error with auctex on Debian - request for reproduction
Hi,
Is there anyone out there who can check
http://tex.stackexchange.com/questions/151696/journal-of-statistical-software-class-breaks-preview-in-auctex
and tell me whether it is reproducible?
I heard from someone on the AUCTeX list, who could not reproduce it, so it
may be Debian specific. I'm using the default TeX Live installation on
wheezy.
The full text of the SO question follows. If you want more details, please
ask. I also attach the output of the preview run.
Please CC me on any reply. Thanks.
Regards, Faheem
#
I'm using `AUCTeX 11.86-11` on Debian stable.
The `jss.sty` style file contained in the [JSS style file zip
archive](http://www.jstatsoft.org/downloads/JSSstyle.zip) breaks AUCTeX's
preview for me even with a simple file. NOTE: the `jsslogo.jpg` file is
also required.
For example,the following file
\documentclass[article]{jss}
\begin{document}
$x$
\end{document}
gives the error
LaTeX found no preview images
The more detailed error is:
ERROR: Illegal parameter number in definition of \MYLATEXline.
--- TeX said ---
to be read again
1
l.386 \gdef\FV@BeginScanning#1^^M
{%
Can any AUCTeX user
1. confirm this
2. tell me what the problem is?
I'll also try the AUCTeX list and update here if a solution is found.
Running `Preview-LaTeX' on `test' with ``pdflatex -ini -interaction=nonstopmode
pdflatex prv_test.ini
\nonstopmode\nofiles\PassOptionsToPackage{active,tightpage,auctex}{preview}\AtBeginDocument{\ifx\ifPreview\undefined\RequirePackage[displaymath,floats,graphics,textmath,sections,footnotes]{preview}[2004/11/05]\fi}
\input test.tex''
This is pdfTeX, Version 3.1415926-2.4-1.40.13 (TeX Live 2012/Debian) (INITEX)
restricted \write18 enabled.
entering extended mode
(./prv_test.ini
LaTeX2e 2011/06/27
Babel v3.8m and hyphenation patterns for english, dumylang, nohyphenation, lo
aded.
(/usr/share/texlive/texmf-dist/tex/latex/carlisle/mylatex.ltx)
(/usr/share/texlive/texmf-dist/tex/latex/tools/.tex File ignored))
No auxiliary output files.
(./test.tex (./jss.cls
Document Class: jss 2013/04/06 2.2 jss class by Achim Zeileis
(/usr/share/texlive/texmf-dist/tex/latex/base/article.cls
Document Class: article 2007/10/19 v1.4h Standard LaTeX document class
(/usr/share/texlive/texmf-dist/tex/latex/base/size11.clo))
(/usr/share/texlive/texmf-dist/tex/latex/graphics/graphicx.sty
(/usr/share/texlive/texmf-dist/tex/latex/graphics/keyval.sty)
(/usr/share/texlive/texmf-dist/tex/latex/graphics/graphics.sty
(/usr/share/texlive/texmf-dist/tex/latex/graphics/trig.sty)
(/usr/share/texlive/texmf-dist/tex/latex/latexconfig/graphics.cfg)
(/usr/share/texlive/texmf-dist/tex/latex/pdftex-def/pdftex.def
(/usr/share/texlive/texmf-dist/tex/generic/oberdiek/infwarerr.sty)
(/usr/share/texlive/texmf-dist/tex/generic/oberdiek/ltxcmds.sty
(/usr/share/texlive/texmf-dist/tex/latex/graphics/color.sty
(/usr/share/texlive/texmf-dist/tex/latex/latexconfig/color.cfg))
(/usr/share/texlive/texmf-dist/tex/latex/ae/ae.sty
(/usr/share/texlive/texmf-dist/tex/latex/base/fontenc.sty
(/usr/share/texlive/texmf-dist/tex/latex/base/t1enc.def)
(/usr/share/texlive/texmf-dist/tex/latex/ae/t1aer.fd)))
(/usr/share/texlive/texmf-dist/tex/latex/fancyvrb/fancyvrb.sty
Style option: `fancyvrb' v2.7a, with DG/SPQR fixes, and firstline=lastline fix
2008/02/07 (tvz)
! Illegal parameter number in definition of \MYLATEXline.
to be read again
1
l.386 \gdef\FV@BeginScanning#1^^M
{%
! Illegal parameter number in definition of \MYLATEXline.
to be read again
1
l.389 \endgroup
! Illegal parameter number in definition of \MYLATEXline.
to be read again
1
l.395 ...be discarded. Hit return to continue.}}
! Illegal parameter number in definition of \MYLATEXline.
to be read again
1
l.402 \gdef\FancyVerbGetLine#1^^M
{%
! Illegal parameter number in definition of \MYLATEXline.
to be read again
1
l.413 \endgroup
! Illegal parameter number in definition of \MYLATEXline.
to be read again
1
l.413 \endgroup
! Illegal parameter number in definition of \MYLATEXline.
to be read again
1
l.425 ...n#1{\@ifnextchar\@nil{\@gobble}{\FV@EOF}}
! Illegal parameter number in definition of \MYLATEXline.
to be read again
1
l.426 ...@nil#1\@empty\else\expandafter\FV@EOF\fi}
! Illegal parameter number in definition of \MYLATEXline.
to be read again
1
l.426
[SOLVED] Re: dpkg error
Hi Javier, On Sat, 18 Apr 2009, Javier Barroso wrote: On Thu, Apr 16, 2009 at 7:53 PM, Faheem Mitha fah...@email.unc.edu wrote: [snip] It seems like /var/lib/dpkg/available is corrupt ? Does aptitude update help ? If not, 'dpkg --clear-avail' , could help. Thanks for the reply. I got some help on IRC and dealt with the problem. The recommended approach was to remove /var/lib/dpkg/available and run dselect update which regenerates it, which worked. Maybe aptitude update would work too. I should have posted to the list saying this query was resolved. Sorry. Regards, Faheem Mitha. -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
dpkg error
Hi, I'm getting the following error on upgrade (debian Lenny). As suggested by the error message, there is some funny stuff in /var/lib/dpkg/available in the kdebase package section. Is this is a known issue? If not, can anyone suggest how to debug this? For the record, I've being using and adminstering Debian since 2001, and don't recall having ever seen an error like this before. Please CC me on any reply. Thanks. Regards, Faheem Mitha. merlin:/home/faheem# apt-get upgrade Reading package lists... Done Building dependency tree Reading state information... Done The following packages have been kept back: linux-image-2.6-686 The following packages will be upgraded: base-files bind9-host dnsutils firmware-iwlwifi gdm gtk2-engines-pixbuf imagemagick libbind9-40 libdns45 libgtk2.0-0 libgtk2.0-bin libgtk2.0-common libisc45 libisccc40 libisccfg40 liblwres40 libmagick10 libpam-modules libpam-runtime libpam0g libpq5 libsmbclient libvolume-id0 libwbclient0 linux-libc-dev openoffice.org openoffice.org-base openoffice.org-base-core openoffice.org-calc openoffice.org-common openoffice.org-core openoffice.org-draw openoffice.org-evolution openoffice.org-filter-mobiledev openoffice.org-gcj openoffice.org-gnome openoffice.org-gtk openoffice.org-help-en-us openoffice.org-impress openoffice.org-java-common openoffice.org-math openoffice.org-officebean openoffice.org-report-builder-bin openoffice.org-style-andromeda openoffice.org-style-tango openoffice.org-writer postgresql postgresql-8.3 postgresql-client-8.3 python-uno texlive-bibtex-extra texlive-fonts-extra texlive-fonts-extra-doc texlive-generic-extra texlive-humanities texlive-humanities-doc texlive-latex-extra texlive-latex-extra-doc texlive-math-extra texlive-pstricks texlive-pstricks-doc texlive-publishers texlive-publishers-doc ttf-opensymbol udev xserver-xorg-core xserver-xorg-video-savage 67 upgraded, 0 newly installed, 0 to remove and 1 not upgraded. Need to get 0B/331MB of archives. After this operation, 3119kB of additional disk space will be used. Do you want to continue [Y/n]? Extracting templates from packages: 100% Preconfiguring packages ... dpkg: parse error, in file `/var/lib/dpkg/available' near line 11824 package `kdebase': `Recommends' field, reference to `kdm': version contains ` ' E: Sub-process /usr/bin/dpkg returned an error code (2) * Package: kdebase Priority: optional Section: kde Installed-Size: 68 Maintainer: Debian Qt/KDE Maintainers debian-qt-...@lists.debian.org Architecture: all Version: 4:3.5.9.dfsg.1-6 Depends: kappfinder (= 4:3.5.9.dfsg.1-6), kate (= 4:3.5.9.dfsg.1-6), kcontrol (= 4:3.5.9.dfsg.1-6), kdebase-bin (= \ 4:3.5.9.dfsg.1-6), kdebase-data (= 4:3.5.9.dfsg.1-6), kdebase-kio-plugins (= 4:3.5.9.dfsg.1-6), kdepasswd (= 4:3.5.9\ .dfsg.1-6), kdeprint (= 4:3.5.9.dfsg.1-6), kdesktop (= 4:3.5.9.dfsg.1-6), kfind (= 4:3.5.9.dfsg.1-6), khelpcenter (\ = 4:3.5.9.dfsg.1-6), kicker (= 4:3.5.9.dfsg.1-6), klipper (= 4:3.5.9.dfsg.1-6), kmenuedit (= 4:3.5.9.dfsg.1-6), konq\ ueror-nsplugins (= 4:3.5.9.dfsg.1-6), konqueror (= 4:3.5.9.dfsg.1-6), konsole (= 4:3.5.9.dfsg.1-6), kpager (= 4:3.5\ .9.dfsg.1-6), kpersonalizer (= 4:3.5.9.dfsg.1-6), ksmserver (= 4:3.5.9.dfsg.1-6), ksplash (= 4:3.5.9.dfsg.1-6), ksys\ guard (= 4:3.5.9.dfsg.1-6), ktip (= 4:3.5.9.dfsg.1-6), kwin (= 4:3.5.9.dfsg.1-6), libkonq4 (= 4:3.5.9.dfsg.1-6), ha\ l | kfreebsd-gnu | hurd, pmount | kfreebsd-gnu | hurd Recommends: kdm (= 4:3.5.9^Fd^O^Fg.^A- ^...@^@^...@\377\373\353\277sts: kdebase-doc-html (GenuineIntelsg.1-6) Size: 38376 Description: base components from the official KDE release KDE (the K Desktop Environment) is a powerful Open Source graphical desktop environment for Unix workstations. It combines ease of use, contemporary functionality, and outstanding graphical design with the technological superiority of the Unix operating system. . This metapackage includes the nucleus of KDE, namely the minimal package set necessary to run KDE as a desktop environment. This includes the window manager, taskbar, control center, a text editor, file manager, web browser, X terminal emulator, and many other programs and components. Homepage: http://www.kde.org -- To UNSUBSCRIBE, email to debian-user-requ...@lists.debian.org with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org
network problem with router
Dear People, I'm having some odd networking problems with my router, The D-Link DGL-4300 (http://www.amazon.com/gp/product/B0006TIA02/002-4711104-7484852). Incoming ssh connections through the router using port forwarding to two different machines with different network cards and different drivers are drastically slow with kernel versions 2.6.18 and 2.6.17 (stock Debian kernels), both initial connections and ongoing traffic. This problem does not show up when running 2.6.16 kernels or earlier (as far as I can tell) on either machine. I'm flummoxed as to what to do about this. Any suggestions gratefully appreciated. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: two network cards on a server
On Thu, 19 Apr 2007, Roberto C. Sánchez wrote: On Wed, Apr 18, 2007 at 09:18:10PM -0400, Faheem Mitha wrote: eth0 Link encap:Ethernet HWaddr 00:E0:81:31:33:8C inet addr:152.3.172.111 Bcast:152.3.173.255 Mask:255.255.254.0 eth1 Link encap:Ethernet HWaddr 00:E0:81:31:33:64 inet addr:152.3.172.60 Bcast:152.3.173.255 Mask:255.255.254.0 You can see here that they are on the same subnet. The problem, as I mentioned previously, is that having two physical interfaces on the same subnet can confuse your system quite badly. What is the output of `/sbin/route -n` ? /sbin/route -n Kernel IP routing table Destination Gateway Genmask Flags Metric RefUse Iface 152.3.172.0 0.0.0.0 255.255.254.0 U 0 00 eth1 152.3.172.0 0.0.0.0 255.255.254.0 U 0 00 eth0 0.0.0.0 152.3.172.1 0.0.0.0 UG0 00 eth1 Somebody suggested that using two network addresses with one network card should be possible. Would this make these problems go away? Thanks.Faheem.
Re: two network cards on a server
On Fri, 20 Apr 2007, Alex Samad wrote:
On Thu, Apr 19, 2007 at 05:15:25PM -0400, Roberto C. Sánchez wrote:
On Thu, Apr 19, 2007 at 05:06:47PM -0400, Faheem Mitha wrote:
/sbin/route -n
Kernel IP routing table
Destination Gateway Genmask Flags Metric RefUse
Iface
152.3.172.0 0.0.0.0 255.255.254.0 U 0 00 eth1
152.3.172.0 0.0.0.0 255.255.254.0 U 0 00 eth0
0.0.0.0 152.3.172.1 0.0.0.0 UG0 00 eth1
Somebody suggested that using two network addresses with one network card
should be possible. Would this make these problems go away?
sorry, what are the problems that you are seeing ?
They were described in the first post to this thread. Namely one of the
two network cards on this machine (the one configured later) behaves
eratically, is only intermittently reachable from outside, and is
constantly broadcasting dhcp requests.
Yes. I have one server (firewall actually) with 8 IPs on one physical
interface. It simplifies things greatly. You want to search for how to
do aliases.
have a look for ip (it part of the iproute package)
Usage: ip addr {add|del} IFADDR dev STRING
ip addr {show|flush} [ dev STRING ] [ scope SCOPE-ID ]
[ to PREFIX ] [ FLAG-LIST ] [ label PATTERN ]
IFADDR := PREFIX | ADDR peer PREFIX
[ broadcast ADDR ] [ anycast ADDR ]
[ label STRING ] [ scope SCOPE-ID ]
SCOPE-ID := [ host | link | global | NUMBER ]
FLAG-LIST := [ FLAG-LIST ] FLAG
FLAG := [ permanent | dynamic | secondary | primary |
tentative | deprecated ]
Er, why is this relevant? Sorry if I am being clueless.
Thanks. Faheem.
two network cards on a server
Hi everyone, I've got a server on which there are two network cards. One of the network cards has been configured, and has been running satisfactorily for a while. I recently plugged in the other network card and started using it for a different IP address. However, I've been having persistent problems with that interface. Namely, some of the time I can't connect to the interface, which seems to go up and down unpredictably, and floods the system logs with DHCP requests. Clearly something is badly wrong here, and I am hoping this is something simple that can be fixed easily. This typically looks like Apr 18 15:16:42 florence kernel: e100: eth0: e100_watchdog: link up, 100Mbps, full-duplex Apr 18 15:16:43 florence dhclient: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 3 Apr 18 15:16:43 florence dhclient: receive_packet failed on eth0: Network is down Apr 18 15:16:45 florence dhclient: DHCPDISCOVER on eth0 to 255.255.255.255 port 67 interval 3 Apr 18 15:16:46 florence dhclient: DHCPOFFER from 152.3.172.1 Apr 18 15:16:46 florence dhclient: DHCPREQUEST on eth0 to 255.255.255.255 port 67 Apr 18 15:16:46 florence dhclient: DHCPACK from 152.3.172.1 Apr 18 15:16:46 florence dhclient: bound to 152.3.172.111 -- renewal in 43200 seconds. Apr 18 15:16:48 florence dhclient: DHCPREQUEST on eth1 to 152.3.250.61 port 67 Apr 18 15:16:48 florence dhclient: DHCPACK from 152.3.250.61 Apr 18 15:16:48 florence dhclient: bound to 152.3.172.60 -- renewal in 43200 seconds. Apr 18 15:16:53 florence kernel: eth0: no IPv6 routers present If any has any ideas about this, please let me know. Please copy me at [EMAIL PROTECTED] Thanks. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: two network cards on a server
On Wed, 18 Apr 2007, Roberto C. Sánchez wrote: On Wed, Apr 18, 2007 at 04:20:49PM -0400, Faheem Mitha wrote: Hi everyone, I've got a server on which there are two network cards. One of the network cards has been configured, and has been running satisfactorily for a while. I recently plugged in the other network card and started using it for a different IP address. However, I've been having persistent problems with that interface. Namely, some of the time I can't connect to the interface, which seems to go up and down unpredictably, and floods the system logs with DHCP requests. Clearly something is badly wrong here, and I am hoping this is something simple that can be fixed easily. Are eth1 and eth0 configured to be on the same subnet? Regards, -Roberto Yes. I can send you the output of ifconfig if useful. The addresses are both in the 152.3.172.* subnet. Faheem.
Re: two network cards on a server
On Wed, 18 Apr 2007, Roberto C. Sánchez wrote: Well, there is your problem. Having two physical interfaces on the same subnet is generally a bad idea: http://lists.debian.org/debian-user/2007/01/msg01633.html http://lists.debian.org/debian-firewall/2005/03/msg00073.html You have to really know what you are doing. It appears that you have overlooked something in your configuration. If you can provide more details, I or someone else might be able to help get you going in the right direction. Regards, -Roberto I'd be very happy to provide details, if you could tell me what details are relevant. Faheem.
Re: two network cards on a server
On Wed, 18 Apr 2007, Roberto C. Sánchez wrote: Well, there is your problem. Having two physical interfaces on the same subnet is generally a bad idea: http://lists.debian.org/debian-user/2007/01/msg01633.html http://lists.debian.org/debian-firewall/2005/03/msg00073.html You have to really know what you are doing. It appears that you have overlooked something in your configuration. If you can provide more details, I or someone else might be able to help get you going in the right direction. Regards, -Roberto Just to be clear, the interfaces are both of the form 152.3.172.*. If for example they were reconfigured so that one was of the form 152.3.171.* and the other was of the form 152.3.172. (say) then would the problem just disappear? Thanks. Faheem.
Re: two network cards on a server
On Wed, 18 Apr 2007, Roberto C. Sánchez wrote: On Wed, Apr 18, 2007 at 06:28:29PM -0400, Faheem Mitha wrote: Just to be clear, the interfaces are both of the form 152.3.172.*. If for example they were reconfigured so that one was of the form 152.3.171.* and the other was of the form 152.3.172. (say) then would the problem just disappear? That largely depends on your subnet mask and how your network is routed. What is the output of `/sbin/ifconfig` with both interfaces up? See below. The problem interface is eth0. The original one, which still seems to work Ok, is eth1. Thanks. Faheem. eth0 Link encap:Ethernet HWaddr 00:E0:81:31:33:8C inet addr:152.3.172.111 Bcast:152.3.173.255 Mask:255.255.254.0 inet6 addr: fe80::2e0:81ff:fe31:338c/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:2362483 errors:0 dropped:0 overruns:0 frame:0 TX packets:3663 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:156143284 (148.9 MiB) TX bytes:617795 (603.3 KiB) eth1 Link encap:Ethernet HWaddr 00:E0:81:31:33:64 inet addr:152.3.172.60 Bcast:152.3.173.255 Mask:255.255.254.0 inet6 addr: fe80::2e0:81ff:fe31:3364/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:14220041 errors:0 dropped:0 overruns:0 frame:0 TX packets:26761605 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:2763292401 (2.5 GiB) TX bytes:35308449413 (32.8 GiB) Interrupt:169 loLink encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:2317579 errors:0 dropped:0 overruns:0 frame:0 TX packets:2317579 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:3617337316 (3.3 GiB) TX bytes:3617337316 (3.3 GiB)
Release.gpg keys on security.debian.org are currently blank
Hi, Release.gpg keys on security.debian.org are currently empty files. This only seems to be causing problems on one machine, but does anyone know what is going on here, and how long it might be before it is fixed? Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
problem with switch
Hello everyone, I've recently encountered an odd problem with a switch and a Dell PC running Debian. I'm using it on a cable modem. When I plug the machine directly into the cable modem, I can get on the net (the cable modem uses DHCP as per usual). However, when I plug the machine into the switch, the machine can't get on the net, in fact it can't get a lease from the DHCP server at all. However, a Mac OS X laptop does not have any problems. Therefore, the problem presumably has something to do with the combination of the switch and the machine. The machine in question has a Broadcom card, and is using Debian's somewhat hacked around with tg3 driver, so I'm wondering if that could perhaps be causing the problem, and if so, whether switching to the bcm5700 would solve it. Does anyone have any ideas or suggestions about how I might resolve this issue? Thanks in advance. Please cc me; I'm not subscribed. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: problem with switch
On Sun, 6 Aug 2006, Andrew Sackville-West wrote: On Sun, Aug 06, 2006 at 02:33:35PM -0400, Faheem Mitha wrote: Hello everyone, I've recently encountered an odd problem with a switch and a Dell PC running Debian. I'm using it on a cable modem. When I plug the machine directly into the cable modem, I can get on the net (the cable modem uses DHCP as per usual). if you can get a dhcp lease directly from your cable modem, then dhcp client is working on the debian box. However, when I plug the machine into the switch, the machine can't get on the net, in fact it can't get a lease from the DHCP server at all. does the switch have an incorporated router? Not certain, but don't think so. It is a fairly generic Netgear switch, which I picked up at Best Buy. Routers and switches are separate things, aren't they? does it supply dhcp service? is that dhcp service turned on? Don't know to both. does it have some sort of configuration that might be preventing you from getting a lease? Don't know, but if so wonder how the laptop below is getting a lease. Presumably it is possible to access the switch configuration and see what is going on. However, a Mac OS X laptop does not have any problems. Therefore, the problem presumably has something to do with the combination of the switch and the machine. is the laptop connecting by wire or wireless? if wireless, you're dealing with different systems... If wired, how does the osx box get ip? is it set for fixed IP? are you saying that the osx box gets a dhcp lease from the switch(router?)? Sorry, I should have been more precise. Yes, the laptop is also connecting via ethernet through the switch (exactly the same configuration as the Dell PC), and it is also using DHCP, but from the dhcp server of the cable isp (I think). I really doubt it is getting it from the switch, because it is certainly a valid IP address. This is what makes this so puzzling, since at least at first glance this seems to point to some problem with the Dell PC rather than the network. The machine in question has a Broadcom card, and is using Debian's somewhat hacked around with tg3 driver, so I'm wondering if that could perhaps be causing the problem, and if so, whether switching to the bcm5700 would solve it. I can't imagine that if it works direct connected to the modem that it won't work through a switch. there is something else wrong in your network, would be my guess. try setting a static ip on the debian box and see if that works. Unfortunately I don't have a static IP address to use, at least not on that cable service. Thanks for your helpful suggestions. I now have some more pointers for trouble-shooting. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
wireless network card
Dear People, I'm currently visiting Salt Lake from out East (Chapel Hill, NC), through the middle of this coming week. I've been trying to find locally a wireless network card that works with Linux to install on my host's laptop. The cards available at the local retail outlets seem to only work with proprietary drivers, and I don't want to use proprietary drivers for such a basic thing. Ideally, I'd prefer to use a card which is supported by a driver already in the kernel, but I'd consider third party drivers too, as long as they were open source. The important thing is that I should be able to pick it up locally. I don't have the time to wait for something to be delivered. So, any ideas where I can pick up one from? I only have a few days to figure something out. Please CC me. I'm not currently subscribed. Thanks. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
creating and serving temporary files with apache
Dear People, I'm fairly new to apache administraction, so I apologise in advance if this an obvious question. I am running Apache, which is running some CGI scripts, which allow a web client (browser) to upload data, process it, and then return the process results to the client in the form of clickable links which correspond to the results. Let us assume for the purpose of this question that I have a CGI script along with other web pages, located in /var/www/data, which needs to write temporary files for the purpose described above. My question is as follows. What is a good place to locate these files, and what permissions should be set on these files? It seems to be clear that allowing apache's user (namely www-data) write permission to /var/www/data is a bad idea, because it would allow an attacker who obtained the permissions of www-data free access to the web pages there. However, it is less clear where these files should be put. First I was thinking of putting them in /tmp, but I am not sure it is a good idea for apache to be serving files from /tmp. Also, we require these files to be preserved over quite long periods of time, and /tmp is cleared on every reboot. I'm now toying with the idea of putting them in say /var/www/data/tmp, where tmp would be owned by www-data (both user and group www-data), and nobody else would have write access. Actually, disabling read access might be a good idea as well. What do people think of that? Any other suggestions/opinions? Thanks in advance. Please cc me, I'm not subscribed. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
UPS for servers?
Hi, I'm looking for a UPS to hook up to a rack containing three servers. I'm not precisely sure of the power consumption, but could find some hard numbers if it became necessary. I want something that will keep all three servers up for at least 1 hr. The rack is a 6U. I want a UPS whose software is supported by Linux, preferably something packaged for Debian, and one that is reasonably priced (under USD 1000). The functionality I would like is that the following sequence of events is possible 1) Power goes out. 2) UPS waits T1 minutes and if the power has not gone back on, informs the machines. 3) The machines send out a broadcast message to users telling them the machines are going to go down in T2 minutes. 4) If the power has not come back in T2 minutes, the machines go down. Otherwise, they send out a broadcast message saying the power has come back on. Does anyone have any idea if ways exist to power up a machine remotely? I'm going to talk to APC, but I don't know how much a sales rep would know about Linux, and he/she certainly would not know about Debian. Recommendations/suggestions appreciated. Please CC me; I'm not subscribed. In any case, I will check periodically via Gmane. Thanks in advance.Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: problems with basic lvm commands on Debian
On Sat, 27 Aug 2005, Allan Wind wrote: On 2005-08-25T23:10:05-0400, Faheem Mitha wrote: [EMAIL PROTECTED]:/dev/mapper# pvdisplay /dev/hda2 No physical volume label read from /dev/hda2 Failed to read physical volume /dev/hda2 How does the disk partition look like (fdisk)? Maybe there is garbage data on hda2 that confuse pvcreate. If I recall correctly the lvm howto has a suitable dd commend to take care of that. Check the lvm archive to see if it is a known issue. You could collapse the partitions into one pv and vg, then create two seperate lv's from it. Well, I've reformatted since I had that problem. I think the reason for these errors was (as someone pointed out on the lvm mailing list) that I had mounted the partitions that I trying to put into the volume group, and was running the system off filesystems mounted on them. So it is not surprising that there were problems. I should figure out how to set a reply to header in Pine. Thanks for replying. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
problems with basic lvm commands on Debian
Dear People, I must be missing something obvious, but I'm having problems with basic commands on Debian Sarge, using lvm2. I installed lvm2 and dmsetup. I'm using the stock Debian kernel 2.6.8-2-386. I can do basic manipulations using /dev/hda1, but not with /dev/hda2. This is on a very ordinary Dell Optiplex, normal x86 architecture. *** [EMAIL PROTECTED]:/home/faheem# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/hda2 37483592300352 35279136 1% / tmpfs 258052 0258052 0% /dev/shm /dev/hda1 918322 14622854704 2% /boot Physical volume /dev/hda2 successfully created [EMAIL PROTECTED]:/dev/mapper# pvcreate /dev/hda1 [EMAIL PROTECTED]:/dev/mapper# pvdisplay /dev/hda1 --- Physical volume --- PV Name /dev/hda1 VG Name testvg PV Size 956.00 MB / not usable 0 Allocatable yes PE Size (KByte) 4096 Total PE 239 Free PE 239 Allocated PE 0 PV UUID sGOTrW-ttC4-m4uq-Cwck-V5s3-knz7-NpydvJ [EMAIL PROTECTED]:/dev/mapper# vgcreate testvg /dev/hda1 Volume group testvg successfully created [EMAIL PROTECTED]:/dev/mapper# vgdisplay testvg --- Volume group --- VG Name testvg System ID Formatlvm2 Metadata Areas1 Metadata Sequence No 1 VG Access read/write VG Status resizable MAX LV0 Cur LV0 Open LV 0 Max PV0 Cur PV1 Act PV1 VG Size 956.00 MB PE Size 4.00 MB Total PE 239 Alloc PE / Size 0 / 0 Free PE / Size 239 / 956.00 MB VG UUID IHrOgz-RWDW-EIAY-m9vD-idgT-1I99-cyoS8n [EMAIL PROTECTED]:/dev/mapper# pvcreate /dev/hda2 Physical volume /dev/hda2 successfully created All good so far. However. [EMAIL PROTECTED]:/dev/mapper# pvdisplay /dev/hda2 No physical volume label read from /dev/hda2 Failed to read physical volume /dev/hda2 [EMAIL PROTECTED]:/home/faheem# vgextend testvg /dev/hda2 No physical volume label read from /dev/hda2 /dev/hda2 not identified as an existing physical volume Unable to add physical volume '/dev/hda2' to volume group 'testvg'. Any idea what is going on here? This is a new installation, and only has the bare bones on it. Please cc me on any reply. Thanks in advance. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: comments about hardware
On Thu, 21 Jul 2005, Lennart Sorensen wrote: One dual core opteron compared to two single core opterons: Each opteron has a memory controller built in that does dual channel memory support. A dual core opteron still only has one memory controller and hypertransport to the chipset. The two cores share it, so two single cores have theoretically twice the memory bandwidth of a single dual core. Of course they also require a dual socket board rather than a single socket board, and you could put two dual core opterons in a dual board and get to use 4 cores total without having to pay for a much more expensive 4 socket board. Two cores in one package may on the other hand have faster access to each other's caches which may be an advantage in some situations, while in others sharing the memory bandwidth could hurt. At the same time the dual core would always have it's memory local, while two single cores half the ram is likely connected to the other cpu so access would have a 1 cycle penalty for access. A decent OS would try to make sure applications are running on the cpu whos ram they are currently in whenever possible. Hi, Thanks for the information. If you have experience of using dual core processors with Debian, I'd be glad to hear of the details. Best regards, Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
comments about hardware
Dear People, My bioinformatics research group at Duke is buying a server, which will mostly be used as a server, particularly for web based services. The idea here is that a user will submit a request for some bioinformatics calculation via a web interface (often using Python or R or similar), the server does the calculation, and returns it as a web page. None of us are experts about recent hardware, so would appreciate any feedback about hardware specs. The following quote is from Monarch Computers. We plan to run Linux on this. It has not yet been decided yet what, but it seems most likely that it will be either some Red Hat variant (Fedora Core, CentOS), or Debian (possibly Ubuntu). Ok, so here are some specific questions. 1) Dual core Opterons first came on the market in April. The sales rep said that AMD Dual Core Opterons did not work with Fedora Core. Since they only install Fedora and SuSE, they had no info about Debian. Any idea what the status is here? How well are they supported, and how stably do they run under Linux? Also, I was told that a dual core Opteron, which is somewhat more than twice the cost two regular Opterons of similar speed, is not equivalent to two regular Opterons in functionality. Can anyone point me to information about this, or offer a comment? 2) I'm wondering if the listed motherboard is the best choice. I see it listed in http://alioth.debian.org/docman/view.php/30192/27/mainboards.html We are looking for the motherboard that has the least known issues. Preferably something that will work right out of the box. Google found me http://lists.debian.org/debian-amd64/2004/09/msg00443.html but would be interested in other reports. The specs are here http://www.tyan.com/products/html/thunderk8spro_spec.html It looks like both the graphics card and the ethernet cards are onboard. Looks like the graphics card is ATI RAGE XL PCI, which supposedly works with the 'ati' driver. Is this under XFree 4.3? The ethernet cards are an Intel Ethernet Pro 100, which supposedly works with the e100 driver and a Gigabit Broadcom which works with the tg3 driver. There seem to be two cards here. Is that correct? I'm kinda allergic to onboard cards. They are often trouble. Has anyone had experience with Debian Sarge installation with this? Does anyone have a board to suggest that they prefer to this? 3) I'm also wondering if peple have thoughts about the RAID setup. The rep said he would be using RAID 1, but I see RAID 10 is listed. I'll have to check on this. Anyway, assuming this corresponds to http://en.wikipedia.org/wiki/Redundant_array_of_independent_disks#RAID_10 with each RAID 1 set as two drives, and 4 RAID 1 sets striped together, does this seem reasonable? Thanks.Faheem. *** ITEM NUM PRICE PER ITEM TOT Monarch Empro Custom 2U Rack S 1.0075.00 75.00 RMC2K2-9I-XPSS,2U,8 Bays,SATA, 1.00 725.00 725.00 AIC 2U Riser Card/Rear Window1.00 112.00 112.00 Tyan S2882G3NR-D Dual Socket94 1.00 394.00 394.00 Amd OSA265FAA6CB Dual Core Opt 2.00 851.00 1,702.00 Thermal Grease, Shin-Etsu G675 2.0014.0028.00 THERMALTAKE A1838 AMD Opteron2.0025.0050.00 WESTERN DIGITAL 250 GB 2500JD1.00 115.00 115.00 3WARE Escalade 9500S-8 - 8-por 1.00 485.00 485.00 RAID 10 Setup1.0025.0025.00 WESTERN DIGITAL 250 GB 2500JD8.00115.00 920.00 SONY DWD-56A 8X4X2.4 DVD RW+/- 1.00129.00 129.00 SUSE Linux 9.3 Professional Ed 1.00 92.00 92.00 24/7 TECH SUPPORT+ONSITE 3 YR. 1.00199.00 199.00 Net Order: 5,051.00 Freight: 75.00 5,126.00 -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: comments about hardware
On Mon, 18 Jul 2005, Ed Tomlinson wrote: Hi, Dual core support is not distro specific. It depends on the kernel used. I believe that debian, with a recent (2.6.12.3+) kernel should be fine. Do you have any personal experience in using this? I'm concerned about stability issues. Also, would there not be some practical difficulty in getting this installed, seeing as with an earlier kernel it probably won't boot, and the official Debian installers all use 2.6.8? Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
bash completion does not work over ssh except as root
Dear People, I have noticed that bash completion only works on a remote machine I have sshed into, if I log in as root. This is obviously a case of something not being sourced correctly, but I am not sure what the problem is. Can anyone tell me how to solve this, or if not, confirm the problem? I use bash completion heavily so it is an annoying bug. I have this in .bash_profile on both the remote and local machines, but it does not solve the problem. # ~/.bash_profile: executed by bash(1) for login shells. # see /usr/share/doc/bash/examples/startup-files for examples. # the files are located in the bash-doc package. umask 022 # the rest of this file is commented out. # include .bashrc if it exists if [ -f ~/.bashrc ]; then source ~/.bashrc fi ... *** I'm tracking Sarge. Please cc me, I'm not subscribed. Thanks. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: bash completion does not work over ssh except as root
On Sun, 12 Sep 2004, Stefan O'Rear wrote: -- BEGIN DATA BLOCK -- # enable bash completion in interactive shells if [ $PS1 -a -f /etc/bash_completion ]; then . /etc/bash_completion fi -- END DATA BLOCK -- You need that or the equivilent in /etc/bash.bashrc or ~/.bashrc (it is commented out by default in Woody). -- The world's most effective spam filter: ln -sf /dev/full /var/mail/$USER Yes, but the comment to that code (which you left out) is # enable programmable completion features (you don't need to enable # this, if it's already enabled in /etc/bash.bashrc). I have it uncommented out in /etc/bash.bashrc. Isn't that enough? If not, I think this is a bug, as it contradicts the comment. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: bash completion does not work over ssh except as root
On Sun, 12 Sep 2004, Stefan O'Rear wrote: I don't really know - I have it in both and it works everywhere. You are right. Once I enable it in .bashrc, it works over ssh, contradicting the comment. I'll file a bug. In woody + unstable libc, gcc. Don't follow you. What is the relevance of libc and gcc? Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
recommendation for digital camera
Dear People, I know this question has been asked before, but the answer is constantly changing.:-) So, I'm looking for a digital camera with the following (tentative) requirements. I'm happy to have these requirements questioned. 1) A cost of $500 or under. 2) Needs to work well with Debian (obviously). I think the best support is obtained by cameras that are seen as USB mass storage devices, right? 3) Preferably uses as little proprietary technology as possible. This may not be a realistic requirement, though. I'm basically trying to avoid the situation when I can only buy an accessory made by the company manufactured by the camera because nothing else is compatible. 4) Something 4 Megapixel or over would be good, I think. I'd also be interested in good internet resources about digital cameras, particularly Linux specific. Thanks in advance. Please CC me, I'm not subscribed. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: help with python bug
On Sat, 5 Jun 2004, John Hasler wrote: Faheem Mitha writes: Can anyone help me with a seemingly Debian-specific Python bug, Debian bug no. 252517 (see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=252517 ) Firstly, can anyone reproduce the problem on Sarge or unstable? I cannot reproduce it on Unstable. Strange. I just upgraded a test installation to unstable, and I was still able to reproduce it. Could someone else try it on sarge or unstable? It is only a few minutes work. Thanks in advance. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: help with python bug
On Sat, 5 Jun 2004, John Hasler wrote: Faheem Mitha writes: Can anyone help me with a seemingly Debian-specific Python bug, Debian bug no. 252517 (see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=252517 ) Firstly, can anyone reproduce the problem on Sarge or unstable? I cannot reproduce it on Unstable. Thank you for taking a look. Hmm, this is puzzling. I'll try this out on a test sid installation and see what happens. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
help with python bug
Dear People, Can anyone help me with a seemingly Debian-specific Python bug, Debian bug no. 252517 (see http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=252517 ) Firstly, can anyone reproduce the problem on Sarge or unstable? Secondly, can anyone suggest a solution or workaround? I have tried the same code on a Redhat machine, and I see no problem. So, this makes me think one of the Debian patches must be causing it. I tried recompiling Python 2.3 with debugging support and then got no end of errors. Any suggestions would be gratefully appreciated. Please CC me. Thanks in advance. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT - trivial programming language
On Sat, 29 May 2004 15:54:48 +0200, Kai Grossjohann [EMAIL PROTECTED] wrote:
Steve Lamb [EMAIL PROTECTED] writes:
Faheem Mitha wrote:
Bob Proulx makes good points elsewhere in this thread. Whether you
like the indentation as syntax feature is really a matter of
taste. Personally, I am ambivalent about it. On the one hand it makes
code more compact. On the other hand, indentation is easily lost
information, for example when cutting and pasting. Also, I have the
habit in emacs of hitting tab to get code to line up. I've done this
for years, and it is a hard habit to break. Unfortunately it has
disastrous consequences in python code.
So have EMACS treat the tab key as the equivolent number of spaces and
write the file out using spaces instead of tabs.
This won't help: in Emacs, TAB computes the right indentation from
the buffer contents. In C-like languages, this is fairly easy: after
a {, indent more, after a }, indent less. There are a lot of
other rules, but that's the basic.
But in Python, it is not possible to compute the indentation from the
buffer contents. This means that just hitting TAB on each line is
either a no-op (then why are we doing this), or it would get the end
of a loop wrong and suchlike.
Yes. There have been a lot of people in this thread you have
misunderstood what I was saying. I think Kai summarised it pretty
well. The bottom line is that in Python whitespace is syntatically
meaningful, in C etc. it is not.
This has the consequence that in C, emacs is able to correctly indent
the code, using the built-in syntax rules it knows about. This is very
useful since it shows up trivial syntax bugs right away.
On the other hand, much of the time emacs does not know what to do
with Python code (when in python mode, that is), since part of the
synax info is encoded in the whitespace. Blindly hitting tab, which
works fine with C/C++ and probably most of the other languages out
there, can really mess up Python code. It does know enough to avoid
indenting if you forget the colon at the end of if statement, for
example, so it *is* of some use.
Hope this clarifies things.
Faheem.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: OT - trivial programming language
On Fri, 21 May 2004 20:33:14 -0400, richard lyons [EMAIL PROTECTED] wrote: That makes two votes for Python (the other was off-list). I've had it in mind to find time to investigate Python -- so I'll have a go at that. Make that 3 votes. :-) I recently started learning Python. I like it. It has a very clean syntax. Bob Proulx makes good points elsewhere in this thread. Whether you like the indentation as syntax feature is really a matter of taste. Personally, I am ambivalent about it. On the one hand it makes code more compact. On the other hand, indentation is easily lost information, for example when cutting and pasting. Also, I have the habit in emacs of hitting tab to get code to line up. I've done this for years, and it is a hard habit to break. Unfortunately it has disastrous consequences in python code. Bob says ruby is pretty good, and I believe him. However, libraries are a make-or-break feature. The nicest language in the world is useless without comprehensive libraries, and python is up there when it comes to libraries. I am not sure how ruby compares. Also, there are a lot of interesting projects forming around python, eg. boost-python and pyx. Debian supports python pretty well. All significant free software projects involving python (that I am aware of) are in Debian. Also, Debian has a nice Python policy which makes things work well. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: lookit does not start on bootup
On Fri, 09 Apr 2004 20:03:36 -0500, dircha [EMAIL PROTECTED] wrote: Faheem Mitha wrote: Thanks for the suggestion. It is possible that this is the problem, and I'll investigate it. S1 does seem rather low. First I'll need to learn the recommended way to manipulate runtime links in Debian. If I remember correctly, there is some tool to do specifically this. update-rc.d Right. Thanks for the reminder. Although modifying the symlinks manually won't cause any problems or inconsistencies either. Ok.Faheem, -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: lookit does not start on bootup
On Fri, 9 Apr 2004 20:59:20 -0400, S.Squarepants [EMAIL PROTECTED] wrote: While there may be a tool for it (I really don't know), that isn't the best way to do it IMHO. What would you do if you ending up sitting at a Fedora or Slackware box in the future that is missing the same tool? Simply: # mv /etc/rc?.d/S00lokkit /etc/rc?.d/S99lokkit (Change numbering accordingly, and the ? applies to the runlevel being manipulated.) Done. Works on every platform. Will always work as long as the init structure that is in-use today remains the standard. If a graphical or other tool makes you more comfortable, by all means use the tool. But I'd suggest learning to do without those tools as well so you can rescue things when the tool isn't there, doesn't work or can't be accessed. Actually, I was thinking of just changing the postinstall script or whatever and rebuilding the package. Hmm, yes, lokkit.postinst has if [ -x /etc/init.d/lokkit ]; then update-rc.d lokkit defaults 01 99 /dev/null If it does solve the problem, I'll follow up to my earlier Debian bug report. I will probably want to install this on other systems, and I don't want to have to manually fiddle with symbolic links all the time. Though you are right, you ought not be overly dependent on your tools. But ask yourself what you would do if one horrible day apt disappeared from Debian. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Intel Pro/100 VE
On Thu, 15 Apr 2004 [EMAIL PROTECTED] wrote: thanx to all who responded to this thread. The problem I feel is now resolved and I will explain how I did it for documentation purposes. the suggestion by Faheem, though I'm sure it works, I did not take it because I did not know enough about kernel recompilation to make it work, I kept running into walls. I keep running into issues about conflicting kernel versions when I try installing it. In addition, since I don't have internet access at this point, using apt-get (I'm pretty sure it's not on the cd, though I could be wrong) to get the package and then recompiling a whole bunch of stuff just isn't worth it. Recompiling the kernel is a useful skill and much easier (with kernel-package) that is may at first appear. The tutorial at http://newbiedoc.sourceforge.net/system/kernel-pkg.html is pretty nice. I corresponded briefly with the author and he really did his homework. In any case I am surprised that just simply installing one of the precompiled Debian kernel images from 2.4.20 or later did not work for you. But perhaps there were other difficulties in doing so. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Intel Pro/100 VE
On Sun, 11 Apr 2004 11:12:05 -0400, Paul [EMAIL PROTECTED] wrote: Hey peoples, I have been trying to get debian to work and have been very unsuccessful. I just got a Toshiba M35-s320 laptop and have been trying to install the network adapter with the 3.0r2 disk image. I did some research and I believe that imy network card, Intel PRO/100 VE is supported in the EEPRO100 driver but everytime I enable it it says insmod failed. I checked my bios and there is no PlugNPlay option so I don't think it's a bios conflict. Intel supplies a e100 driver but I, for the life of me cannot seem to be able to download it. Any ideas? You are aware that the e100 driver exists as a standalone module source package in Debian, right? I would prefer using this one over eepro100, since I have had some issues in the past with eepro100. I've had no problems compiling e100 as an external module with a variety on kernels, using kernel-package. Also, eepro100 is in the kernel as of 2.4.20, so you could just use a more recent kernel. *** Package: openafs-modules-source Priority: extra Section: net Installed-Size: 4464 Maintainer: Sam Hartman [EMAIL PROTECTED] Architecture: all Source: openafs Version: 1.2.11-1 Depends: bison, flex, debhelper, libpam0g-dev, libncurses5-dev, kernel-package, e2fslibs-dev Filename: pool/main/o/openafs/openafs-modules-source_1.2.11-1_all.deb Size: 4489072 MD5sum: b088707d130b502894a42a3e0f282d65 Description: The AFS distributed filesystem- Module Sources AFS is a distributed filesystem allowing cross-platform sharing of files among multiple computers. Facilities are provided for access control, authentication, backup and administrative management. . This package provides source to the AFS kernel modules. HTH. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: lookit does not start on bootup
On Mon, 5 Apr 2004 00:58:55 -0500, S.Squarepants [EMAIL PROTECTED] wrote: I have no idea if this will help you or not, but I once had a problem with some other things that failed to work prior to the network being up. I doubt that's the case here, but it still might rely on something else that needs to load first. Try changing the S numbers to something higher. Maybe start at S99, and if that works try it again at S50 or something till you get it in the right range. Redhat used to have the same problem (don't know if Fedora fixed it) with pcmcia. It would start /after/ networking, so the network wouldn't come up because the card needed for it wasn't initializing until later on in the boot process. HTH, but I have no idea whether it will or not. Thanks for the suggestion. It is possible that this is the problem, and I'll investigate it. S1 does seem rather low. First I'll need to learn the recommended way to manipulate runtime links in Debian. If I remember correctly, there is some tool to do specifically this. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: lookit does not start on bootup
On Sun, 04 Apr 2004 15:19:39 -0500, dircha [EMAIL PROTECTED] wrote: I recall that lokkit never worked for me either. That prompted me just to learn how to use iptables manually, so I never figured out why lokkit was failing. After bringing it up manually, try: # iptables -L to be sure that it isn't just failing and suppressing the output, or failing and redirecting the ouput elsewhere. No, iptables -L gives the right results. If the rules are being loaded when you bring it up manually, it's really hard to say what the problem might be, which is why I suspect it is just silently failing. If you could copy to the list the output of $ ls /etc/rc*.d, $ cat /etc/modules, # iptables -L, $ lsmod, and the contents of the rules script generated by lokkit, I could get a better idea of what is going on. At least these are the places I would look if I were going about diagnosing the problem on one of my own systems. Ok. Output follows. Thanks. My hunch is that if it is silently failing when you bring it up manually, that the problem is that there are kernel modules not being loaded which are needed by iptables. No, that is not the problem. * $ ls /etc/rc*.d * /etc/rc0.d: K01kdm K01xdm K11cron K14ppp K15fetchmail K19aumix K19setserial K19spamassassin K20apache K20apache2 K20athcool K20bastille-firewall K20cupsys K20exim K20inetd K20lpd K20lprng K20makedev K20rsync K20ssh K20timidity K20udftools K20xfs K20xprint K21alsa K23ntp-server K25hwclock.sh K30etc-setserial K55usbmgr K75hdparm K89atd K89hotplug K89klogd K89shorewall K90sysklogd K99lokkit S20sendsigs S30urandom S31umountnfs.sh S35networking S40umountfs S90halt /etc/rc1.d: K01kdm K01xdm K11cron K14ppp K15fetchmail K19aumix K19spamassassin K20apache K20apache2 K20athcool K20bastille-firewall K20cupsys K20exim K20inetd K20lpd K20lprng K20makedev K20rsync K20ssh K20timidity K20udftools K20xfs K20xprint K21alsa K23ntp-server K55usbmgr K89atd K89klogd K90sysklogd K99lokkit S11hotplug S20single S21aumix /etc/rc2.d: K11anacron S01lokkit S10sysklogd S11hotplug S11klogd S14ppp S15usbmgr S19spamassassin S20alsa S20apache2 S20athcool S20bastille-firewall S20cupsys S20exim S20inetd S20lpd S20lprng S20makedev S20rsync S20ssh S20timidity S20udftools S20xfs S20xprint S21aumix S23ntp-server S89anacron S89atd S89cron S91apache S99fetchmail S99kdm S99rmnologin S99stop-bootlogd S99xdm /etc/rc3.d: K11anacron S01lokkit S10sysklogd S11hotplug S11klogd S14ppp S15usbmgr S19spamassassin S20alsa S20apache2 S20athcool S20bastille-firewall S20cupsys S20exim S20inetd S20lpd S20lprng S20makedev S20rsync S20ssh S20timidity S20udftools S20xfs S20xprint S21aumix S23ntp-server S89anacron S89atd S89cron S91apache S99fetchmail S99kdm S99rmnologin S99stop-bootlogd S99xdm /etc/rc4.d: K11anacron S01lokkit S10sysklogd S11hotplug S11klogd S14ppp S15usbmgr S19spamassassin S20alsa S20apache2 S20athcool S20bastille-firewall S20cupsys S20exim S20inetd S20lpd S20lprng S20makedev S20rsync S20ssh S20timidity S20udftools S20xfs S20xprint S21aumix S23ntp-server S89anacron S89atd S89cron S91apache S99fetchmail S99kdm S99rmnologin S99stop-bootlogd S99xdm /etc/rc5.d: K11anacron S01lokkit S10sysklogd S11hotplug S11klogd S14ppp S15usbmgr S19spamassassin S20alsa S20apache2 S20athcool S20bastille-firewall S20cupsys S20exim S20inetd S20lpd S20lprng S20makedev S20rsync S20ssh S20timidity S20udftools S20xfs S20xprint S21aumix S23ntp-server S89anacron S89atd S89cron S91apache S99fetchmail S99kdm S99rmnologin S99stop-bootlogd S99xdm /etc/rc6.d: K01kdm K01xdm K11cron K14ppp K15fetchmail K19aumix K19setserial K19spamassassin K20apache K20apache2 K20athcool K20bastille-firewall K20cupsys K20exim K20inetd K20lpd K20lprng K20makedev K20rsync K20ssh K20timidity K20udftools K20xfs K20xprint K21alsa K23ntp-server K25hwclock.sh K30etc-setserial K55usbmgr K75hdparm K89atd K89hotplug K89klogd K89shorewall K90sysklogd K99lokkit S20sendsigs S30urandom S31umountnfs.sh S35networking S40umountfs S90reboot /etc/rcS.d: README S02mountvirtfs S05bootlogd S05keymap.sh S07hdparm S10checkroot.sh S18hwclockfirst.sh S20module-init-tools S20modutils S30checkfs.sh S30etc-setserial S30procps.sh S35devpts.sh S35mountall.sh S35mountkernfs S36discover S36hotplug S38pppd-dns S39dns-clean S39ifupdown S40hostname.sh S40networking S40shorewall S45mountnfs.sh S46setserial S48console-screen.sh S50hwclock.sh S51ntpdate S55bootmisc.sh S55urandom S70screen-cleanup S70xfree86-common S75sudo * /etc/modules * # /etc/modules: kernel modules to load at boot time. # # This file should contain the names of kernel modules that are # to be loaded at boot time, one per line. Comments begin with # a #, and everything on the line after them are ignored. usb-uhci input usbkbd keybdev emu10k1 usbmouse agpgart parport parport_pc isa-pnp hid input keybdev usbkbd #added 7th June 2002 by Faheem apm #added 17th February 2004 by
lookit does not start on bootup
Dear People, I have a minor but annoying problem with lokkit. It does not start at bootup. The runlevel look normal eg. etc/rc0.d/K99lokkit etc/rc1.d/K99lokkit etc/rc2.d/S01lokkit etc/rc3.d/S01lokkit etc/rc4.d/S01lokkit etc/rc5.d/S01lokkit etc/rc6.d/K99lokkit but I get errors at bootup which don't appear to be captured to any file so I can't reproduce them here. I can start it manually by Chrestomanci:~# /etc/init.d/lokkit start Starting basic firewall rules: lokkit. I reported this as http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=219686 but did not get a response. Is anybody else seeing this problem? Thanks in advance. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
slrn drawing characters problem in Konsole
Dear People, I'm running sarge. I recently noticed (I don't reboot my machine very often and only discover breakage when I reboot) that the red threading in slrn was now replaced by red blocks. I'm not sure what the problem is, but it seems likely the problem was with Konsole and/or X 4.3, since X 4.3 had just arrived in testing. Or maybe that was just a coincidence. The problem seems to be visible in most fonts, and is certainly present in my usual font, which is the unicode option. So, can anyone using slrn reproduce this problem, and more importantly, suggest a fix? I'd be inclined to report this as a Konsole bug unless I hear something to the contrary. Thanks in advance. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Mouse behavior with kernel 2.6.X ( running 2.6.4 and 2.6.3 on another machine)
On Fri, 19 Mar 2004 04:02:46 +0200, Micha Feigin [EMAIL PROTECTED] wrote: On Thu, Mar 18, 2004 at 02:35:22PM +, n.v.t n.v.t wrote: Hi, I'm struggling witht his for a while. I have been running 2.6. But I don't like the mouse behavior at all. It's *to* fast compared to 2.4. How can I get the 2.4 mouse behavior back? If you are using both a ps2 mouse (/dev/ps2) and usb (/dev/input/mice), or at least both interfaces for some reason (common on laptops), then you need to use only /dev/input/mice on 2.6 since all mice are piped there emulated as ImPS2, or try using /dev/input/mouse[n] I've been having the same problem, with a PS/2 and USB mouse installed in tandem. Looks from this info like I need to choose between running 2.4 and 2.6. Any suggestions about how I could dualboot between both? Please CC me, I'm not on the list. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian on Dell
On Mon, 08 Mar 2004 10:15:06 +0100, Marc Koenders [EMAIL PROTECTED] wrote: The config you are contemplating makes me turn green in envy. Well, it is just going to be my office machine. I'm not paying for it, neither is it mine. Left to myself I would have used an independent vendor anyway. In any case, they may not go with my spec. [snip] Installing Debian is something you dont do at 02:00 at night. You gotta make sure you are awake and alert. Read the messages Debian pops. Even if your only option is OK or Continue. Or else. I decided to do install from the netinstall iso (woody) and since the 2.4 kernel somehow did not recognise my usb keyboard i had to go for the standard kernel and do without my usb mouse till after install. Since I, being cocky from many successfull Redhat, Suse and Mandrake installs on other machines, did not follow my advice to be awake while installing i ran into some nice unexpected trouble. dselect is probably a nice tool but eh IMHO even the FreeBSD netinstall has a more userfriendly installer. Yes, installing Debian can be (ahem) interesting at times. The new sarge installer (debian-installer) is a quantum leap in improvement, though. Still with some help at some critical points (Thnx Jens) I ended up with a nice sleek fit-like-a-glove os on a machine (and yes i admit. it will be never be as cool as the configuration you are gonna get =) ) that I could not have built better. I wish you alot of fun with your new machine and I do hope you decide to go for Debian. It needs some time to be tweaked but that is part of the fun IMO. Thanks! It has been a long time since I used anything but Debian (except under the pressure of extreme necessity). Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Debian on Dell
On Thu, 26 Feb 2004, Faheem Mitha wrote: Dear People, I just started a job at Duke. I've been told to put together a machine quote for my office. I want to run Debian on it. I tried to persuade my employer to use Monarch, but was told I have to use Dell. I'm not sure of the best way to go about putting together a machine from Dell which has good general Linux support as opposed to Red Hat support. I've taken a look at Dell's Precision line of workstations, but it looks less than ideal. For one things the descriptions are less than specific about what hardware is included. Also, there is not as much choice as I would like. I was hoping to somehow actually be able to do a significant amount of customization without breaking the bank. I'd greatly appreciate hearing about people's experiences of success with Debian on Dell, and related advice, suggestions, and working configurations. Thanks in advance. I want to thank everyone on debian-user who responded to my earlier message. I've worked out a tentative spec, which I post below. Here follow some comments/questions. 1) One thing I am wondering about is the 8X DVD+RW/+R, Data Only. I have no experience with something like this in Linux. Can one burn and play DVDs. CD-R/CD-RW's with this thing under Linux? I was wondering if it would be better to get separate DVD (read-only drive) and CD-RW drives to be on the safe side. 2) They are also being rather coy about the ethernet card. I assume (educated guess based on a Trilug post from Daniel Chen and other info) that there is an onboard Intel card (which works with the e1000 driver), I did hear something about a Broadcom card being used too, though. 3) I managed to get the Nvidia GeForce4 MX 440 card on a Dell Optiplex GX270 to work under X. I could not manage to work it with the nv X driver (as of 4.2 in testing) but the proprietary nvidia driver (ndvidia) worked. I hope it will be the same with this nVidia, Quadro NVS 280. Daniel Chen's message seems to confirm this. A question: I have heard that the nvidia kernel modules are binary. How come they seem to work with pretty much any kernel I try? Usually binary kernel modules (in my experience) are very sensitive to the version of the kernel being compiled against). At any rate, it seems part of the kernel driver is actually being compiled. The documentation says Since the Linux kernel does not support a binary driver interface, we provide for rebuilding these files on the target machine (or distribution) and then linking with the binary version of the NV kernel driver. but I'm not sure what this means. 4) I opted for SCSI instead of SATA drives, since it seems that the Linux support for the (presumably on-board) SATA controller might be problematic. Anybody with experience on this? A friend of mine said the SCSI controller on Precisions was Adaptec but Daniel said it was LSI, which is presumably well supported by the mptfusion kernel modules. I'd prefer Adaptec, though. 5) I'd welcome suggestions on changes in the configuration below to reduce cost while impacting functionality as little as possible. Thanks.Faheem. *** Dell Precision Workstation 360n Intel=AE Pentium=AE 4 Processor 2.80GHz, 1MB / 800 MHz FSB Operating System: Red Hat Linux WS (V.3) with one year RHN subscription Chassis: Small Mini-Tower Keyboard: Entry Level Quietkey Keyboard, PS/2, (No Hot Keys) Memory: 2GB,DDR400 SDRAM Memory,ECC (4 DIMMS) Mouse: Dell USB 2-Button Optical Mouse with Scroll Monitor: Dell UltraSharp?1901FP flat panel(19.0 viewable), HeightAdjustableStand Graphics Card: nVidia, Quadro NVS 280, 64MB, dual monitor VGA capable Boot Hard Drive: 36GB Ultra 320 SCSI, 1 inch (10,000 rpm) 2nd Hard Drive: 36GB Ultra 320 SCSI, 1 inch (10,000 rpm) Floppy Drive: 3.5 inch 1.44MB Floppy Drive Controller Primary: U320 SCSI Adapter with RAID 0 at No extra charge Sound: Sound Blaster Audigy II with onboard 1394 Speakers: Dell Two Piece Stereo System CD-ROM, DVD and Read-Write Devices: CD-ROM, DVD and Read-Write Devices: 8X DVD+RW/+R, Data Only -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Debian on Dell
Dear People, I just started a job at Duke. I've been told to put together a machine quote for my office. I want to run Debian on it. I tried to persuade my employer to use Monarch, but was told I have to use Dell. I'm not sure of the best way to go about putting together a machine from Dell which has good general Linux support as opposed to Red Hat support. I've taken a look at Dell's Precision line of workstations, but it looks less than ideal. For one things the descriptions are less than specific about what hardware is included. Also, there is not as much choice as I would like. I was hoping to somehow actually be able to do a significant amount of customization without breaking the bank. I'd greatly appreciate hearing about people's experiences of success with Debian on Dell, and related advice, suggestions, and working configurations. Thanks in advance. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
checking whether partition mounted as ext3
Dear People, Just wondering if anyone knows of a easy and definitive way to determine whether a specific mounted partition is ext2 or ext3, and if ext3, whether is mounted as ordered data or journal. Currently, I look at the boot messages, but they are not always clear. Thanks in advance. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Linux compatability for IBM desktop machines
On Sat, 17 Jan 2004 12:59:29 -0500, Mark Roach [EMAIL PROTECTED] wrote: On Thu, 2004-01-15 at 14:38, Faheem Mitha wrote: Dear People, I am writing to ask whether anyone has any Linux experience with something similar to the two desktop machines described at [...] From the descriptions given, I'm not sure what to make of Integrated 10/100/1000 MB ENET or Integrated 10/100100 MB ENET. (I'm guessing the latter was a typo and should be the same as the former). Does anyone have any idea about this? Most of the onboard 10/100/1000 nics I've seen in desktop pc's are Broadcom chips. Debian's kernels do not come with that driver by default, but knoppix does. You might try using knoppix as your compatibility tester... Hi. Thanks for responding. It seems likely (but not certain) from various bits of info I've gotten that this is the Intel PRO/1000 Gigabit Ethernet Card, supported by the e1000 driver. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Linux compatability for IBM desktop machines
On Fri, 16 Jan 2004 18:55:09 +0530, Sanjay Chigurupati [EMAIL PROTECTED] wrote: Hi, About the desktops, 1) check what NIC its using: netgear normally has linux drivers available 2) soundcard : which make is it? 3) Intel : which model? 4) Nvidia Quadro 4: might be available ( check their site nvidia.com). They make good drivers for Linux. just have to check if the model is already supported in the available driver version. instructions are provided too. Thanks for responding. Supposedly Quadro4 is supported by the nv driver in X 4.3, though the info on the site is contradictory. Can anyone confirm this? About the laptops: We're not planning to buy the laptop. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Linux compatability for IBM desktop machines
On Fri, 16 Jan 2004 06:56:56 -0600, Kent West [EMAIL PROTECTED] wrote: Faheem Mitha wrote: Dear People, I am writing to ask whether anyone has any Linux experience with something similar to the two desktop machines described at (http://www.unc.edu/cci/index/ccishop/departmental.html) See http://www.unc.edu/cci/index/ccishop/fullconfiguration.html for more detail. I went to http://www.ibm.com, then Products and Services, then Personal Computing/Desktops, then ThinkCentre M Series, then Customize on the Value item, which let me know that this Value item is the M Series M50, like your General Corporate Model. That might give you a bit more info on the hardware. Yes, good suggestion, thanks. This confirms what other people have told me, that it is likely the Intel PRO/1000 Gigabit Ethernet Card, supported by the e1000 driver. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: checking whether partition mounted as ext3
On Sun, 18 Jan 2004 18:00:51 -0500, Paul Morgan [EMAIL PROTECTED] wrote: On Sun, 18 Jan 2004 13:51:28 -0700, Doug Holland wrote: On Sun 18 Jan 2004 1:16 pm, Faheem Mitha wrote: Dear People, Just wondering if anyone knows of a easy and definitive way to determine whether a specific mounted partition is ext2 or ext3, and if ext3, whether is mounted as ordered data or journal. Currently, I look at the boot messages, but they are not always clear. Thanks in advance. Faheem. Run mount at the command line, with no arguments, and it'll tell you which filesystems are mounted with which fs types. I'm not sure how to tell which mode an ext3 partition is using, though I think running fsck may tell you. Running mount, as you suggest, will tell you with which options the FS was mounted, including the journaling mode. It just seems to echo what is in /etc/fstab. I wanted something which actually went and checked the partitions. BTW, I was under the impression that it was necessary to specify the mounting of root (/) at boot time, in the grub menu or some such, rather than /etc/fstab. I was just looking at stuff which suggested this may not be necessary. Can anyone point me to a definitive reference on this? Thanks. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Can't make xconfig for kernel 2.6.test9
On Tue, 13 Jan 2004 10:31:34 -0600, Ramasubramanian Ramesh [EMAIL PROTECTED] wrote: On Sun, Jan 11, 2004 at 05:25:45PM -0600, Ramasubramanian Ramesh wrote: | Hi, | | I seem to have a problem doing make xconfig on the kernel-2.6.test9 | sources | obtained from sarge. I suspect qt installation problem based on the ld | output. But you can see from apt-get output (at the end) libqt-dev is | installed. From /usr/share/doc/kernel-source-2.6.0/debian.README.gz [...] make xconfig also requires [...] libqt3-mt-dev and g++ = 3.0 for the new 2.6 kenel versions [...] I suggest trying to isntall libqt3-mt-dev instead of libqt-dev. I would also recommend getting the final release, instead of a pre-release test version. According to libqt-dev package, it replaces libqt3 and conflicts libqt3. Also there are no libqt3 in testing/sarge. Package: libqt3-mt-dev(IN TESTING) Version: 3:3.2.1-6 Replaces: libqt-mt-dev, libqt3-dev (= 3.0.5-4), libqt3-helper, libqt3-headers (= 3:3.1.1-3) Depends: xlibs-dev (= 4.2.1), libmng-dev (= 1.0.3), libpng12-0-dev, libjpeg62-dev, zlib1g-dev, libfreetype6-dev, libc6-dev, libqt3c102-mt (= 3:3.2.1-6), libqt3-headers (= 3:3.2.1-6), qt3-dev-tools (= 3:3.2.1-6), xlibmesa-gl-dev | libgl-dev, xlibmesa-glu-dev | libglu1-mesa-dev | libglu-dev, libxft2-dev, libxrender-dev, libxcursor-dev, libaudio-dev, xlibs-pic Recommends: libqt3-compat-headers All the above dependencies are in testing. Just try installing it and see. So I installed libqt-dev. In any case I think something is wrong with libqt-dev in testing and I have asked the debian owner of this package and waiting for a response. I read the debian.README.gz for 2.6.test9 and it only talks about ncurses-dev and tkX.X-dev and I have both yet no make xconfig. I have now turned my attention to make menu config and it will have to do until I know if there is something wrong with my libqt-dev or some other install. BTW, there are no 2.6.0 or 2.6.1 in testing as far as I can tell. I did apt-get update still the above two do not show up (there is test11 though). 2.6.0 source package are in unstable. Get it from there. As already mentioned, the readme for that package says to use libqt3-mt-dev (which is in testing along with its dependencies) for make xconfig, which works for me. If you have tried some other qt dev package already, remember to clean your source tree. With kernel-package fakeroot make-kpkg clean should work. HTH. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Linux compatability for IBM desktop machines
Dear People, I am writing to ask whether anyone has any Linux experience with something similar to the two desktop machines described at (http://www.unc.edu/cci/index/ccishop/departmental.html) See http://www.unc.edu/cci/index/ccishop/fullconfiguration.html for more detail. Our department (UNC-CH) is thinking of purchasing a machine for use with Debian. Specifically, potential problems seem to be with a) NIC b) Video card. From the descriptions given, I'm not sure what to make of Integrated 10/100/1000 MB ENET or Integrated 10/100100 MB ENET. (I'm guessing the latter was a typo and should be the same as the former). Does anyone have any idea about this? As for the video card, it would also be useful to have more information. The web page says Intel Extreme Graphics 32MB and Open AGP slot NVIDIA QUADRO-4 64mb respectively. As far as the former goes, http://www.intel.com/solutions/index.htm?iid=Corporate+Header_Solutions seems to be relevant. However, it seems to be a binary X module + needs kernel support. As far as the latter goes, I think Nvidia also has binary drivers. I don't think the nv driver in X supports this card. Am I correct? I see that Debian packages the nvidia driver, and the description explicitly mentions the Quadro chipset. I used these drivers once and they worked OK, but does anyone have any direct experience with the Quadro chipset, specifically how it works with recent 2.4 kernels? Also, has anyone any experience of using this with a flat-panel display? Thanks in advance. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: problems with rsa authentication (for use with mpich)
On Sun, 12 Oct 2003 15:01:41 -0400, Roberto Sanchez [EMAIL PROTECTED] wrote: In your case it is the key contained in ~/.ssh/id_dsa.pub (or id_rsa.pub, if you use RSA authentication) from the host on which the login attempt will originate. Is it actually necessary to use the key? In other words, cannot one opt to not do DSA or RSA based authentication and just do host- based authentication (I believe that is what it is called)? I don't see any mention of this in the documentation. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt
On Mon, 13 Oct 2003 13:28:30 +0200, wjl [EMAIL PROTECTED] wrote: Hi group, since today I have a strange behaviour of apt 0.5.14 in unstable: when trying to install a package or doing an 'apt-get dist-upgrade', I get: Reading Package Lists... Done Building Dependency Tree... Done Segmentation fault RESULT=139 I'm not sure if this is in the buglist already - there is one segfault described there, but that was still with 0.5.13... A hint anyone? Why not try doing a backtrace with gdb to see where the problem lies? Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: problems with rsa authentication (for use with mpich)
On Sun, 12 Oct 2003, Pigeon wrote: On Sat, Oct 11, 2003 at 06:13:34PM +, Faheem Mitha wrote: Dear People, I am trying to enable rsa authentication for use with mpich. However, I am having some problems. Let us say I have two hosts, foo and bar. I'm trying to log in from foo to bar using rsh passwordless authentication. snip it doesn't work and falls back to password Have a look in /home/faheem/.ssh/authorized_keys on bar. You'll probably find the line ends in '...gnf5t78t= [EMAIL PROTECTED]' or something like that. Change it to '...gnf5t78t= /home/faheem/.ssh/identity' - doesn't matter if /home/faheem/.ssh/identity doesn't exist. That will probably make it work. The only file in that directory is known_hosts. I'm not sure why keys would enter into this at all anyway. As I understand it, the whole idea is to just stick the trusted hosts into .rhosts, and you're away. This seems to be a weird bug that doesn't affect most people, just selects occasional random victims to confuse the @$#% out of, like me, and one or two others on this list. Please feel free to cc any replies to me at [EMAIL PROTECTED] I feel free, but do you actually _want_ CCs, or just don't mind them? I want replies. This is force of habit from Usenet, where I have been told off for asking for CCs before. So I have learned to express myself more euphemistically. But I guess debian-user is more easy going than Usenet. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: problems with rsa authentication (for use with mpich)
On Sun, 12 Oct 2003, Roberto Sanchez wrote: Faheem Mitha wrote: The only file in that directory is known_hosts. I'm not sure why keys would enter into this at all anyway. As I understand it, the whole idea is to just stick the trusted hosts into .rhosts, and you're away. You want the key placed into the file ~/.ssh/authorized_keys. For example, I have placed the key from each of my client machines into the server, allowing me to login without password. I use the file I mentioned above, and it Just Works (TM). Which key is this exactly? Also, are you sure this is not what is called rsh rsa authentication? I was just looking for rsh authentication. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
problems with rsa authentication (for use with mpich)
Dear People, I am trying to enable rsa authentication for use with mpich. However, I am having some problems. Let us say I have two hosts, foo and bar. I'm trying to log in from foo to bar using rsh passwordless authentication. I added foo to the file .rhosts in my home directory on bar. I also changed the default on bar's ssh server (/etc/ssh/sshd_config) from # rhosts authentication should not be used #RhostsAuthentication no (changed from default by Faheem on 11th #October 2003) RhostsAuthentication yes # Don't read the user's ~/.rhosts and ~/.shosts files #IgnoreRhosts yes (changed from default by Faheem on 11th October #2003) IgnoreRhosts no I also changed the default on foo on the client side (/etc/ssh/ssh_config)) to # RhostsAuthentication no (changed from default by Faheem 11th October 2003) RhostsAuthentication yes However, I still get [EMAIL PROTECTED]:/usr/share/doc/ssh$ rsh -v bar OpenSSH_3.6.1p2 Debian 1:3.6.1p2-9, SSH protocols 1.5/2.0, OpenSSL 0x0090703f debug1: Reading configuration data /etc/ssh/ssh_config debug1: Rhosts Authentication disabled, originating port will not be trusted. debug1: Connecting to bar [...] port 22. debug1: Connection established. debug1: identity file /home/faheem/.ssh/identity type -1 debug1: identity file /home/faheem/.ssh/id_rsa type -1 debug1: identity file /home/faheem/.ssh/id_dsa type -1 debug1: Remote protocol version 2.0, remote software version OpenSSH_3.6.1p2 Debian 1:3.6.1p2-9 debug1: match: OpenSSH_3.6.1p2 Debian 1:3.6.1p2-9 pat OpenSSH* debug1: Enabling compatibility mode for protocol 2.0 debug1: Local version string SSH-2.0-OpenSSH_3.6.1p2 Debian 1:3.6.1p2-9 debug1: SSH2_MSG_KEXINIT sent debug1: SSH2_MSG_KEXINIT received debug1: kex: server-client aes128-cbc hmac-md5 none debug1: kex: client-server aes128-cbc hmac-md5 none debug1: SSH2_MSG_KEX_DH_GEX_REQUEST sent debug1: expecting SSH2_MSG_KEX_DH_GEX_GROUP debug1: SSH2_MSG_KEX_DH_GEX_INIT sent debug1: expecting SSH2_MSG_KEX_DH_GEX_REPLY debug1: Host 'bar' is known and matches the RSA host key. debug1: Found key in /home/faheem/.ssh/known_hosts:4 debug1: ssh_rsa_verify: signature correct debug1: SSH2_MSG_NEWKEYS sent debug1: expecting SSH2_MSG_NEWKEYS debug1: SSH2_MSG_NEWKEYS received debug1: SSH2_MSG_SERVICE_REQUEST sent debug1: SSH2_MSG_SERVICE_ACCEPT received debug1: Authentications that can continue: publickey,password,keyboard-interactive debug1: Next authentication method: publickey debug1: Trying private key: /home/faheem/.ssh/identity debug1: Trying private key: /home/faheem/.ssh/id_rsa debug1: Trying private key: /home/faheem/.ssh/id_dsa debug1: Next authentication method: keyboard-interactive debug1: Authentications that can continue: publickey,password,keyboard-interactive debug1: Next authentication method: password Can anyone help me? Please feel free to cc any replies to me at [EMAIL PROTECTED] Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
rsh authentication (Re: problems with rsa authentication (for use with mpich))
On Sat, 11 Oct 2003 18:13:34 + (UTC), Faheem Mitha [EMAIL PROTECTED] wrote: Dear People, I am trying to enable rsa authentication for use with mpich. However, I am having some problems. Sorry. That should have been rsh authentication. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: touchpads in Linux
On Sat, 20 Sep 2003 15:47:21 -0700, Wendell Cochran [EMAIL PROTECTED] wrote: Subsequent discussion in [EMAIL PROTECTED] reached the lout's boss, who firmly disavowed bias against Linux. Just then two late-comers to the discussion asked why all the fuss; just plug in a Cirque, they told me, away you go. Right they were, too. That's nice. However, could you give me more details? Are you using it as a normal PS2 device in X? Are you using the Glidepoint protocol? Could you possibly send me your X config setup for the mouse? Are you able to use all the features of the Easycat in Linux? I was considering getting the Smartcat, but I am not sure whether all the features are supported. These tablets sound like they might be quite a good thing. I am surprised more people are not using them. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
touchpads in Linux
Dear people, I'm trying to find a good touchpad (as a standalone device) to use with Linux in the place of a regular mouse. After spending quite a lot of time surfing the web, I am more confused than ever. The best general reference I found http://www.tifaq.org/mice/touchpads.html The main touchpads available all seem to have the name Glidepoint. The major manufacturer of these appears to be Cirque, which has the Easycat, Smartcat etc. However, Cirque has also licensed the technology to Alps, which makes their own Glidepoint. Strangely enough, the information I was able to find with regard to Linux was mostly about the Alps version, though the Cirques seem to be the most popular. I found traces of a driver someone has written for the Cirque *cat series, called Kitcat, but the main site seems to be defunct. To further confuse things, there appears to be something called the Adesso Smartcat. I am not sure what this is. If you are using one of these products under Linux, I'd be greatly obliged if you could tell me how well it works, and whether you would recommend it for use under Linux. Thanks. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [OT]: CVS replacement
On Fri, 08 Aug 2003 07:26:28 +0200, Xavier Maillard [EMAIL PROTECTED] wrote: Since ages, I wanted to replace CVS here. I firstly tried to find out all the most interesting candidates for this. I have now a list of 2 items: subversion and tla. I tried to setup tla but didn't achieve totally this task. So now I am facing a dilemna: what is the most appropriate software to finally replace my old CVS (which hasn't satisfied me) ? For the little I tried tla, it seems smaller and easier to setup but as I didn't try at all subversion, this may be wrong. So I claim your help in making my choice: what would you use to replace CVS ? Subversion claims to be the CVS replacement but tla seems also interesting to me. Does anybody have feedback/recommendations for the 2 solutions ? Hi. Like other people who replied here, I use subversion (since late April), and am very happy with it on the whole. Bear in mind if you start using it that it is still alpha, so there are lots of bugs, which are continually being fixed. The subversion developers and user community are very nice and friendly people. Subversion is actually the first version control system I've ever used. I tried cvs and could not get along with it. By the way, does anyone here use tla? If so, care to summarise the advantages/differences between tla and subversion from your perspective? In Colin Walter's online diary/weblog, he mentioned that he uses tla, which he prefers over subversion, but he didn't give any details. I've seen no signs of other Debian people using it. I looked at it briefly, but was a bit put off by the fact that it enforces such strict rules for naming things. Eg. archives have to start with an email address. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt and changelogs
On Tue, 5 Aug 2003 19:36:26 -0400, David Corbin [EMAIL PROTECTED] wrote: On Monday 04 August 2003 20:54, Hugh Saunders wrote: On Mon, Aug 04, 2003 at 06:10:19PM -0500, Matt Peter wrote: or install apt-listchanges It fails for me, with this: The following information may help to resolve the situation: The following packages have unmet dependencies: apt-listchanges: Depends: python (= 2.2.2) but 2.1.3-3.2 is to be installed Depends: python-apt but it is not going to be installed E: Broken packages Is this a local problem of mine, or debian problem? This is fixable. See a post called Fun with python-apt by Anthony DeRobertis ([EMAIL PROTECTED]) to debian-devel which describes how to fix the problem. Worked for me. The google group url is http://groups.google.com/groups?hl=enlr=ie= UTF-8selm=20030614054006%245aa9%40gated-at.bofh.it (sorry, had to break up the url, silly slrn). Alternatively, you could backport apt 0.5.8 (which seems quite stable) to sarge. This should make the problem go away. Also worked for me. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
pam authentication (Re: setting up an openafs server on Debian)
Hi, Thanks for your helpful reply. On Thu, 24 Jul 2003, Todd Pytel wrote: On Thu, 24 Jul 2003 16:06:37 -0400 (EDT) Faheem Mitha [EMAIL PROTECTED] wrote: Does a tutorial for AFS server installation on Debian exist anywhere? My impression is no. There's a decent write-up in the docs for one of the AFS packages - I don't remember which one specifically. Those docs assume that you'll be setting AFS with MIT Kerb 5, which is recommended these days, so they won't quite apply to your university network. But in any case, AFS isn't something that you'll just pick up in a day, especially if you're not familiar with Kerberos already. If you have the machines to spare, I would strongly recommend setting up a private Kerberos realm before you get into AFS. That is a good suggestion. I'll try setting up Kerberos first before doing anything else. In any case, there is another, hopefully bite-sized task which has been on my todo list for awhile. This is, to enable integrated afs token fetching and login onto my Linux machines. My users don't like typing klog, and they can't remember it... :-) I took a look at the Openafs quick start Unix guide and it says the following. *** Place the AFS entry below any entries that impose conditions under which you want the service to fail for a user who does not meet the entry's requirements. Mark these entries required. Place the AFS entry above any entries that need to execute only if AFS authentication fails. Insert the following AFS entry if using the Red Hat distribution: auth sufficient /lib/security/pam_afs.so try_first_pass ignore_root * The /etc/pam.d/login file for my test machine (Debian) follows. This is pretty much the default file; I've hardly modified it at all. I am not sure where exactly to place this line. I get the general idea but am afraid that if I put it in the wrong place I will not be able to log in. It would be extremely nice if someone using openafs Debian would tell me what line(s) they are using and where they put it. Then I too could use it with confidence that it will work. I hope it will not involve rearranging the entries, though. I basically want the login to behave as follows (if possible). If it cannot contact the afs server it should still log me into the local machine without a long delay (a few seconds should be Ok). Otherwise I want everything to behave as normal. The fetching token behaviour should be like klog. The Debian package I'm using is libpam-openafs-kaserver, which I think it the correct one. At any rate, it contains the file /lib/security/pam_afs.so. Thanks in advance. Faheem. * # # The PAM configuration file for the Shadow `login' service # # NOTE: If you use a session module (such as kerberos or NIS+) # that retains persistent credentials (like key caches, etc), you # need to enable the `CLOSE_SESSIONS' option in /etc/login.defs # in order for login to stay around until after logout to call # pam_close_session() and cleanup. # # Outputs an issue file prior to each login prompt (Replaces the # ISSUE_FILE option from login.defs). Uncomment for use # auth required pam_issue.so issue=/etc/issue # Disallows root logins except on tty's listed in /etc/securetty # (Replaces the `CONSOLE' setting from login.defs) auth requisite pam_securetty.so # Disallows other than root logins when /etc/nologin exists # (Replaces the `NOLOGINS_FILE' option from login.defs) auth requisite pam_nologin.so # This module parses /etc/environment (the standard for setting # environ vars) and also allows you to use an extended config # file /etc/security/pam_env.conf. # (Replaces the `ENVIRON_FILE' setting from login.defs) auth required pam_env.so # Standard Un*x authentication. The nullok line allows passwordless # accounts. auth required pam_unix.so nullok # This allows certain extra groups to be granted to a user # based on things like time of day, tty, service, and user. # Please uncomment and edit /etc/security/group.conf if you # wish to use this. # (Replaces the `CONSOLE_GROUPS' option in login.defs) # auth optional pam_group.so # Uncomment and edit /etc/security/time.conf if you need to set # time restrainst on logins. # (Replaces the `PORTTIME_CHECKS_ENAB' option from login.defs # as well as /etc/porttime) # accountrequisite pam_time.so # Uncomment and edit /etc/security/access.conf if you need to # set access limits. # (Replaces /etc/login.access file) # account required pam_access.so # Standard Un*x account and session accountrequired pam_unix.so sessionrequired pam_unix.so # Sets up user limits, please uncomment and read /etc/security/limits.conf # to enable this functionality. # (Replaces
Re: setting up an openafs server on Debian
On Fri, 25 Jul 2003, David Z Maze wrote: (It'd be interesting to know what your actual goals are here. You're probably not going to be able to use your disk to randomly add volumes to the cs.unc.edu cell. I've found it useful to set up a personal cell before on the AFS is better than NFS mantra, but accessing it from outside meant reconfiguring machines to know about my Kerberos realm and AFS cell, which was a pain. It seems like my current research group at MIT would benefit from having AFS, but deploying that would be pretty tricky, and we're sufficiently xenophobic and demanding that we don't we to use the main athena.mit.edu AFS cell, so we'd have to come up with hardware and admins to run our own cell. AFS server in a box would be neat, but I don't think anyone makes one.) Well, I'm just trying to set up a private cell, largely for experimentation at this point. If this works out, I might start using it for my private use. Also, our department uses the university AFS, which only allows a small amount of space per person. I was considering whether it would be feasible to set up a private AFS server just for my dept. Nobody knows anything about this, so it would be just me administering it. I'm not sure how practical this latter goal would be in the long term, though. It would probably require a dedicated system administrator. Incidentally, the AFS cell we use is called isis.unc.edu. I've been looking at the Debian documentation. The configuration session is quite enlightening. I think I might be able to get a working system by just following the steps. I think I should do a little technical reading first, though. I've also been writing to the openafs mailing list, which has been giving me reams of useful information. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
setting up an openafs server on Debian
Dear People, I just sent the following, rather clueless, and somewhat Debian-specific message to the openafs-info mailing list. If anyone has any insights to share they would be welcome. Does anyone here have any experience setting up an openafs server on Debian and would be willing to share his experiences? If it is not already obvious, I know almost nothing about kerberos and precious little about afs. You can cc me if you want, though I can check replies via gmane. Thanks. Faheem. -- Forwarded message -- Date: Thu, 24 Jul 2003 15:45:59 -0400 (EDT) From: Faheem Mitha [EMAIL PROTECTED] To: [EMAIL PROTECTED] Subject: some simple openafs questions Dear People, I'm ashamed to admit that I have been administering openafs clients on Linux for over two years, but still lack a basic understanding of how openafs actually works. So here are some very simple and probably rather silly questions. They are regarding things where I find the online documentation not completely clear. I should say that I use openafs on Debian (testing), which does so much of the configuration that it makes it easy to be ignorant, though the choice of course is mine. In what follows, the documentation refers to that obtained by following the Documentation link from openafs.org, which seems to be mostly IBM's own documentation. Also, I'm using AFS on the University of North Carolina campus. 1) When using an afs client, the command `klog' fetches tokens from the campus server. Am I correct in thinking that this fetching involves use of kerberos on the campus server? I don't have kerberos installed on my client machine, though I have seen descriptions which involve installation of kerberos on the client machine. Is kerberos not required at the client end? 2) I'm considering trying to install a Openafs server on a Debian machine. I am not completely clear from the documentation whether it is actually nececssary to install and configure kerberos (kerberos 5 seems to be the preferred version). Parts of the documentation suggest that one could use the `afs authentication system', whatever this is. Adding to my confusion is that the openafs debian packages openafs-dbserver and openafs-fileserver do not mention kerberos even as a recommends. If it is not necessary, is it still desirable to use kerberos? Does a tutorial for AFS server installation on Debian exist anywhere? My impression is no. Thanks in advance. Please cc me on any reply. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: X sudden brokenness
On Tue, 24 Jun 2003 01:32:58 -0500, Stephan Sauerburger [EMAIL PROTECTED] wrote: Laugh.. I solved it. Somehow, my XF86Config-4 must have been replaced after the last apt-get upgrade. I noticed it was having trouble loading GLcore; that's when I changed it to nvidia in the Modules section. However, I failed to notice that the Device section in the file I just sent the list had: Drivernv instead of: Drivernvidia Thanks and sorry for the bandwidth usage.. =oP If you don't want xserver-xfree86 to overwrite your current X config file on future upgrades, I suggest you do dpkg-reconfigure xserver-xfree86 and when it asks if you want to manage your config file using Debconf say no. Otherwise (I think) it may (will?) overwrite your config file (with info in the debconf database) the next time an upgrade happens. You can get debconf to play nice with your local changes, but this is slightly more complicated. See the documentation. Anyway, since I posted it, might as well let my 2 sub-questions still stand: 1. Is it possible to reinstall a package, including all its dependencies, forcing placement of all their files, without doing a remove first? You said earlier that However, they only get like 70-something KB of data. They should be getting dozens of megabytes worth of files with all their dependencies, shouldn't they? Isn't there a way to force reinstallation of absolutely everything, without apt-get remove'ing first? The package *is* being reinstalled, it is just that some stuff may come from the local cache in /var/cache/apt/archives, depending on how recently you cleaned it. Apt looks there first. However, only the package itself will be reinstalled. If you really want to force reinstallation of all dependencies you'll have to specify them explicitly on the command line. apt-get --reinstall install ... 2. If you have your distro and all your packages installed under Unstable, is there a clean, easy way to revert back to Stable or Testing? There is a high probability that you could screw up your system by an attempted downgrade. Debian does not support this officially, so no provisions are made for this. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [tir@igidr.ac.in: Problems with xlibs.]
On Thu, 29 May 2003 22:16:04 +0530, Tirthankar C. Patnaik [EMAIL PROTECTED] wrote: lewis:/var/cache/apt/archives# apt-get -f install Reading Package Lists... Done Building Dependency Tree... Done Correcting dependencies... Done The following extra packages will be installed: xlibs 1 packages upgraded, 0 newly installed, 0 to remove and 530 not upgraded. 11 packages not fully installed or removed. Need to get 0B/1253kB of archives. After unpacking 81.9kB will be freed. Do you want to continue? [Y/n] (Reading database ... 106921 files and directories currently installed.) Preparing to replace xlibs 4.1.0-14 (using .../xlibs_4.2.1-6_i386.deb) ... Unpacking replacement xlibs ... dpkg: error processing /var/cache/apt/archives/xlibs_4.2.1-6_i386.deb (--unpack): trying to overwrite `/usr/X11R6/lib/X11/app-defaults', which is also in package wordnet Errors were encountered while processing: ^^^ /var/cache/apt/archives/xlibs_4.2.1-6_i386.deb ^^^ E: Sub-process /usr/bin/dpkg returned an error code (1) lewis:/var/cache/apt/archives# How do I get over this problem with xlibs? Try removing wordnet, and then try again. apt-get remove --purge wordnet You can try reinstalling wordnet after you are done with everything else. Hmm. I don't have this installed, and packages.debian.org does not show /usr/X11R6/lib/X11/app-defaults as one of the files included in wordnet. In any case, try that and if you have further problems, get back to the list. You can cc me if you wish. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
re: New Creating Custom Kernels newbiedoc
On Wed, 26 Mar 2003 [EMAIL PROTECTED] wrote: Thank you for your feedback. Just one more. I think that you need tk*-dev in section 2.2, not just tk*, to do make xconfig, since this amounts to doing a compilation. Faheem -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
re: New Creating Custom Kernels newbiedoc
On Wed, 26 Mar 2003 [EMAIL PROTECTED] wrote: Thank you for your feedback. 1) The pcmcia option merits special mention, since by default it is turned on (at least for the kernels I have used) and you explicitly need to turn it off if you aren't using it, otherwise the compilation will halt before it begins. You could give directions on how to turn it off. I don't understand what you mean here. A kernel won't refuse to compile just because you have included an option and you don't have the hardware for it. By way of discussing Third-party modules I used PCMCIA as an example. I mentioned that PCMCIA kernel support must be turned off to use the pcmcia-cs package. Or did you mean something else entirely? If so please let me know. I guess I was not precise enough. :-) What I meant to say was that by default PCMCIA support is turned on, but apparently you still need to pick some more specific suboption or something (I don't use this option so I'm not sure what it is). So, when compiling the kernel, if you are not using PCMCIA, it will halt for further user input at this stage unless PCMCIA is turned off completely, which can be very confusing to someone who doesn't know what it means. It keeps happening to me (with the standard Debian sources). I'm surprised that you haven't (unless you actually do use PCMCIA). Hmm. Strange. I am not currently able to reproduce this problem with either the 2.4.17 sources or the 2.4.19 source (which is what I have installed currently), though I have run into it repeatedly in the past. Ok, please ignore the above comment. 2) By the way, isn't it better to do the kernel image and the third party modules separately? Usually I do a fakeroot make-kpkg modules-clean before fakeroot make-kpkg modules-image which is what the kernel-package README recommends. That may have been true at one time; I actually do it all at once. I've done dozens of kernels on several machines and never had it hiccup for that reason. This is probably true if you are only doing it once. But if you are recompiling the kernel after changing options (whether using the same kernel sources or different kernel sources) and using the same module sources, you need in general to clean the modules directory in between, otherwise old object files etc. can mess things up. So in general I think it is best to add the modules clean option. If you don't agree, you could ask Manoj what he thinks. 3) You could mention the --added_modules option in 8.3 for use with modules-image or modules-clean. It is in the man page, but I didn't notice its existence for quite some time. I may do that. The idea is to cover just enough, but not too much; I'm looking for that balance point. True, you don't want to try to cover every possibility. But it is no big deal. If you are talking about compiling third party modules, you could just mention that you can select which of the modules to compile by using the option. It is only one extra line. A couple of extra things. 1) You could mention in section 2.2 that you can use different versions of gcc to compile, and that you can set the CC environment variable to choose which version of gcc to use. This is relevant because of the upcoming gcc transition to 3.2. Some people may prefer to continue using 2.95 for a bit. 2) Also, it might be worth seeing if you can get your tutorial distributed officially with kernel-package. The only thing is I am not sure how Debian feels about the GNU FDL, but I think there are issues. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: New Creating Custom Kernels newbiedoc
On Wed, 26 Mar 2003 [EMAIL PROTECTED] wrote: On Wed, 26 Mar 2003 13:43:32 -0500 (EST) Faheem Mitha [EMAIL PROTECTED] wrote: On Wed, 26 Mar 2003 [EMAIL PROTECTED] wrote: Thank you for your feedback. Just one more. I think that you need tk*-dev in section 2.2, not just tk*, to do make xconfig, since this amounts to doing a compilation. I worked from a clean installation of Woody, without even X. I installed just enough to do menuconfig. Then I installed just enough to get X. Then I tried xconfig. make whined about not finding wish. apt-get install wish told me that wish is included in either tk8.0 or tk8.1 or tk8.3 (which is why the package section reads that way). Installing tk8.3 was enough to do xconfig. Hm. Really? Well, this sounds very thorough. You should tell Manoj about this, since the README.gz in kernel-package mentions tk*-dev. Just a couple more things. In 2.2 you say 8. bin86 (for building 2.2.x kernels on PCs) But I thought that bin86 was an x86 thing, and not just for 2.2 kernels. Also, in 3.2 why don't you just suggest using dpkg -l to check the versions, instead of telling people to read the changelog? It is true that if you want to make sure that the version you have is newer than the one mentioned, you could check the Debian changelog to see if the version mentioned was in it, but I almost never do that in practice. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: New Creating Custom Kernels newbiedoc
On Tue, 25 Mar 2003 15:38:55 -0500, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: A new version of Creating Custom Kernels With Debian's Kernel-Package System is available at: http://newbiedoc.sourceforge.net/system/kernel-pkg.html I'm interested in feedback, especially about mistakes or useability issues. Thank you, Kevin Nice job. A couple of quick comments. 1) The pcmcia option merits special mention, since by default it is turned on (at least for the kernels I have used) and you explictly need to turn it off if you aren't using it, otherwise the compilation will halt before it begins. You could give directions on how to turn it off. 2) By the way, isn't it better to do the kernel image and the third party modules separately? Usually I do a fakeroot make-kpkg modules-clean before fakeroot make-kpkg modules-image which is what the kernel-package README recommends. 3) You could mention the --added_modules option in 8.3 for use with modules-image or modules-clean. It is in the man page, but I didn't notice its existence for quite some time. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
gmane.org (was Re: default editor)
On Tue, 18 Mar 2003 21:15:28 -0800, Will Yardley [EMAIL PROTECTED] wrote: [ Tried to post to linux.debian.user, but apparently this doesn't work at the moment, unfortunately. ] Try the public access newsserver gmane.org. For more information see the web page of the same name. The group corresponding to this mailing list is gmane.linux.debian.user. This is a mail to news gateway (both directions: MANE apparently stands for MAil to NEws And back again) so it does everything you could want. Very useful if you use news; I hope it doesn't go away. BTW, this is an initiative by the author of Gnus. Faheem (wondering if he should make a donation to gmane). -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: gmane.org (was Re: default editor)
On Wed, 19 Mar 2003 21:04:25 + (UTC), Faheem Mitha [EMAIL PROTECTED] wrote: On Tue, 18 Mar 2003 21:15:28 -0800, Will Yardley [EMAIL PROTECTED] wrote: [ Tried to post to linux.debian.user, but apparently this doesn't work at the moment, unfortunately. ] Try the public access newsserver gmane.org. For more information see the web page of the same name. The group corresponding to this mailing list is gmane.linux.debian.user. This is a mail to news gateway (both directions: MANE apparently stands for MAil to NEws And back again) so it does everything you could want. Very useful if you use news; I hope it doesn't go away. BTW, this is an initiative by the author of Gnus. Oh yes, and the server gmane.org runs Debian. See http://gmane.org/host.php. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 'apt-cache search' question
On Sun, 16 Mar 2003 10:46:46 -0500, Rick Pasotto [EMAIL PROTECTED] wrote: There was recently a question about which .deb added GIF capability to the gimp. Using 'apt-cache search gimp gif' did not return the correct answer. The man page for apt-cache says: search search performs a full text search on all available package files for the regex pattern given. It searchs the package names and the descriptions for an occurance of the string and prints out the package name and the short description. The problem seems to be that the word 'gimp' appears only in the package name and the word 'gif' appears only in the description. Shouldn't the search treat the package name and the description as a single string? Doing so would have given the correct answer for the above query. I assume you mean gimp1.2-nonfree (or gimp1.3-nonfree)? I looked at the source code for apt-cache search but it is not clear what it is doing. I agree with you that combined searching of the package name and the description is desirable. I think you should report this as a bug against apt. I did a search, and I don't see anything currently like this in the bug database. In fact, there only seem to be a few (a couple or so) bugs against apt-cache search. If you don't want to file a bug I can do so. Let me know. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: 'apt-cache search' question
On Sun, 16 Mar 2003 22:04:40 -0500, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote: On Mon, 17 Mar 2003 02:20:21 + (UTC) Faheem Mitha [EMAIL PROTECTED] wrote: the gimp. Using 'apt-cache search gimp gif' did not return the correct answer. The man page for apt-cache says: search search performs a full text search on all available package files for the regex pattern given. It searchs the package names and the descriptions for an occurance of the string and I'm pretty sure it already does what the man page says it does. When I use apt-cache search sometimes I get a whole lot of stuff that isn't what I'm looking for, because of the package description search. I think the problem is that in the command line apt-cache search gimp gif the phrase gimp gif is not a regular expression. What happened here is that apt-cache searched for the pattern gimp and stopped after it saw the space. It didn't even see gif. Hmm. Interesting theory. This is not what apt-cache search actually does, though. This is also not what the man page says it does. Regular expressions is a highly detailed technical topic, which I couldn't begin to describe in a mail message. If you can find the O'Reilly book Mastering Regular Expressions you'll find it to be excellent. I don't know any others, but that doesn't mean they don't exist. Don't embarrass yourselves by filing a bug on this. :) I'm happy to embarrass myself. :-) So, is anyone going to file a bug, or shall I? Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: buy or build computer?
On Sun, 9 Mar 2003 13:55:32 -0500, Peter Christensen [EMAIL PROTECTED] wrote: My five-year-old Gateway Pentium 200 MHz died recently. (It won't boot from the hard drive or a rescue disk, and it won't go into bios-setup mode.) I don't think it's fixable, and anyway, it was so slow that it's probably time to replace it. Temporarily I'm using a borrowed computer with Win95. Yuck! For my next computer I want to make sure that everything is compatible with Linux. I searched this list and found a few posts about buying computers. They were a little old (one or two years), so I'm wondering if the situation has changed. A few people recommended the AMD Athlon processor over Pentiums. And Matrox for video, Soundblaster or Ensoniq for sound. Any thoughts on this? I've heard that computers nowadays are built with the cheapest possible components, so I was wondering if building it myself would be a good idea. It might not be much cheaper than buying one from Dell or Gateway, but if the result was a better quality machine it might be worthwhile. So far I've only had to replace broken components in my Gateway, such as the hard drive and CDrom, also added memory. Building a computer would be a challenge, but I think I'd enjoy doing it... I would definitely advise against buying a computer from a mainstream vendor, eg IBM, Dell, etc. However, there is a third option which you are not considering, which is buying a computer from a small place, which will build it according to your specifications. This is the way I went. I went to resellerratings.com, and eventually settled on a place called Envision Computer Solutions (http://envisioncs.net/), which was highly rated. I had some technical problems with the order (not the owner's fault) but eventually got them settled, and the computer in question is now humming away quietly near me. I would definitely buy from this place again, and I am very picky when it comes to service. The owner, Todd Keller, is pretty neutral when it comes to Linux. He does not know anything about it, but neither does he freak out when it is mentioned, like the big resellers supposedly do (I have no personal experience with this). He is generally very cooperative and service-oriented. Bear in mind that if you assemble the computer yourself: a) There is a chance you may damage something if you don't know exactly what you are doing. At the very least a book or two is required, and no book can stop you from making mistakes. You might want to take a look at the book PC Hardware in a Nutshell, which I have a copy of, and the corresponding web site, http://hardwareguys.com/ I found it useful in educating me about the issues. However, I would not necessarily take their specific picks too seriously. These choices are often quite subjective, and the best choice changes all the time anyway. In particular, I am not convinced that PC Power and Cooling power supplies are all that great. b) You'll have to research the components in addition to putting them all together. Just the research by itself is quite a bit of work. I spent several weeks on understanding the issues. Of course hardware doesn't change that much so you won't have to do this very often, but you will need to do this once. Bear in mind of course you want things that are well supported by Linux. So, it might be an idea to get someone to actually assemble it for you. Otherwise, it could be an expensive (not to mention time-consuming) learning experience. I think, however, most people in this group would disagree. You can see what they think from two earlier threads on debian-user which I started. 1) On Mon, 2 Jul 2001 From: Faheem Mitha [EMAIL PROTECTED] Subject: buying a computer (http://lists.debian.org/debian-user/2001/debian-user-200107/msg00318.html) 2) On 25 Feb 2002 From: Faheem Mitha [EMAIL PROTECTED] Subject: hardware quote comments? (http://lists.debian.org/debian-user/2002/debian-user-200202/msg04267.html) The above hardware selection is rather different from my final hardware selection, which I can send you if you want. Also, I eventually did not go with this vendor, since Envision was rather cheaper and I wasn't willing to pay a big premium to have Debian preinstalled. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
adding mime support for viewing ps.gz attachments
Dear People, /usr/lib/mime/packages/gv has only the following lines application/postscript; /usr/bin/X11/gv %s; test=test -n $DISPLAY; description=postscript application/ghostview; /usr/bin/X11/gv %s; test=test -n $DISPLAY application/pdf; /usr/bin/X11/gv %s; test=test -n $DISPLAY However, gv views ps.gz (postscript files compressed with gzip) just fine, presumably uncompressing them on the fly. Can anyone tell me what the appropriate line would be to add to tell applications to view ps.gz files using gv as well? Just sticking in ps.gz in the place of postscript does not seem to do the trick. If I can figure out how to do this, I'll make this a wishlist bug for gv. Thanks. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: kernel-package gcc configuration etc. (Re: How do I do a sid upgrade and not get the gcc 3.2 stuff?)
On 25 Feb 2003 02:14:41 +0100, Mario Vukelic [EMAIL PROTECTED] wrote:
On Mon, 2003-02-24 at 21:03, Faheem Mitha wrote:
So, you can still compile kernels with it, though I agree that there
might be some problems if everything insists on calling gcc.
Set the env variable then
Hmm. Does
anyone know if it is currently possible to configure what version of
gcc to use with kernel-package?
I just use
[EMAIL PROTECTED]: /usr/src/linux # GCC=gcc-2.95 make-kpkg yada yada
Thanks. Are you sure you don't mean CC? I can't find this documented
anywhere in kernel-package docs, but there is the following line in
/usr/bin/make-kpkg.
my $cc = $ENV{'CC'};
I don't know perl, but it looks like it is sourcing the CC environment
variable.
Perhaps this should be added to the man page (kernel-pkg.conf
perhaps)? What do you think?
Faheem.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
kernel-package gcc configuration etc. (Re: How do I do a sid upgrade and not get the gcc 3.2 stuff?)
On Mon, 24 Feb 2003 13:53:15 -0500, John Covici [EMAIL PROTECTED] wrote: I was trying to do a sid upgrade and it keeps trying to give me the gcc 3.2 compiler, but I want to stay with the 2.95 -- the new ones are broke and don't compile kernels properly. gcc 2.95 will still be on your system. faheem ~apt-show-versions -a gcc-2.95 gcc-2.951:2.95.4-11woody1 install ok installed gcc-2.951:2.95.4-11woody1 stable gcc-2.951:2.95.4-11 testing gcc-2.951:2.95.4-15 unstable gcc-2.95/stable uptodate 1:2.95.4-11woody1 It just won't be the default gcc any longer. faheem ~apt-show-versions -a gcc gcc 2:2.95.4-17 install ok installed gcc 2:2.95.4-14 stable gcc 2:2.95.4-17 testing gcc 3:3.2.2-0 unstable gcc/testing uptodate 2:2.95.4-17 So, you can still compile kernels with it, though I agree that there might be some problems if everything insists on calling gcc. Hmm. Does anyone know if it is currently possible to configure what version of gcc to use with kernel-package? BTW, I don't see any kernel compilation Debian bugs reported for gcc. If you have a problem, report it and perhaps someone will fix it. They can't fix it if they don't know it exists. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Hen and egg with tktable-dev
On 22 Feb 2003 17:12:27 +0100, Svenn Are Bjerkem [EMAIL PROTECTED] wrote: Hi, want to install tktable and tktable-dev and use tk8.4-dev tktable-dev need tk8.3-dev tk8.4-dev provide tk-dev tk8.3-dev conflict tk-dev So if I want tktable-dev I have to deinstall tk8.4-dev with which I want to develop. You might be able to recompile tktable to use tk8.4. You'll have to change the control file to reflect this, but my guess is it will work Ok, even though the current package does not mention tk8.4 explicitly. BTW, there is a 2.8 version availahle, though the most recent version in Debian is 2.7. The project home page says for 2.8 WHAT: Tktable v2.8, 2D editable table/matrix widget WHERE: http://tktable.sourceforge.net/ Tktable binaries are part of ActiveTcl: http://www.ActiveState.com/Tcl REQUIREMENTS: Tcl/Tk8.0+, compiling is required Compiles everywhere Tk does! It appears that tktable does not really depend on tk8.3. So, you should be set. Tktable has rather modest build-depends, specifically debhelper ( 3.0.0), tcl8.3, tcl8.3-dev, tk8.3, tk8.3-dev, rman, debhelper Yes, debhelper is mentioned twice (may be a typo). Download the source (apt-get source tktable) or whatever, change all the 8.3 references to 8.4 in the control file and (possibly) the .dsc file, install the build dependencies, and you should be good to go. I recommmend getting devscripts and using debuild binary. You might want to pin your new tktable using apt preferences (my preferred method) or otherwise make sure it is not upgraded by apt. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: dpkg cache broken after system crash
On Sat, 22 Feb 2003 21:21:47 +0100, Marcio Rosa da Silva [EMAIL PROTECTED] wrote: brain:~ 4 # apt-get install linuxdoc-tools Reading Package Lists... Done Building Dependency Tree... Done The following extra packages will be installed: libsp1 sp The following NEW packages will be installed: libsp1 linuxdoc-tools sp 0 packages upgraded, 3 newly installed, 0 to remove and 0 not upgraded. Need to get 0B/1894kB of archives. After unpacking 6176kB will be used. Do you want to continue? [Y/n] (Reading database ... dpkg: serious warning: files list file for package `libpisock++0' missing, assuming package has no files currently installed. dpkg: error processing /var/cache/apt/archives/libsp1_1.3.4-1.2.1-28_i386.deb (--unpack): failed in buffer_read(fd): files list for package `pilot-link': Is a directory Errors were encountered while processing: /var/cache/apt/archives/libsp1_1.3.4-1.2.1-28_i386.deb Processing was halted because there were too many errors. E: Sub-process /usr/bin/dpkg returned an error code (1) You could try reinstalling libpisock++0, pilot-link and other packages for which you get error messages, assuming dpkg is still working. This should be harmless in any case. You might consider doing a complete purge before reinstalling. Perhaps apt-get remove --purge pilot-link libpisock++0 #ignore package-integrity-related errors you get here and then apt-get install pilot-link libpisock++0 # and other packages that got removed above I'd also take a look at the output of apt-get -f install though this is more about dependency checking. It looks like some Palm-related tools you were using got messed up, perhaps because they were in use when your system froze. If the above fails, try reinstalling all your Palm-related tools. And set up a journalling file system to avoid this happening again. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
[OT] volunteer work
Dear People, I've marked this OT, but really it is probably less OT than many posts that have appeared on this mailing list recently. :-) I was wondering if any people on this group do volunteer work using their knowledge/expertise with Debian/Free Software to help deserving causes. I'm particularly thinking of progressive groups here. Here I'm defining progressive groups broadly to mean any group whose aims are similar to those of Free Software, namely promoting freedom, community etc.. Another possibility would be helping your local municipality or town with their free software needs, hence (theoretically) saving everyone money in taxes etc. The reason I am asking is that I am thinking about getting involved in something like this in the near future as a way to make a contribution, and I was wondering how difficult it would be. Is there really a market for this kind of thing or not? I suppose I am assuming here the existence of groups that can afford access to the net but cannot afford the cost of full time computer staff/system adminstrators to keep their desktops/servers running. I expect this situation exists, but I wonder what difficulties might arise here. Helping ones local government might be more difficult than this, since they are more likely to want to go with the mainstream options. I'd be interested in hearing about people's experiences. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
problems setting up ppp connection with ispwest
Dear People, Anyone with experience setting up a ispwest ppp account on Debian? I thought I would give them a try since they offer a 15 day free trial account, but I'm having problems getting set up with pppconfig. My current ISP worked right out of the box with PAP. I suppose they support PAP, but I tried with and without static dns, and no dice. I tried talking to their technical support, but they say we are compatible with linux but don't offer support. Ha ha. At the moment, I'm not asking for debugging help, but asking if there is anyone who has been down this road before and knows how to do it. Thanks. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Help debugging package problems (again)
On Sun, 2 Feb 2003 17:43:29 -0800 (PST), Bill Moseley [EMAIL PROTECTED] wrote: [back to list] On Sun, 2 Feb 2003, Faheem Mitha wrote: ii libc6 2.3.1-9 GNU C Library: Shared libraries and Timezone data As I am sure hordes of people will shortly tell you, you have upgraded to the libc6 in unstable. In general, not such a great idea. Now some of the packages in testing are going to be uninstallable for you, and you may have other problems. Be careful about upgrading packages from unstable. In general you don't want to upgrade base packages, particularly libc and c++ libraries and perl. locales in testing depends on a particular value of libc6 That's probably what happened. So a program has a bug fix or a feature that's required and it's in unstable what options do I have? Compile the deb sources from unstable into a deb which you then install with dpkg -i (if possible). Yes, this is not ideal, but the other option is just to stay with stable or track unstable, and both those options have their own problems. Some of the Debian developers are on record as saying they don't think testing is such a great idea, since you don't get security updates in testing etc. In some cases it is possible to recompile packages from unstable locally, and it this is possible, do it. But they might have too many dependencies that are only in unstable for such an approach to be practical. The new apt-src by Joey Hess (only in unstable currently) might help with that. I'll take a look. At some point it almost seems easier to just build from source. I've been trying to avoid that as much as possible. But maybe it's worse to break things my mixing up testing and unstable. Just to clarify, when I say build from sources, do you mean compile the deb sources? Because this is what I mean. Apparently apt-src gives the long-awaited functionality of compiling from sources and recursively compiling the build dependencies if they are not sufficiently up to date. (This is a real pain to handle manually.) Sort of ports-like functionality. So once it is available, it will make things easier. I'm not aware of any backports to testing, unfortunately. You might try downgrading libc to testing. Warning: this might break your system. Apt-get will give you horrible warnings if you try to do this, but if your system is still mostly testing, it may well work. Try apt-get install libc6/testing Using apt-get -s install libc6/testing says it's going to remove over 200 packages. bumby:~# dpkg -l | grep ^ii | wc -l 487 bumby:~# apt-get -s install libc6/testing | grep Remv | wc -l 219 Doesn't make sense. It's removing packages that are in testing. For example: Remv aumix-gtk (2.7-26 Debian:testing) I think that downgrade would really break things Hmm. This may be because you may now have testing packages which depend on packages which are now installed as the unstable version (since many dependencies just say = something). If the unstable packages are now pulled out (since you are not telling apt to replace them with the testing version) then the testing packages get pulled out as well. This is just speculation, but you could take a look at the dependency structure and see. Check and see how many packages you now have from unstable. A better option might be to force a mass downgrade using apt preferences. If you put /etc/apt/prefences Package: * Pin: release a=testing Pin-Priority: 1001 Then all the packages in testing will have sufficiently high priority to overcome the downgrade prevention barrier. Since this is allowing the option to back up all unstable packages to testing versions, it may result in the removal of less packages. Whether it breaks the system or not would depend on whether it is going to mess with the base packages. Certainly there are no guarantees. You are presumably aware that mass downgrading is *not* officially supported. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [apt] Disabling upgrade to insecure packages
On Sun, 12 Jan 2003 18:15:46 +0100, Vincent Lefevre [EMAIL PROTECTED] wrote: On Sun, Jan 12, 2003 at 15:46:36 +, Faheem Mitha wrote: Really? This is not what the apt_preferences man page says. In particular Each package may be pinned to a specific version and each Packages file has a priority for every package inside. The highest priority assigned to a package is the one that is used. I admit I haven't looked at the source code, so I don't know how this is implemented, but it seems clear to me. Also, this is not just theoretical. I have a bunch of X 4.1 packages installed, and the lines in the apt preferences file prevent them from being upgraded to the more recent 4.2 version. I thought this is what you were after. Correct me if I was mistaken. But what if you want to allow the upgrade to 4.3 (or higher) when it becomes available? This is the problem. You can remove those lines once 4.3 is available. There is no way for apt to know which version you want to upgrade to unless you tell it. If you want to just make a particular version (ie. 4.2) uninstallable, then pin it to a sufficiently low priority, as I have outlined elsewhere in the thread. This is generally not such a good strategy, though, since minor upgrades will break this. Apt can't read your mind, or do magic, you know. You need to tell it what you want it to do. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [apt] Disabling upgrade to insecure packages
On Sun, 12 Jan 2003 11:12:40 +0100, Vincent Lefevre [EMAIL PROTECTED] wrote: On Sun, Jan 12, 2003 at 06:05:47 +, Faheem Mitha wrote: On Fri, 10 Jan 2003 19:04:37 +0100, Vincent Lefevre [EMAIL PROTECTED] wrote: Well, if one could put on hold a particular version of a package (given by the user), it would be better than nothing. Is there a way to do this? Of course. Just put in an entry in /etc/apt/preferences with sufficiently high pin number. Eg. I've put all my X packages on hold because I am running testing but have problems with 4.2. So I have Package: xfree86-common Pin: version 4.1.0-16 Pin-Priority: 1001 etc. etc. See also apt_preferences. HTH. Faheem. No, this won't work. This would put on hold *any* version after 4.1.0-16. This is not a particular version. Really? This is not what the apt_preferences man page says. In particular Each package may be pinned to a specific version and each Packages file has a priority for every package inside. The highest priority assigned to a package is the one that is used. I admit I haven't looked at the source code, so I don't know how this is implemented, but it seems clear to me. Also, this is not just theoretical. I have a bunch of X 4.1 packages installed, and the lines in the apt preferences file prevent them from being upgraded to the more recent 4.2 version. I thought this is what you were after. Correct me if I was mistaken. See also that I can't upgrade to 4.2 (which is in testing) or any more recent version even if I try. Chrestomanci:/home/faheem# dpkg -l xfree86-common [...] ||/ Name VersionDescription +++-==-==-=== pi xfree86-common 4.1.0-16 X Window System (XFree86) infrastructure Chrestomanci:/home/faheem# apt-get install xfree86-common Reading Package Lists... Done Building Dependency Tree... Done Sorry, xfree86-common is already the newest version. 0 packages upgraded, 0 newly installed, 0 to remove and 0 not upgraded. Chrestomanci:/home/faheem# apt-get install -t unstable xfree86-common Reading Package Lists... Done Building Dependency Tree... Done Sorry, xfree86-common is already the newest version. 0 packages upgraded, 0 newly installed, 0 to remove and 357 not upgraded. I would like something like the opposite: give a low score to some version, but I don't know how to do this, because the generic rules seem to take the precedence in this case. But the other versions would by default have lower scores, so would not be installed in favour of your pinned version. The version installed is always the version with the highest pin number. Note that 1000 is the downgrade prevention barrier, so with the above Pin even more recent versions should be removed in favour of the pinned version. If I am missing something, let me know what it is. In any case, perhaps you could just try it and see if it works. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [apt] Disabling upgrade to insecure packages
On Sun, 12 Jan 2003 11:12:40 +0100, Vincent Lefevre [EMAIL PROTECTED] wrote: On Sun, Jan 12, 2003 at 06:05:47 +, Faheem Mitha wrote: On Fri, 10 Jan 2003 19:04:37 +0100, Vincent Lefevre [EMAIL PROTECTED] wrote: Well, if one could put on hold a particular version of a package (given by the user), it would be better than nothing. Is there a way to do this? Of course. Just put in an entry in /etc/apt/preferences with sufficiently high pin number. Eg. I've put all my X packages on hold because I am running testing but have problems with 4.2. So I have Package: xfree86-common Pin: version 4.1.0-16 Pin-Priority: 1001 etc. etc. See also apt_preferences. HTH. Faheem. No, this won't work. This would put on hold *any* version after 4.1.0-16. This is not a particular version. I would like something like the opposite: give a low score to some version, but I don't know how to do this, because the generic rules seem to take the precedence in this case. If you want a particular version not to be installed, then give it a Pin lower than 100 and it will never be installed if there is another installable version in your sources. I've just read through this thread again, and I'm a little confused as to what you want. But apt preferences are good things to know about in any case. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: Home LAN email setup. Do you love yours?
On Sat, 11 Jan 2003 07:10:46 -0800 (PST), Bill Moseley [EMAIL PROTECTED] wrote: Searchable Archives? I should find a better method of archiving messages, but haven't needed that yet (since Eudora is fine with large folders). A system that auto-archived old read messages into a searchable archive would be great -- but I'd need a way to select and reply to an old archived message. That is, bring a message or thread out of the archive and into the mail system. If I could keep my mail box size down (yet still available) then IMAP might be the way to go. But I'd need a system where I can search for an email, and it would search both the current mail and the archived mail. Is there such a thing? I've been meaning to look into this in more detail, but consider a combination of hypermail and something like htdig. This will require a significant amount of configuration, though. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: [apt] Disabling upgrade to insecure packages
On Fri, 10 Jan 2003 19:04:37 +0100, Vincent Lefevre [EMAIL PROTECTED] wrote: On Fri, Jan 10, 2003 at 11:36:27 -0600, Jamin W. Collins wrote: There is nothing (that I know of) in the packaging system to know which packages are secure and which are not. You would need to put the package on hold during the insecure times and remove it from hold after. Well, if one could put on hold a particular version of a package (given by the user), it would be better than nothing. Is there a way to do this? Of course. Just put in an entry in /etc/apt/preferences with sufficiently high pin number. Eg. I've put all my X packages on hold because I am running testing but have problems with 4.2. So I have Package: xfree86-common Pin: version 4.1.0-16 Pin-Priority: 1001 etc. etc. See also apt_preferences. HTH. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: USB Mouse not working
On Sat, 4 Jan 2003 00:20:59 -0800 (PST), suresh kumar sharma [EMAIL PROTECTED] wrote: Hi, I have been trying to install a logitech optical mouse on my laptop through USB,but it does not work I have loaded input,usbcore,usb-uhci,and hid modules .I get light on the mouse and when I check dmesg I can see that it recoganises the mouse. input0: USB HID v1.10 Mouse [Logitech USB-PS/2 Optical Mouse] on usb1:2.0 but it does not work . I am using kernel 2.4.20 and have CONFIG_HID_HIDINPUT option as yes . Putting linux usb into Google will give you a number of links. The second link is The Linux USB sub-system. I recommend to your attention the first three sections of Ch 2. The section USB Human Interface Device (HID) Configuration covers USB mice. It is an excellent, clear and thorough exposition including detailed instructions, and will tell you everything you need to know. If something is not clear, ask the list, or any general Linux forum. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: ntpd; local clock reference 127.127.1.0
On Tue, 31 Dec 2002 18:43:19 -0500, Brenda J. Butler [EMAIL PROTECTED] wrote: On Tue, Dec 31, 2002 at 05:51:45PM -0500, Brenda J. Butler wrote: I have a dial-up system that is going to be a server for my home network. I want my inside machines to sync to seal (the ss20) and seal to sync to outside machines when the link is up, and to just provide time to the inside machines when the link is down. Another question: If I do get this to work, will the ntpd on seal continue to correct the time even when the outside link is down and the local clock is being used as reference? Have you considered using chrony instead? I'm on a dialup, and after attempts to get ntp to work for me, I discovered the existence of chrony and went with it. It works well, is easy to set up and is designed for use on dialup systems. There was some discussion of it on debian-user recently, I believe. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: MS Intellimouse Explorer
On Sun, 22 Dec 2002 16:08:52 -0500, Trey Sizemore [EMAIL PROTECTED] wrote: I liked the idea of Knoppix and downloaded it to try. I noticed that my Intellimouse Explorer (USB/optical) does not work however. Side to side motion moves the mouse up and down (up and down does nothing) and clicking the mouse buttons moves the cursor to the right incrementally. The mouse works fine in Red Hat and Libranet, but I have the same mouse issue in Mandrake 9.0. I have posted to both forums over the past weeks with no results (other than switch to a PS/2 mouse) and searched many FAQs. Any help in resolving this issue is greatly appreciated. I have the Intellimouse Explorer, and I get the same behaviour with the default bootup. But have you tried booting Knoppix with knoppix wheelmouse at the bootprompt? This enables IMPS/2 protocol for wheelmice. If this doesn't work, you could try the expert option. This gives you the option to interactively configure stuff. I haven't got around to trying either of these myself, and I've just lent my Knoppix CD to someone. See http://download.linuxtag.org/knoppix/knoppix-cheatcodes.txt for a list of boot-time options. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: MS Intellimouse Explorer
On Tue, 24 Dec 2002 02:52:52 + (UTC), Faheem Mitha [EMAIL PROTECTED] wrote: On Sun, 22 Dec 2002 16:08:52 -0500, Trey Sizemore [EMAIL PROTECTED] wrote: I liked the idea of Knoppix and downloaded it to try. I noticed that my Intellimouse Explorer (USB/optical) does not work however. Side to side motion moves the mouse up and down (up and down does nothing) and clicking the mouse buttons moves the cursor to the right incrementally. The mouse works fine in Red Hat and Libranet, but I have the same mouse issue in Mandrake 9.0. I have posted to both forums over the past weeks with no results (other than switch to a PS/2 mouse) and searched many FAQs. Any help in resolving this issue is greatly appreciated. I have the Intellimouse Explorer, and I get the same behaviour with the default bootup. But have you tried booting Knoppix with knoppix wheelmouse at the bootprompt? This enables IMPS/2 protocol for wheelmice. I just googled for knoppix intellimous and found someone who said knoppix wheelmouse doesn't work and that the cursor just runs mindlessly on the screen, ending at the far right where it stays no matter what I try. Oh well. I guess I should try these things before recommending them. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: many packages need to be upgraded
On Sun, 15 Dec 2002 21:20:06 -0500, Michael Waters [EMAIL PROTECTED] wrote: On Mon, Dec 16, 2002 at 02:04 +, Colin Watson wrote: In the case of gcc-2.95: gcc-2.95 (2.95.4.ds13-11woody1) stable; urgency=low * Upload to woody-proposed-updates: - Fix profiling for arm. - Fix internal compiler errors on s390. - Update Pascal release candidate 7 to final release. - Build the protoize-2.95 package again just in case that somebody relies on the protoize-2.95 and unprotoize-2.95 binaries. - No other changes to other compilers and architectures. -- Matthias Klose [EMAIL PROTECTED] Fri, 13 Sep 2002 23:58:04 +0200 Is there a way of showing this info for upgraded packages before downloading them? You could go to packages.qa.debian.org and type in gcc-2.95 (for example). It will then list for you the most recent changelogs, in order of increasing age, along with other information. You can see what version you are about to downgrade by doing apt-get -s upgrade or whatever. This is probably not the best way of doing this, but I don't currently know of a better option. It doesn't seem to show the version above, though. Possibly because it is a woody backport? It only lists the gcc-2.95 2.95.4.ds13-11 changelog. HTH. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: strange problem compiling source rpms
On Fri, 13 Dec 2002 10:31:18 -0700, Bob Proulx [EMAIL PROTECTED] wrote: --mP3DRpeJDSE+ciuQ Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Faheem Mitha [EMAIL PROTECTED] [2002-12-12 21:57:35 +]: rpm is trying to mkdir /usr/src/rpm/SOURCES, of course this fails if /usr/src/rpm does not exist. It's dumb code and a bad error message. Actually it does not try to make the directories. It just fails if the directories are not premade. Oh, I see. I thought something like install -d was the conventional way to do this. Inside of rpm it is a C program and it should make the directories using C code. The 'install -d' is for makefiles which are shell scripts. I see. But in that case the error message does seem rather misleading. Why does it complain about not being able to make directories if it is not trying to make them? (I haven't examined the internals of rpm). If you are building rpms and want to do so in your home directory this I suggest this. File $HOME/.rpmmacros: %_topdir /home/replay-with-your-login-name-here/src/rpms Then make all of the directories it wants in ~/src/rpms/RPMS, etc. and you can have a private build area. This is they way I build my own rpms. No collisions with other processes in the shared /usr/src area. I would have thought the only way there might be problems would be if multiple users were to try to build rpms (as user) in /usr/src/rpm, and then only if they were trying to build the same package. There would then be problems with files with different permissions colliding with each other. Is this correct? I recently ran into this problem with kernel building, and I'm now using /usr/local/src/login-name as a build area for packages (eg kernel) that I am building myself on a multiuser system. I prefer this to /home/replay-with-your-login-name-here/src/ Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: strange problem compiling source rpms
On Thu, 12 Dec 2002 14:37:58 -0500, Joey Hess [EMAIL PROTECTED] wrote: Faheem Mitha wrote: No. Am I supposed to create it? Isn't this something that should be done automatically, by a postinst script of whatever? If I am supposed to create it, I think it should be documented. The directory is supposed to be included in the package but was left out by mistake. I'm fixinig it now. Great, thanks. Does the user running rpm have permission to create it? (I suspect not to both) Um. Yes. It is root. (See the #?) Actually, even my user account (faheem) has permission to create directories in /usr/src. I've compiled kernels there as user. This is what is so puzzling about this error message. rpm is trying to mkdir /usr/src/rpm/SOURCES, of course this fails if /usr/src/rpm does not exist. It's dumb code and a bad error message. Oh, I see. I thought something like install -d was the conventional way to do this. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
strange problem compiling source rpms
Dear People, I recently tried to install a source rpm (as root) on a Debian system. I got laplace:/home/faheem# rpm -Uvh plugger-4.0-17.src.rpm error: cannot create %sourcedir /usr/src/rpm/SOURCES But this works on my home computer Chrestomanci:/home/faheem# rpm -Uvh plugger-4.0-6.src.rpm 1:plugger ### [100%] Ok, so they are not exactly the same version, but very similar. Can anyone explain what is going on? I'll been pulling my hair for a while trying to figure out why root cannot create /usr/src/rpm/SOURCES etc. Both laplace and my home computer Chrestomanci are supposedly tracking sarge, and have the exact same version of rpm. I see that I have the entire /usr/src/rpm directory hierachy installed on my system, and I am perfectly sure I did not create it myself. However, the rpm Debian package installation script does not seem to install it either, so it is a mystery how it got there. Did rpm simply create it as it needed it? In case anyone is wondering, I am trying to install the src rpm in order to compile it, an operation I had earlier successfully done on my home computer. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: X problem: mouse pointer shifted
On Wed, 11 Dec 2002 16:06:27 +0100, Jens Grivolla [EMAIL PROTECTED] wrote: Hi, I have a problem on my Acer TM210 laptop with integrated ALi (Trident) graphics (sorry, I don't have the exact name, it doesn't seem to show anywhere) on XFree86 4.2.1.1, but also on older versions. Regularly, after some time, the mouse pointer will shift and then be shown around 100 pixels to the right of the actual position (i.e. it will stop before reaching the left screen border and mouse clicks are to the left of the shown pointer position). It will sometimes switch back after a few hours. Has anybody experienced those problems, and is there a fix? Yes. From a post of mine dated 16th February 2002 in Debian user. I think you are the victim of the following bug, since your card is a trident. http://sdb.suse.de/en/sdb/html/wessels_trident_mouse.html Try putting Option sw_cursor in your /etc/X11/XF86Config-4 in the video card section. Somebody really ought to add this to the X documentation. It has been reported at least once to the Debian bug tracking system. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: strange problem compiling source rpms
On Wed, 11 Dec 2002 15:53:27 -0800 (PST), nate [EMAIL PROTECTED] wrote: Faheem Mitha said: Dear People, In case anyone is wondering, I am trying to install the src rpm in order to compile it, an operation I had earlier successfully done on my home computer. I am no RPM expert but anytime I wanted to compile a source rpm I did rpm --rebuild filename.src.rpm its worked well for almost every package, is this not the right way ? Yes, but before it can compile, it has to first install the source rpm. after it's compiled it usually dumps a binary rpm somewhere in /usr/src/rpms. Yes, this is what should happen in theory. However, this does not work either, at least on the new computer. I get similar errors, ie. laplace:/home/faheem# rpm --rebuild plugger-4.0-17.src.rpm Installing plugger-4.0-17.src.rpm error: cannot create %sourcedir /usr/src/rpm/SOURCES error: plugger-4.0-17.src.rpm cannot be installed It works fine on my home computer. I'll try this on another computer and see if I can reproduce this behaviour. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt error, Dynamic MMap ran out of room
On Thu, 28 Nov 2002 10:36:02 -0500 (EST), Walter Tautz [EMAIL PROTECTED] wrote: E: Dynamic MMap ran out of room E: Error occured while processing ifplugd (NewFileVer1) E: Problem with MergeList /var/lib/apt/lists/ftp.de.debian.org_debian_dists_unstable_main_binary-i386_Packages E: The package lists or status file could not be parsed or opened. I am running woody with some packages from testing Take a look at http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=166758 The bottom line is you need to modify apt.conf. The syntax should be APT::Cache-Limit 4194304; (from /usr/share/doc/apt/examples/configure-index.gz). This is the current default, Jason Gunthorpe suggests changing it to 16777216 in the bug report page above. See /usr/share/doc/apt/examples/configure-index.gz and man apt.conf also. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Re: apt error, Dynamic MMap ran out of room
On Thu, 28 Nov 2002 10:36:02 -0500 (EST), Walter Tautz [EMAIL PROTECTED] wrote: E: Dynamic MMap ran out of room E: Error occured while processing ifplugd (NewFileVer1) E: Problem with MergeList /var/lib/apt/lists/ftp.de.debian.org_debian_dists_unstable_main_binary-i386_Packages E: The package lists or status file could not be parsed or opened. I am running woody with some packages from testing Take a look at http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=166758 The bottom line is you need to modify apt.conf. The syntax should be APT::Cache-Limit 4194304; (from /usr/share/doc/apt/examples/configure-index.gz). This is the current default, Jason Gunthorpe suggests changing it to 16777216 in the bug report page above. See /usr/share/doc/apt/examples/configure-index.gz and man apt.conf also. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
commercial scientific programs on Debian
Dear People, I was wondering whether how well (or at all) some commercial scientific programs run on Debian, say Sarge. Specifically, I was wondering about Gauss, Mathematica, Matlab, SAS, Splus. I think that Mathematica runs ok, but I'm not sure about the others. I suppose in most cases these programs are only available as binaries, so there would have to be compatibility between these binaries and the Debian shared libraries. Also, most of these commercial programs only support the commercial Linux distributions like Redhat/SuSE. If anyone has specific information about these programs, please let me know, and for simplicity let us assume that I am talking about the most recent versions. Thanks. Faheem. -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
