Re: [OT?] Loadbalancer on Debian
On Mon, Jun 20, 2005 at 08:37:08AM +0200, Niclas S?derlund wrote: Hi all. Im searching for a free alternative to byuing an expensive loadbalancer appliance. I have two identical servers on the inside and need third machine on the outside doing a round-robin of a couple of ports to the two servers on the inside, keeping states and such so that the traffic is flowing unhindered and that new connections are established in equal numbers to the two servers. Is this possible with Debian and if so, what application does this? Yes. www.linuxvirtualserver.org will do what you describes. The kernel does the load balancing, and the tools are in Debian, though if you compile your own kernel you may need to compile your own tools (which is easy and works fine in Debian.) --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How popular is Debian (popularity contest)
On Sun, Aug 08, 2004 at 07:45:34PM +0800, John Summerfield wrote: George Roman wrote: right is a relative word. debian is the right choice for me, if others (people who use other dist) prefer to reinstall their systems with each new release it is their business. Which distros need to be reinstalled for each new release? Slackware used to; I don't know if it still does. Patrick V. told me at Comdex ~8 years ago that he wasn't interested in in-place upgrades. Maybe he's changed his mind? In any case, it prompted us to switch to Debian at some point. Nice guy, though. --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
No more Alt/Meta in xterms
A couple of weeks ago the alt and meta (windows) keys stopped working normally in an xterm. They work in Emacs (when it has its own, separate window), and at a text-mode console, and I can use meta-tab to switch between windows in wmaker, but in an xterm window they don't. Running Sid. Yesterday's update to 4.3.0.dfsg.1-6 didn't help. The odd thing is, they do _work_, I just have to hit them both at once to get the effect of hitting meta. For example, I'm editing this in emacs within an xterm (i.e., not emacs in its own separate window), and to rewrap the paragraph I have to type Alt-Meta-q, not just Meta-q. The problem showed up with the X update from a couple of weeks ago; until then everything was working fine and I didn't change anything. Anyone having a similar problem? Suggestions to where to look to fix it? Many thanks, -- Pete Harlan [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: mutt - saving message to disk (not mail folder)
On Wed, Nov 19, 2003 at 08:37:29PM +, p wrote: On Wed, Nov 19, 2003 at 09:22:01PM +0200, Micha Feigin wrote: How do I save a mail message to disk with mutt? I want to save it to disk, not another mail folder (saving to a different mail folder appears quite clearly in the docs but couldn't find save to disk). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] // | cat [file name] Does that differ from 's' (save), then entering the filename? (Besides the fact that will overwrite an existing file while 's' will append, and the fact that 's' will mark the mail for deletion (use 'C' instead of 's', I believe, if you don't want that.)) --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
perl_tie_hash: Can't locate object method TIEHASH...
Sorry if this was already discussed recently. I just subscribed to the list and couldn't find anything in the archives. After the most recent perl 5.8.2-2 update, mod_perl scripts on our apache server die with: [Mon Nov 17 13:28:19 2003] [notice] Apache/1.3.27 (Unix) mod_perl/1.29 configured -- resuming normal operations [Mon Nov 17 13:28:19 2003] [notice] Accept mutex: sysvsem (Default: sysvsem) perl_tie_hash: Can't locate object method TIEHASH via package Apache::Table at /usr/share/perl/5.8.2/CGI/Cookie.pm line 77. [Mon Nov 17 13:28:22 2003] [error] Can't call method FETCH on an undefined value at /usr/share/perl/5.8.2/CGI/Cookie.pm line 77. This is a self-compiled Apache+mod_perl setup under Debian Unstable. It worked before the fix. I've recompiled to no avail, updated CGI, Apache, etc. from cpan, no change. Any help would be appreciated! Thanks, -- Pete Harlan [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: perl_tie_hash: Can't locate object method TIEHASH...
Following up to myself, a workaround for this particular problem is to
patch Cookie.pm (Version 1.24) as follows:
--- Cookie.pm.orig 2003-11-17 14:04:13.0 -0600
+++ Cookie.pm 2003-11-17 14:06:20.0 -0600
@@ -73,7 +73,7 @@
my $r = shift;
$r ||= eval { Apache-request() } if $MOD_PERL;
if ($r) {
-$raw_cookie = $r-headers_in-{'Cookie'};
+$raw_cookie = $r-header_in ('Cookie');
} else {
if ($MOD_PERL !exists $ENV{REQUEST_METHOD}) {
die Run $r-subprocess_env; before calling fetch();
I don't know if that's the right fix (I suspect there's a deeper
problem somewhere with headers_in), but it appears to work for us.
--
Pete Harlan
[EMAIL PROTECTED]
On Mon, Nov 17, 2003 at 01:56:09PM -0600, Pete Harlan wrote:
Sorry if this was already discussed recently. I just subscribed to
the list and couldn't find anything in the archives.
After the most recent perl 5.8.2-2 update, mod_perl scripts on our
apache server die with:
[Mon Nov 17 13:28:19 2003] [notice] Apache/1.3.27 (Unix) mod_perl/1.29
configured -- resuming normal operations
[Mon Nov 17 13:28:19 2003] [notice] Accept mutex: sysvsem (Default: sysvsem)
perl_tie_hash: Can't locate object method TIEHASH via package
Apache::Table at /usr/share/perl/5.8.2/CGI/Cookie.pm line 77.
[Mon Nov 17 13:28:22 2003] [error] Can't call method FETCH on an undefined
value at /usr/share/perl/5.8.2/CGI/Cookie.pm line 77.
This is a self-compiled Apache+mod_perl setup under Debian Unstable.
It worked before the fix. I've recompiled to no avail, updated CGI,
Apache, etc. from cpan, no change.
Any help would be appreciated!
Thanks,
--
Pete Harlan
[EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Integrated Broadcom nic on Dell times out with dhcp
and get a DHCP address, but the network remains broken with errors such as: b44: eth0: Link is down b44: eth0: Link is up at 10 Mbps, half duplex. b44: eth0: Flow control is off for TX and off for RX. [...] Using the default interface 'eth0'. Basic registers of MII PHY #1: 1000 782d 0040 6360 0de1 0021 0004 2001. Basic mode control register 0x1000: Auto-negotiation enabled. You have link beat, and everything is working OK. Your link partner is generating 10baseT link beat (no autonegotiation). End of basic transceiver information. Have you tried using mii-tool to set speed and duplex, rather than letting it get auto-negotiated? The description above says (no autonegotiation) but you've got auto-negotiation enabled for the card. Some device/card combos work well with auto-negotiation enabled, and others work horribly. Maybe setting it yourself would help. (Which to set it to is probably a matter of trial and error; there are only four combos.) HTH, -- Pete Harlan [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: PGP Signatures
On Wed, Jan 29, 2003 at 04:07:04PM +, Colin Watson wrote: On Wed, Jan 29, 2003 at 03:21:53PM +0100, mess-mate wrote: On Wed, 29 Jan 2003 07:07:30 -0500 Seneca [EMAIL PROTECTED] wrote: | In ~/.gnupg/gnupg.conf, uncomment or add | keyserver-options auto-key-retrieve. ?? there is no gnupg.conf !! Did I missing somewhat ? For ~/.gnupg/gnupg.conf, read ~/.gnupg/options. From the gpg manpage: OPTIONS Long options can be put in an options file (default ~/.gnupg/gpg.conf). I had .gnupg/options, but gpg appears to use whichever is there. If they're both there, it opens gpg.conf and ignores options. --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: How to make Alt GNU Emacs Meta?
When you do dpkg-reconfigure xserver-xfree86, you have a chance to input keyboard options. Input altwin:meta_win, which is mentioned in the dialog box that asks for options, to get the potato behavior or Alt and Windows keys. HTH, --Pete On Sat, Jan 11, 2003 at 03:52:00PM -0500, Richard Cobbe wrote: Lo, on Saturday, January 11, Bob Proulx did write: Adam [EMAIL PROTECTED] [2003-01-10 22:28:14 -0800]: In the current stable release of debian, GNU Emacs uses the Windows key as Meta instead of Alt. I am told this is not true for other linux distributions or other releases of debian. How can I fix this? If I have to I can make an .xmodmap but I would rather not. If there is some package I can upgrade to testing or unstable I am willing to do that. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: functional languages (was: Politics of Java)
On Mon, Dec 16, 2002 at 02:17:05PM -0500, David Teague wrote: ... Craig and others Having undesirable featuers such as maintaining state or having dynamic scoping, does not make a language not be functional. The I'll agree to disagree on that semantic point. (You could say that you've written a functional _program_ in language X, but I wouldn't say that makes X a functional language.) Unfortunately they added some ketchup to the caviar by introducing the state maintaining looping, sequencing, assignment, and binding of functions and data to identifiers to Scheme. Assignment of course. But what non-FP looping does Scheme have, and why is binding and sequencing non-FP? (Sequencing isn't useful in a FP language, but there's nothing imperative about it.) I believe these and I/O are the only imperative features of scheme, and (less certain) believe they are the only imperative features of lisp. Er... Assignment/mutation and I/O are the only imperative features of any language. (Statements (as opposed to expressions) don't count: You could add them to a FP language without making it non-FP, you would just never have a use for them. (Of course, if I understand your use of functional language, you could add anything imaginable to a language and still call it functional...)) --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: putty logout hang
On Mon, Dec 16, 2002 at 02:53:47PM -0800, Charlie Reiman wrote: -Original Message- From: Brooks R. Robinson [mailto:[EMAIL PROTECTED]] Sent: Monday, December 16, 2002 2:39 PM To: debian Subject: RE: putty logout hang | Hi, | Whenever I login to my woody server and run a command | and place it in the background, I can't logout from the putty | window properly. | For example: | I'll use putty (ssh), login and run an rsync command over the network. | (This command may take 10 minutes) | So I ^Z it, and then bg it to the background. | I then use the exit command and I logout but the putty window hangs. ... I believe this behavior is intentional. It is. the reason but I'll take a wild guess and say that as long as the child process is attached to a tty, ssh won't let go of the parent. I suspect if That's true. Something like this might work (bash, sh) ls /dev/null /dev/null 21 It does. If you want to run something in the background and log out of your ssh session, you have to make sure to redirect stdin, stdout and stderr (or your application could close them), because ssh(d?) will wait there as long as input could be required or output could arrive for you to see. It's doing you a favor. This is new in Woody; Potato's (ancient) openssh packages, which only did protocol 1, didn't wait. --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: functional languages (was: Politics of Java)
On Fri, Dec 13, 2002 at 07:45:34AM -0800, Craig Dickson wrote: Kirk Strauser wrote: At 2002-12-13T14:52:51Z, Johann Spies [EMAIL PROTECTED] writes: Yes. So is Ocaml and I think Scheme also. Since Scheme is a Lisp derivative, yes, it's also a functional language. Scheme is a functional language; but I hesitate to call Lisp functional. In fact, Scheme was originally developed as a simplification of Lisp to focus on functional programming. Lisp can be used in a functional way, but also has imperative features. In fact, by modern standards, even Scheme is not a purely functional language; its expressions can have side effects, variables can be reassigned, etc. Lisp and Scheme are not functional languages. A functional languge is one that doesn't support mutating data; Lisp and Scheme very much do. You can write Lisp and Scheme expressions that don't mutate anything and call them functional; you can do that in Fortran too. But the languages aren't functional, they're imperative. (Lisp and Scheme have larger, more useful functional subsets than Fortran (or C), but that's doesn't merit calling them functional.) Haskell is functional. --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT: functional languages (was: Politics of Java)
On Fri, Dec 13, 2002 at 03:13:57PM -0800, Craig Dickson wrote: Pete Harlan wrote: Lisp and Scheme are not functional languages. A functional languge is one that doesn't support mutating data; Lisp and Scheme very much do. I certainly agree about Lisp. With Scheme, it's a bit trickier, especially since the history is that Scheme was first invented to be a Lisp-like language for programming with functions using recursion, first-class functions, dynamic scoping, and continuations -- essentially Lisp with the most non-FP features thrown out, plus dynamic scoping and continuations, which were not features of Lisp, and are very common in s/dynamic/lexical/g; --Pete (Deleting the other eight paragraphs I wrote because it's so off topic... ;) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: how to set /etc/hosts to force loopback connect
On Fri, Nov 08, 2002 at 04:05:51PM +0800, Patrick Hsieh wrote: Hello list, I'd like to make my debian box to loopback connect when it tries to connect to a specific domain name, say www.foo.com. So I configured /etc/hosts and put 127.0.0.1 www.foo.com in it. When I try to ping www.foo.com, it exactly goes to 127.0.0.1. But when I use telnet www.foo.com 80, it will not connect to 127.0.0.1. Telnet uses DNS before it looks in /etc/hosts. It's a bug in the name resolving routines having to do with ipv6. Some people don't think it's a bug. I don't know why. bugs.debian.org may have the anser somewhere. HTH, --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ifup/ifdown are not idempotent
On Fri, Oct 11, 2002 at 11:58:13AM +0200, martin f krafft wrote: check this series of commands: Are you sure there's not an old /etc/init.d/network or other startup script that brings the interface up initially using ifconfig? In my experience, when ifup brings up an interface, ifdown brings it down. Does your /etc/network/interfaces list anything else to do when the interface is brought up/down that might be causing the messages? On one of our boxes, I'd forgotten that an interface entry ran a firewall script before coming up; moving the firewall script made it so ifup didn't bring the interface up at all. That doesn't appear to be your problem, but a long-forgotten entry in one of those scripts can cause some puzzling behavior. --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian: abandon ship?
On Wed, Jun 05, 2002 at 12:47:17PM -0500, Manoj Srivastava wrote: Ian == Ian D Stewart [EMAIL PROTECTED] writes: Ian Speaking only for myself, it was the condescending tone adopted Ian by one of the developers (don't remember the fellow's name; he Ian was the one ranting about about his $250,00/hr fee) more than Ian the actual content that I found offensive. While stating that Ian you don't give a rip about the users may be intelectually Ian honest, one should not be surprised when such statements Ian endanger userbase loyalty. That would be me. To bring in the context that you have elided, this is exactly what I said: Manoj Telling me how to spend my time comes with the obligation of Manoj helping me pay my mortgage. My posted rates are $250 an hour. You have a strange definition of condescending. Are you, then, opposed to this sentiment? Can we call on you and tell you how to spend your time? manoj He didn't say he was opposed to the sentiment. He was saying that by stating the obvious, you were talking down to the other person. Granted, in an exchange in which both parties think the other side isn't seeing plain logic, it's a fuzzy line between making yourself crystal clear and condescension. --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: geforce4 420 graphics card problem
The X folks have put in support for the Geforce4, but the X server in Debian testing (or unstable, last time I checked) doesn't yet have those changes. So for now either compile X yourself, or use the proprietary driver from NVidia (directions for installation are available at their website). The latter requires installing a binary kernel module, which will taint your kernel (making getting help when it crashes more difficult). The free driver will come out in time, but it doesn't do as good a job at 3d accelleration as the proprietary one. (I presume it does some, but I don't know.) --Pete On Tue, Jun 04, 2002 at 05:16:39PM +1000, Shaun Nykvist wrote: Hi all I am having difficulties trying to get a geforce4 420 card to run under debian. I am fairly new to debian, could someone direct me in setting up this card under debian. I am using 2.4.18 kernel and packages from testing Cheers Shaun -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Remote admin of intermittently-connected machine?
We had a headless server on the other side of the world that connected intermittently with a dynamic IP over a 28.8 dialup line. It would email me its IP address when it connected; at that point I had five minutes to ssh into that IP address and stop it from disconnecting, and do whatever admin work was needed. The drawbacks to this method in your case are (a) you can only do administration when you're logged in when they connect, and (b) your friends may not enjoy having you know every time they connect to the net. The advantage is that it allows for interactive administration, it's foolproofedly simple to set up, and your friends don't have to do anything themselves. --Pete On Wed, May 29, 2002 at 07:07:05PM +0100, Karl E. Jorgensen wrote: Since I set up a Debian/GNU Linux box for a couple of friends, I find myself having to administer it remotely. Which is OK, as I did install ssh on it. But the box is only connected to the internet via a pay-as-you-go dial-up connection. And: I only want them to be connected to the internet when *they* like it (it's their phonebill after all). Which probably doesn't always coincide with when I want to log in and do things. So, this weird scheme occured to me: It is possible to do *some* of the admin via email? (not TCP/IP over SMTP/POP3...) I'm thinking something along the lines of being able to email commands to the box (GNUPG signed and encrypted of course), and have it email me back the output (again, signed encrypted). Similar to at(1), but remotely and via email. This won't work for interactive commands, but *should* work for the basics (i.e. apt-get --download-only --yes install so-and-so etc.) Writing something like the above is fraught with caveats and security holes, so I'd rather not write the lot from scratch. Has anybody else been in the same situation? What was your solution? Ideas and comments are welcome! -- Karl E. J?rgensen [EMAIL PROTECTED] www.karl.jorgensen.com Today's fortune: If you're not part of the solution, you're part of the precipitate. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: safe_mysqld hangs ssh connection every time. (woody)
If you ssh somewhere using openssh protocol 2 and launch a background task that hasn't closed stdout and stderr, your ssh session will hang when upon logout, until the launched app closes those descripters (e.g., when it finishes). Perhaps /etc/init.d/mysql redirects stdout/stderr to /dev/null? Ours didn't, until we added it, but perhaps yours already does. F-secure's ssh client doesn't have this problem (or, as it's seen by many, feature), nor did openssh protocol 1. --Pete On Tue, May 28, 2002 at 04:39:43PM -0700, justin cunningham wrote: Why does this happen? If I start mysql /etc/init.d/mysql start this is not a problem though safe_mysqld --with or without options it hangs. Connecting via ssh terminal emulation is linux. -justin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Applications not using hosts file for name resolution
This is a wierd but correct behaviour of IPv6 resolution. There are a huge number of bug reports about it. Do you know the rationale behind IPv6 considering this 'correct'? If you could explain it for the benefit of myself and everyone else who gets bit by this and searches the Debian archives, I'd be most grateful. It's patently a bug, if only in the IPv6 spec. Overriding DNS is half of what we use the hosts file _for_... Many thanks, --Pete (Not the guy who makes Pete's Wicked Ale) On Tue, May 28, 2002 at 09:38:00AM +1000, John wrote: Pat Colbeck wrote: Hi I have noticed something starnge about my Woody system. I have to maintain a hosts file due to some firewall and external DNS weirdness for some of the hosts in the office. If I ping them then they resolve via the hosts file but applications like telnet and postfix seem to be using DNS (thus getting the wrong address). Any ideas why ? Here is my resolv.conf file from /etc order hosts bind domain esc.azlan.co.uk nameserver 10.44.69.102 nameserver 10.44.65.120 This is a wierd but correct behaviour of IPv6 resolution. There are a huge number of bug reports about it. Disable IPv6 to get the 'normal' behaviour. John P Foster ( who is still NOT the guy who makes Foster's Lager) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Nvidia GeForce2 [was Re: woody]
Nvidia cards require their proprietary driver to work under X. No, I believe this isn't true. Use the 'nv' driver (included with XFree86 and open source (obviously)) and you should be okay. If you want full 3D acceleration (essentially as good as their Windows driver), then you can use their proprietary driver, but otherwise you don't need it. I have a Geforce4 card at home, which the 'nv' driver doesn't (yet) recognize, so I had to use the proprietary one to get X going at all. But the nv driver supports through the Geforce3's: (II) NV: driver for NVIDIA chipsets: RIVA128, RIVATNT, RIVATNT2, RIVATNT2 (A), RIVATNT2 (B), RIVATNT2 (Ultra), RIVATNT2 (Vanta), RIVATNT2 M64, RIVATNT2 (Integrated), GeForce 256, GeForce DDR, Quadro, GeForce2 GTS, GeForce2 GTS (rev 1), GeForce2 ultra, Quadro 2 Pro, GeForce2 MX, GeForce2 MX DDR, Quadro 2 MXR, GeForce 2 Go, GeForce3, GeForce3 (rev 1), GeForce3 (rev 2), GeForce3 (rev 3) --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Recommended tape backup software - tape vs disk - raided
On Tue, May 21, 2002 at 01:05:46AM -0700, Alvin Oga wrote: i wouldn't give backups to people ( when the backups contain user passwds and ( financial data or other sensitive stuff... Encrypt the backup, so you don't have to worry about it as much. Yeah, some folks have the passphrase, but if you accidentally leave it on the subway you don't have to worry about some random stranger making any sense of it. --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: C's pointer arithmetic (Was Re: Setting effective UID for a shell script)
On Thu, May 16, 2002 at 09:02:25AM +0200, Perceval Anichini wrote: When you write argv + 1, the compiler will understand : compute the address of argv, and add one time the size of the type which is pointed by argv. I remember to you that argv[1] = argv + 1. Brackets are only syntactic sugar, which allows us not seeing we are dealing with pointer arithmetic. This could hardly be more offtopic, but argv [1] == *(argv + 1) not (argv + 1). (argv [2]), or argv + 2, would have been what the original coder was trying to express. Cheers, --Pete -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: SCSI Transfer rate
In my case it was the wrong adapter between my disk and cable. (It was an 80=68-pin converter thingy.) Replacing it with the right one solved the problem. HTH, --Pete On Sun, Apr 21, 2002 at 05:54:05PM -0400, Robert Webb wrote: Hi all, I have a Symbios SCSI controller built onto my MB. The bios and debian on boot up report 40 MegaBit transfer rate. But further down in the kern.log I noticed the following: sym53c876-0-0,*: FAST-5 WIDE SCSI 10.0 MB/s (200 ns, offset 15) SCSI device sda: hdwr sector= 512 bytes. Sectors= 924 [4340 MB] [4.3 GB] sym53c876-0-1,*: FAST-5 WIDE SCSI 10.0 MB/s (200 ns, offset 15) SCSI device sdb: hdwr sector= 512 bytes. Sectors= 924 [4340 MB] [4.3 GB] sym53c876-0-2,*: FAST-5 WIDE SCSI 10.0 MB/s (200 ns, offset 15) SCSI device sdc: hdwr sector= 512 bytes. Sectors= 8380080 [4091 MB] [4.1 GB] sym53c876-0-3,*: FAST-5 WIDE SCSI 10.0 MB/s (200 ns, offset 15) SCSI device sdd: hdwr sector= 512 bytes. Sectors= 8380080 [4091 MB] [4.1 GB] which states that it is only 10 MB/s. Is there something I need to change to correct this or is this normal?? Thanks, Robert Webb -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Serverworks USB controller
Same problem here. I have a Nikon Coolpix 5000 camera, and a Serverworks motherboard (Thunder 2500). usb-ohci is the only host controller module that worked at all, and it 'worked' just as shown below. The camera connects fine in Win98, so it's possible to talk to it from the Serverworks USB controller, but not from 2.4.19-pre2 or -pre6. I tried compiling 2.5.8-pre3 last night to see if it worked, but the compilation failed. --Pete On Fri, Apr 12, 2002 at 04:29:29PM +0100, Tom Playford wrote: Hi, I've been having some problems with a serverworks USB controller in an NEC server. Inserting a USB device, in this case an OV511 webcam, gives this output: 12 16:20:51 marvin kernel: hub.c: USB new device connect on bus1/1, assigned device number 8 Apr 12 16:20:54 marvin kernel: usb_control/bulk_msg: timeout Apr 12 16:20:54 marvin kernel: usb-ohci.c: unlink URB timeout Apr 12 16:20:54 marvin kernel: usb.c: USB device not accepting new address=8 (error=-110) Apr 12 16:20:55 marvin kernel: hub.c: USB new device connect on bus1/1, assigned device number 9 Apr 12 16:20:58 marvin kernel: usb_control/bulk_msg: timeout Apr 12 16:20:58 marvin kernel: usb-ohci.c: unlink URB timeout Apr 12 16:20:58 marvin kernel: usb.c: USB device not accepting new address=9 (error=-110) And the device is not functional. I've never had this sort of problem with other USB controllers before. Is it just a simple matter of no driver support or is it more complicated? I'm using kernel 2.4.19-pre6. lspci -vv gives: 00:0f.2 USB Controller: ServerWorks OSB4/CSB5 OHCI USB Controller (rev 04) (prog-if 10 [OHCI]) Subsystem: ServerWorks OSB4/CSB5 OHCI USB Controller Control: I/O+ Mem+ BusMaster+ SpecCycle- MemWINV+ VGASnoop- ParErr- Stepping- SERR+ FastB2B- Status: Cap- 66Mhz- UDF- FastB2B+ ParErr- DEVSEL=medium TAbort- TAbort- MAbort- SERR- PERR+ Latency: 64 (2ns max), cache line size 08 Interrupt: pin A routed to IRQ 16 Region 0: Memory at fb102000 (32-bit, non-prefetchable) [size=4K] I don't feel brave enough to go for the LKML yet ;) Thanks Tom playford -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Alt as Meta Key in Emacs21 and XFree 4.1.0.1
This will swap your left windows and alt keys, so you can use Alt for Meta and (in wmaker and probably other window managers) the windows key to switch windows. 1. Use xkeycaps to rearrange the keyboard how you'd like, and save the modifications in a file such as ~/.xmodmap. Mine looks like this: remove Mod1 = Alt_L remove Mod4 = Meta_L keycode 0x73 = Alt_L keycode 0x40 = Meta_L add Mod1 = Alt_L Alt_L add Mod4 = Meta_L 2. Put this in your .xsession file: xmodmap ~/.xmodmap Of course there are probably a dozen other ways to do the same thing; this works for me. -- Pete Harlan [EMAIL PROTECTED] On Mon, Apr 08, 2002 at 01:56:30PM -0800, Sean 'Shaleh' Perry wrote: On 08-Apr-2002 Holger Rauch wrote: Hi! I noticed that in Debian Woody the Alt key does not work as Meta key when using Emacs 21 under X (XFree 4.1.0.1). What do I have to put in my .emacs file to get it to work? (I was looking at the FAQ using Help-Emacs FAQ, but found nothing there.) Searching with google revealed: -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Which window-manager for a kiosk?
You can launch your startup app from /etc/inittab. I used to work for a company that installed kiosks, and this is how we did it. We ran our own app, not a browser, and our X needs were very minimal. No window manager, no windows except our one, no keyboard, no mouse (touchscreen only), etc. Just run X or startx or whatever right out of inittab, giving the app as the startup argument. The default runlevel put us in kiosk mode. If we wanted a more normal box, for maintenance or development, we'd switch runlevels. It worked very well. Best of luck, --Pete On Tue, Mar 19, 2002 at 10:19:44PM +0100, Karsten Heymann wrote: On Tue, Mar 19, 2002 at 11:41:35AM -0600, Kent West wrote: Which window manager will give me the ability to size/move/close/etc windows, but without the ability to start other programs, etc. I want a kiosk style setup, on a low-memory machine. ICEWM has that menu at the bottom with the Start menu, so it's unsuitable. I tried twm, but I have to manually place the default app (Galeon) when it starts or when a new window is opened by the web site being visited, and I would prefer the window to just take the entire screen or at least not require manual placement. Any suggestions? I suggest (all untested): - use no window manager - start galeon in fullscreen mode from .xinitrc or .xsession - map away the F11-Key with xmodmap - maybe disable C-A-F[1-12] - maybe disable C-A-Backspace - disable most mime stuff to prevent the start of external viewers - have galeon automatically restart if it exits That should keep the system quite closed. As I didn't test that, suggestions are welcome! Greets, Karsten
Re: a grep question
Well since you don't show what command you actually typed, it's hard to tell you what you did wrong. But this might give you what you're looking for: find /etc -type f | xargs grep -H '10\.' where /etc is the root of whatever tree you want, obviously, and -type f tells find to only list files, not directories. xargs runs the command you give it, using xargs's standard input to get the list of arguments to that command. grep's -H is so you know which file the match is from (in case there's only one argument handed to grep). BTW, when grep (or cat or just about any unix command) appears to hang, try typing a Control-D character---it's probably waiting on its standard input, which Control-D will terminate. HTH, --Pete On Tue, Mar 05, 2002 at 12:52:44PM -0800, justin cunningham wrote: Hi, I read through man on find and grep and am trying to search for an ip in some files contained in folders but every time I type in grep options it just hangs. What am I doing wrong? Conversely i can go into /site.com/cgi-bin then cat any.cgi | grep 10.0.0.1 and will get the desired result but instead of doing this for 'every file' in 'every folder' I want to search for the 10.ip in the files from the site's root directory. Hope this is clear. Thanks, justin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: fork: Resource temporarily unavailable
I leave gnomeicu running all the time and my process table get filled with defunct gnomeicu processes. I have to stop/restart gnomeicu to get rid of them. I've had this fork: ... message happen when I had only around 300 processes running. /proc/sys/kernel/threads-max is the only relevant knob I've found, and it is set at 4095. Similarly, we could be out of file descriptors, but /proc/sys/fs/file-max (which presumably caps the number of open files?) is set at 8192; I'm sure the 300 processes didn't each have 25+ files open. Does anyone know how to increase the number of allowed processes? --Pete
lprng: pages/min varies with # items queued?
We're seeing something that's baffling me. We queue up a hundred pages to our locally-attached HP postscript printer, and when the first pages begin printing they print as fast as you'd expect from the printer (12ppm or somesuch), but as the queue grows the pages come out slower and slower (slower than 1ppm at the worst), and then when the queue is shorter again at the end it speeds up again. The pages are essentially the same; it's not a function of the content of the pages. The behavior is the same whether we connect the printer through the parallel or serial port, so it's not a cable/interface issue. It's a Woody system. It has started doing this in the past month or so; the same hardware/software printed all pages quickly before that, for years. Many thanks for any ideas; it's driving us insane. --Pete
Re: OT: mysql vs. postgresql
This would be true if it was true, but it isn't. MySQL is really unsuited to multiple readers, unless the reads are trivial select-one-row-by-id jobs. Flame bait. MySQL has been great for us, with scads of multiple readers with complex queries. Your mileage may differ, hence try them both and choose the one you like the best that works for you. Here's an anecdote from the Geocrawler people, who found that Neither database is a stationary target. Someone's experiences a year ago aren't necessarily relevant today. In particular, MySQL has introduced its InnoDB backend table handler that addresses just about every concern a person could have about its other table handlers, and then some. I'm sure PostgreSQL has improved also. We haven't moved to InnoDB because we haven't had to, despite MySQL being unsuitable for our needs cough. --Pete
Re: Non X-windows GUI web browser?
These guys are great, but if you *must* use a GUI broswer, and also only want webmin bound to 127.0.0.1 (Good Idea) you can use ssh to tunnel connections there from a client with an X browser on it. That's true, but be careful with ssh tunnels; without firewall rules, they'll tunnel any other machine's connection to your local host's port 80 (say), into the firewall, not just your own machine's connections. --Pete
Re: XFree86 4
means everything works. To make it your default: ln -s /etc/X11/X /usr/bin/X11/XFree86 That should read: ln -s /usr/bin/X11/XFree86 /etc/X11/X --Pete
Re: [users] Re: Time to fight for our beloved DEB format!
The reason why the run levels are specified was to handle cases where An LSB application may wish to have some kind of daemon which is only running when X11/xdm is running. Now, we could have added yet another level of indirection (in computer I'm probably missing something here, but isn't the main purpose of knowing the meaning of runlevels so the start/stop script links can be put in the right runlevel-specific directories? So if Debian wants to define its runlevels differently than the standard, then it just has to adjust its LSB-install program to translate from one LSB runlevel to its local runlevel, and install the link in whatever directory it wants. In other words, the level of indirection can take place within the Debian lsb-installation manager, rather than within the LSB spec itself. Then the admin who cares customizes /etc/LSB.conf can map the runlevels however she wants. Does the LSB have other reasons for specifying runlevels that couldn't be handled via a mapping in the installer? -- Pete Harlan [EMAIL PROTECTED]
Testing for Kylix compatibility fails (potato)
Borland's pre-install script for Kylix tests for a libc bug that it claims exists in Potato. I've included the output of its test script below. Does anyone know of a workaround/bugfix? I searched but didn't find anything about this. Presumably it's fixed on Woody/Sid, but I didn't test it there. Thanks, -- Pete Harlan [EMAIL PROTECTED] Borland Kylix System Compatibility Test Checking loaderThis will test whether libc installed on your system has a design flaw of the dynamic loading of shared objects. By dynamicly loading shared object 'a' that depends on shared objects 'b' and 'c', then dynamically loading shared object 'b' which depends on 'c', shared object 'c' will remain in memory when 'a' and 'b' are closed. This is what is in memory now: Name Address Handle --- 0x 0x400136f0 /lib/libdl.so.2 0x40019000 0x40013c60 /lib/libc.so.60x4001e000 0x4001d000 /lib/ld-linux.so.20x4000 0x40013420 Loading shared object lib-a.so And this is what is now in memory Name Address Handle --- 0x 0x400136f0 /lib/libdl.so.2 0x40019000 0x40013c60 /lib/libc.so.60x4001e000 0x4001d000 /lib/ld-linux.so.20x4000 0x40013420 /var/home/erik/tmp/borpretest/lib-a.so0x40015000 0x0804af80 /var/home/erik/tmp/borpretest/lib-b.so0x40017000 0x0804b228 /var/home/erik/tmp/borpretest/lib-c.so0x400fb000 0x0804b4d0 Getting address of 'a_function' and calling it Now loading shared object lib-b.so Again, this is what is in memory Name Address Handle --- 0x 0x400136f0 /lib/libdl.so.2 0x40019000 0x40013c60 /lib/libc.so.60x4001e000 0x4001d000 /lib/ld-linux.so.20x4000 0x40013420 /var/home/erik/tmp/borpretest/lib-a.so0x40015000 0x0804af80 /var/home/erik/tmp/borpretest/lib-b.so0x40017000 0x0804b228 /var/home/erik/tmp/borpretest/lib-c.so0x400fb000 0x0804b4d0 Getting address of 'b_function' and calling it Now this is where things begin to break down. The lib-b.so handle will be closed first, then the lib-a.so handle will be closed. Expected behaviour should be that the objects in memory return to the state prior loading lib-a.so Closing b_handle Name Address Handle --- 0x 0x400136f0 /lib/libdl.so.2 0x40019000 0x40013c60 /lib/libc.so.60x4001e000 0x4001d000 /lib/ld-linux.so.20x4000 0x40013420 /var/home/erik/tmp/borpretest/lib-a.so0x40015000 0x0804af80 /var/home/erik/tmp/borpretest/lib-b.so0x40017000 0x0804b228 /var/home/erik/tmp/borpretest/lib-c.so0x400fb000 0x0804b4d0 So far so good Closing a_handle Name Address Handle --- 0x 0x400136f0 /lib/libdl.so.2 0x40019000 0x40013c60 /lib/libc.so.60x4001e000 0x4001d000 /lib/ld-linux.so.20x4000 0x40013420 /var/home/erik/tmp/borpretest/lib-c.so0x400fb000 0x0804b4d0 WHOOPS!!! Lookie there! lib-c.so is still hanging around! This is bad... Thanks for your time FAILED Checking kernel = 2.2OK Checking libc = 2.1.2Major version is 2 Minor version is 1 Revision version is 3 OK Checking libjpeg = 6.2.0Checking /usr/lib/libjpeg.so.62 /usr/lib/libjpeg.so.62 resolves to libjpeg.so.62.0.0 Checking /usr/lib/libjpeg.so.62.0.0 Major version is 62 Minor version is 0 Revision version is 0 OK This system is not compatible with Borland Kylix. Please see the documents in this directory for information on how to upgrade your system.
Apt keeps giving me 400 Bad Request
I'm upgrading slink-potato. Each time I run 'apt-get dist-upgrade' (after having initialy run 'apt-get update') it tries and fails to grab a number of packages before giving up; each one looks like this: - Need to get 109MB/140MB of archives. After unpacking 51.9MB will be used. Do you want to continue? [Y/n] Get:1 http://http.us.debian.org potato/main libtiff3g 3.4beta037-8 [77.2kB] Err http://http.us.debian.org potato/main freetype2 1.2-6.1 400 Bad request Get:2 http://http.us.debian.org potato/main eterm 0.8.9-9 [467kB] Err http://http.us.debian.org potato/main imlib-progs 1.9.7-2 400 Bad request - My sources.list looks like deb http://http.us.debian.org/debian potato main contrib non-free deb http://non-us.debian.org/debian-non-US stable non-US Each run of apt-get does make about 5mb of progress. Intel, 28.8 modem link, plenty of disk space/ram, etc. What dumb thing am I doing wrong? Thanks in advance, -- Pete Harlan [EMAIL PROTECTED]
Re: Diamond Viper V770 Ultra
I am trying to help someone setup their Diamond Viper V770 Ultra under Linux (not sure exactly which distro they use, but I'm a debian person myself). They are having problems with X-windows (what else?) description of problem :- 1) Window appears to be four times it's correct size. 2) Unable to move the viewport to see the whole X-Windows Screen. If anyone has any ideas it would be greatly appreciated (I have got them to try all the normal possibilities, but they haven't worked for them, I am thinking that they need a X-server upgrade...) I just set up a non-Ultra TNT2 for someone using Slink and an upgrade to XFree86 3.3.4 did the trick. If they're using Debian Slink, add this to your sources.list: deb http://samosa.debian.org/~branden xfree86-334-slink/ Good luck, --Pete
Re: 2.2.* on slink
Check out this page for the list of things you need to update from potato: http://www.debian.org/releases/stable/running-kernel-2.2 Appletalk disk-serving to our Macs stopped working around 2.2.10 or so. I went back to 2.0.37 and it works again. The above URL didn't mention netatalk. It's the only thing we noticed breaking when using 2.2 with slink. Good luck!
All packages Obsolete/local in dselect?
Hi, All of my installed packages show up as Obsolete/local packages when using dselect: --- Obsolete/local Required packages in section base --- *** Req base adduser 3.8 none *** Req base ae 962-21.1none *** Req base base-files 2.1.0 none *** Req base base-passwd 2.0.3.3 none ... I've seen this before when switching to the 'apt' method in dselect, and going back to 'ftp' has always solved it, but no more. I deleted my /var/lib/dpkg/methods/ftp/ packages files and redownloaded them but to no avail. I'm running slink, upgraded from hamm. I've seen this happen with 'apt' on several machines, but only just saw it from 'ftp' too. Any ideas how I would go about resolving this? Thanks, -- Pete Harlan [EMAIL PROTECTED]
Re: email from cracklib cron
I'm not clear on when its ok to upload a fix to a stable distribution. This is clearly not a security issue or even a serious problem. Would it still be appropriate to upload to slink? It's the eternal debate about what the word 'stable' means in a software distribution. One definition is, non-experimental, but up to date. Another is, Changes only for security reasons. Debian seems to use the latter definition. Perhaps the package developer has some say; I'm not sure how that works. -- Pete Harlan [EMAIL PROTECTED] http://www.artselect.com/ http://www.mealsforyou.com/
Re: Do you still need screensavers these days?
Of course, better off wasn't defined... What if better off means using less electricity? What if it means saving on wear tear? Saving on wear tear uses less electricity, because it takes a ton of electricity to manufacture a new monitor. I've heard that it takes more electricity to make a new monitor than that monitor will use in its lifespan. Getting back to the subject line, we manufactured kiosks for a while and after a few months the monitors had the main screen's buttons burned into the phosphors, so yes, burn-in is still an issue as of a couple of years ago. --Pete
Re: X-windows not working anymore...
Question. I recently upgraded my kernel from 2.0.36(I think) to 2.2.1. I am running slink(frozen) and now, my x-windows doesn't work. If I This happened to me, and .xsession-errors now says /etc/X11/Xsession: line 47: syntax error near unexpected token `default)' The bit of /etc/X11/Xsessions that reads case $1 in failsafe) if grep -q ^allow-failsafe $optionfile; then if [ -x /usr/bin/X11/xterm ]; then exec xterm -geometry +1+1 else echo Xsession: unable to launch failsafe X session: xterm not found. exit fi fi default) ;; perhaps should have a ;; on the line before the default) line. (Haven't tried this myself yet.) Good luck, -- Pete Harlan [EMAIL PROTECTED]
Re: Slow system: problem isolated, need help with solution
Ok, it turns out that pppd is somehow affecting my computer. Everything works just fine and fast without pppd running, but as soon as I attempt to dial out, things slow down. How do I go about it? Andrew I had something like this happen to me; it turned out my external modem was plugged into a bad serial port on my motherboard. Switching to the other port fixed it. It was noticeable by checking /proc/interrupts; the bad port was generating gazillions of interrupts per second. Handling them was bogging my machine down. Good luck, --Pete
Re: Automatic font-lock in emacs
How do I configure my emacs so that it will detect the file type(*.c, *.h, *.tex, *.pro ) and automatically starts the font lock... Add this: (global-font-lock-mode t) to your .emacs file. (This is for emacs 20.3; don't know about other versions.) For more on this, look for help(*) on font-lock-mode. --Pete [*] C-h f font-lock-mode
Re: moving the root partition (esp. /dev)
I'm trying to move all the files in my root partition to a different disk parition. I had no problems using cp -ar to do this with the /usr tree, but /dev is giving me problems. I can't seem to copy the devices Don't give it the -r option; -a alone will suffice. In particuar, -a includes the -R option, which is what you wanted instead of -r (which make all non-directory files into regular files: man cp). -- Pete Harlan [EMAIL PROTECTED]
Re: Virtual Console Messed Up When Switching From X
I have the same problem, also with an S3 ViRGE/DX card. I can produce it by stopping and starting xdm. Running setfont (or is that seT7onT?) restores the vc console fonts, but it's probably a bug in the SVGA's server for this chipset. Perhaps it's fixed in 3.3.3... -- Pete Harlan [EMAIL PROTECTED]
Re: Backspace in xterm (again) [FIXED]
Pete Does Alt-backspace work for anyone in an xterm (deleting the previous Pete word on bash input, for example)? This worked fine in Bo and before. ... and now it works fine! Thanks to everyone for making a wonderful system where all the bad things go away and nice surprises show up around every corner. After being disconcerted that with the latest slink synch I lost all my X clients (xdm, xterm, etc.), and then re-finding them in the dselect list, and installing them, I found that Alt-backspace works properly again in an xterm. -- Pete Harlan [EMAIL PROTECTED]
Re: WARNING! Do NOT upgrade sysklogd to 1.3-29
The latest slink update to sysklogd 1.3-29 is SERIOUSLY broken. I installed it and could no longer become root. If you have installed it, Perhaps then a new version should be uploaded; this version would fail during install with a brief explanation, so we don't have to hear people who don't read the list complain that their system is broken. -- Pete Harlan [EMAIL PROTECTED] http://www.artselect.com
Odd dselect/apt/slink behavior?
Hi, I upgraded from bo to hamm (via cheapbytes's fixed cd_autoup.sh) and then to slink, installed 'apt', selected it as my dselect method, and now almost all of my regular packages are in the Obsolete/local category. I've seen this other times I've done this, and switching back to the ftp method makes everything go back to normal. What's wrong? In addition to most packages being mis-categorized, there are many missing (e.g., emacs of any flavor, and elm). My /etc/apt/sources.list file is unmodified, and looks like: # Use for a local mirror - remove the ftp1 http lines for the bits # your mirror contains. # deb file:/your/mirror/here/debian stable main contrib non-free # See sources.list(5) for more information, especial # Remember that you can only use http, ftp or file URIs deb http://ftp1.us.debian.org/debian unstable main contrib non-free deb http://non-us.debian.org/debian-non-US unstable non-US When I've installed freshly from the CD, rather than updating, I haven't had this problem. Any suggestions? Thanks, -- Pete Harlan [EMAIL PROTECTED]
Re: Odd dselect/apt/slink behavior?
Torsten writes: Probably the dselect database are not updated by apt-get. Try the apt select method in dselect and do an update (using the apt method). Are the packages still shown as obsolete? Yes, they're still obsolete. The update succeeded, to all appearances. --Pete
Re: mixed text and postscript
| I'd like to be able to print both text and PostScript using one | printcap entry. HP recommends setting up two separate printcap ... | this with LPRng? I'm using an HP LaserJet 4000N with a JetDirect | card. Nothing wrong with Magicfilter, but it doesn't have to be that magic either. I've used a filter consisting of: #!/bin/bash # # psAndTextFilter # # Prepares incoming files for a postscript printer; if the input is already # postscript, just passes it along. # We must send a Control-D character after the end of any documents # that go to the HP printer. enscript --pass-through --no-header -M Letter -T 4 -o - 2/dev/null echo -n '' for a 4000T. Note that that's a real control-d, not two characters in the final echo. Not all printers need that; HPs seem to. You will want to change the media type for enscript to suit your needs. -- Pete Harlan [EMAIL PROTECTED]
Re: getting sendmail to relay...
slackware machine to relay all the mail for irony.org. here's what it logs: Sep 21 19:47:19 irony sendmail[3703]: TAA03703: ruleset=check_rcpt, arg1=[EMAIL PROTECTED], [EMAIL PROTECTED] [192.168.13.4], reject=550 [EMAIL PROTECTED]... Relaying denied Does anybody know the arcane wonders of sendmail well enough to tell me what to put into the conf file? I don't really want to change MTAs...but I do need a main mail hub. Any help would be appreciated www.sendmail.org is your friend here, but the file you're looking for is /etc/mail/relay-domains. Put 'irony.org' there, reload, and you should be ready to roll. This assumes sendmail 8.9.x; you didn't say what version of Debian you installed. It matters, since 8.8.x (a) relays by default and (b) often uses different mechanisms for enabling relaying if you do disable it. -- Pete Harlan [EMAIL PROTECTED]
Re: IP Firewalling/Forwarding baffles me
If some kind soul could help, I'd appreciate it. :-) The HOWTO instructions don't seem to work correctly (but more than likely, my lack of understanding is at fault), so I plea for help. :-) You don't say which HOWTO, but the IP-Masquerade mini-HOWTO is excellent. Print that, print the ipfwadm manpage, read both a couple of times and I think it will solve all your problems. Worked for me. I think part of my problem may be how the 'hidden' machine's configured (default router, domain, etc.), but the HOWTO didn't really cover that. . . . The Networking howto (whatever it's called) ought to, or the system administrator's guide, or lots of different books on system administration. Good luck, -- Pete Harlan [EMAIL PROTECTED]
Re: Backspace in xterm (again)
After some experimentation, I found that the following works if you add it to /etc/X11/Xresources: *VT100.Translations: #override KeyBackSpace: string(\177)\n\ MetaKeyBackSpace: string(\033\177)\n\ ... Didn't produce any change for me. bash-2.01$ bind -p | grep back=20 ... \e\C-h: backward-kill-word \e\C-?: backward-kill-word And that's why. Bash is expecting a literal ESCAPE, and doesn't really care what key you use to get it. Emacs running inside an xterm doesn't see Metas either. Nothing in an xterm acts like it's seeing a Meta-backspace, even though it sees metas for every other key combo (afaik). /Somewhere/ something is special-casing Meta-backspace, and that ought to be fixed. I'd be willing to live with a workaround such as you proposed but it doesn't work for me. Does anyone have any idea which layer might be interfering with M-backspace? Another odd thing is that rxvt works (sort of---alt and meta are switched). Old binaries of xterm /don't/ work, though I don't have a statically-linked one to try out. Maybe it's a glibc bug somewhere? No, because remote Bo xterms, which work when sitting at that machine, don't work when displaying on my Slink machine. I'm stymied. -- Pete Harlan [EMAIL PROTECTED]
Re: Backspace in xterm (again)
Pete Does Alt-backspace work for anyone in an xterm (deleting the previous Pete word on bash input, for example)? This worked fine in Bo and before. What kind of keyboard do you have? On my 104 key (windows 95) keyboard, the windows key produces Meta and the Alt key produces Alt when in X. Meta-backspace deletes the previous word in bash in an xterm. Alt-backspace just deletes one character. Using an rxvt it is the other way around. I've tried a couple of keyboards; I've got a 104-key W95 keyboard currently. I get the same behavio[u]r as you, except nothing gets back-word-delete in bash. Meta /is defined/, because I can Meta-D to delete a word forward, or M-F to skip a word forward. It's only Meta-backspace that fails (afaik). I've fiddled with xkeycaps (I switched the meanings of the 'windows' key and the 'Alt' key, for example, so Meta is in the 'proper' place), to no avail. Outside X, or in Emacs, all works normally. rxvt does what you describe: It works, but it's reversed (alt acts as meta and meta is ignored.) I wonder what the difference between your setup and mine could be... -- Pete Harlan [EMAIL PROTECTED]
Re: Backspace in xterm (again)
Just to followup on my own message, it's probably not a bug in xterm because bo's xterm binary misbehaves identically when copied to slink. -- Pete Harlan [EMAIL PROTECTED]
Re: Backspace in xterm (again)
I too have this problem few weeks ago. What I did was: 1. Run xf86config 2. When it asks me for Do you want to use XKB? say yes. 3. Choose a preconfigured keymaps. I selected then: 1 Standard 101-key, US encoding With this one, no more fiddling with xmodmap or anything, Backspace the Delete key simply works. I did a fresh install of hamm and slink, including the above, and backspace and delete work fine, but Alt-backspace in an xterm[-debian] just does a backspace (instead of a meta-backspace, as it does (and should) on a text console or in an Emacs window.) Does Alt-backspace work for anyone in an xterm (deleting the previous word on bash input, for example)? This worked fine in Bo and before. -- Pete Harlan [EMAIL PROTECTED]
Re: Sendmail and virtualdomains
Yup. For 'dbm' substitute 'hash' and it should work. I suppose the precompiled debian doesn't have dbm support turned on. (You have to use 'hash' both in the FEATURE macro and when you create the database with makemap hash /etc/virtusertable /etc/virtusertable Good luck, -- Pete Harlan [EMAIL PROTECTED] http://www.artselect.com today I tried to reconfigure mySendmail 8.8.8 (hamm distribution) to set up some virtualdomains. I followed instructions I've seen at www.sendmail.org/virtual-hosting.html and added line FEATURE(virtusertable, `dbm /etc/virtusertable')dnl to my /etc/mail/sendmail.mc file (at the end of file, to preserve my configuration). Then I run sendmailconfig and Ooops ... get error message Class dbm not supported or sth like this. Am I doing sth wrong ? How can I set up virtualhosting properly ? All advices appreciate, thanx, Adam Heczko, [EMAIL PROTECTED] Technical University of Silesia Gliwice, Poland
Re: [pppd] error: link not 8-bit clean
Wrong passwords produce the link is not 8-bit clean message from pppd. For some reason improper authentication doesn't (can't?) get reported as such. I missed the beginning of this thread, so perhaps this has been mentioned. Good luck, -- Pete Harlan [EMAIL PROTECTED]
Re: Apt how, why, where
E kernel-image package offer the option of copying the kernel to the E place loadlin expects it in your setup? I would figure that ... There is no real standard place where loadlin users keep the images; and any hardwired solution is unlikely to satisfy more than a ... Comments? manoj Isn't there a config file in /etc? I accept the defaults when installing my custom-made kernel packages, but even there it would be easier if I could specify it once in the conf file. This should make the loadlin people happy, and the ask-questions-all-at-once install folks too. Just a thought. Bloat good. -- Pete Harlan [EMAIL PROTECTED] PS: Anyone know why Meta-backspace doesn't work in an xterm in hamm/slink? It works on the console, or when emacs has its own X window, but not on the xterm[-debian] command-line or emacs when it's in an xterm.
Re: Bizarre getty issue
Frederic: try installing mingetty (uses less mem and CPU anyway), then
change all the lines in inittab to use mingetty. This should both solve
the problem and help you w/ resources.
...
packages are done installing. I'm curious, is getty, mingetty, etc, even
slightly necessary? Realistically, I wont be using a modem... this was
Getty is the process that prints Login: (or whatever) on your
virtual consoles; it's used for normal logins, not (usually) modem
lines. ('mgetty' is more commonly used for modem lines; it has a lot
of nice features for managing them.)
Replacing 'getty' with a different getty is fine, but you should also
try to understand what problem caused init to be unable to run the
particular getty process.
The disabling for 5 minutes is just init's way to not waste a lot of
resources when something it's supposed restart keeps dying. The
solution is to find out (by looking at /etc/inittab) why it's trying
to run something, why that thing is dying, and fixing the situation
(perhaps by taking the line out of inittab.)
--
Pete Harlan
[EMAIL PROTECTED]
Meta-Backspace in Xterms?
Hello, After a clean install of Hamm, and also after moving to Slink, Alt-backspace now sends a simple DEL rather than M-DEL when I'm in an xterm[*](either at the command-line or when running emacs -nw.) Other alt commands (e.g., Alt-D for delete-forward) do work. [*] i.e., xterm-debian. It does work (sends M-DEL) when emacs is in its own window. Anyone know a fix offhand? It's amazing how much you grow to rely on the simplest of things sometimes. Thanks, -- Pete Harlan [EMAIL PROTECTED] PS: My system is behaving terribly after the reinstall. Freezes for ten seconds aren't uncommon, and got a complete lockup earlier. Never had any problems before the reinstall. I assume I'm the only one. Sigh.
Re: Is this expected ssh behavior?
probably enough to close this bug report. You wouldn't happen to know how to get bash to read all of it's appropriate login scripts and then executing a command would you (this is for over an ssh connection)? From the ssh(1) manpage: Additionally, ssh reads /etc/environment and $HOME/.ssh/environment, and adds lines of the format VARNAME=value to the environment. Some systems may have still additional mechanisms for setting up the environment, such as /etc/default/login on Solaris. So if you want to do anything beyond setting environment variables in that login script you're hosed, but I set up the environment I want in ~/.ssh/environment and it works. Good luck, -- Pete Harlan [EMAIL PROTECTED] -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: emacs (in color mode)
I have just got the new version of emacs (which I have never used before) and would like to enable colour highlighting of my C \ tex Put (global-font-lock-mode t) in your .emacs file, prepare yourself for a lot of colour, and restart emacs. (Assuming you're running it in an X window. If there's a way to colour text on a console or in an xterm, I don't know it.) Good luck, -- Pete Harlan [EMAIL PROTECTED] http://www.artselect.com -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: sendmail forward Virtual dom - user accounts???
is there anyone out there that can give me a little instruction as to how to get sendmail to forward wildcard mail for a domain to a user account? for www.sendmail.org is a good source of information about things like this. There's a FAQ there that I've found quite helpful a number of times. The main page has a link to http://www.sendmail.org/virtual-hosting.html which is a step-by-step guide to doing just what you want, if I understand you correctly. The only hitch in their instructions is that I had to use hash instead of dbm for the database type; for some reason the sendmail I had didn't grok dbm, but hash worked fine. Good luck, -- Pete Harlan [EMAIL PROTECTED] http://www.artselect.com -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: Debian Package Manager Worthless Junk???
Bob Nielsen writes: (To be fair, I haven't used Red Hat since 4.2 and it may have improved since then, but they severely mismanaged the conversion to glibc.) He who lives in a glass house should not throw stones, methinks... Debian 1.3.1 is a year old. Six months ago 2.0 was announced as Near Completion, when it was nearer inception than completion. I'm not ragging on the Debian team, just saying lighten up on Red Hat a little. We're all on the same side, eh? They chose to risk leaping before looking, while Debian risked hesitating. Was the latter more prudent? Maybe. Are my Debian 1.3.1 systems prehistoric? Yes. Is that bad? Sometimes. -- Pete Harlan [EMAIL PROTECTED] -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: SOLUTION? how to hold mail in another machine temporarily
is the next most preferred contact. Host B knows that it is not the final destination (because of the MX record for host A) so it queues the mail in /var/spool/mqueue (if it's a linux box). Host B will attempt to deliver that mail for as long as the timeout period allows. That's correct. I just thought I'd add that you don't want to list host A in your sendmail.cw file (the list of domains for whom you want to accept mail). Since you said you have to put aliases on host B, it made me wonder if you were doing that. You don't want sendmail to accept the mail, per se, you just want it to relay it [eventually] to A. (Note also that the MX record for B should have a higher number (corresponding to a lower priority) than the MX record for A.) -- Pete Harlan [EMAIL PROTECTED] http://www.artselect.com -- Unsubscribe? mail -s unsubscribe [EMAIL PROTECTED] /dev/null
Re: emacs *very slow* to launch
Netscape in X is another matter though. Would setting the mode to store the image in swap speed stuff up? Netscape maintains its large cache in ~/.netscape/cache, which it checks somehow on startup. This, I believe, contributes more to its slow launch than its bulk does. -- Pete Harlan [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Stronghold Apache
(I'm assuming 1.2.6 since 1.3 is still beta) with 128bit encryption. They have binaries for red hat 5.0 but I want to use debian 2.0 since it's what I know and I like it. anyone had any experience with this? anyone know of another way (other We've done it with [the nearly prehistoric] Debian 1.3, and it works fine. Don't know about 2.0. -- Pete Harlan [EMAIL PROTECTED] I burnt hamm on a CD and now the house smells like bacon and melted plastic. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Default Beep
difference between my puter and everyone elses at work!). In the interests of helping /even/ those who like to use vi in text- only modes ;), a look at drivers/char/console.c and an strace of setterm -blank 20 shows the following: Set the pitch of the beep: E[10;1000] where 'E' is an escape, the brackets are regular square brackets, the '10' is the command to set the beep, and '1000' is the hz of the beep tone. (E.g.: % perl -e 'print \033[10;1000];' % (now console has a different beep) ) If you wish to change the duration of the beep, the command is 11 instead of 10, and its argument is the duration in milliseconds. Happy beeping, -- Pete Harlan [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Default Beep
There is an app that lets you do this. I read it in one of the Debian packaged howtos. It is the same program that lets you change all kinds of low level stuff. Like console colors and what not. Chris wrote: I just had a weird though - is it possible to change the default beep in linux? You know - the one you get in vi all the time ;) If you use X, you can put something like xset b 50 1760 20 in your .xsession or .xinitrc file. man xset. If you're not using X, and your biggest complaint is the beep, then there's something wrong with your brain (but then you already said you use vi ;) HTH, -- Pete Harlan [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Setup w/large hard drives
Could someone give me a tutorial on how one should deal with large hard drives with Linux? I could've sworn I was on the right track Someone please correct me if I'm wrong, but I think that you can do the following, if you don't want to split your disk up into more partitions than necessary: Device Filesystem Size Mountpoint Contains - /dev/sda1 ext25mb /kernels kernels only /dev/sda2 ext28gb /Everything /dev/sda3 swap100mb none Swap space lilo.conf then looks something like - boot=/dev/sda root=/dev/sda2 install=/boot/boot.b map=/boot/map vga=normal delay=50 read-only image=/kernels/vmlinuz label=linux image=/kernels/vmlinuz-33 label=linux-33 It's my guess that this would work, though I haven't tried it; I'm planning on doing so the next time I set up a machine with a single large drive, so let me know if it doesn't ;) You'd have to make sure you copied your kernels into the right place, of course. But it saves you the pain of having a bunch of partitions that never turn out to be the right size. -- Pete Harlan [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Funneling a donain's mail to one account.
On Fri, 8 May 1998, R. Chris Ross wrote: ... Something like this: church.org local:pastor I'd probably do it with a virtusertable entry: @church.org pastor@local.domain Yup, that's the way to go. Read all about it at: http://www.sendmail.org/virtual-hosting.html The only thing I had to do differently from their instructions was to use the 'hash' database type rather than 'dbm' (there may have been a way to use 'dbm', but I don't know it.) So the lines would look like FEATURE(virtusertable, `hash /etc/virtusertable')dnl and makemap hash /etc/virtusertable /etc/virtusertable Now all you have to remember to do is to rerun the makemap command when you edit virtusertable, something I invariably forget... -- Pete Harlan [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Books or Debian wrong? and other stuff
And as long as I am here, I have noticed that the escape charactor in kermit does not work ^\. Neither does there seem to be anyway to exit dosemu other than killing the process. Kermit is full of bugs, and hamm does not have a current version. Kermit is easy to download, compile and install, and works well. Ctrl-\ is the escape character, but you have to follow it with another character to cause anything to happen. For example, Ctrl-\ C will get you back to the kermit prompt, and Ctrl-\ ? will give you a list of other escaped commands. --Pete, who has downloaded, compiled and installed A LOT of things by hand, thanks to using Debian's prehistoric stable 1.3. I've got so much stuff in /usr/local now, my computer tipped over... Debian? That's like Slackware, only they never come out with a new CD. ;) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: squake/multiplayer
when I'm playing multiplayer quake (1.09) I can't move forwards or backwards.. I can in single player mode.. My opponents are win95 tcp/ip with quake 1.06.. is there an incompatibility? IIRC, there is indeed an incompatibility between 1.09 and earlier quakes. You or they need to down- or upgrade. --Pete, who may not recall correctly. -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: catch 22
I hope you can help me find out the origin or how we got this catch 22 phrase? I would appreciate a respomse from you. I got to you page when I did a search on catch 22 Thank you in advance. Ofori Boateng Searching for catch 22 brought you to the Debian pages because of this thread. That's the origin of the phrase. -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: PGP
I notice that MIT now has a Linux build for PGP5.0 US version. I know that there is a PGP Debian package Does anyone know what's up with Phil, anyway? How can he, of all people, expect anyone to trust cryptography without source? Sorry for the off-topic. I hope nobody bothers to package this up for Debian unless source is released. There, now it's on-topic. And please correct me if source for 5.0 is available... -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: [SOLVED] Annoying boot-up messages
Bingo! I had added a custom keymap. Added the appropriate lines to boot.dpkg-new, mv'ed it to boot, rebooted, and Voila! Thanks much. FYI, normaly you only have to change /etc/kbd/default.map. No need to change /etc/init.d/boot. Ahhh, but I'm a relative newbie to Linux and to bash, so I let /etc/init.d/boot load the default keymap first, and then my custom keymap so if I screwed up the change to the script at least I'd have a usable For this sort of stuff, create a file /etc/rc.boot/local and throw whatever you want in there. It will get run as part of the bootup process, will get run only once, and won't get overwritten when you upgrade your box. In ours I set the keyboard speed and rewrite /etc/issue to put the current kernel version in the login prompt. -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Diagrams
Is there any package that is good for DB diagrams (something close to a case program but just the for the diagrams) ? MetaPost will draw fine diagrams. The program is mp and is included with teTeX. It's language based (i.e., not an interactive drawing program); unfortunately it's based on a brain-dead language. It's what Knuth's using for the diagrams in his new books. Its benefits include nice typesetting, easy to include (postscript) output in TeX, and it knows about connecting boxes with arrows so you can move the boxes without losing the connection. Its drawbacks include a language that is even more poorly-designed than TeX, bad error messages, and lack of visual tools. (Someone should put the good concepts there into a proper programming language.) There's probably something better out there for what you need. -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
DSFG Free?
configurable, solid, and DFSG free. It also has support for all the I have slipped into a parallel universe where GPL seems to have been replaced by DFSG. Unfortunately, www.debian.org and www.gnu.org reside in my former universe. A web search turns up many wonderful pages written in languages I can't yet read. Would someone please have mercy on a poor soul who is seeing DFSG everywhere and has no idea what it stands for, and enlighten him? Many thanks, -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Update: mounting MS-DOS fs onto Linux
One possible unconfirmed cause (Okay, a guess) Is the Win95 partition OSR2? (Fat32?) To my knoweldge, the vfat filesystem doesn't yet support the alterations made by OSR2. Matt I have a bleeding-edge Windows95 OSR2 installation (October1997 OEM) and (Debian) Linux reads it fine as vfat. It isn't FAT32; that's something else again, not dependent, I believe, on OSR2. --Pete -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Update: mounting MS-DOS fs onto Linux
One possible unconfirmed cause (Okay, a guess) Is the Win95 partition OSR2? (Fat32?) To my knoweldge, the vfat filesystem doesn't yet support the alterations made by OSR2. Matt I have a bleeding-edge Windows95 OSR2 installation (October1997 OEM) and (Debian) Linux reads it fine as vfat. It isn't FAT32; that's something else again, not dependent, I believe, on OSR2. --Pete Were you able to mount -t vfat /dev/hda1 /mnt your Win95 partition without any special work? Yes, exactly. [Eg, I am assuming you *do not* run mkfs over the Win95 partition.] Right. --Pete -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Book for writing shell scripts.
I am looking for suggestions on a good book for writing shell scripts. O'Reilly publishes a book on bash and another on awk sed. I like Kernighan and Pike, The Unix Programming Environment. It's dated (like ten years or so), and doesn't cover Bash extensions, and covers a lot more than shell programming, but all of that is good; its programs are very portable. The thing I like best is the Unix mindset, given here better than anywhere else I've seen. It's not a step-by-step intro to the details of shell programming, but when I was learning Bash and wanted to do something in it, I'd just flip open the book and follow their example. For details, the Bash manpage tells all. I only had to read it ten times or so before understanding what all of it meant ;) Like most people, I'm recommending what I have myself done; not having learned shell programming any other ways, I can't comment on them. -- Pete Harlan [EMAIL PROTECTED] This year will go down in history. For the first time, a civilized nation has full gun regisration. Our streets will be safer, our police more efficient, and the world will follow our lead into the future! --Adolph Hitler, 1935 -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Quake problem
I have a problem with quake on a new machine. It's a p166 MMX with an ET6000 pci vid card, 32M RAM. running squake often times crashes out when starting with the following error: svgalib: caught signal 11: segmentation fault the kernel is 2.0.30 and all packages are from stable as of a week ago. Someone reported that this went away when their kernel was compiled for 386 instead of Pentium. You might try this. It seems to be associated with MMX machines. I noticed there's an squake1.1 out there. Maybe it's fixed there? Though if recompiling the kernel fixes it, maybe it's not a quake problem. -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: best Linux video card
S3 has signed the Open Hardware certification for the Virge. Thanks Bruce What's the Open Hardware certification? What other vendors have/haven't signed it? What other chipsets has S3 signed it for, or not signed it for? Why did Debian's version numbering change? (Just kidding about that last one ;) Thanks, -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: best Linux video card (Open Hardware)
S3 has signed the Open Hardware certification for the Virge. ... What's the Open Hardware certification? What other vendors Look at http://www.debian.org/OpenHardware/. It's a way for manufacturers to promise that they'll publish specs for their hardware. Bruce did this along with a lot of other non-M$ folks. -- Pete Harlan, answering his own question having been emailed the answer by Bruce. [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: StarOffice3.1
I want to install StarOffice, but it require libc.5.4.4 or newer, while the last stable version is 5.4.33-6 or something. Where do I get it? Is 5.4.33 is 29 versions newer than 5.4.4. You're okay. Think of it as base infinity, where the infinigits are in base ten. Or maybe that's not the clearest way to think of it. -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: pppd
Now my link seems to be slow - is there a possibility to watch the throughput (like x.xx kb/s or sim.)? pppstats isn't perfect, but it does what you ask. -- Pete Harlan [EMAIL PROTECTED] http://www.mymenus.com/ The Best Recipe Site on the Web -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: scroll-back buffer
How can I clear the scroll-back buffer when a use logs out (other than changing tty's)? Echo four thousand spaces from .bash_logout? --Pete -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: How does mgetty differ from uugetty ?
My problem is that I could not find the section that describes the features that are supported by uugetty but not by mgetty. Can you tell me what are uugetty is horribly broken; this feature is not supported by mgetty. -- Pete Harlan [EMAIL PROTECTED] (This is not just a flip answer: It's a meaningful flip answer. Any 'features' uugetty has that mgetty doesn't can't meaningfully be called features unless uugetty works properly, which it doesn't. My knowledge of this is a few years old; perhaps someone has picked up maintenance of uugetty and gotten it working well. It's more likely that the documentation for mgetty is out of date, and should more properly not even refer to uugetty. The specific way it was broken when I looked at it was that it didn't properly share the modem with outgoing programs. If you want only to use the modem for incoming lines, and you find something that uugetty gives you that mgetty doesn't, then perhaps it's the program for you.) -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: is the Creative Labs AWE64 GOLD Soundcard supported?
How well/is the AWE64 GOLD Sound card supported under Linux? ..or will it act/function like the SB16? The alternative to doing backflips getting this card to work under Linux is to pay $20 to 4front-tech for their commercial driver for Linux. It's hassle-free, works, and is good for five years of upgrades. But it's not as cool as getting it to work for free (though you'll need to spend the $20 anyway on a hat after you tear your hair out). -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: fw:[oclug] moving an open file
In Slackware, I could be downloading a file in minicom then go to another VC and move the file to another directory without interrupting the file's growth. In debian, I can't do that anymore. Is it possible that you're (under Debian, but not Slackware) moving the file across volumes? In such a case mv, if it works at all (I've seen this behavior at least once in my life), can move by copying and deleting. --Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
[Q] Mail spool dir okay over nfs?
Hi, Is it completely safe to use, say, elm, to process the /var/spool/mail directory over nfs in Debian 1.3.1? I know at some time there was a problem with different locking conventions resulting in lost or scrambled mailboxes; have they been long resolved? The host runs sendmail, or sometimes smail, and the client will use elm or mail to read/delete email. Thanks, -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: screenshot in xfree86
Does anyone know how to take a screenshot of the X desktop? xwd will grab the data. It's then an issue of converting the format to one you like (netpbm package is usually the answer for this sort of thing). -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Stable means not-changing?
For the most part, it means non-changing. While it would be nice to
fix each package with a problem, doing so always runs the risk of breaking
other packages on the system. Verifying the integrity of the system as a
Perhaps this has been taken a little too much to heart; I keep
updating my system thinking one or two packages must have had some
fixes (security being my major concern), but nothing's changed. It's
better than having a lot of minor Foo-23.deb -- Foo-24.deb updates,
but gives the impression that stable means abandoned.
E.g. bash-2.0, which was found to be buggy almost immediately
(granted, not with a security issue, but it broke other packages).
Under 1.1 and 1.2 these things were fixed right away, which led me to
think that security issues would be address equally quickly; 1.3.1
makes a person wonder.
Note that I am not actually complaining about 1.3.1, just trying to
point out what it looks like to someone used to frequent stable
updates. For all I know 1.3.1 hasn't changed because, except for
bash, it's perfect. And I'm certainly glad that you're not having us
download a whole new X just for a nitpick change to xdm's login
screen...
Finally, the analogy to kernel development comes to mind; though it's
progressing now, for long periods the broken 2.0.30 kernel saw little
public attention from many key kernel gurus, which puzzled the masses
who were used to Linus poring over every oops on linux-kernel.
whole is far more difficult than verifying a single package. For
In the case of Bash-2.0, which broke a lot of scripts anyway[*],
turning it into Bash-2.01 would only have been an improvement.
They are especially rare in this case because Hamm marks such a major change
(with libc6 and all). Thus, fixes are very hard to propogate back to Bo
Granted this changes the picture. It's probably better to focus on
2.0 and get it out before Christmas than to coddle 1.3.1 and delay
Hamm a year.
--Pete
[*] Does anyone know where there was a doc explaining that { foo }
suddenly had to become { foo; } when upgrading to Bash-2.0? That
only choked on about a hundred of my scripts that had worked fine
under 1.14 (or whatever it was)...
--
TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to
[EMAIL PROTECTED] .
Trouble? e-mail to [EMAIL PROTECTED] .
Re: Background NFS Mounts
I'm having a problem with my computer hanging for large amounts of time when it is not connected to the network because of some NFS mounts it does. You can use the noauto option, and spawn a background job to mount them in a minute. E.g., fstab: -- flakey:/ /nfs/flakey nfs defaults,soft,noauto 0 0 /etc/rc.boot/local: (or wherever you put these things) --- # Mount flakey soon. echo mount /nfs/flakey | at now + 1 minute Works for us; if you don't want to be emailed when it fails you can redirect 2/dev/null. G'Luck, -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Linux kernel 2.0.31????
When is 31 going to be finished? There are already pre-31 patches ... with it, the BEST thing you can do at this point is install 2.0.29 and ... Out of curiosity what's so evil about 2.0.30? It is the first kernel I've tried since 0.99pl8 that I couldn't keep running. We had immediate trouble with our network card, a 3c590, due to a broken driver. Other folks have had trouble with it under low memory conditions, due to some changes that were supposed to speed things up from 2.0.29. If you run badblocks -w on a partition (note: this will wipe out your partition!) you will exhaust your buffers and experience problems on 2.0.30. We've had no problems whatsoever running pre-patch-7 applied to 2.0.30; I'd recommend at least testing it, because if you don't find problems with it it will soon become 2.0.31. (After a couple of known minorkinks are worked out.) -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Unknown PCI device (8086:7100 and others)
It seems as if the mother board is a bit too new, 7100-7113 and 1300 is unknown. I also had trouble like this (I don't remember the specific unknown numbers; the machine was a new Dec Venturis FX-2 with Pentium MMX). It disappeared when I installed Linus's (then-) latest patch, pre-2.0.31-6. It seems that he included a lot more pci devices there. The real question is Can I ignore the warnings? And the answer from me is, I wish I knew! Our machine appeared to work fine despite the unknown devices. --Pete -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
[HELP] smail stuck from cron?
Hi, ps -axf shows this: 0:00 /usr/sbin/cron 0:00 \_ /USR/SBIN/CRON 0:00 \_ (sh zombie) 0:00 \_ /usr/sbin/sendmail -FCronDaemon -odi -oem -or0s root The sendmail process (really smail) stays stuck like this forever (afaik). I do have a non-empty crontab, so an entry probably had output it's trying to mail to me. Any way I can tell which one? Any other ideas? The machine is runnning Debian 1.2, kernel 2.0.29, and has plenty of free resources (ram, disk, processes). Thanks, -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
Re: Can dpkg Repair Damaged Packages?
Bruce Writes: MD5 checksums suitable for use with md5sum -cv are in /var/lib/dpkg/info/*.md5sums . Unfortunately, not _all_ packages produce md5sums files, but you'll be able to check the ones that do. On our server, only ckermit provides such a file. Perhaps it should be a required part of every Debian package? And Jim Pick Writes: Klee Dienes [EMAIL PROTECTED] has a package called dpkgcert that really helped me out - it examines every file in you system, and points out which ones differ from ones supplied in packages. Unfortunately, it needs a database of certificates that contain the per-file MD5 sums of all the packages (which I got from Klee). I haven't heard from him for a while, so I don't know where you can get the certificates. This would be just what I would like (plus a reinstall option to dpkg, so any found problems could be corrected). It would be nice if the package could pull the md5sums out of the packages themselves, rather than rely on a separate (possibly out of sync) database. And as a last resort, such a utility could generate the md5sums itself from the packages as it worked. Does anyone have dpkgcert.deb? If it's unavailable, such a thing can't be that hard to write... Thanks, -- Pete Harlan [EMAIL PROTECTED] -- TO UNSUBSCRIBE FROM THIS MAILING LIST: e-mail the word unsubscribe to [EMAIL PROTECTED] . Trouble? e-mail to [EMAIL PROTECTED] .
