root\ primus/sbin/=/sbin\ primus/usr/=/usr \ primus/var/=/var -- Rob -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
