Re: Re: Gnome 2.14 crashes in Etch after upgrade
I noticed the message below on the list, and have the same problem. I manually restored parts of my .gconf directory, noting which things cause the crash. It was a time consuming task, the biggest nuisance is having to reboot after every crash. I have reduced it to: ~/.gconf/%gconf-tree.xml ~/.gconf/desktop/gnome/font_rendering/ ~/.gconf/desktop/gnome/peripherals/ font_rendering contanis one %gconf.xml file peripherals contains an empty %gconf.xml file, and a keyboard and mouse direcory, the former with 4 more subdirectories and a %gconf.xml, the latter with only a %gconf.xml I can't remember trying to restore only subdirectories within the peripherals directory, might have worked, but it was getting too cumbersome. font_rendering's %gconf.xml: type="string"> grayscale medium One of these two entries must crash Gnome then? Am I correct if I assume %gconf-tree.xml is just a flat-file version of the directory structure, meaning it's only font_rendering and something in peripherals that caused problems, on my machine? Pity no release critical bug was filed before this filtered through to testing. Don't want to deal with these issues this year... ;) Hugo On Fri, Apr 28, 2006 at 02:08:06PM +0100, Chris Lale wrote: I dist-upgraded Etch (Testing) today. Attempting to open a Gnome session caused Gnome to abort and made other virtual terminals unusable (they all had a screen with a coloured pattern and no prompt or text). This was true for any existing user with a gnome configuration. Newly-created users did not have this problem. I can solve the problem by deleting each user's ~/.gconf file (mv .gconf .gconf.old). Each user must then login and rebuild their preferences (eg theme, panel applets etc) from scratch. Is there an easier way? -- Jabber: [EMAIL PROTECTED] Google Talk: [EMAIL PROTECTED] -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: need proxy server that can auth to parent proxy server
Thanks. I was hoping to get away with something slightly lighter than squid, but it seems I might just have to use squid... Hugo -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
need proxy server that can auth to parent proxy server
Hello! I have an app here that can handle a proxy, but not proxies that require authentication. I want to run a proxy proxy on my machine, that this app can connect to, and which will then connect to another proxy requiring auth - the only function needs to be to perform the additional auth step. Any ideas what proxy can do this? Does not *have* to be in Debian, I'll consider packaging it if it isn't. Thanks, Hugo van der Merwe signature.asc Description: Digital signature
Re: Disappearing files ...
On Wed, Mar 10, 2004 at 10:00:02PM +0200, wrote: > I recently noticed I've got a serious case of missing files on my > gateway/firewall/webserver. Running debsums indicates it is only files > in the /usr/share/doc/ tree that are missing. I've tracked this to [03/Dec/2003:07:58:35 +0200] looking at apache logs, thanks to htdig's indexing I've got daily accesses, and files seemed to start going missing during the indexing. Yea, that's a very long time ago. I never noticed, the machine wasn't in public use for two months, since Feb a few people have hit 404's accessing documentation, but no one ever tells the webmaster... In fact, the machine was probably completely disconnected at the time, syslog shows pppd running on the 8th and the 20th (December). I did work via a NAT gateway (dialup) at some point though, for shortish periods of time, but that didn't do port forwarding either. I feel this was therefor not malicious, and will no longer worry about that. Still completely puzzled, but that is just too bad. I'm also curious why my logs (e.g. /var/log/syslog) are no longer rotated. That one also dates back to 3 December ... curiouser and curiouser... (I did notice the non-rotation a long time ago, but didn't worry about it. The curious thing is the date...) Thanks, Hugo van der Merwe signature.asc Description: Digital signature
Disappearing files ...
I recently noticed I've got a serious case of missing files on my gateway/firewall/webserver. Running debsums indicates it is only files in the /usr/share/doc/ tree that are missing. What could have caused all these files to disappear? I remember hearing something about a broken tar in unstable, but that shouldn't strike a mostly-stable machine so hard? It runs Debian stable, except for a couple of odd packages from testing and unstable, mostly documentation, but also a few other things like, including quixote, mod_python, scgi, probably apache, aptitude... It has an uptime of 45 days, running an oldish kernel, 2.4.20, which I know has some security issues, but hoped that with no local users, and limited public services, with up-to-date other packages (e.g. ssh), that it isn't remotely exploitable. It has a nice 75GB hard drive, with basically one partition (except for /boot, it's an old Celeron 300A overclocked to 450, needs a boot partition), formatted with reiserfs. I have reinstalled some packages already, though I reinstalled mostly documentation packages. (I pressed L in aptitude on the documentation section.) Any ideas? Amongst other searches, I've searched the debian-user archives for "usr/share/doc", didn't find anything. Sure, I should have been more thorough in my searching, but I have to finish a project for tomorrow... Any ideas will be greatly appreciated, thanks, Hugo van der Merwe signature.asc Description: Digital signature
ulimit, bash and ash
I just noticed that in bash "ulimit -u" is the same as ash's "ulimit -p", while bash has another meaning for "ulimit -p" ... This makes writing scripts quite difficult, I'd say you cannot then use ulimit in /bin/sh scripts, only in scripts specifically for bash or ash. Just a thought, posted in case anyone care to comment. Thanks, Hugo van der Merwe signature.asc Description: Digital signature
exim4 and "local delivery"
I have two machines, configured with exim4 to use a smarthost. However, I want to be able to send an email from one local machine to another, avoiding the smarthost. What is the easiest way of doing this? I've tried setting some of the values prompted for during the debconf configuration, but they don't seem relevant? Before exim4, I did: through_firewall: driver = domainlist transport = remote_smtp route_list = "*.localnet $domain byname" in the ROUTERS CONFIGURATION section. What would be the equivalent exim4 setting? I unfortunately don't have the time right now to figure this out for myself, and hope someone here already knows the answer. Thanks, Hugo van der Merwe -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
USB storage device, devfs, and /dev/scsi/*
I've got a USB storage device (what should one call these things? I hate calling it a "flash disk" since I feel it's not a disk (disc?) ;) It worked great at one point. However, recently, it still seems to register fine, etc etc, but the mountable device no longer appears in my devfs /dev/. How can I debug this, what should I look for, what might be missing? Thanks, Hugo -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Looking for tips on ppp problem: LCP ConfReq
On Sun, Jun 29, 2003 at 04:00:54PM +0200, Hugo van der Merwe wrote: > On Thu, Jun 26, 2003 at 07:33:46AM +0200, Thomas Krennwallner wrote: > > What are the modem manufacturers? Do you connect with a US Robotics on > > the gateway and with the internal modem of the laptop? > > Same modem, both cases, a Dynalink external 56K modem. I ahve an old > US Robotics 28800 as well though, I could give that a try... I think I'm going to give up on that machine for now. I connected the dynalink to my normal desktop machine, with, again, identical configuration, and it worked without a hitch. (OK, this time not the identical software, since I run sid on this machine.) The US Robotics 28800 modem I have, doesn't want to talk to the gateway machine. (It doesn't respond as required to the AT* commands sent to it.) However, plug that modem into the laptop, and again, no problems. So I now believe it must be something else, e.g. serial problems, not modem/ppp configuration. I will maybe investigate the problem more some other day, in the mean time I'm going to dial with my sid box. Thanks, all. Hugo van der Merwe -- The earth's surface is accelerating outward at 9.8 meters per sec^2 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: Looking for tips on ppp problem: LCP ConfReq
On Thu, Jun 26, 2003 at 07:33:46AM +0200, Thomas Krennwallner wrote: > Hi! > > On Wed Jun 25, 2003 at 11:46:00PM +0200, Hugo van der Merwe wrote: > > gateway/firewall/webserver. I had to move it to a modem for a while, > > however, it refuses to connect. The laptop, however, with pretty much > > an identical setup, connects without a hitch. So at the moment I've > > got the laptop set up as gateway so that I can send this. > > What are the modem manufacturers? Do you connect with a US Robotics on > the gateway and with the internal modem of the laptop? Same modem, both cases, a Dynalink external 56K modem. I ahve an old US Robotics 28800 as well though, I could give that a try... > I had a similar problem yesterday. I have a US Robotics Sportster > Message Plus. All went well except those pesky hangups without any > special reason. > > So I found a french (sic!) page and I only could read some thoughts on > USR AT commands. So I added AT&F1 before the ATDT stuff and voilà it > worked without a hassle. My init string is currently "ATQ0 V1 E1 S0=0 &C1 &D2 S11=55 +FCLASS=0" - IIRC that was what wvdialconf recommends (though I don't use wvdial). I will also try adding &F1, thanks... > I don't know if you use a USR modem but I think your problem lies within > some special AT commands to setup your modem properly. Maybe there are > some helping codes in the Modem-HOWTO. OK, my plan of action: 1) Add &F1 for Dynalink 2) Try US Robotics modem instead of Dynamlink 3) Try &F1 addition for US Robotics 4) Check modem howto Thanks, Hugo van der Merwe -- The earth's surface is accelerating outward at 9.8 meters per sec^2 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Looking for tips on ppp problem: LCP ConfReq
I've got two machines (well, more, but only two are relevant to the problem) here, both running woody. One a laptop, one a ... well, can't really call it a desktop, it's usually my gateway/firewall/webserver. I had to move it to a modem for a while, however, it refuses to connect. The laptop, however, with pretty much an identical setup, connects without a hitch. So at the moment I've got the laptop set up as gateway so that I can send this. The connection doesn't get any further than LCP ConfReq. Apart for stuff in ip-up.d, ip-down.d, the .bak files, and pap-secrets, /etc/ppp is identical on both machines. (Um, also options.ttyXX differs, but that's just a template for options.tty??, right?) Only difference in /etc/chatscripts is a "ABORT DELAYED". What else could it be? I'd suspect if the connection gets as far as it does, that the serial setup is good? Any ideas that I should pursue? Thanks, Hugo van der Merwe Jun 25 22:04:57 baboon pppd[31601]: pppd 2.4.1 started by hugo, uid 1000 Jun 25 22:04:58 baboon chat[31603]: abort on (BUSY) Jun 25 22:04:58 baboon chat[31603]: abort on (NO CARRIER) Jun 25 22:04:58 baboon chat[31603]: abort on (VOICE) Jun 25 22:04:58 baboon chat[31603]: abort on (NO DIALTONE) Jun 25 22:04:58 baboon chat[31603]: abort on (NO DIAL TONE) Jun 25 22:04:58 baboon chat[31603]: abort on (NO ANSWER) Jun 25 22:04:58 baboon chat[31603]: abort on (DELAYED) Jun 25 22:04:58 baboon chat[31603]: send (ATQ0 V1 E1 S0=0 &C1 &D2 S11=55 +FCLASS=0^M) Jun 25 22:04:59 baboon chat[31603]: expect (OK) Jun 25 22:04:59 baboon chat[31603]: ATQ0 V1 E1 S0=0 &C1 &D2 S11=55 +FCLASS=0^M^M Jun 25 22:04:59 baboon chat[31603]: OK Jun 25 22:04:59 baboon chat[31603]: -- got it Jun 25 22:04:59 baboon chat[31603]: send (ATDT0860007249^M) Jun 25 22:04:59 baboon chat[31603]: expect (CONNECT) Jun 25 22:04:59 baboon chat[31603]: ^M Jun 25 22:05:32 baboon chat[31603]: ATDT0860007249^M^M Jun 25 22:05:32 baboon chat[31603]: CONNECT Jun 25 22:05:32 baboon chat[31603]: -- got it Jun 25 22:05:32 baboon chat[31603]: send (\d) Jun 25 22:05:33 baboon pppd[31601]: Serial connection established. Jun 25 22:05:33 baboon pppd[31601]: using channel 7 Jun 25 22:05:33 baboon pppd[31601]: Using interface ppp0 Jun 25 22:05:33 baboon pppd[31601]: Connect: ppp0 <--> /dev/ttyS0 Jun 25 22:05:34 baboon pppd[31601]: sent [LCP ConfReq id=0x1] Jun 25 22:05:37 baboon pppd[31601]: sent [LCP ConfReq id=0x1] Jun 25 22:05:40 baboon pppd[31601]: sent [LCP ConfReq id=0x1] Jun 25 22:05:43 baboon pppd[31601]: sent [LCP ConfReq id=0x1] Jun 25 22:05:46 baboon pppd[31601]: sent [LCP ConfReq id=0x1] Jun 25 22:05:55 baboon last message repeated 3 times Jun 25 22:05:57 baboon pppd[31601]: Hangup (SIGHUP) Jun 25 22:05:57 baboon pppd[31601]: Modem hangup Jun 25 22:05:57 baboon pppd[31601]: Connection terminated. Jun 25 22:05:58 baboon pppd[31601]: Exit. dhclient log entries snipped. -- The earth's surface is accelerating outward at 9.8 meters per sec^2 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Found Answer: Re: ATI Radeon OpenGL results in black screen/window
My apologies for not researching properly, a more thorough google search quickly and easily turned up info about current Debian packages being broken, and also the link: http://dri.sourceforge.net/snapshots/README.Debian Will try this out soon, thanks. Hugo van der Merwe On Wed, Mar 26, 2003 at 09:45:39AM +0200, hugo wrote: > glxinfo claims: > > direct rendering: Yes > > When I have "direct rendering: No" because of starting a second X pgp0.pgp Description: PGP signature
ATI Radeon OpenGL results in black screen/window
glxinfo claims: direct rendering: Yes When I have "direct rendering: No" because of starting a second X session, 3D works fine. With direct rendering: Yes, however, I get only a black screen, or a black window. glxgears outputs a good framerate to the console, and makes my machine warmer, but still, only black. I've just tested with Knoppix, my display card *is* still working. It's one of the earlier ATI Radeon models, and it used to work just fine in Debian too. Any ideas on what I must do to get it working again? I've got a custom kernel 2.4.20, I've got radeonfb loaded, that shouldn't make a difference, I've got "radeon" module loaded too, and agpgart, with: CONFIG_AGP=m # CONFIG_AGP_INTEL is not set # CONFIG_AGP_I810 is not set # CONFIG_AGP_VIA is not set CONFIG_AGP_AMD=y # CONFIG_AGP_AMD_8151 is not set # CONFIG_AGP_SIS is not set # CONFIG_AGP_ALI is not set # CONFIG_AGP_SWORKS is not set CONFIG_DRM=y # CONFIG_DRM_OLD is not set Which I'd expect to work, since I have: 00:01.0 PCI bridge: Advanced Micro Devices [AMD] AMD-760 [IGD4-1P] AGP Bridge (According to lspci.) Ideas? I'm running Debian unstable, updated today, with local mirror, so I'm at most a day, likely not two, behind. Thanks, Hugo van der Merwe -- The earth's surface is accelerating outward at 9.8 meters per sec^2 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgp0.pgp Description: PGP signature
Re: Two network cards of same type, which eth0, which eth1?
> > I am going to be installing two new network cards soon, most likely > > I'll be getting that cheap junk that is the rtl8139. How do I choose > > which one must be eth0? Module parameters to the 8139too module? (In > > the past I simply loaded the correct module first. > > THey are detected in PCI order... so the lowest numbered PCI slot first, > and then in order. So this depends on your motherboard. In my case they > are numbered from low to high in the case. Check your mobo > documentation. Thanks, that would probably be fine, but I decided to rather go with one rtl8139 and one smc, for convenience. Have yet to put them in, will probably only do so tomorrow. Thanks a lot, Hugo van der Merwe -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Two network cards of same type, which eth0, which eth1?
Hello, I am going to be installing two new network cards soon, most likely I'll be getting that cheap junk that is the rtl8139. How do I choose which one must be eth0? Module parameters to the 8139too module? (In the past I simply loaded the correct module first.) Alternatively I could go to the trouble of buying two different kinds of network cards, then I'd be open to recommendations: something not too expensive, and of course, something that works very well with Linux ;) TIA, Hugo van der Merwe -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Re: fetchmail is confusing me - SOLVED
> Does anyone know what this means (specifically "retained", and > fetchmail not passing it on to my local MTA like other emails): > > I'm guessing that fetchmail thinks you've already seen this message, > and so isn't downloading it. Try passing it the `-a' option. This is what fetchmail -v -v reveals: <...SNIP...> fetchmail: POP3> RETR 1 fetchmail: POP3< +OK 513 octets reading message [EMAIL PROTECTED]:1 of 7 (513 octets) About to rewrite From: Mail System Internal Data <[EMAIL PROTECTED]> Rewritten version is From: Mail System Internal Data <[EMAIL PROTECTED]> fetchmail: retained fetchmail: POP3> RETR 2 <...SNIP...> Mail System Internal Data. Thanks for the help, Hugo van der Merwe -- The earth's surface is accelerating outward at 9.8 meters per sec^2 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 msg29973/pgp0.pgp Description: PGP signature
fetchmail is confusing me
Hello all, I've got fetchmail from latest unstable. Does anyone know what this means (specifically "retained", and fetchmail not passing it on to my local MTA like other emails): $ fetchmail -v mail.adept.co.za fetchmail: 6.1.2 querying mail.adept.co.za (protocol POP3) at Mon 10 Feb 2003 07:10:43 SAST: poll started fetchmail: POP3< +OK ready fetchmail: POP3> CAPA fetchmail: POP3< +OK Capability list follows fetchmail: POP3< TOP fetchmail: POP3< USER fetchmail: POP3< LOGIN-DELAY 0 fetchmail: POP3< EXPIRE 0 fetchmail: POP3< UIDL fetchmail: POP3< RESP-CODES fetchmail: POP3< AUTH-RESP-CODE fetchmail: POP3< X-MANGLE fetchmail: POP3< X-MACRO fetchmail: POP3< X-LOCALTIME Mon, 10 Feb 2003 07:11:39 +0200 fetchmail: POP3< . fetchmail: POP3> USER hugo fetchmail: POP3< +OK Password required for hugo. fetchmail: POP3> PASS * fetchmail: POP3< +OK hugo has 1 visible message (0 hidden) in 513 octets. fetchmail: POP3> STAT fetchmail: POP3< +OK 1 513 1 message for hugo at mail.adept.co.za (513 octets). fetchmail: POP3> LIST fetchmail: POP3< +OK 1 visible messages (513 octets) fetchmail: POP3< 1 513 fetchmail: POP3< . fetchmail: POP3> RETR 1 fetchmail: POP3< +OK 513 octets reading message [EMAIL PROTECTED]:1 of 1 (513 octets) fetchmail: retained fetchmail: POP3> QUIT fetchmail: POP3< +OK Pop server at stb-mail-01.adept.za.net signing off. fetchmail: 6.1.2 querying mail.adept.co.za (protocol POP3) at Mon 10 Feb 2003 07:10:53 SAST: poll completed fetchmail: normal termination, status 1 I'll use a -v -v next time, for now, I want to know what "emails" these are that never reach me. Please CC me, thanks. Hugo van der Merwe -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Apache configuration and host names?
I've tried setting ServerName, turning UseCanonicalName on and off, as well as modifying /etc/hosts. My web server is still redirecting to "baboon.localnet" rather than it's publically visable address. I commented out baboon.localnet in hosts, there is no reference to localnet in httpd.conf, and I stopped and started apache (didn't restart)... Why does it still redirect to baboon.localnet? Or might it be my client, mozilla? I removed baboon.localnet from the /etc/hosts file on the client machine and restarted mozilla to make sure it isn't that either. What else can I try? Thanks, Hugo van der Merwe ps. Please CC. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
process accounting paused...
Is process accounting paused for high loadlevels too? (I believe it is paused when you run out of disk space...) I've been seeing pauses lately, was curious what might be causing them. Thanks, Hugo van der Merwe (ps. I would appreciate a CC, thanks.) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
kernel bug!
Every day my server "dies", presumably doing cron job stuff. So I went
through the cron.daily items, and ran into:
baboon:/etc/cron.daily# ./modutils
Assertion failure in journal_forget() at transaction.c:1217: "!jh->b_committed_data"
kernel BUG at transaction.c:1217!
invalid operand:
CPU:0
EIP:0010:[]Not tainted
EFLAGS: 00010282
eax: 0058 ebx: c25c3dc0 ecx: c2dea000 edx:
esi: cb4df000 edi: c1612380 ebp: c9291b40 esp: c2debd6c
ds: 0018 es: 0018 ss: 0018
Process rm (pid: 2456, stackpage=c2deb000)
Stack: c01ff160 c01ff6e1 c01ff140 04c1 c01ff6f0 c0fbef20 c8af87e0
c8af87e0 cb4df094 c014e9db c0fbef20 c1612380 0003 c41d1000 c0fbef20
00290413 0003 c41d1000 c0fbef20 c015061f c0fbef20 c8af87e0
Call Trace:[] [] [] [] []
[] [] [] [] [] []
[] [] [] [] [] []
[]
Code: 0f 0b c1 04 40 f1 1f c0 83 c4 14 53 e8 bd 02 00 00 c7 43 14
find: rm terminated by signal 11
baboon:/etc/cron.daily# mount
/dev/hda1 on / type ext3 (rw,errors=remount-ro)
proc on /proc type proc (rw)
/dev/hda5 on /var type ext3 (rw)
/dev/hda7 on /tmp type reiserfs (rw)
/dev/hda8 on /usr type ext3 (rw)
/dev/hda9 on /home type ext3 (rw)
I'm surprised I only have one reiserfs, I thought I converted all the
smaller partitions. Oh well.
This is definately what's killing my server. (It's still half alive, I
could continue working in an ssh session, but cannot create new
sessions, cannot access apache, can access dictd... if I run ps ax in
the ssh session, it "freezes" here:
2423 tts/0S 0:00 /bin/sh ./modutils
2424 tts/0S 0:00 /bin/sh /sbin/insmod_ksymoops_clean
2428 tts/0D 0:00 find /var/log/ksymoops -type f -atime +2 -exec rm {}
which is clearly related to the processes going balistic. I will try
running an fsck. Any ideas? What should I do, what should I investigate?
hugo@baboon[1]:~$ uname -a
Linux baboon 2.4.19 #1 Thu Oct 17 14:05:10 SAST 2002 i686 unknown
Thanks,
Hugo van der Merwe
Please CC...
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
kill no longer functioning, curious?
I noticed today that not everything responds to SIGTERM anymore... SIGKILL, SIGSTOP and SIGCONT still work, but SIGTERM does not.. in some cases. I can kill most apps that are running straight from the command line. I can however not kill e.g. bash, or ogg123/mpg321 if they're started by my "music daemon"... (when run from the command line they're fine.) Any ideas why/how this happened? I'm running unstable, updated yesterday or the day before. Thanks, Hugo van der Merwe -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
apache-ssl, libapache-mod-ssl
What is the difference (when would one use what) between libapache-mod-ssl, and apache-ssl? Is one of these "new", since the "crypto-in-main" transition? (I only noticed the -mod-ssl one today, I've used apache-ssl in the past.) Thanks, Hugo van der Merwe -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
Trident display card
Hello, A friend has a display card identified by lspci as: 00:0b.0 VGA compatible controller: Trident Microsystems TGUI 9660/968x/968x (rev d3) I tried the xfree4 "trident" driver, but it didn't display quite right. (The display was ... um, "fragmented", though the mouse cursor moved fine, iirc.) Any ideas about this card? I guess I could also try 3.3.6. And what XFree4 driver is the "general SVGA" one, if there is one? I saw "VGA", which also gave some trouble (again: iirc), is "VGA" the generic driver? Might it be that the driver is fine, I just have to play with vid modes etc? Any tips appreciated, Hugo van der Merwe
Re: ssh session doesn't terminate until programs are closed
> see http://www.openssh.com/faq.html#3.10 which (for shell) recommends a > stdout is your alternative anyway (proper daemonizing). > If coding in C, use daemon(3). Thanks a lot. Hugo van der Merwe
ssh session doesn't terminate until programs are closed
Hello, I was wondering about an ssh session not closing until all the programs started in it are closed. How can I change this behaviour? Or maybe better, how can I modify my program to not be one of the programs it waits for? (Currently I fork() once and only the child branch continues. This does look like it backgrounds the program fine...) Thanks, Hugo van der Merwe
installing on local hard drive, without boot floppies ?
Hello, I want to install on a local drive, without rebooting with boot floppies (I want to keep the system running, while installing to a new drive). I can use debootstrap to set up a chroot, but how can I find out what other things I need to do, which would have been done by boot-floppies, but won't be now? I have a good idea of all the ... "important" stuff, but am afraid I will miss something. If boot-floppies source was laid out really nicely, it probably would have been easy to find out what I'd be missing. But unless I miss something, it is a bit too complex for me to go to the trouble of finding out that way. So out of interest, will the new installer that is being worked on, make alternative methods of install like this a lot easier? (As it is still being, "designed" in some ways, maybe this is something that can be kept in mind if it is not already ...) Thanks, Hugo van der Merwe
rebuilding kernels without make-kpkg clean?
Hello, What can one do if one just built a kernel, immediately realizes one forgot e.g. your sound driver, and want to rebuild the kernel. I am a bit hesitant to just delete the stamp files. I just don't want to recompile things that are already compiled... The docs are quite adamant that one should run make-kpkg clean unless you know what you are doing. So I want to find out what I'm doing... ;) Any ideas? Thanks, Hugo van der Merwe
Re: (not-debian-)programs not finding libraries on Debian system...
> > Sorry for CC'ing, not sure you are subscribed to debian-user, > > I am. obeying mail-followup-to again ... now I know > Indeed. But there have been Red Hat releases in the past on which > applications where compiled to use a hardwired library search path by > default. This program might have been compiled on a RH system, but it depends on the source, not the compiler environment, I guess... ? > > Did not find any "RPATH" in objdump's output. > > Then I'm out of ideas. Sorry. Turns out "export LD_LIBRARY_PATH=/usr/X11R6/lib/" helped. Apart from that, I've found a version with some more bug fixes, and source-only, which I compiled, and seems to not suffer from this problem. So thanks for all the help, it's now solved. Hugo van der Merwe
Re: (not-debian-)programs not finding libraries on Debian system...
On Fri, Sep 28, 2001 at 08:16:44PM +0200, J.H.M. Dassen (Ray) wrote: Sorry for CC'ing, not sure you are subscribed to debian-user, and I'm taking this off debian-devel. (As suggested.) > > I have compiled a program here that is looking for X libraries in > > /usr/lib... > > On a Red Hat system by any chance? I made a mistake, these binaries were actually not compiled by me. (I compiled a lot of stuff lately, I got a little confused ;) So maybe it was *compiled* on RedHat. It still shouldn't have hardcoded paths, though, should it? > > I don't want to make symlinks for all of these, that's just plain ugly. > > What else can I do, put them in ld.so.preload or something? > > It sounds like you compiled it with a -rpath option. Check if "objdump > --all-headers spice3" has an "RPATH" setting in it. If it does, try to > compile without it. If that's not feasible, edit the binary and make the > first byte of the RPATH data into an ASCII NUL. Did not find any "RPATH" in objdump's output. Thanks, Hugo van der Merwe
Re: (not-debian-)programs not finding libraries on Debian system...
Taking this to debian-user, as suggested.
> > I have compiled a program here that is looking for X libraries in
> > /usr/lib... I don't want to make symlinks for all of these, that's just
> > plain ugly. What else can I do, put them in ld.so.preload or something?
>
> You could tell us a bit more. What debian version are you using?
Sid. Come to think of it, there were a number of programs I compiled,
this wasn't one of them... these binaries are precompiled. So it wasn't
compiled on my system. Does this make a big difference?
> Did you install the xlib6 package from the oldlibs section (which is dependend
> on libc5)? What about a recompile?
xlib6 is not installed... I can try that, but all the libraries it
complained about was in /usr/X11R6/lib/, one for one I made symlinks in
/usr/lib, and it got past each of those then ... after the 7th symlink I
decided, no, I'm polluting my system, there must be a cleaner way.
> > $ strace ./spice3
> > execve("./spice3", ["./spice3"], [/* 32 vars */]) = 0
> > old_mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1,
> > 0) = 0x40007000
> > mprotect(0x4000, 21406, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
> > mprotect(0x8048000, 1176796, PROT_READ|PROT_WRITE|PROT_EXEC) = 0
> > stat("/etc/ld.so.cache", {st_mode=S_IFREG|0644, st_size=59404, ...}) = 0
> > open("/etc/ld.so.cache", O_RDONLY) = 3
> > old_mmap(NULL, 59404, PROT_READ, MAP_SHARED, 3, 0) = 0x40008000
> > close(3)= 0
> > stat("/etc/ld.so.preload", 0xb988) = -1 ENOENT (No such file or
> > directory)
> > open("/usr/lib/libXaw.so.6", O_RDONLY) = -1 ENOENT (No such file or
> > directory)
> > open("/lib/dpkg -", O_RDONLY) = -1 ENOENT (No such file or directory)
> > write(2, "./spice3: can\'t load library \'li"..., 43./spice3: can't load
> > library 'libXaw.so.6'
> > ) = 43
> > _exit(16) = ?
> Strange. There should be a ld.so.cache on your system.
> Try to run ldconfig as root and check if /etc/ld.so.cache is created.
There is an ld.so.cache. Not an ld.so.preload. (the strace shows this if
I am reading it correctly.) I reran ldconfig, but that didn't help.
Thanks,
Hugo van der Merwe
Re: locales, character sets, the like...
> Easiest thing to do is set the LANG variable in one or both of > /etc/environment or ~/.bashrc. The default character set should be "C" > (aka ASCII). For instance, I have LANG="en_US" (which translates to > iso-8859-1 for the character set, and changes collating slightly). > Choose whatever is appropriate for your language/nationality... Thanks. I played with the charset variable in mutt, didn't seem to do much, except without it an "ê" comes out as a "?", while with it, (iso-8859-1 or "C"), and without LANG environment variable, it comes out as \352. However, if I set LANG to en_US (will play with other settings later), then the charset variable doesn't matter and "ê" comes out right. Thanks! What is the "LOCALE" environment variable for? Lastly I assume that if I set LANG to e.g. af_ZA, that this will impact language choice in certain applications, but there will be a neat fallback to... what, maybe en_US if something isn't specifically set right in af_ZA? > > and an e^ (ê) comes out as "\352". (What is 352 anyway? Should be one > > byte? So what, octal?) The "source" of the email is encoded as "=EA", > 8**2 * 3 + 8**1 * 5 + 8**0 * 2 == 234 (base 10) Hmm, which matches 16*14+10==234 ... dem, must have made a calculation error previously. Stupid. Thanks, Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpMhwL7kjR3N.pgp Description: PGP signature
locales, character sets, the like...
Hello, 'scuse me, this is irritating me too much at the moment. I did some research to find it out myself, but clearly didn't look in the right places, while I can continue looking, maybe someone can save me time during my test week and just point in the right direction? I am trying to figure out the LOCALE stuff and character set settings. Where can I find out what and how to setup my locale (does that include language settings etc.)? Secondly, how does one control the character set? I assume it defaults to an iso-8859-1 character set? I have tried setting mutt's "$charset" to iso-8859-1, but an email with: Content-type: TEXT/PLAIN; charset=iso-8859-1 and an e^ (ê) comes out as "\352". (What is 352 anyway? Should be one byte? So what, octal?) The "source" of the email is encoded as "=EA", this is as I remember it from my pine days, and it did work well in pine. (I set my charset to iso-8859-1 as well, iirc.) Thanks to anyone that can shed light on this topic during these bleak hours, Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpUNwDcRAtQr.pgp Description: PGP signature
Re: AMD Athlon optimised, customized kernel, on Woody, and kernel oopses
> > Some more info: compiled on my own unstable machine, also got serious > > problems. Then I recompiled on my machine optimising for 486 instead of > > Athlon, sounds like there was no problem. So AMD optimization is broken? > > That's it. I've seen numerous problems related to the Athlon > optimizations being reported to the Linux Kernel Mailing List > and I do not even dare to compile my own kernels (for AMD or > Intel, for that matter) with anything above Pentium classic. I hear Pentium Classic optimisations make programs slower for other processors... makes me wonder what I should optimize for... I have had no problems with my Celeron and Optimizations for PPro/Celeron/P2 though. Should I try optimising for K6? Will this help with K7/Athlon? It is interesting to note that the kernel-image-2.4.9-k6 image is labelled as for K6/K7/Athlon, so this should be fine then... (I see 2.4.9 entered "local" mirror today. Goodie) Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpUaMf4tkQ8S.pgp Description: PGP signature
Re: AMD Athlon optimised, customized kernel, on Woody, and kernel oopses
> I am now compiling him a kernel on my machine (running unstable), hoping > this might sort it out. Might it be GCC3 issues? (What compiler is used > when using make-kpkg btw? And when compiling .deb's from source? > (gcc2.xx or gcc3)) I guess it might be testing's prerelease gcc3 > (1:3.0-0pre010403, vs 1:3.0.1-0pre010811) Some more info: compiled on my own unstable machine, also got serious problems. Then I recompiled on my machine optimising for 486 instead of Athlon, sounds like there was no problem. So AMD optimization is broken? Would make-kpkg be using gcc3 or an older gcc? This might also be 2.4.8 issues, I realize? Hope 2.4.9 makes it into Debian soon, or local mirror soon if it is already in Debian. Ideas? Thanks Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpibYVaIJKOb.pgp Description: PGP signature
Re: SOLVED: Re: gnumeric and gnome-print
> The solution: > > apt-get install gsfonts Oh, followed by a dpkg-reconfigure libgnomeprint-data, or the "perl /usr/share/doc/libgnomeprint-data/run-*" thingy. (What an odd place to put an installation script, eh?) > I think gnumeric should at least recommend this package, ? This isn't really gnumeric's problem, is it? Rather, libgnomeprint-data should recommend the gsfonts package? Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpnEwtmFRJgH.pgp Description: PGP signature
SOLVED: Re: gnumeric and gnome-print
> > > This I can confirm by running gnome-font-install -d : > > > > > What is gnome-font-install missing? Any ideas? > > > > Does running /usr/share/doc/libgnomeprint-data/run-gnome-font-install help? > > Ack, anyone got a fix for this Gnumeric problem? The solution: apt-get install gsfonts I think gnumeric should at least recommend this package, ? Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpb5M8xovHB5.pgp Description: PGP signature
Re: gnumeric and gnome-print
On Sun, Aug 12, 2001 at 10:31:10PM +0200, J.H.M. Dassen (Ray) wrote: > perl /usr/share/doc/libgnomeprint-data/run-gnome-font-install \ > /usr/bin/gnome-font-install /usr/share /usr/share > > (see the postinst of libgnomeprint-data) Thanks, but it didn't work... Rather verbose output: atlantis:/home/hugo# perl /usr/share/doc/libgnomeprint-data/run-gnome-font-install /usr/bin/gnome-font-install /usr/share /usr/share Info: ghostscript was found and has fonts in these directories: Info: /usr/local/lib/ghostscript/common Info: /usr/local/lib/ghostscript/5.50 Info: /usr/local/lib/ghostscript/fonts Info: /usr/local/share/ghostscript/common Info: /usr/local/share/ghostscript/5.50 Info: /usr/local/share/ghostscript/fonts Info: /var/lib/defoma/gs.d/dirs/fonts Info: /usr/share/ghostscript/common Info: /usr/share/ghostscript/5.50 Info: /usr/share/ghostscript/fonts Info: /usr/lib/ghostscript/common Info: /usr/lib/ghostscript/5.50 Info: /usr/lib/ghostscript/fonts Info: running "/usr/bin/gnome-font-install --debug --afm-path=/usr/share/fonts/afms --pfb-path=/usr/share/fonts/pfbs --target=/usr/share/fonts/fontmap2 --assignment=ghostscript,/usr/local/lib/ghostscript/common --assignment=ghostscript,/usr/local/lib/ghostscript/5.50 --assignment=ghostscript,/usr/local/lib/ghostscript/fonts --assignment=ghostscript,/usr/local/share/ghostscript/common --assignment=ghostscript,/usr/local/share/ghostscript/5.50 --assignment=ghostscript,/usr/local/share/ghostscript/fonts --assignment=ghostscript,/var/lib/defoma/gs.d/dirs/fonts --assignment=ghostscript,/usr/share/ghostscript/common --assignment=ghostscript,/usr/share/ghostscript/5.50 --assignment=ghostscript,/usr/share/ghostscript/fonts --assignment=ghostscript,/usr/lib/ghostscript/common --assignment=ghostscript,/usr/lib/ghostscript/5.50 --assignment=ghostscript,/usr/lib/ghostscript/fonts --fontmap-path=/usr/share/fonts". /usr/share/fonts/pfbs is not a directory Assigned ghostscript : /var/lib/defoma/gs.d/dirs/fonts Assigned ghostscript : /usr/share/ghostscript/5.50 Reading fontmap... Done Verifying fontmap entries Done Scanning source maps... Trying srcfile /usr/share/fonts/adobe-urw.font Trying srcfile /usr/share/fonts/urw-urw.font Trying srcfile /usr/share/fonts/debian.font Trying srcfile /usr/share/fonts/groff.font Trying srcfile /usr/share/fonts/tex.font Trying srcfile /usr/share/fonts/X.font Done Scanning directories: Done Sorting fonts... Done Sorting Type1 aliases... Done Building fonts... Done Something went wrong - NO FONTS atlantis:/home/hugo# cat /usr/share/fonts/fontmap2 atlantis:/home/hugo# The bug reports also weren't very helpful, though I think the person that filed #107797 has the same problem Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpJfXmYaQP2S.pgp Description: PGP signature
Re: hard drive problems... misconfiguration?
> it is a good idea to visit the manufacturers' site. download an utility > and test the drive for errors. seagate and western digital have pretty > good utilities that run off a floppy after you boot using a dos bootdisk. Thanks for all the replies, I've downloaded a drive testing program from Fujitsu's site, and found an error identified by a cryptic error code, and instructions to send it to the manufacturer for fixing. I've sent off a mail requesting more info on what this failure actually is... but unlike GNU/Linux support, Fujitsu Support doesn't function on Sundays. It's wait and see... Thanks everyone, Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgphMJTGLz5zl.pgp Description: PGP signature
Re: Dysfunctional ne (RTL8019AS)
Thanks for your replies... I set "PNP OS" back to "No" in the BIOS (setting it to Yes was what seemed to fix the problem last time, weird...), then I reserved IRQ's 3 and 4 for serial, 5 for my NE2K, 12 for PS/2 mouse (when I get round to fixing that one again), and booted GNU/Linux again. The RTL shifted from IRQ 12 to 11, and the NE appears to work fine on IRQ 5. I guess the BIOS just configured the network card better than it was configured when it left it for GNU/Linux. Makes sense. Thanks, Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpb7umvJqpl8.pgp Description: PGP signature
Re: gnumeric and gnome-print
> > Does running /usr/share/doc/libgnomeprint-data/run-gnome-font-install help? > No it doesn't, because the above file doesn't exist! ;^) I believe it does. At least on my system. Needs to be executed with perl. > Ack, anyone got a fix for this Gnumeric problem? I see there are a number of bug reports that are related. I've requested them, will mail again when I find something relevant. Hugo -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpHSPvMIzUOF.pgp Description: PGP signature
Re: gnumeric and gnome-print
> Does running /usr/share/doc/libgnomeprint-data/run-gnome-font-install help? Usage: run-gnome-font-install installer datadir srcdir What does it want as parameters? Hugo -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpoe7ABpqR6z.pgp Description: PGP signature
gnumeric and gnome-print
Hello, gnumeric is failing, claiming not being able to find a default font. This I can confirm by running gnome-font-install -d : atlantis:/home/hugo# gnome-font-install atlantis:/home/hugo# gnome-font-install -d Reading fontmap... Done Verifying fontmap entries Done Scanning source maps... Done Scanning directories: Done Sorting fonts... Done Sorting Type1 aliases... Done Building fonts... Done Something went wrong - NO FONTS atlantis:/home/hugo# What is gnome-font-install missing? Any ideas? Thanks, Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpqSdvanKqWu.pgp Description: PGP signature
hard drive problems... misconfiguration?
Hello,
This isn't exactly Debian specific, but maybe someone here knows what
the problem is ... ?
I have an old P120 laptop with a 1Gig drive. This drive is acting up
quite a lot lately. At the bottom of the email is an example of the log
messages.
I am wondering if it may be LBA related... the BIOS's autodetection did
set LBA to true, 32-bit IO to false. (I've set this to true in the past,
but when I got these messages I reverted to the autodetected false.)
hdparm's output:
/dev/hda:
multcount= 0 (off)
I/O support = 0 (default 16-bit)
unmaskirq= 0 (off)
using_dma= 0 (off)
keepsettings = 0 (off)
nowerr = 0 (off)
readonly = 0 (off)
readahead= 8 (on)
geometry = 527/64/63, sectors = 2125220, start = 0
I had lba32 in lilo.conf at some point, but I removed this. (Is this
only relevant at boot time, for lilo btw? I.e. if lilo has no problems
with it I can have lba32 in there anyway?) I also have "compact" in my
lilo.conf, so that it boots a little faster. Is this hardware failure,
or misconfiguration?
Thanks,
Hugo van der Merwe
hda: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
hda: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=524418, sector=524355
ide0: reset: success
hda: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
hda: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=524418, sector=524355
hda: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
hda: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=524418, sector=524355
hda: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
hda: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=524418, sector=524355
hda: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
hda: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=524418, sector=524355
ide0: reset: success
hda: read_intr: status=0x59 { DriveReady SeekComplete DataRequest Error }
hda: read_intr: error=0x10 { SectorIdNotFound }, LBAsect=524418, sector=524355
end_request: I/O error, dev 03:01 (hda), sector 524355
EXT2-fs error (device ide0(3,1)): ext2_write_inode: unable to read inode block
- inode=28353, block=65544
--
To send me private (non-world-readable) mail, GPG encrypt it.
1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698
pgpT5JCEpAisg.pgp
Description: PGP signature
Dysfunctional ne (RTL8019AS)
Hello, I recently replaced my old SB16 with a Sound Blaster Live, which caused quite some reassignments of IRQ's etc. I have two network cards in the machine, one that "ne2k-diag" claims is an RTL8019AS (bought as an ne2000 compatible, and served faithfully until now). (The other is an rtl8139, and I use the 8139too driver.) I've now shifted the ne to IRQ 5, it's base is still 0x300 - it seems to work fine, but if I try to use it, I get e.g. the following: [EMAIL PROTECTED]:~/znobackup$ ssh chimp NETDEV WATCHDOG: eth1: transmit timed out NETDEV WATCHDOG: eth1: transmit timed out ... Should I simply go on and try different IRQ's? (IRQ 7 is for parallel port, right? So that won't work?) Or how else can I diagnose this problem? Oh, another thing, my rtl8139 causes the following when the network is heavily congested: Aug 8 22:22:52 baboon kernel: NETDEV WATCHDOG: eth0: transmit timed out Aug 8 22:22:52 baboon kernel: eth0: Tx queue start entry 407299 dirty entry 407295. Aug 8 22:22:52 baboon kernel: eth0: Tx descriptor 0 is 2000. Aug 8 22:22:52 baboon kernel: eth0: Tx descriptor 1 is 2000. Aug 8 22:22:52 baboon kernel: eth0: Tx descriptor 2 is 2000. Aug 8 22:22:52 baboon kernel: eth0: Tx descriptor 3 is 2000. (queue head) Aug 8 22:22:52 baboon kernel: eth0: Setting half-duplex based on auto-negotiated partner ability 0000. Any ideas? Thanks a lot, Hugo van der Merwe -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgposxl7umLPJ.pgp Description: PGP signature
Re: PS/2 mouse problems... IRQ conflict?
On Sun, Jul 01, 2001 at 08:43:04PM +0200, Joost Kooij wrote: > 1st question: Are you also using gpm? Try turning that off and see if > problems persist. I have gpm running, but configured it to use only the serial mouse. (As I don't want to use the repeater, and ps/2 mouse cannot be used by both X and gpm at the same time, while this does work with a serial mouse.) Hugo -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgp63sWx5Fu4X.pgp Description: PGP signature
PS/2 mouse problems... IRQ conflict?
Hello, I've had problems with X input... basically, after doing some stuff, the keyboard and PS/2 mouse would stop working altogether... (the serial mouse still worked...) until I figured out to disable the PS/2 mouse. I am now operating without it. I find that if I enable it, after some work, input hangs. Any ideas? I'd think this might be an IRQ conflict or something, except that /proc/interrupts seems to have nothing on IRQ 12 if the mouse isn't enabled. How can I debug this problem? What must I look at? Thanks in advance, Hugo van der Merwe (A CC: will speed up replies... thanks...) -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpnCMq6qbsRq.pgp Description: PGP signature
modular kernels
Hello, I believe one of the advantages of making lots of modules when compiling a kernel, is not having to reboot when you need a new feature, as long as that feature can be compiled as a module. How can I compile a new module and install it without rebooting? (I'm using make-kpkg for everything so far...) Thanks, Hugo van der Merwe p.s. I'll appreciate a CC: -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgppZP2haaXCo.pgp Description: PGP signature
Re: Ensoniq 5880 AudioPCI - module loads, but no sound...
> Try > > $ fuser -v /dev/dsp > > It'll show you the program that has currently opened the sound device. > > Most likely, you have 'esd' running as part of your Gnome session. > To fix, > > - run all your audio apps in ESound mode > - disable "Startup Sound Server" in your Gnome control panel > > - Hari I'll try that. But I did check for any running programs... (Sommer with "ps ax") (sommer is Afrikaans, but I don't find nice English alternative ;) (I was not running X either.) alsa works... but not great... in Quake3 the sound in e.g. the intro sequence plays for a short while, then stops. Later, it comes back again. This looks like a more technical problem. Likely related to why OSS won't work? Hugo -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpxsXrjlt5b3.pgp Description: PGP signature
Re: Ensoniq 5880 AudioPCI - module loads, but no sound...
Just have to mention, alsa appears to be working great. Still odd that the normal es1371 module doesn't work. Thanks anyway, Hugo On Fri, May 04, 2001 at 03:04:31PM +0200, hugo wrote: > A friend has a new PC (Athlon 900MHz) with a motherboard with onboard > "SoundBlaster PCI128" (or should I call it Ensoniq 5880 AudioPCI...) > Oh, and I'm using kernel 2.4.2.. I'm gonna try alsa now. -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpuDWFBSfVLU.pgp Description: PGP signature
Ensoniq 5880 AudioPCI - module loads, but no sound...
A friend has a new PC (Athlon 900MHz) with a motherboard with onboard "SoundBlaster PCI128" (or should I call it Ensoniq 5880 AudioPCI...) Relevant Output of lspci: $ /sbin/lspci | grep 5880 ; /sbin/lspci -n | grep 5880 00:0e.0 Multimedia audio controller: Ensoniq 5880 AudioPCI (rev 03) 00:0e.0 Class 0401: 1274:5880 (rev 03) The odd thing is, the es1371 module loads without any apparent problems (Output of dmesg: es1371: version v0.27 time 17:21:08 May 3 2001 es1371: found chip, vendor id 0x1274 device id 0x5880 revision 0x03 PCI: Found IRQ 10 for device 00:0e.0 es1371: found es1371 rev 3 at io 0xd000 irq 10 es1371: features: joystick 0x0 ac97_codec: AC97 codec, id: 0x:0x (Unknown) ) and the mixer appears to work (I used gom -it), but when I try to play something (an mp3 with mpg123, a wav or .au with sox's "play", esd's test by simply running "esd"), the relevant program "freezes" and no sound comes out. (By "freezes" I mean it doesn't have any progress - it is not the case that they are playing but the sound is muted.) A search on google didn't turn anything up (other than that 1274:5880 works with the es1371 module...) - maybe I need to refine my searching techniques though... ;) Any ideas? Oh, and I'm using kernel 2.4.2.. I'm gonna try alsa now. Thanks, Hugo van der Merwe ps. I'll appreciate a CC:. -- To send me private (non-world-readable) mail, GPG encrypt it. 1024D/60715698: 5F2E 8EC2 E0A4 5D25 0569 F281 4A6C D76D 6071 5698 pgpiTijQU0ZvZ.pgp Description: PGP signature
/dev/ttyS0 permissions changing...
Hello,
Something on my machine is periodically changing /dev/ttyS0's
permissions to 640, whereas it needs to be 660 to use the modem... I
have now switched back to using ttyS1, but on laptops with only one
serial port this is not an option, so I'm still wondering what it is
that does this. I also want to know out of pure interest. Any ideas?
There's nothing obvious (at least to me at this point in time) in
cron.{d,daily,weekly,monthly}... I suppose I should try and find out
exactly when this happens, but that will take a couple of days... as it
doesn't happen daily.
Thanks,
Hugo van der Merwe
Re: dpkg --getselections, apt, and update-menus
Mmm. I'm reading via the web site, as I am not subscribed to this list... makes replying a little harder. Maybe I ought to specifically ask CC:'s to me. > >> getselections is supposed to be current. A package marked deinstall is > >> removed, but not purged. Or at least it is expected to be. > > > >On my machine, it definately doesn't work like that. I just tested it. > >Changed a selection in dselect, but didn't do an "Install", the > >selection change was evident in dpkg --get-selections. > > I don't think that's what Sean meant. If, in dselect, you use '-' to > mark a package for removal - or use 'apt-get remove' - then a package's > configuration files won't be removed, and some state about the package > will remain in the system. However, if you use '_' in dselect, or > 'apt-get --purge remove', then the package's configuration files will be > removed too, and /var/lib/dpkg/status will forget that it ever existed. Um, that's kind of not what I meant either. I wasn't replying to the remove/purge thing, rather just stating the fact that --get-selections does not coincide with the current state of the machine. Rather with the future desired state of the machine. > >> a) # dpkg --get-selections|sed -e 's/deinstall$/install/'|dpkg > >--set-selections > >># apt-get dselect-upgrade > > > >Ah yes... sed. The problem with this is, it will also install the > >packages that are not installed (that are really currently > >"deinstalled"). > > No, completely uninstalled (purged or never installed) packages don't > show up in 'dpkg --get-selections' at all. I was here referring to packages that have been removed, but not purged. They show up in --get-selections, and will be installed by these commands. All I was aiming to do was get the selections to match with the current state of the machine. I think the best way to go about his is sed'ing through the status file. Thanks, Hugo van der Merwe
Re: dpkg --getselections, apt, and update-menus
> getselections is supposed to be current. A package marked deinstall is > removed, but not purged. Or at least it is expected to be. On my machine, it definately doesn't work like that. I just tested it. Changed a selection in dselect, but didn't do an "Install", the selection change was evident in dpkg --get-selections. > There is no other > good way for an outside program to ask dpkg about the state of the system. I realise this is the problem. Which is why one might rather look at updating dpkg, rather than update-menus... (Oh, my machine is a very up-to-date woody, well, as up to date as the local mirror, which seems to update a couple of times a week, but not necessarily every day.) > a) # dpkg --get-selections|sed -e 's/deinstall$/install/'|dpkg > --set-selections ># apt-get dselect-upgrade Ah yes... sed. The problem with this is, it will also install the packages that are not installed (that are really currently "deinstalled"). Maybe one should use sed on /var/lib/dpkg/status...? Replacing lines containing "Status: install ok deinstall" or something like that with "Status: install ok install"... It does sort of feel wrong to go this "low level" though. > b) run the dselect update as root, but actually browse as a user A little extra trouble, but probably the best solution thus far. Infact, I just got the idea, and tested it, to simply do: # dselect update ; su hugo -c "dselect select" I could just make this a "listnewpackages" script or alias... > c) dselect has a 'revert' key binding, but it only works for that run of > dselect. So if you exit, there is no help. Yup... a "global revert" would be usefull. > Personally, I think console apt should have some way of showing new packages. > if a package is marked 'new' and you have the sort order set to show new items > first, it would emulate dselect fairly well. If possible, this way should be avoidable if it makes console-apt any slower, I think. But personally, on my machine, I would also love to see this. Thanks, Hugo van der Merwe
dpkg --getselections, apt, and update-menus
Lately I've been using console-apt for installations, but I tend to run dselect, just to see what packages are new. As such, everytime I ran it, a couple of package selections changed due to upgrades, depencencies, etc. I didn't care about it, as I don't use dselect anymore. I found it rather odd that many things were no longer on my system's menu system. That's when I studied how update-menu's generates the menu. I noticed it used dpkg --get-selections, which I ran myself, and noticed the hundreds of "deinstall" selections, and realised what the fault was. I just went through dselect and selected everything that was installed, that fixed up my system. A couple of ways I could have avoided this: - Is there some other easy way to see what packages are new? I would prefer to not use dselect at all, but this is something I really like. I thought of making this a wishlist bug for console-apt, but I don't think it will "fit in" well with the current (simple/unbloated/fast) design of capt (I like capt being as fast as possible, for use on old computers, so maybe `list-new-packages` should be a seperate program). One needs a backup of the previous package lists to do this. I think dselect did this with /var/lib/dpkg/available[-old], does using only apt/capt update these files, or possibly not the -old one? - Maybe update-menus should rather look at what is currently installed, rather than what the selections for the next "install/remove" operation will be. It does seem to make more sense to me. - This is maybe the best/most useful solution at this point: some easy way to sync the selections with what is currently installed? i.e. "undo" all selections. How can one do this? I admit to not have researched this properly: I haven't checked mailing lists for previous discussions, and haven't checked the BTS, so I apologize if this has already been discussed. Thanks, Hugo van der Merwe
fullscreen apps (in X) - windowmanager support?
Hello, The best way (as far as I know the only way) to get an app to full screen in X, is to drop the resolution (assuming you have many modelines defined) to the resolution of the app, or the smalelst resolution a little larger, then to "position" the display correctly. The big problem is accidently moving the mouse, and thereby shifting the screen. Some apps support locking the mouse into the window, some apps do all this automatically. Is there a way to lock the mouse into a window without that app specifically supporting it internally? That would help a lot to begin with, but ultimately I'd like to have/write a little program/script, possibly built into a window manager (I'll temporarly switch to another WM if necessary, but I'd prefer support for this in windowmaker ;), that, given a window handle or id or whatever its called in X, will automatically drop the resolution to the best size, lock the mouse cursor into the window, and if possible, "position" the display correctly. Any ideas? Thanks, Hugo van der Merwe
Re: network configuration - after install
On Wed, Oct 11, 2000 at 01:38:22PM +, Jamil Geor wrote: > Hi, > You could just use modconf to install the network card module that > you need. > > Jamil Thanks. It appears, however, that I didn't make the question very clear. Just installing the module is not going to get the network configured: /etc/hosts, /etc/resolv.conf, /etc/network/interfaces... these files have all not been set up as necessary, and I think I may be missing a couple, which is why I'm looking for something similar to "pcnetconfig" (which is for pcmcia network cards). Thanks, Hugo van der Merwe > On Mon, 9 Oct 2000, Hugo van der Merwe wrote: > > pcnetconfig does this job. Is there an easy way to get the networking > > configured after a base install? (I'm considering going through the
network configuration - after install
Hello, I am doing another debian install, and was uncertain about the network card type. I decided to skip that part of the install, hoping to find an easy way to do it after the basic system is up and running. On a laptop, pcnetconfig does this job. Is there an easy way to get the networking configured after a base install? (I'm considering going through the installation menu again.) Thanks, Hugo van der Merwe ps. Is /dev/cdrom created when the module is loaded, or some other time by some other script?
Re: pseudo-image kit: where to get binary-*.list
On Fri, Sep 08, 2000 at 02:16:41PM +0200, [EMAIL PROTECTED] wrote: > > I got them from http://cdimage.debian.org/cd-images/2.2_rev0/i386/ . > > Cheers > Matthias Thanks, Hugo
pseudo-image kit: where to get binary-*.list
Hello, I have d/l'ed the pseudo image kit, and need the binary-i386-?.list's... however, the site mentioned in the readme, http://www.uk.debian.org/debian-cd/cd-images, does not exist. Where do I find the .list files? Please CC to me. I do not subscribe to debian-user, and have limited time for web access today. Thanks. Hugo van der Merwe
gdk-imlib1 version
Hi, For the past week at least, I have kept a couple of packages on hold (I run unstable), as they depend on gdk-imlib1 (>= 1.9.8.1), while on my mirror the version of gdk-imlib1 is 1.9.8-4. I was tempted to submit a bug report, but realised it might be only my mirror, and that someone else will quickly submit a report if this is really so. However, this has still not been fixed, and I checked a second South African (that's where I live) mirror, which also only has 1.9.8-4 available. Then I checked ftp.debian.org, which also has version 1.9.8-4. I suppose I should submit bug reports against packages that depend on gdk-imlib1 (>= 1.9.8.1)? This strikes me as funny though, because there have been updates to some of these packages, without this being fixed. I cannot believe something like this can slip through the cracks so easily, which is why I'm so hesitant to submit a bug report. Thanks, Hugo van der Merwe
Re: Improper shutdown - weird results
On Thu, 10 Aug 2000, John Foster wrote: > Just a guess but did he have the filesystem mounted when he ran fsck :-) Aha, that could be... where else would he have found fsck? OK, so how does one go about running fsck without mounting root? I suppose that's supposed to be on a set of rescue disks, to be loaded into a ramdisk? That makes me wonder, is there a good chance that it would have messed up other files as well? Later, Hugo
Improper shutdown - weird results
Hi, A friend of mine, for whom I installed Debian, recently accidentily turned off his computer before the shutdown process was complete. When he booted back up, the whole thing was messed up... (He had to run fsck manually, and probably said "y" to all questions.) It turned out that his /etc/group file was incorrect (out dated, or something weird, but still recognisable), his /etc/shadow file contained what should have been the current /etc/group, his /etc/passwd contained what should have been /etc/shadow! Needless to say, he couln't log in. (There was some error message that looked something like "::: No such file or directory" - not that I'm 100% sure of the formatting, this was all told over the telephone...). After moving the files around (in single user mode), and copying /etc/passwd- to /etc/passwd, he could log in again, but instead of the login prompt: [EMAIL PROTECTED]:~$ he had: I have no [EMAIL PROTECTED]:~$ (Or so he explained over the telephone...) I suggested deleting the user and recreating it, out of laziness. (Oops, I just realize that that probably took him out of the groups as well, but he'll figure that out quickly and fix it.) What I was wondering, firstly, is how these files could have been moved around... he logged out all the users, and it didn't sound like he was running anything strange just before it. The second thing is, why the "I have no name"? Anyone have any idea where that info is supposed to be? I'd think it simply uses the login name... Anyway, just sharing an "incomplete shutdown experience"... Hugo van der Merwe
SoundBlaster Vibra PCI 128
Hello,
Anyone have any ideas how to get a SoundBlaster Vibra PCI 128 working? (Or
if it is at all possible?)
Secondly, I get the feeling it should be possible to recompile just the
modules, after possibly selecting more or less modules in "make
menuconfig" (or make config of course), without recompiling the whole
kernel. How does one go about doing this using make-kpkg, or does one have
to use more "conventional" tools directly to be able to do this?
Thanks,
Hugo van der Merwe
ps. I get the feeling most people (incl. me) have a habit of automatically
CC:'ing to the author. Is it still necessary to mention that I would
appreciate it? ("it" = CC'ing to me at [EMAIL PROTECTED]) Thanks.
Re: Accton Cheetah PCI network card
> Its an RTL8139... Thanks, works like a dream. (The "Quick Installation Guide" claimed it was an EN1207D, which according to the Ethernet Howto should use de4x5 or tulip, so I tried those with all combinations of parameters I could think off... Should this be reported as a "bug" in the howto's? Also, that is IMHO the meanest part of the debian install - determining which network card module to load for a network install. (Doesn't linux autodetect this kind of thing once fully setup? Can this kind of functionality not be included in the installation disks?)) Hugo van der Merwe
Accton Cheetah PCI network card
Hi, Anyone know if there is support for an Accton Cheetah PCI network card? I got stuck with the install when I had to load a module for the network card. (Network install.) Thanks, Hugo van der Merwe (I'll appreciate a CC to [EMAIL PROTECTED])
Bogus BogoMips with new kernel?
Hello, I recently (this weekend) reïnstalled potato on a laptop, pentium 120, which used to have a couple of months old version of potato on it. It used to have a bogomips rating of about 48, if I remember correctly. Now, suddenly, with the default kernel (2.2.17...pre6 somethingorother, it used to run 2.2.13, I think, or something around there), I get a bogomips rating of 240! This certainly looks bogus to me. Any ideas? Please ensure that replies are cc'ed to me ([EMAIL PROTECTED]), as I cannot handle all the traffic on debian-user (I rather go for debian-laptop). Thanks, Hugo van der Merwe
Re: Funny format CD's...
> Yes. There is a kernel module called cdfs. Find it on the Web and > compile it and copy it to /lib/modules//fs. > Once installed, you can mount a cd with > mount -t cdfs /dev/ Thanks. I'll have a look at it. From what I have learned from HOWTO's etc., it appears that it is possible to write tracks seperately, e.g. I can write one track with $ cdrecord dev=0,0,0 -nofix -audio track1.cdr then later write another $ cdrecord dev=0,0,0 -nofix -audio track2.cdr then when I'm done: $ cdrecord dev=0,0,0 -fix My first attempt at doing this was this script: if [ "$#" = "1" ] ; then for i in `cat $1` ; do mpg123 -r 44100 --stereo -q -s $i | \ cdrecord -v dev=0,0,0 -audio -pad -swab -nofix - done fi After taking out the disk to make sure it is nearly full, I put it back and ran: $ cdrecord dev=0,0,0 -fix Cdrecord 1.8 (i686-pc-linux-gnu) Copyright (C) 1995-2000 Jörg Schilling Using generic SCSI-3/mmc CD-R driver (mmc_cdr). Driver flags : SWABAUDIO Starting to write CD/DVD at speed 4 in write mode for single session. Last chance to quit, starting real write in 1 seconds. cdrecord: Input/output error. close track/session: scsi sendcmd: retryable errorCDB: 5B 00 02 00 00 00 00 00 00 00 status: 0x2 (CHECK CONDITION) Sense Bytes: 70 00 05 00 00 00 00 13 00 00 00 00 72 03 00 00 Sense Key: 0x5 Illegal Request, Segment 0 Sense Code: 0x72 Qual 0x03 (session fixation error - incomplete track in session) Fru 0x0 Sense flags: Blk 0 (not valid) cmd finished after 0.017s timeout 480s Any ideas? Was my script messed up, or was this a freak occurance? I didn't notice anything strange while writing the tracks, except something about secsize being -1... Thanks, Hugo van der Merwe
Funny format CD's...
Is it possible to mount the 7th track of a CD? The first 6 being audio tracks, and the seventh the first and only data track? I believe it is possible to write another track to a CD-R device later, as long as the disk hasn't been fixated. I was just wondering if it was possible to add a data track after adding audio tracks... Is it otherwise possible to rip a specific track, but not making it an audio track? i.e. suppose I write a data track to track 7, I'd like to rip it back to the CD image it was... I'd guess cdparanoia won't work, 'cause it extracts audio tracks to wave files... the last thing I can think of, is dumping the whole CD, and then cutting off the first part and possibly the last part, so that you'll only be left with track 7... Any ideas? Thanks Hugo van der Merwe
Re: Suspend and X
> I set replies to go to debian-laptop. And you added debian-user to the addresses? > I actually posted one solution to the debian-laptop list this morning. I > force the laptop to switch to vt 1 when it suspends. Add the following > script to /etc/apm/event.d/01chvt (be sure it's executable): How important is that specific filename? I assume it is simply to determine in which order the scripts are executed... any specific reason why it should be executed early? > #!/bin/sh > # change to vt1 when suspending to keep X from locking > > if [ "$1" = suspend ]; then > chvt 1 > fi > > Of course after suspending the system comes back up on vt 1, but I can > easily switch back to X if it keeps the system from crashing. I still have my case statement in there, to not switch if it is currently on tty[1-9]. It seems to work very well if I use "apm -s" as root, but if I close the laptop display, it doesn't switch and it locks up... any ideas? Hugo van der Merwe
New HP Printers...?
Hello, I am looking at buying a new printer, and had my eyes on a HP DeskJet 710C, which I would definitely have bought if I only ran Windows. However, I wish to use this printer on a machine that runs potato exclusively. I had a look at the options of magicfilterconfig, and the only hp or deskjet options seem to be "deskjet, dj500, dj500c, dj550c". Do I just use one of these? I was sure debian would support the newer printers. A comparison of the /etc/magicfilter/dj* filters (3 of them, I've got version 1.2-36 installed, I don't think there is a later one, but it has been a couple of days since my last apt-get update) shows that the only difference is the parameters passed to gs. gs --help outputs 4 lines containing "dj", in those four lines, the following seem possibilities: djet500 cdeskjet cdjcolor cdj550 cdj500 djet500c hpdj cdj670 cdj850 cdj890 cdj1600 I suppose I could experiment with the cdj670 and cdj850 drivers, but I do not wish to spend money on a printer I cannot be sure about. What would you advise? What makes a good Linux printer, that isn't too expensive (similar in price and/or specs to the HP dj710c)? Is creating my own new magicfilter filter based on one of the others the best bet, followed by pointing the if= parameter in /etc/printcap to that filter? The differences between the dj500c and dj550c magicfilters are greater than that between dj500 and dj500c (which has hardly a difference at all). Parameters to gs: In dj500c the driver is selected with -sDEVICE=cdj500, whereas in dj550c there is no -sDEVICE parameter, it has "@cdj550.upp". Any ideas why, or what the @ means? The dj550c filter also does not have a -r parameter, is the default for gs 600, or is the default built into the "cdj550.upp" driver, maybe? Last but not least I noticed that the 500 filter makes use of djscript, but the 550 filter does not. I'd think it would be a good idea to make the 550 filter also make use of djscript, should I file a Severity:wishlist bug report? (BTW, I use a 560c at home, using the 550 filter.) Thanks in advance, Hugo van der Merwe ps. I would appreciate it if you could CC replies to me ([EMAIL PROTECTED]), as I usually read the list through deja.com, but not very often. (Hmm, I hope I'm not asking a question that was answered yesterday, I suppose the right thing for me to do would be to do a search in deja.com, but I'm offline now...) Are there any other ways to read/post to this list except deja.com and the mailing list (possibly digest to cut down on the incredible number of emails)? I'll have a look at the archives on the Debian website when I get the chance.
Package differences - bla_all.deb vs. bla.deb
What is the difference between the locales_2.1.2-5_all.deb package file and the locales_2.1.2-5.deb package file? I assume packages with a _i386 "extention" are specifically for the i386 architecture. That makes me think the _all package is for all architectures, but then what is the package without any "extention" for? Is there some place where I can see this kind of thing? I didn't notice it in the debian-policy documents. Maybe I didn't look in the right place? I would prefer a pointer to where I can get this information, so that I can easily get the complete information. I will not mind a direct answer though. Thanks for the list's eagerness to help, Hugo van der Merwe
RE: pcmcia net card does not show up in "ifconfig" listing.
On Thu, 30 Sep 1999, Pollywog wrote: > > On 30-Sep-99 Hugo van der Merwe wrote: > > I realise something is probably not configured correctly. I installed > > pcmcia-source, and compiled it with dpkg-kpkg, then installed the > > resulting .deb file. Everything seems to be working fine, except that I > > expected the network card in the slot to show up when I run ifconfig. It > > does not. > > > What happens when you do (as root): > > /etc/init.d/network start ? /etc/init.d/network contains the following 3 lines: #! /bin/sh ifconfig lo 127.0.0.1 route add -net 127.0.0.0 Needless to say, it doesn't do much... but it is identical to the script on the other laptop. Clearly the pcmcia stuff is not found in that script. /etc/init.d/pcmcia start does not solve the problem either. sum pcmcia and cksum pcmcia gives the same result on both laptops, so I assume the scripts are identical, so the fault is not in that script either. Most of the configuration stuff for pcmcia seems to be in /etc/pcmcia. Shall I start comparing files in that directory? Hugo van der Merwe
Re: eth0: unknown interface...
On Fri, 1 Oct 1999, Matthew Dalton wrote: > It may not help, but I installed debian on my laptop using a laplink > cable and it was not too difficult to set up. I was planning on using an > ethernet card, but I couldn't get one right away... so I spent $10 on a > laplink cable instead. I installed most of it by directly connecting the laptop hard drive to my desktop. Packages can be installed using chroot. The only thing that must be changed to make it work when it is in the laptop, is /etc/fstab, unless I missed something... Hugo van der Merwe
Re: pcmcia net card does not show up in "ifconfig" listing.
On Thu, 30 Sep 1999, Martin Fluch wrote: > On Thu, 30 Sep 1999, Hugo van der Merwe wrote: > > > I realise something is probably not configured correctly. I installed > > pcmcia-source, and compiled it with dpkg-kpkg, then installed the > > resulting .deb file. Everything seems to be working fine, except that I > > expected the network card in the slot to show up when I run ifconfig. It > > does not. > > Did you do a make-kpkg modules_images. There should show up a second > deb-package called pcmcia-... (I don't remember exactly). I did that yes. And installed the package. Is there any place to specify what drivers must be compiled (like in make menuconfig when using something like an NE2000 on a desktop)? > If the box says 'Windows 95 or better', it should run on Linux, right? > - anonymous Hugo van der Merwe
pcmcia net card does not show up in "ifconfig" listing.
I realise something is probably not configured correctly. I installed pcmcia-source, and compiled it with dpkg-kpkg, then installed the resulting .deb file. Everything seems to be working fine, except that I expected the network card in the slot to show up when I run ifconfig. It does not. How can I determine whether everything is working correctly? What must I do to make the card (an Accton network card) show up when "ifconfig" is run? I recompiled the kernel using the same options (under Network Device Support) as on the laptop on which the card works. BTW, I'm running Debian 2.1 (slink), kernel 2.0.36. I am not very experienced (Although I have got this card working on another laptop, as well as a well working desktop computer), so even tips that seem "obvious" can be useful. (Yes, the card does seem to be inserted completely, and I hear two beeps when inserting it (one lowish, and one an octave higher). ) Thanks in advance, Hugo van der Merwe ps. Please CC: replies to [EMAIL PROTECTED] - it is the only way to be 100% certain that replies will reach me.
Re: eth0: unknown interface...
> I have installed the debian base system on a laptop, but now I need to > install the rest over a network (from another brilliant working Debian > machine). For this to work, I must get the pcmcia network card working... I just realised, I think I didn't make use of the menu item "Configure PCMCIA Support". Is there an easy way of configuring PCMCIA support for installation once I've already installed the base system? Can I run dbootstrap again? Where do I get it (the boot disks aren't mountable, right?) Or is it easier to reinstall from scratch? > I installed the same kernel package I installed on anther laptop, and the > pcmcia-modules package that was compiled with it, and am using the same > network card. However, even with this kernel, eth0 is still an unknown > device. How does Debian Linux get to know eth0? I though it was only in > the kernel, but clearly it isn't. Or what modules must I load with > modprobe? As I just mentioned, I think the pcmcia card isn't properly configured/initialised. Thanks in advance, Hugo van der Merwe -- ps. Please ensure replies are also sent to [EMAIL PROTECTED] - I do not subscribe to the mailing list, and I mistrust the news server I read the list through.
eth0: unknown interface...
Hello, I have installed the debian base system on a laptop, but now I need to install the rest over a network (from another brilliant working Debian machine). For this to work, I must get the pcmcia network card working... I installed the same kernel package I installed on anther laptop, and the pcmcia-modules package that was compiled with it, and am using the same network card. However, even with this kernel, eth0 is still an unknown device. How does Debian Linux get to know eth0? I though it was only in the kernel, but clearly it isn't. Or what modules must I load with modprobe? Thanks in advance, Hugo van der Merwe -- ps. Please ensure replies are also sent to [EMAIL PROTECTED] - I do not subscribe to the mailing list, and I mistrust the news server I read the list through.
Re: Installing debian to another hard drive
> On Tue, Sep 28, 1999 at 10:42:15PM +0200, Hugo van der Merwe wrote: > > Hello, > > > > How can I install the debian "base" system to another hard drive in the > > same computer? I have a working Debian / DOS+Win95 dual boot system, and > > would like to install Debian to a laptop, whose floppy disk drive is > > disfunctional, and has no CD-ROM drive. > > > > I have connected the laptop's hard drive to my Debian machine, using an > > adapter. There must be some way to install Debian to this hard drive, > > using my desktop machine. How can I do this? Will the normal > > install-from-dos-partition procedure work? I tried the following: I did a normal install from my dos partition (from dos), and installed the base system, and made the disk bootable, BUT it was all done on /dev/hdc, the laptop hard drive. Then I placed the disk back in the laptop, and booted it, and received errors, fstab was referencing /dev/hdc - I changed all those references (two) to /dev/hda, now it boots fine. However, the laptop cannot ping other computers, and it cannot ping itself at 192.168.1.8, which it should, but it can ping itself at 127.0.0.1 route displays only one entry, namely 127.0.0.0 * 255.0.0.0 U 0 0 1 lo I believe it is supposed to display more. I need to get ftp/nfs working to install more stuff (I don't mind using floppy disks to get it working though). Where is the network configuration stuff stored, or is there a nice net config program? Thanks, Hugo ps. Please cc: to [EMAIL PROTECTED], as I am not subscribed to the list, and sometimes miss a reply when reading the newsgroups...
Installing debian to another hard drive
Hello, How can I install the debian "base" system to another hard drive in the same computer? I have a working Debian / DOS+Win95 dual boot system, and would like to install Debian to a laptop, whose floppy disk drive is disfunctional, and has no CD-ROM drive. I have connected the laptop's hard drive to my Debian machine, using an adapter. There must be some way to install Debian to this hard drive, using my desktop machine. How can I do this? Will the normal install-from-dos-partition procedure work? Thanks in advance, Hugo van der Merwe ps. I would appreciate if replies could be CC:'ed to [EMAIL PROTECTED] I am not subscribed to the list, and sometimes my ISP's news server messes up...
Installing Debian to different hard disk
Hello, How can I install the debian "base" system to another hard drive in the same computer? I have a working Debian / DOS+Win95 dual boot system, and would like to install Debian to a laptop, whose floppy disk drive is disfunctional, and has no CD-ROM drive. I have connected the laptop's hard drive to my Debian machine, using an adapter. There must be some way to install Debian to this hard drive, using my desktop machine. How can I do this? Will the normal install-from-dos-partition procedure work? Thanks in advance, Hugo van der Merwe ps. I would appreciate if replies could be CC:'ed to [EMAIL PROTECTED] I am not subscribed to the list, and sometimes my ISP's news server messes up...
dselect and contrib/non-free - DWN-6/22
> Wichert Akkerman has proposed splitting non-free and contrib out to a > new server, nonfree.debian.org. The impetus for this is to emphasize > that they are not part of Debian, since "the distinction isn't as > visible as it used to be; advances in searching in the distribution > and tools like apt make it very hard to see when something is in main > and when not". The proposal will come up for a vote soon and is > generating lots of discussion. One thing that might help is if dselect does not sort purely alphabetical when sorting by section. contrib/* should not be between comm and devel, but rather be apart from main. Same thing for non-free. Just another thought, Hugo van der Merwe
Re: ftp/telnet fails since upgrade
On Thu, 17 Jun 1999, William Ono wrote: > On Wed, 16 Jun 1999, Hugo van der Merwe wrote: > > > having ALL: ALL in hosts.deny and ALL: 192.168.1.* in hosts.allow worked > > perfectly. hosts.allow was supposed to override hosts.deny. When did this > > change? > > You want "ALL: 192.168.1." without the star at the end. > > The behaviour you mention is AFAIK still the correct one. Ah, yes. I have now restored my hosts.deny to its former meanness, and removed the star at the end of 192.168.1., as well as add 127.0.0. without which "telnet localhost" didn't work. > hth. It did, thanks. Hugo van der Merwe
Re: ftp/telnet fails since upgrade
On Tue, 15 Jun 1999, Bruce Chapman wrote:
> >longer ftp or telnet to my machine. ("421 Service not available, remote
> >server has closed connection", and "Connection closed by foreign host").
>
> In /etc/hosts.deny try commenting out the line ALL: PARANOID
The problem was not ALL: PARANOID, it was ALL: ALL. (Which I placed there
myself.) Funny, in my redhat days, and Debian 2.1 days as well, I believe,
having ALL: ALL in hosts.deny and ALL: 192.168.1.* in hosts.allow worked
perfectly. hosts.allow was supposed to override hosts.deny. When did this
change?
Thanks for everyone's help, do you think I should post a summary of
suggestions for someone with a similar problem in the future? I'll do that
this weekend, or Friday.
I have also installed dhis, and it works perfectly, circumventing my ISP's
firewalls as well... No more hiding behind them, I now have to worry about
my own security. What do most people place in their hosts.allow and
hosts.deny files?
Hugo van der Merwe
Re: ftp/telnet fails since upgrade
> >longer ftp or telnet to my machine. ("421 Service not available, remote
> >server has closed connection", and "Connection closed by foreign host").
>
> In /etc/hosts.deny try commenting out the line ALL: PARANOID
Thanks. That didn't help with the ftp/telnet problem, but I think it
solved another problem that crept up today.
Hugo van der Merwe
Re: Configuring exim (or rather fetchmail and procmail)
> > But I am not sure it is the right thing to do, why wouldnt you want to > > receive > > mail directly? > > Because I am going to want to use procmail soon, and it's easier > to configure it to only work through fetchmail, and because if I > try to fetch mail with POP3 and SMTP a collision sometimes occurs. >From my experiences with procmail, fetchmail and exim: I use fetchmail to retrieve mail from pop3 servers, then fetchmail passes the message to exim (or that's what I believe). Here follows my fetchmailrc, procmailrc, ip-up and ip-down scripts for anyone who wishes to configure procmail and fetchmail: ~/.fetchmailrc: defaults interface "ppp0/196.2.0.0/255.255.0.0" #very nice for a ppp link. poll pop3.mweb.co.za protocol POP3 user username_here password password_here fetchall The user can then run fetchmail to retrieve mail. A system wide /etc/fetchmailrc is also useful, so that mail is retrieved periodically by the system, for all users: set daemon 300 #check mail every 5 minutes (300 seconds) defaults interface "ppp0/196.2.0.0/255.255.0.0" poll pop3.mweb.co.za protocol POP3 #download mail from hugovdm account to user hugo user hugovdm is hugo password password_here fetchall poll mail.fnmail.com protocol POP3 user anneke is anneke password password_here fetchall poll mail.fnmail.com protocol POP3 user mayvdm is mayvdm password password_here fetchall /etc/ppp/ip-up.d/fetchmail-up #!/bin/sh test -r /etc/fetchmailrc && \ fetchmail --syslog --invisible --fetchmailrc /etc/fetchmailrc /etc/ppp/ip-down.d/fetchmail-down #!/bin/sh test -x /usr/bin/fetchmail && \ fetchmail -q I then created a .forward file for myself. I believe exim looks at the .forward file to see what to do with mail, so whether using exim or fetchmail makes no difference to the working of procmail. (Someone correct me if I'm wrong.) ~/.forward: | /usr/bin/procmail This just pipes all mail to procmail. ~/.procmailrc: PATH=/usr/local/bin:/usr/bin:/bin MAILDIR=$HOME/mail LOGFILE=$MAILDIR/procmail-logfile :0: * [EMAIL PROTECTED] Debian-Changes :0: * [EMAIL PROTECTED] Debian-User This worked for me. I haven't used it lately though (I renamed .forward to disable it). Notes: :0 - marks beginning of recipe, :0: - tells procmail to use a lock file with same name as folder. * - pass expression to internal egrep ^ - start of a line . - any character * - any number >= 1 of . (.* represents any number of any characters). Debian-Changes - file matching mail in this folder. If a mail message matches no recipies, it is filed in $DEFAULT, which is, by default, $ORGMAIL, which is, by default, /var/spool/mail/$LOGNAME. See man procmailrc for more info. (And man fetchmail (not man fetchmailrc).) HTH someone, Hugo van der Merwe
Re: ftp/telnet fails since upgrade
> For the ftp problem, take a look at /etc/inetd.conf and see if the ftp > line has been commented out. If so, remove the octothorpe (#). There > might be more than one of these lines; you should only enable one. These two lines are uncommented in inetd.conf: ftp stream tcp nowait root/usr/sbin/tcpd \ /usr/sbin/in.ftpd telnet stream tcp nowait root/usr/sbin/tcpd \ /usr/sbin/in.telnetd (They are both on one line. The \ is just to wrap the mail message.) Maybe I should try upgrading to 2.2.9? Upgrading to 2.0.36 from 2.0.34 solved my sound problems back in my 2.0 days. Hugo van der Merwe
Re: ftp/telnet fails since upgrade
> >I have, however, recently run into a problem. Since I upgraded, I can no
> >longer ftp or telnet to my machine. ("421 Service not available, remote
> >server has closed connection", and "Connection closed by foreign host").
> >I'm not sure if this started happening when I installed potato, or when
> >I installed kernel 2.2.7. I think it was the latter though.
>
> It happened when you upgraded to Potato.
>
> Potato seperates a lot of the services that used to be in one package
> into several packages...you want to install the telnetd package to be
> able to telnet to your box. you'll probably want to look at the
> packages under net carefully to find out what other services you're
> missing...
telnetd is installed (and there is a /usr/sbin/in.telnetd file on my
system). I have installed nearly all packages under "Standard". (As well
as "Req" and "Imp".)
inetd is running. (Don't know if it is configured correctly though...)
This is the output I get:
# telnet localhost
Trying 127.0.0.1...
Connected to localhost
Escape character is '^]'.
Connection closed by foreign host.
#
Thanks for your replies,
Hugo van der Merwe
ftp/telnet fails since upgrade
Hello!
I have been using Debian for a while now (before that I used RH5.1). I
really prefer Debian, its packaging system, and the fact that it works
so nicely from a shell. (You don't need to load X before configuring
stuff.) IPMasquerading and the like was a lot easier to set up.
I have, however, recently run into a problem. Since I upgraded, I can no
longer ftp or telnet to my machine. ("421 Service not available, remote
server has closed connection", and "Connection closed by foreign host").
I'm not sure if this started happening when I installed potato, or when
I installed kernel 2.2.7. I think it was the latter though.
My hosts.allow and .deny files stayed the same, and I can still telnet
from this machine to others. (Not to itself.)
Any ideas?
TIA,
Hugo van der Merwe
Too many posts for my mailbox
My mailbox cannot handle the number of posts this list generates, especially not when I'm away for the weekend or even week. What should I do to participate in this mailing list without clogging my mailbox? Please send a reply directly to [EMAIL PROTECTED], because I'm now going to temporarily unsubsribe from this mailing list, until I find an alternate solution. TIA, Hugo van der Merwe
